Home

COSMO Documentation Part VII: User's Guide

1. V P WMO descr type mnemonics meaning need need 0 31 002 int MEDRE Extended delayed descriptor replicat 3 03 050 Wind data at a pressure level 3 03 052 Wind data at a height level opt opt 0 04 086 int n NLTPD Time displacem since launch time s need need 0 08 042 int n MEVSS Extend vertic sounding significance opt need 0 07 004 float n MPN Pressure vertical location Pa need opt 0 07 009 int n NHHH Geopotential height gpm opt opt 0 05 015 float n MLADH Latitude displacem since launch site opt opt 0 06 015 float n MLODH Longitude displace since launch site need need 011001 int n NDNDN Wind direction degree true need need 0 11 002 float n NFNFN Wind speed m s 0 31 001 int MDREP Delayed descriptor replication factor 3 03 051 Wind shear data at a pressure level 3 03 053 Wind shear data at a height level 0 04 086 int n NLTPDO Time displacem since launch time s 0 08 042 int n MEVSSO Extend vertic sounding significance 0 07 004 float n MPNO Pressure vertical location Pa 007009 int n NHHHO Geopotential height gpm 0 05 015 float n MLADHO Latitude displacem since launch site 0 06 015 float n MLODHO Longitude displace since launch site 0 11 061 float n NVBVB Absolute shear in 1 km layer below 0 11 062 float n NVAVA Absolute shear in 1 km layer above m s
2. Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 134 Variables controlling the LHN algorithm itself along with the temperature increments Name Type Definition Purpose Comments Default reference precipitation lhn qrs LOG use of the vertically averaged precipitation flux as reference TRUE precipitation i e this is used as the model quantity instead of model surface precip to compare with the observed precip rate rqrsgmax REAL threshold in terms of fraction of the maximum of the precip 0 0 itation flux in a model column which defines the uppermost model level used to compute the vertically averaged precip flux which is then deployed as reference precipitation grid point search lhn search LOG search for an appropriate nearby model heating profile TRUE grid point search applied to those grid points where the observed and modelled rain rates differ strongly fac lhn search REAL threshold of the ratio of observed and simulated rain rate 5 0 above which the grid point search is performed rlhn search INT max radius in number of grid pts for grid point search 10 scaling factors of the latent heating profile fac lhn up REAL upper limit Qmaz of the factor o which scales 2 0 fac lhn down REAL lower limit Qmin the model latent heating profil
3. Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 119 Several of the subsequent blocks of descriptions of NAMELIST parameters contain variables which are arrays of 4 elements These always relate to the following physical quantities 1 horizontal wind u v 2 surface pressure ps this relates to station pressure observations and to pressure observation and analysis increments at the lowest model level 3 temperature T 4 humidity f Nudging coefficients Name Type Definition Purpose Comments Default gnudg 4 REAL nudging coefficients in 1 s 8 10 1 2 10 6 10 7 6 107 for radiosonde data gnudgar 4 REAL nudging coefficients in 1 s 6 10 0 8 107 5 for aircraft data gnudgsu 4 REAL nudging coefficients in 1 s 6 10 5 1 2 10 0 6 1074 for surface level data gnudggp REAL nudging coefficient in 1 s 0 for GPS derived IWV Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 120 Temporal weights Name Type Definition Purpose Comments Default ltipol ltipsu tipolmx tipmxsu LOG LOG REAL REAL linear temporal interpolation linear interpolation in time between pairs of collocated upper air reports one report in the past and one in the future
4. flags on latitude longitude date time altitude Further details related to these elements are given in the AOF documentation in appendix A of Documentation Part IV Implementation Guide In file YUAOFEX the header is repeated in octal numbers Then each observation level of the report body is also given first in digital numbers and then in octal numbers For each obser vation type used currently the first elements are pressure in 0 1 hPa or geopotential m wind direction wind speed m s and temperature 0 1 K For TEMP radiosonde reports this is followed by dew point 0 1 K height m 1000 and two flag words For PILOT and aircraft reports the same applies except that there is no value for dew point For SYNOP reports dew point is followed by pressure tendency 0 1 hPa 3h sea surface temperature 0 1K the weather group word and general cloud group word a pressure level code flag and two further flag words and optional groups eckckckckckck KA ck ck ck ck ck ck ck k k ck ck ck ck k ck ck ck ck ck ck k k k kckckckck ck ck ck ck ck ck ckck ck ck ck kk kk kk OF 2147483647 14425 88773930 22 17777777777 34131 522512452 26 40 50 2147483647 17777777777 522512452 2147483647 18863 20020109 88773930 2147483647 0 LIUPUPPUVTET 44657 114275615 17777777777 0 2140 2147483647 4134 17777777777 eckckckckckck ck ckck ck ck ck ckck ck ck ck kckck ck k ck ck ck RARA RARA ck ck koc k
5. Figure 8 17 Example file YUVERIF Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 173 report header 1 te noo Sl te d mM m E 13 14 basic report type gt 0 multi level report number of vertical levels 0 complete synoptic surface level report SYNOP short surface level report with cloud and precipitation 2 very short surface level report without cloud and precip 3 upper air single level report 4 GPS report on integrated water vapour IWV station identity longitude of observing station 1 100 deg latitude of observing station 1 100 deg mi m observation time relative to initial verification hour see VOF header min station altitude height of model orography at station location m observation type see Figure 8 12 code type see Figure 8 12 station characteristics bit pattern as octal number see below report flag word reasons for status gt 1 bit pattern as octal number see below report status 0 active report i e used by nudging or LETKF 1 single level aircraft set passive because used as part of a multi level rep 2 passive report i e not used by nudging or LETKF threshold quality control QC flag for extrapolated surface pressure from multi level radiosonde report 0 active data used value ok 1 active data used value not ok 2 only passive d
6. Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 132 The following variables and are relevant only if 1surfa TRUE 2 dimensional analyses derived from observations Name Type Definition Purpose Comments Default 2 m temperature analyses lt2m LOG 2 m temperature field is analysed if true FALSE ht2a REAL time in hours of first 2 m temperature analysis 999 relative to initial time of model run ht2i REAL time increment in hours between successive 999 2 m temperature analyses 2 m relative humidity analyses lrh2m LOG 2 m relative humidity field is analysed FALSE hh2a REAL time in hours of first 2 m relative humidity analysis 999 relative to initial time of model run hh2i REAL time increment in hours between successive 999 2 m relative humidity analyses 10 m wind analyses lffi0m LOG 10 m wind speed is analysed FALSE hffa REAL time in hours of first 10 m wind analysis 999 relative to initial time of model run hffi REAL time increment in hours between successive 999 10 m wind analyses precipitation analyses lprecp LOG precipitation is analysed FALSE hprc REAL time in hours of analysis of precipitation sum 999 relative to initial time of model run raintp REAL time period in hours over which precipitation is summed up 12 ldiasa LOG diagnos
7. Reading and writing Ready Files ytrans in CHAR Directory path for reading ready files ytrans out CHAR Directory path for writing ready files nincwait INT Seconds to wait until the next read attempt for a lateral bound 0 ary file if the corresponding ready file is not available nmaxwait INT Maximum seconds to wait until abort if the ready file for the 0 next lateral boundary file is not available Controlling the Soil Moisture Analysis 20 nsma stat INT Status for soil moisture analysis nsma stat is coded in the O PDS of the grib ed fields for soil moisture content in the local use area 0 this is a normal soil moisture field this is a field derived by soil moisture analysis Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 12 DATABASE Specification of Database Job 140 7 12 DATABASE Specification of Database Job This Namelist group has been eliminated in Version 4 25 since the direct writing to the database system has never been used Instead of using the file based GRIB IO there is also a capability to read and write GRIB files directly from a database system However this option works only if you work within the DWD IT environment which provides a very special but not portable interface csodaban to a commercial database system To use this option you should be familiar with the DWD database int
8. 6 4 Observation Input Files 58 use WMO descriptor type mnemonics meaning 3 02 036 Cloud with bases below station level 3 02 047 Direction of cloud drift 3 02 048 Direction and elevation of cloud 3 02 037 State of ground snow depth ground minimum temperature opt 0 20 062 int ME State of ground w or w o snow opt 0 13 013 float NSSS Total snow depth 0 12 113 float MTGTGH Ground min temperat past 12 hrs 3 02 055 Icing and ice 3 02 057 SHIP marine data 3 02 056 Sea surface temperature depth 3 02 021 Waves 3 02 024 Wind waves 3 02 023 Swell waves 3 02 043 Basic synoptic period data 3 02 060 SHIP period data E 3 02 038 Present and past weather opt 0 20 003 int NWW Present weather opt 0 04 024 int MGGTP Time period in hours Opt 0 20 004 int MWI1 Past weather 1 Opt 0 20 005 int MW2 Past weather 2 3 02 039 Sunshine data 3 02 040 Precipitation measurement 0 07 032 float MHOSEN3 Height of sensor above ground marine deck platform precip 1 02 002 Replicate next 2 descript twice opt 0 04 024 int 2 MGGTPI Time period in hours opt 0 13 011 float 2 MRRR Total precipitation total water equivalent of snow He Extreme temperature data 0 07 032 float MHOSENA Height of sensor above ground marine deck platf temper opt 0 07 033 float MHAWAS2 Height
9. Boundary field required for cloud ice rain and snow if appropriate NAMELIST parameters are set qi specific cloud ice content qr specific rain content qs specific snow content If the NAMELIST parameter LGEN is set to TRUE see 7 2 the input of initial and boundary data is skipped and artificial data are generated instead in the routines gen ini data and gen bound data Note that these routines have to be edited and modified by the user to generate appropriate conditions for the case under consideration Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files 53 6 4 Observation Input Files This section describes the format of the NetCDF file observation input to the COSMO model and what has to be done if these files are used as input to the nudging type assim ilation scheme inside COSMO or to writing NetCDF or ASCII i e YUVERIF feedobs files for verification purposes or to perform surface analyses based on observations e g of 2 m temperature The observation input described here currently relates to conventional observations only and it does not describe the gribbed input for the latent heat nudging An alternative observation input for the conventional observations is given by the AOF file Here a single binary file has to be read of which the format is described in a separate documentation available from christoph schraff dwd de The AOF interface is n
10. Table notes 1 Only either the pair of variables MID and NIIT or the single variable YDDDD is strictly needed to exist If both exist and have non missing values for a certain report then the values of the pair MIT and NIIT are used Only one of the variables MHOBNN and MHOSNN is strictly needed to exist If both exist and have non missing values for a certain report then the values of MHOBNN are used n in the variable type definition means that this variable has an additional dimension used here for the vertical levels and hence several values may be present in one report If the corresponding replication factors MEDRE or MDREP are zero for all reports in the NetCDF file then the corresponding multi dimensional variables do not need to exist and probably will not exist in the NetCDF file Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files 64 6 4 2 Observation types with templates proposed or approved by WMO The BUFR templates which the NetCDF files described in this sub section are based on are described in the lower part of http www wmo int pages prog www WMOCodes TemplateExamples html respectively for GPS zenith total delay and water vapour data at http egvap dmi dk support formats egvap_bufr_v10 pdf e BUOY File name cdfin_buoy The template follows the proposed WMO descriptor TM 3 08 008 The table below li
11. The EFR method for the temperature prediction is replaced by a direct solution of the heat conduction equation The effect of freezing melting of soil water ice is included The process of snow melting is changed A time dependent snow albedo is introduced The multi layer concept avoids the dependence of layer thicknesses on the soil type Additionally it avoids the use of different layer structures for the thermal and the hydrological sections of the model The lake model FLake Basic Namelist settings lsoil TRUE llake TRUE FLake Fresh water Lake is a lake model parameterisation scheme capable of pre dicting the surface temperature in lakes of various depth on the time scales from a few hours to many years see http lakemodel net for references and other information about FLake It is based on a two layer parametric representation assumed shape of the evolving temperature profile and on the integral budgets of heat and kinetic energy for the layers in question The same concept is used to describe the tempera ture structure of the ice cover An entrainment equation is used to compute the depth of a convectively mixed layer and a relaxation type equation is used to compute the wind mixed layer depth in stable and neutral stratification Both mixing regimes are treated with due regard for the volumetric character of solar radiation heating Simple thermodynamic arguments are invoked to develop the evolution equation
12. The last 12 lines in Figure 8 8 relate to single observed quantities or to single observation levels of multi level reports The lapse rate and wind shear messages deliver station identity observation time threshold value actual value in K resp m s and the pressure range of the rejected levels By default see NAMELIST parameter maxmlv a maximum of 100 observation levels are allowed in a multi level report see second but last message Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 161 STATION 04339 OBS LOCATION OUT OF DOMAIN 222 0 70 5 2 0 HRS STATION 004 OBS LOCATION OUT OF DOMAIN 10 5 70 8 0 1 HRS STATION EU0324 OBSERVATION TOO OLD 0 6 FORECAST HRS STATION 11142 HEIGHT 647 DIFF TO MODEL 937 TOO LARGE 0 0 HRS STATION 11916 HEIGHT 2007 DIFF TO MODEL 1294 TOO LARGE 0 0 HRS STATION 01455 HEIGHT 255 DIFF TO MODEL 410 TOO LARGE 0 0 HRS STATION LCAR METO HEIGHT 16 DIFF TO MODEL 172 TOO LARGE 1 0 HRS STATION ARDE LPT_ HEIGHT 1497 DIFF TO MODEL 2033 TOO LARGE 1 0 HRS STATION ASDEO09 SEA OBS IS LOCATED OVER MODEL LAND AT 10 0 5320 1 6 HRS STATION ASDE0O9 NO SURFACE LEVEL REPORT DERIVED FROM TEMP PILOT STATION DBBI SEA OBS IS LOCATED OVER MODEL LAND AT 8 5 53 5 7 0 0 HRS STATION 61933 STA HEIGHT FLAGGED 0 5 HRS STATION 62934 STA HE
13. 0 460 W SO 1 1 338 W SO ICE 1 0 234 T_SO 2 0 181 W SO 2 2 710 W SO ICE 2 0 000 T_SO 3 0 545 W SO 3 8 211 W SO ICE 3 0 000 T_SO 4 0 678 W SO 4 25 668 W SO ICE 4 0 000 T SO 5 2 592 W_SO 5 26 953 W_SO_ICE 5 0 000 T_SO 6 5 600 W_SO 6 353 938 W_SO_ICE 6 0 000 T SO 7 8 717 W SO 7 1061 812 W SO ICE 7 0 000 T SO 8 9 795 W SO 8 3185 500 Ww SO ICE 8 0 000 Temperatures T2M 1 191 Winds U10M 0 832 dgr C TD2M 79 592 m s V10M 0 584 TMIN2M 1 191 VBMAX10M 1 206 TMAX2M 1 191 Solar radiation SOBT 0 000 Thermal radiation THBT 232 998 w m 2 SOBS 0 000 w m 2 THBS 104 491 Photosynt active Rad PABS 0 000 Surface albedo ALB 17 108 Precipitation rates and amount Cloud Cover mm d mm RAIN_GSP 0 000 0 000 CLCH 4 701 SNOW_GSP 0 000 0 000 CLCM 0 000 RAIN_CON 0 000 0 000 CLCL 0 000 SNOW_CON 0 000 0 000 CLCT 4 701 TOTAL 0 000 0 000 Figure 8 2 Example file M_stationname with long grid point output Part VII User s Guide 4 28 Section 8 Model Output 8 1 ASCII Output for the Forecast Model 151 LON geographical longitude in of the grid point FC Coriolisparameter in 1074571 SOIL TYPE type of the soil keys 0 9 PS unreduced surface pressure in hPa DPSDT tendency of surface pressure in hPa h For all main levels k 1 ke_tot the following parameters are printed Pmain pressure in hPa of the main levels T temperature in C Q
14. 3 002 3 020 3 031 3 035 3 032 3 025 53 013 2 991 2 977 2 953 2 926 2 894 2 854 2 807 2 750 2 686 R OLO 24539 0H SUN 17 02 2008 18 UTC 8 nergy 1 12 J kg kinetic energy 207 08 0H SUN 17 02 2008 18 UTC vamx m s 47 647 47 693 47 724 47 749 47 769 47 784 47 797 47 807 47 817 47 826 47 834 47 842 47 850 47 859 47 868 47 879 47 894 47 910 47 917 47 921 47 926 47 933 47 941 47 948 47 955 47 963 47 971 47 980 47 990 47 997 48 001 48 005 48 017 48 033 48 045 wamx m s 0 800 0 801 0 810 0 828 0 845 0 860 0 869 0 874 0 877 0 880 0 883 0 883 0 883 0 881 0 885 0 977 0 946 0 975 002 L 029 L 056 084 La DTL 138 164 1 189 212 233 1 249 264 1 275 1 285 4291 1 296 1 300 wa850 cm s 3 231 3 763 3 1982 3 946 4 025 4 151 4 216 4 292 4 342 4 388 4 415 4 432 4 444 4 448 4 453 4 455 4 460 4 465 4 471 4 478 4 485 4 494 4 504 4 513 4 522 4 532 4 541 4 550 4 559 4 566 4 574 4 582 4 587 4 593 4 598 19 02 2008 J kg 19 02 2008 wa500 wa300 cm s cm s 3 940 54151 5 798 7 087 4 657 5 440 4 945 5 979 4 739 5 595 4 887 5 730 4 806 5 588 4 855 5 617 4 819 5 595 4 814 5 567 4 815 5 583 4 814 5 563 4 823 5 581 4 822 5 574 4 836 5 592 4 844 5 586 4 859 5 597 4 877 5 600 4 898 5 617 4 918 5 641 4 935 5 661 4 957 5 680 4 981 5 696 52003 5 708 5 027 5a 116 5 054 5 728 5 080 5 744 5 1
15. Cloud cover at saturation in statistical cloud diagnostic 0 5 clc_diag 0 1 q crit REAL Critical value for normalized over saturation 4 0 q crit 1 10 crsmin REAL Minimum value of stomatal resistance used by the BATS 150 0 approach for vegetation transpiration itype trvg 2 crsmin 50 200 qcO REAL Cloud water threshold for autoconversion 0 0 qi0 REAL Cloud ice threshold for autoconversion 0 0 entr_sc REAL Mean entrainment rate for shallow convection 0 0003 Introduced in Version 4 5 thick_sc REAL limit for convective clouds to be shallow in Pa Recom 25000 0 mended values for thick_sc thick_sc 10000 0 45000 0 Introduced in Version 4 18 mu_rain REAL Shape parameter of the rain drop size distribution 0 0 Reasonable values are mu rain 0 0 5 0 Introduced in Version 4 5 ATTENTION In Version 4 21 the default value has been changed from 0 5 to 0 0 rain_n0_factor REAL To reduce the evaporation of rain drops 1 0 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 5 TUNING Parameters for tuning dynamics and physics 106 Name Type Definition Purpose Comments Default cloud num REAL Cloud droplet number concentration 5 0 E 08 vOsnow REAL Factor in the terminal velocity for snow This was a local 25 0 variable in the subroutines hydci pp and hydci pp gr be fore but had different values To reproduce the resul
16. Comments Default hincbound ydirbd lchkbd yvarbd llb gi llb qr qs llb gg lbdana ytunitbd nlgw bd lbd frame npstrframe ilevbotnoframe REAL CHAR LOG CHAR LOG LOG LOG LOG CHAR INT LOG INT INT Interval in hours between two consecutive sets of boundary data boundary update frequency Within this time inter val boundary values are interpolated linearly in time Only multiples of 0 25 are possible Directory path of the boundary data files Checking the boundary data for min max values has been eliminated in Version 4 12 If TRUE take cloud ice values contained in the lateral boundary data as boundary condition for cloud ice oth erwise cloud ice boundary conditions are set in the model If TRUE take rain and snow values contained in the lateral boundary data as boundary condition If TRUE take graupel values contained in the lateral boundary data as boundary condition If TRUE use analysis data as lateral boundary conditions Time unit indicator in the boundary data file name see Section 6 2 Number of prognostic soil water levels in boundary data This parameter is only used for the old soil model Normally the boundary data are defined on the full 3 D model domain IF 1bd frame TRUE it is assumed that the boundary data are only defined on a lateral frame e g when using boundary data obtained from ECMWTF IFS Numbe
17. FALSE ladv deep LOG To use all metric advective terms FALSE Parameters for the semi implicit time integration scheme Name Type Definition Purpose Comments Default ikrylow si INT Dimension of Krylow space for elliptic solver GMRES 20 iprint si INT Selects output of print statistics of the elliptical solver 0 gt 1 Print statistics in file YUSOLVER 0 No statistics are printed 0 Print statistics on the screen maxit si INT Maximum number of iterations for elliptic solver 200 eps si REAL Precision limit for the convergence of the solution from the 1078 elliptic solver Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 3 DYNCTL Parameters for the Adiabatic Model 87 Parameters for the Runge Kutta scheme Name Type Definition Purpose Comments Default irunge_kutta irk_order iadv order ieva order itheta adv ltadv limiter intcr max lva impl dyn itype fast waves INT INT INT INT INT LOG INT LOG INT Parameter to select a special Runge Kutta scheme 0 This option has been eliminated in Version 4 23 Associated with this option was the src 2timelevel dynamical core which has been removed from the code 1 Use new RK scheme from module src runge kutta 2 use new TVD RK scheme Order of the Runge Kutta scheme Can be in the range from 1 3 The value 3 is recommended Order
18. Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 184 Next two examples for the temperature correction balancing the surface pressure nudging are given one for a grid point with surface pressure close to 1000hPa and one for an elevated grid point The information provided at each model level consists of the height in m the temperature correction in K hPa the full pressure and the hydrostatic pressure in hPa the quotient between the hypothetic height increment at upper levels and at the surface the quotient between the resulting upper air pressure increment and surface pressure increment and the model level index The lines containing dt dt2 dtdeh describe that at timestep 0 the length of the timestep is halved outside the nudging but set to the original length temporarily within the nudging Before that the size of the coarse grid used to speed up the Poisson solver is indicated Messages are also written whenever two dimensional surface analysis based on Synop obser vations are written to Grib files t2m relates to 2 m temperature r2m to 2 m relative humidity fff to 10 m wind speed and prc to precipitation The lines with maybe not written to YUVERIF indicate reports that are flagged not to have been written to the VOF or NetCDF feedobs files at the time when they are deleted internally for the reason of being too old to be used any
19. User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 183 Next see Figure 8 24 the path name of the NetCDF feedobs or feedback file is given At timestep zero and after each time box defined by tconbox the number of reports and number of individual observations that are currently written to the feedobs file are also indicated as well as the numbers of reports and obserations already having been written at previous timesteps It follows a line with NUMBER OF SINGLE OR PAIRS OF REPORTS WITH OBS INCREMENTS This provides the total number of increment reports resp stations in case of temporal linear interpolation of observations which have been used to compute the observation increments for the purpose of nudging or of writing a NetCDF feedobs or VOF file at the timestep given at the beginning of the line In the subsequent lines it is specified how many of these increment reports are multi level upper air single level surface level surface pressure and scatterometer increment reports The multi level reports are further distinguished between radiosonde TEMP PILOT balloon wind profiler radar VAD wind aircraft RASS tem perature and GPS humidity from ZTD resp IWV observations increment reports An additional entry is added for preparing the use of retrievals from satellite radiance data this number is always zero here This whole block of lines is also written at tconbo
20. dt refers to the large time step 10 0 size At associated with the Leapfrog or Runge Kutta time integration of the slow modes Since vertical advection and diffusion is calculated implicitly the large time step can be determined from the linear stability criterion for horizontal advection As At lt 2 Umax where As is the minimum grid spacing in physical space and Umax 18 the maximum absolute horizontal velocity during the forecast For operational purposes Umar is estimated as a high jet stream velocity of about 120 m s The num ber N of small time steps Ar per Leapfrog time intervall 2At N A is calculated automatically inside the pro gram using the CFL stability criterion for horizontal acous tic wave propagation to estimate Av With the above value for Umax we get 7 small steps per Leapfrog intervall or 4 steps per Eulerian forward time step in the Runge Kutta scheme Note Both too small values Ns lt 3 and too large val ues N gt 20 for the number of small steps may have a detrimental impact to the solution Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 2 RUNCTL Parameters for the Model Run 80 Basic control parameters The basic control parameters allow to switch on and of main features and or packages of the model system Name Type Definition Purpose Comments Default lphys ldiagnos ldfi luseobs leps luse_rttov
21. tion Depending on which implementation is used the corresponding macro has to be set when compiling the COSMO Model for example with DRTTOV7 RTTOV7 to use the RT TOVT library Note that the COSMO Model does not use the official RT TOV7 library but a modified one which takes care of using this library within parallel programs It also contains some optimizations for vector processors RTTOV9 to use the RT TOVO library RTTOV10 to use the RT T OV10 library To use RTTOV9 or RTTOV10 another module developed at DWD is necessary as interface to this library mo_rttov_ifc f 90 This module is available from DWD The RTTOV model can compute radiances and brightness temperatures of several instru ments which are located on diverse satellites T he implementation provided in the COSMO Model only can compute the values for the following two instruments and channels specified in Table 7 8 For every channel the following 4 products can be computed e cloudy brightness temperature e clear sky brightness temperature cloudy radiance e clear sky radiance Note that the values for METEOSAT satellites can be computed but be aware of the location of these satellites Only MSG 2 is located such that it is useful for Europe Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 7 SATCTL Controlling the Synthetic Satellite Images 111 Table 7 8 Instruments and channels for use in the CO
22. yvarzl CHAR List of variables for z levels zlev REAL Z levels yvarsl CHAR List of satellite channels for output ngrib INT List of output steps given in time steps 99999 ncomb INT Triplets for building ngrib 99999 hgrib REAL List of output steps given in hours 0 0 hcomb REAL Triplets for building ngrib 0 0 Controlling the output domain and time unit indicator ytunit CHAR Time unit indicator in the output data file name see Sec NES tion 6 2 ydomain CHAR Domain type full or sub indicator in the output data file f name see Section 6 2 slon REAL Left longitude for a subdomain startlon_tot slat REAL Bottom latitude for a subdomain startlat_tot elon REAL Right longitude for a subdomain endlon_tot elat REAL Upper latitude for a subdomain endlat_tot Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 14 GRIBOUT Controlling the Grib Output 145 Additional parameters for controlling the output Name Type Definition Purpose Comments Default ysystem CHAR This switch has been eliminated in Version 4 28 File database or ecfs ydbtype CHAR This switch has been eliminated in Version 4 28 Type of the database yform write CHAR Specifies the format of output files grbi grbi default write GRIB1 data with DWD s GRIB library apil write GRIB 1 data with ECMWPF s grib api api2 write GRIB
23. 3 2 with the boundary values psz po z 0 and Tsz To z 0 for the pressure and temperature at mean sea level z 0 then yields the vertical profiles of the reference state psrexp Est 1 1 He if 840 poz l ps exp gd if820 3 3 Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 2 Differential Form of Thermodynamic Equations 11 20gz Tis Tous e SL For the three parameters psr T s and B which define the basic state we use the default values ps 1000hPa Ts 288 15K and 8 42K The variable names in the programs are pOsl psr t0s1 Tsz and dtO1p 5 resp The model equations are formulated with respect to a rotated lat lon grid with coordinates A p The rotated coordinate system results from the geographical Ag Yg coordinates by tilting the north pole see Part I of the Documentation Dynamics and Numerics In the vertical we use a generalized terrain following height coordinate where any unique function of geometrical height can be used for transformation Since doesn t depend on time the A o C system represents a non deformable coordinate system where surfaces of constant are fixed in space in contrast to the pressure based coordinate system of most hydrostatic models where the surfaces of constant vertical coordinate move in space with changing surface pressure The transformation of the model equations from the orthogonal
24. A v z system to the non orthogonal terrain following A p C system is given by the three elements of the inverse Jacobian matrix 7 Oz Oz Oz Jy Ja Jo Jj Je J vVG 3 4 A 13 a p 23 Ce C 33 OC VG The terrain following system of the COSMO Model is defined to be left handed i e the value of the coordinate increases with decreasing height z from the top of the model to the surface Thus J is always negative and equal to the negative absolute value VG det 7 of the determinant of the inverse Jacobi matrix 3 2 Differential Form of Thermodynamic Equations By transforming the primitive hydro thermodynamical equations to the A v C coordinate system and subtracting the basic state we achieve the following set of prognostic model equations for the three components u v and w of the wind vector the perturbation pressure p the temperature T and the humidity variables q Ou WU 1 Op Ja Op Ou uu _ My at Vu y Bg fv usc Vs ac Ov u 1 8p Jo Or IL m My a TV Vot mr tang fu 55 g Bii 1 Op E v Vw Gac 3 5 Op pr Y VE gpow cpa cva pD ot DCud Oq Uv l f Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 3 Horizontal and Vertical Grid Structure 12 Tv Vgl SM Ma p Here the continuity equation has been replaced by an equation for p In Eqs 3 5 a is the radius of t
25. Definition Purpose Comments Default y scalar advect CHAR String to specify the special horizontal advection type BOTT2 only for the Runge Kutta dynamical core This param STRANG eter replaces the 2 former parameters yef adv qx and 1sl adv qx The following table shows the possible settings of y scalar advect and how the old settings can be re placed only for options that were also available before SL3 MF SL Advection with tricubic interpolation mul tiplicative filling option 1sl_adv_qx TRUE yef adv qx SL3 SFD SL Advection with tricubic interpolation se lective filling diffusion option Bott2 Bott4 Bott 2nd 4th order finite volume scheme 1S1 adv qx FALSE yef adv qx Bott x Bott2_XYZYX Bott4_XYZYX the same with modified sequence of Strang splitting Bott2 Strang Bott4_Strang Bott 2nd 4th order finite volume scheme with Strang splitting z y 2x y z Bott2_Strang B Bott4 Strang B the same but only works in the lowest k offset levels of the at mosphere k offset is hardcoded with 5 vanLeer vanLeer type advection 1sl_adv_qx FALSE yef_adv_qx vanLeer vanLeer Strang vanLeer type advection with Strang splitting PPM PPM type advection 1sl_adv_qx FALSE yef_adv_qx PPM PPM Strang PPM type advection with Strang splitting MPDATA not yet available work at MeteoCH NOTE these strings are now case insensiti
26. LEVEL EVENTS APPLY TO MULTI LEVEL DATA ONLY THE ORDER OF ALL EVENTS MATCHES THE ORDER OF THE CHECKS EXCEPT FOR EVENT 8 LEVEL REJECTED NUMBER OF LEVELS EXCEEDING ODR SIZE LEVEL REJECTED PRESSURE PILOT PRESSURE AND HEIGHT MISSING LEVEL REJECTED PRESSURE PILOT HEIGHT FLAGGED LEVEL REJECTED TOO MANY SURFACE LEVELS LEVEL REJECTED PILOT HEIGHT LEVEL OUTSIDE RANGE OF MODEL LEVELS LEVEL REJECTED PRESSURE 9 HPA OR LEVEL BELOW STATION HEIGHT LEVEL REJECTED SIGNIFICANT LEVEL ABOVE A SPECIFIED LIMIT LEVEL REJECTED REDUNDANT LEVEL IN REPORT NOT ACTIVE YET PRESSURE TEMP HEIGHT MISSING PRESSURE TEMP HEIGHT FLAGGED PRESSURE BAD REPORTING PRACTICE PRESSURE HEIGHT DISTANCE TO OROGRAPHY OR STATION HEIGHT TOO LARGE PRESSURE TENDENCY FLAGGED OR ABSOLUTE VALUE 40 HPA 3H TEMPERATURE MISSING TEMP AT SIGNIFICANT TEMPERATURE LEVELS ONLY TEMPERATURE FLAGGED TEMPERATURE lt 90 C OR gt 60 C P lt 700HPA gt 20 C ETC TEMPERATURE AT 2M HEIGHT OR HEIGHT DISTANCE TO OROGRAPHY TOO LARGE TEMPERATURE TEMP ONLY LAPSE RATE TOO LARGE 1 2 34 5 6 7 8 9 011 12 13 14 15 16 17 18 0 0 00 0 0 0 0 129 20 2257 0 36 0 0 200 O 0 0 00 0 0 0 0 4303 3 Q 111 0 2569 0 0 219 Q 0 0 00 0 0 0 0 1 4 0 0 0 0 0 0 0 38 1 00 02163402 0 4212 77 0 0 519 93 0 0 18 0 0 00 0 0 0 0 1038 0 0 0 01038 0 0 0 0 ENTS DEFINITIONS CONTINUED HUMIDITY MISSING TEMP AT SIGNIFICANT LEVELS BELOW 300 HPA LEVEL HUMIDITY FLAGGED HUMIDITY DEWPOINT lt
27. The following parameters apply only for the RTTOV9 sat long 01 REAL Position of first satellite longitude 999 0 sat long 02 REAL Position of second satellite longitude 999 0 extrp type INT Type of extrapolation above highest model level 0 e 0 constant e 1 linear e 2 extrapolate towards a climatological value iceshape INT To specify whether ice particles are 1 e 1 hexagonal e 2 or ice aggregates iwc2effdiam INT Type of conversion of ice water content to effective diameter 4 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 8 INICTL Parameters for the Model Initialization 113 7 8 INICTL Parameters for the Model Initialization When the COSMO Model starts with interpolated data from a coarse grid driving model GME IFS or the COSMO Model itself the initial data will typically contain unbalanced information for the mass and wind field This will give rise to spurious high frequency oscil lations of high amplitude during the first hours of integration dynamical adaptation Thus the initial data should be modified by an initialization procedure in order to the unbalanced gravity and sound wave components to a realistic level For this purpose a digital filtering initialization scheme DFI has been implemented By default the initialization consists of a 1 hour adiabatic backward integration followed by a 1 hour diabatic forward integration of the model INICTL
28. Z V T Q 1100 0 0 o 1100 0 0 0 STA 08007 OBTYP 6 BLACKLISTED Z V T Q 0 0 1100 900 0 0 0 0 STA 08007 OBTYP 6 BLACKLISTED Z V T Q 0 0 600 0 0 0 0 0 STA 10678 OBTYP 6 BLACKLISTED Z V T Q 0 0 200 0 1100 0 0 0 PRECIPITATION AMOUNT EXCEEDS LIMIT DATUM REJECTED STID 08226 RR 914 0 TR 12 PRECIPITATION AMOUNT EXCEEDS LIMIT DATUM REJECTED STID 07649 RR 18 0 TR 1 03507 Fog with precip weather 127 vis 300 flags 0 0 10488 Fog with precip weather 51 vis 200 flags 0 0 10544 Fog with precip weather 77 vis 300 flags 0 0 LAPSE RATE EU8943 0 8 THRESHOLD VALUE 0 5 0 69 P 332 391 LAPSE RATE 22550 2 5 THRESHOLD VALUE 1 5 2 10 P 565 549 WIND SPEED SHEAR 03920 3 0 THRESH VALUE 42 6 44 0 P 250 4 200 WIND SPEED SHEAR 03496 3 0 THRESH VALUE 39 9 50 0 P 400 300 SINGLE LEV REP 06458 PRESSURE TENDENCY 4506 FLAG 0 SINGLE LEV REP 07600 NO ACCEPTED DATA IN REPORT MULTI LEV REP 03501 130th LEVEL BUT ODR SIZE IS 100 Figure 8 8 Example file YUREJCT Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 162 8 2 4 YUQUCTL Nudging Data Rejected by Quality Control YUQUCTL see Figure 8 9 lists the data which are rejected by the quality control at the observation time Note that as the quality control procedures are applied to each observation several times during the individual assimila
29. and these integrations can be switched on off Name Type Definition Purpose Comments Default imin_integ INT Starting i index of the cuboid 1 nboundlines l integrals LOG To switch on off the computation of integrals FALSE imax integ INT Ending i index of the cuboid ie tot nboundlines jmin_integ INT Starting j index of the cuboid 1 nboundlines jmax_integ INT Ending j index of the cuboid je tot nboundlines kmin integ INT Starting k index of the cuboid 1 kmax integ INT Ending k index of the cuboid ke tot Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 7 SATCTL Controlling the Synthetic Satellite Images 110 7 7 SATCTL Controlling the Synthetic Satellite Images In August 2003 a project was started between DWD and the DLR Institute for Atmospheric Physics which aimed at the generation of synthetic satellite images within the COSMO Model The RTTOV Radiative Transfer model for TIROS Operational Vertical sounder model Version 7 is used to compute radiances for satellite infrared or microwave nadir scan ning radiometers from an atmospheric profile of temperature variable gas concentrations and cloud and surface properties In Version 4 18 also the use of RT T OV Version 9 has been implemented and since Version 4 26 also RTTOV Version 10 can be used The different RT TOV libraries are included in the COSMO Model by conditional compila
30. corrections for the radiation Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 4 PHYCTL Parameters for the Diabatic Model 97 Name Type Definition Purpose Comments Default ico2_rad INT Parameter to choose a special COs concentration scenario 0 0 constant COs concentration of 360 ppm default for weather prediction 1 A1B scenario only CO is considered 2 A1B scenario effective CO is considered i e CO2 CH4 N30 3 B1 scenario only CO is considered 4 B1 scenario effective CO is considered i e CO2 CH N20 5 6 7 Scenario RCP2 6 8 Scenario RCP4 5 9 Scenario RCP6 10 Scenario RCP8 5 1co2_stab LOG Option to perform simulations with stabilized GHG forc FALSE ings iy co2 stab LOG To define the year when GHG stabilization begins this is 2001 only in effect if 1co2 stab TRUE lemiss LOG Option to use an external surface emissivity map if set to FALSE TRUE If lemiss is FALSE default a constant surface emissivity is assumed itype aerosol INT Switch to choose the type of aerosol map 1 1 Tanre As it was up to Version 4 10 in the model Constant aerosol distributions are given for rural ur ban desert areas and the sea 2 Tegen A monthly aerosol climatology is used for sul fate drops total dust organic black carbon and sea salt itype_albedo INT Switch to choose the type of solar surface
31. email address ied yncglob references CHAR URL report etc 12 ncglob realization INT number of the realisation of the experiment 1 The yncglob parameters are written into the netCDF output as global attributes see example header output below 5 2 3 netCDF Header Section The following is a typical content of the header section of a netCDF formatted output file Such a listing can be produced by using the ncdump command ncdump h 1ffd1979010200 nc in this example The dimensions of the variables are written in the C language order i e the last coordinate comes first Example float T time level rlat rlon in C reads float T rlon rlat level time in F90 netcdf 1ffd1979010200 dimensions rlon 101 rlat 107 srlon 101 srlat 107 level 20 leveli 21 height 2m 1 height 10m 1 soil 9 soill 10 time UNLIMITED 1 currently bnds 2 variables char rotated_pole rotated_pole long_name coordinates of the rotated North Pole rotated pole grid mapping name rotated latitude longitude rotated pole grid north pole latitude 32 5f rotated pole grid north pole longitude 170 f float rlon rlon rlon standard name grid longitude rlon long name rotated longitude rlon units degrees float rlat rlat rlat standard name grid latitude rlat long name rotated latitude rlat units degrees float srlon srlon srlon sta
32. end input io amp IOCTL lgen FALSE lasync io FALSE ngribout 1 yform write grbl amp DATABASE amp GRIBIN hincbound 1 0 lchkini TRUE lchkbd TRUE lbdana FALSE lana q TRUE llb gi TRUE lana rho snow TRUE lana qr gs FALSE ydirini gtmp routfor dat initial ydirbd gtmp routfor dat boundaries amp GRIBOUT hcomb 0 0 48 0 1 0 lanalysis FALSE lcheck TRUE 1_p filter TRUE l1 z filter TRUE lwrite const TRUE yvarml U Vy VW VT tov 0c QI OR OS y P Kp UPS T_SNOW TETT SO W_SO W_SNOW Y OV S Wr RAIN GSP SNOW GSP RAIN_CON SNOW_CON U 10M V 10M V T 2M V TD 2M TMIN 2M TMAX 2M n VMAX 10M TCM 1 ICH CLCT LOBO ELE Y TOC E TOT TOQV TKE y WCL r T CL FRESHSNW RHO SNOW W ICE H SNOW t betasw 0 4 epsass 0 15 hd corr q 0 5 hd corr t 0 75 hd dhmax 250 itype hdiff 2 lcond TRUE lspubc TRUE itype_lbcqx 1 end input dyn cat INPUT PHY end input phy amp PHYCTL gsp TRUE lprogprec TRUE ltrans prec TRUE itype gscp 3 rad TRUE nradcoarse 1 lradf avg FALSE hincrad 1 0 lforest TRUE ltur TRUE lexpcor TRUE ltmpcor FALSE lprfcor FALSE lnonloc FALSE imode turb 1 icldm turb 2 r imode_tran 1 itype_synd 2 itype_tran 2 icldm_tran 0 end_input_dia stationlist_tot 0 0 50 050 8 0 0 52 220 14 0 0 47 800 10 600 Frankfurt Flug
33. for each analysis variable analysis date and correction scan Figure 8 28 The squared dis tance zdistm in km to that point the distance scaled by the scan radius zrormx and the horizontal weight wwa are written up for each observation influencing that point SURFACE ANY 2 D ANALYSIS RUN 2M TEMPERATURE ANALYSIS 2M RELATIVE HUMIDITY ANALYSIS PRECIPITATION ANALYSIS 10M WIND SPEED ANALYSIS ANALYSIS DATE MON 16 04 2012 06 UTC Total precipitation amount in the last 12 hours Precipitation analysis parameters Weights function W H R V H where H R RMAX 2 R 2 RMAX 2 R 2 V H 0 5 COS H HMAX PT 0 5 HMAX MAX MOD ORO 400 m Analysis function A SUM W D SUM W Number of scans 3 Scan 1 radius 40000 m Scan 2 radius 70000 m Scan 3 radius 110000 m No smoothing 2m temperature analysis sucessive correction parameters Weights function W H R V H where H R RMAX 2 R 2 RMAX 2 R 2 V H HMAX 2 H 2 HMAX 2 H 2 HMAX MAX MOD ORO 2 7 300 m Increment function I SUM W D SUM W Number of scans 3 Scan 1 radius 200000 m Scan 2 radius 100000 m Scan 3 radius 50000 m 2m Rel Humid Analysis Sucessive Correction parameters Weights function W H R V H where H R RMAX 2 R 2 RMAX 2 R 2 V H HMAX 2 H 2 HMAX 2 H 2 HMAX MAX MOD ORO 3 0 300 m Increment function I SUM W
34. framework and runs on distributed memory machines using domain decomposition Nudging or Newtonian relaxation consists of relaxing the prognostic model variables towards prescribed values within a predetermined time window see e g Davies and Turner 1977 Stauffer and Seaman 1990 In the present scheme nudging is per formed towards direct observations which is more appropriate for high resolution ap plications than nudging towards 3 dimensional analyses Stauffer and Seaman 1994 A relaxation term is introduced into the model equations and the tendency for a prognostic variable w x t is given by A60 Fou Gus Y Wal vet 813 K obs F denotes the model dynamics and physical parameterizations v the value of the kt observation influencing the grid point x at time t xj the observation location Gy the constant so called nudging coefficient and W an observation dependent weight which Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 6 Data Assimilation 28 usually varies between 0 and 1 Neglecting the dynamics and physics and assuming a single observation with a constant weight W equal 1 the model value at the observa tion location relaxes exponentially towards the observed value with an e folding decay rate of 1 G corresponding to about half an hour The observational information is provided to the nudging scheme in the form of NetCDF observation input files which are describe
35. lroutine 1_cosmo_art l pollen llm LOG LOG LOG LOG LOG LOG LOG LOG LOG LOG Main switch to include physical parameterizations If FALSE the corresponding Namelist input from PHYCTL see Section 7 4 is skipped and no physical parameterizations will be computed except cloud condensation and evaporation by saturation ad justment To supress this process set 1cond FALSE in in put group DYNCTL see Section 7 3 Main switch to include diagnostic calculations Main switch to run a model initialization by using the digital filtering scheme Main switch to use observations during the model run for data assimilation purposes To perform a free model forecast luseobs is set to FALSE Main Switch to run the model in ensemble mode Main switch to compute synthetic satellite images To indicate an operational run if TRUE This variable is used to set GRIB2 meta data Main switch to compute the Aerosol and Reactive Tracer mod ule The COSMO Model has to be compiled with DCOSMOART to activate this module in the source code of the COSMO Model The source code of the ART module itself can be ob tained from the Karlsruhe Institute of Technology Main switch to compute the Pollen module The COSMO Model has to be compiled with DPOLLEN to activate this mod ule in the source code of the COSMO Model The source code of the Pollen module itself can be obtained from the Karlsruhe Institu
36. soil moisture analysis SMA It is possible to perform any combination of these four processes and the first group of NAMELIST variables in NUDGING decides which of them will be performed Switches on the main processes Name Type Definition Purpose Comments Default lnudge LOG On off switch for nudging FALSE lverif LOG On off switch for verification FALSE i e for writing a VOF or NetCDF feedobs file see mveripr llhn LOG On off switch for latent heat nudging LHN FALSE lsurfa LOG On off switch for deriving 2D analyses from observations FALSE Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 115 cat INPUT ASS end input ass amp NUDGING inudge TRUE hnudgsta 0 0 hnudgend 7 0 tconbox 180 0 lverif TRUE lverpas TRUE mveripr 3 llhn TRUE llhnverif TRUE lhn wweight TRUE rqrsgmax 0 4 radar in ycdfdir itype obfile 2 hversta 0 001 hverend 3 0 nwtyp 7 niwtyp 1 4 2 1 1 3 1 iwtyp 0 132 133 136 137 1 4 2 9 823 830 930 12 kwtyp 1 1 1 1 2 2 2 2 1 khumbal 100 mruntyp 2 ntpscor 1 ptpstop 400 0 luvgcor TRUE mpsgcor 1 ltipol TRUE tipolmx 3 0 wtukrsa 3 0 wtukrse 1 0 ltipsu TRUE tipmxsu 1 0 wtuksua 1 5 wtuksue 0 5 wtukara 1 5 wtukare 0 5 msprpar 1
37. the COSMO Model With libgribi a data can be packed in unpacked from grib code This library also contains C routines to write data to and read it from disk The Grib library is available from DWD and is provided together with the source code for the model package A short guide for the installation is included in the tar file of the Grib library DWD still uses a Grib file format where all records are starting and ending with additional bytes the so called controlwords An implementation of the Grib library is prepared that also Part VII User s Guide 4 28 Section 4 Installation of the COSMO Model 4 1 External Libraries for the COSMO Model 30 deals with pure Grib files that do not have these controlwords But still we guarantee correct execution only if the controlwords are used To ensure this you have to set the environment variable export LIBDWD FORCE CONTROLWORDS 1 4 1 2 libnetcdf a Since Version 3 18 input and output of data can also be done in the NetCDF format Net work Common Data Format Using NetCDF requires an external library libnetcdf a The source code of this library can be downloaded from http www unidata ucar edu Usage of the NetCDF library can be controlled by conditional compilation and setting the macro NETCDF If this macro is not set during compilation the parts of the source code that do use NetCDF calls are not compiled and the library will not be linked to the binary NOTE The usage
38. to represent tangential Cartesian geometry with constant or zero Coriolis parameter Organization of the Documentation For the documentation of the model we follow closely the European Standards for Writing and Documenting Exchangeable Fortran 90 Code These standards provide a framework for the use of Fortran 90 in European meteorological organizations and weather services and thereby Part VII User s Guide 4 28 Section 1 Overview on the Model System 1 3 Organization of the Documentation 7 Table 1 2 COSMO Documentation A Description of the Nonhydrostatic Regional COSMO Model Part I Dynamics and Numerics Part II Physical Parameterization Part III Data Assimilation Part IV Implementation Documentation Part V Preprocessing Initial and Boundary Data for the COSMO Model Part VI Postprocessing Part VII User s Guide facilitate the exchange of code between these centres According to these standards the model documentation is split into two categories external documentation outside the code and internal documentation inside the code The model provides extensive documentation within the codes of the subroutines This is in form of procedure headers section comments and other comments The external documentation is split into seven parts which are listed in Table 1 2 Parts I III form the scientific documentation which provides information about the theo retical and numerical form
39. 0 0 0 0 2 0 PILOT SHIP D 0 0 0 o B 0 0 DO 8 0 0 0 0 0 0 0 PILOT Mobile o 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Wind Profiler Eu 0 0 0 0 T OQ 0 0 87 0 41 0 0 0 0 0 0 RASS SODAR Eu 0 0 0 0 0 0 0 0 36 0 0 0 0 0 0 0 0 Wind Prof RASS U 0 0 0 0 D p 0 0 0 0 0 0 0 0 0 0 0 RADAR VAD Wind Pr 0 0 0 0 96 0 0214 584 0 17 0 0 0 0 0 0 0 SATEM 0 Scatterometer ASCAT scatteromet 0 0 0 177 0 0 0 0 0 0 0 0 0 0 0 0 0 QuickScat scatter 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 GPS GPS by METO o 0 0 46 92 0 0301 0 0 2005 0 0 0 0 0 0 GPS by LPT_ 0 0 0 36 144 0 0116 28 O 188 0 0 0 0 0 0 GPS by LPTR 0 0 0 0 400 0 0425 71 0 607 0 0 0 0 0 0 0 0 12 0 0133 0 0 4 0 0 0 0 0 0 GPS by BKG_ 0 0 Figure 8 13 Example file YUSTATS third part incomplete Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 169 for each individual report why a certain number of the reports of a certain code type is set passive or rejected Further report events relate mainly to the processing of aircraft reports The DATA EVENTS tables in the forth part Figure 8 14 provide similar statistical information on the reasons for rejecting parts of reports i e either complete observation levels or single observations 1 DATA 0 1o01 0 NO Aa 0 SYNOP SHIP TEMP Land GPS by LPTR 0 SYNOP SHIP TEMP Land GPS by LPTR 1 DATA EV SYNOP Manual La SYNOP Automatic2623 EVENTS DEFINITIONS
40. 1 2 6 273 3 0 58 946 62 TROTE LB TE Q q0sseexa9 0309 295 11 86 48 37 244 2 2 60 AIREP EU5331 M 0 2 1 5 273 8 0 59 955 75 AIRES 9 5 3 1 0 desees 305 295 11 85 48 37 244 2 2 61 AIREP EU5331 M 0 3 0 4 274 3 0 60 963 80 420 2 5 1 2 0 10 keevo 308 295 11 82 48 36 244 2 2 61 AIREP EU5331 OBS CODE TYPE 2 244 STA MOD HEIGHT 443 LON 11 82 LAT 48 36 G P 308 295 HOUR 2 6 u v E rh p z v err t er rh er z er lev 0 3 0 4 274 3 0 60 963 8 420 2 5 1 2 0 10 384 052 L 5 273 8 0 59 955 8 490 2 5 31 1 0 10 384 eT 28 6 273 3 0 58 946 6 S70 2 5 1 1 O lQ 384 0 3 2 6 273 2 0 56 938 7 640 2 5 1 1 0 10 384 0 4 2 6 272 8 0 60 930 8 TIO 2 5 Lal QLQ swe 384 0 5 2 5 272 5 0 58 923 0 780 2 5 1 1 0 10 384 0 3 2 1 270 8 0 62 913 0 870 2 5 1 1 0 I0 384 0 6 228 5 270 8 04 53 902 0 970 2 5 31 1 0 10 384 0 4 2 6 270 0 0 52 890 1 1080 2 5 1 1 0 10 384 0 1 3 1 269 5 0 54 882 5 1150 2 5 1 0 0 10 384 AIREP EU5331 OBS CODE TYPE 2 244 STA MOD HEIGHT 481 LON 12 06 LAT 48 38 G P 311 295 HOUR 2 5 u v t rh p z v err t er rh er z er lev 0 6 3 5 269 3 0 58 882 5 1150 2 5 1 0 0 10 w 3984 0 3 2 6 269 0 0 62 881 4 1160 2 5 1 0 0 10 384 0 7 3 5 268 0 0 56 869 7 1270 2 5 140 0 10 w 394 cO 3 1 267 0 0 63 858 1 1380 2 5 1 0 0 10 384 0 2 4 1 266 5 0 62 848 7 1470 2 5 1 0 0 10 384 0 4 5 1 265 5 0 66 840 4 1550 2 5 1 0 0 10 384 GPS COMO METO
41. 17 timestep 0 obs satu obs IWV adjusted IWV model w satu extra time it rated repor re ice from model cloud rated polat STA hr erat lev ted trieved bias qv obs IWV ice model quot p int p rep ZIMM 0 5 I D deep 5 12 5 72 5 72 6 24 6 24 10 61 0 6 914 65 Figure 8 19 Example file YUPRINT second part related to observations and observation increments Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 180 ps spread 63112 weights 0 97 0 03 0 58 wght incr 52 83 2 9 25 9 dist 171 T 0 9 ps spread 63101 weights 0 97 0 03 0 91 wght incr 47 35 12 6 10 9 dist 163 T 0 9 End of ps ana i weight sqr 3 55 net weight 0 000409 ana incr 0 3988 1 0000 ps ana incr 18 126 0 399 mul k1 01400 3 p z obs v cut 0 8253 400 9165 100000 400 1597 10474 no k air kl EU1337 1 23720 height at obs cutoff 10949 7722 14176 k range 13 6 thresh RH k 40 40 285 385 10142 34 mx omy k 1 k 1 000 0 735 RH incr 0 045 0 092 airmy EU3080 131 84 1 omyk1 2 0 011 0 000 vCS fac 1 000 0 935 dz 2571 omym 0 716381 sfc kl 06610 2 94994 height at obs cutoff 563 9 400 0 k range 41 sfc kl 06610 3 94994 height at obs cutoff 563 9 572 4 k range 40 sfc cor 06610 4 35 167 103 140 spr par ob cutof 564 8 eee ee at k 564 480 402 omyk 06610 4 3 0 972 0 028 1 000 0 972 0 028 269 27 T 0 00 563 9 563 9 1 0 zuwi om u 06610 0 0419 0 7773
42. 174 The report flag word indicates all reasons why a whole report is not used actively in the assimilation nudging or LETKF This flag word is given by the following bit pattern a flag is true if the bit is set to 1 the report flag is equal to the quality check flags for reports in the NetCDF feedobs file see Feedback File Description bit numbers of report flag word 0 passive report type at observation location blacklisted or not on whitelist suspicious location or date time location not in valid area location not in valid height range incorrect surface land ice etc 10 redundant report 11 flight track error flag 12 report merged into another report e g aircraft single level into multi level report 13 thinning 19 no active observations in report QANE Note that the report flag word is written only if the observations are read from NetCDF observation input files If the observations are read from an AOF file a different flag word not described here is used Report Body The regular report body contains all the observed values and the quality flags for the indi vidual data It has 22 entries for complete synoptic surface level reports the last 6 of which are written to the VOF in a second line For the other basic types of reports the last few entries are omitted in such a way that the body length is as follows body length 22 for complete synoptic surface level reports 16 for short surface
43. 2 data with ECMWI s grib_api ncdf write NetCDF data nprocess ini INT Generating process identification for initial data 999999 nprocess bd INT Generating process identification for boundary data 999999 n num INT This switch has been eliminated in Version 4 28 Counter for nests nrbit INT Number of bits per value for grib packing Usually 16 bits 16 give sufficient accuracy nunit of time INT Indicates the Unit of Time for Grib Code 1 lcheck LOG Checking the output data for min max values TRUE lanalysis LOG If TRUE analysis output files are written when the model FALSE runs in nudging mode lwrite const LOG If TRUE constant fields are written as grib output at TRUE initial time luvmasspoint LOG Enables interpolation of horizontal winds to mass grid FALSE points for output lp filter LOG Logical switch if TRUE the fields on pressure levels are FALSE digitaly filtered to remove small scale noise lz filter LOG Logical switch if TRUE the fields on height levels are FALSE digitaly filtered to remove small scale noise itype vertint INT To specify the type of the vertical interpolation used to 1 interpolate values to p and or z level 1 Cubic tension splines previous method 2 Linear interpolation new in Version 4 17 l fi ps smooth LOG This switch has been renamed see below because what really is smoothed here is the mean sea level pressure lfi pmsl smooth LOG L
44. 35 40 40 40 40 40 4 0 TO IX 10 7 4 4 5 5 6 7 8 T 12 10 7 5 5 db 6 7 8 9 vi for humidity 4 element is given as a function of observation error background error and stability for height thickness thresholds see Scientific Documentation Part III qccsu 4 REAL QC thresholds f at observation time for surface level data 12 0 units for as for qcc 500 0 12 0 0 7 QC thresholds yt for GPS IWV data qcciq REAL constant part of the thresholds at observation time in mm 1 0 qcsiq REAL fraction of the IWV of the saturated model temperature profile 0 15 which is added to the QC threshold Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 127 Quality weights for observations Name Type Definition Purpose Comments Default doromx 4 REAL cut off and Gaussian radius in m for reduction factor e 100 0 which is part of quality weight e for surface level 150 0 observations 7 150 0 fp Az E 2 150 0 oromx f Az lt doromx e e for p only fpAz doromx e 1 for the other variables f Az gt doromx e 0 applied to all variables where Az Zobs 2mod difference in m between station height zops and model orography Zmoa for Ps 2moa height of lowest model level and f 4 for p if Az 0 i e if extrapolating fp 1 otherwise gt surface level data are
45. AIRCRAFT REPORT 11 REDUNDANCY BETWEEN 2 MULTI LEVEL OR 2 SINGLE LEVEL REPORTS 12 FLIGHT TRACK SUSPICIOUS OR EXAGGERATED COLOCATION 13 THINNING OF DENSE AIRCRAFT FLIGHT TRACK 14 SURFACE LEVEL FROM MULTI LEVEL REP REDUNDANT AGAINST OTHER REPORT 15 ONE MULTI LEVEL REPORT MADE FROM SINGLE LEVEL REPORTS 16 SINGLE LEVEL REPORT PUT IN MULTI LEVEL REPORT AND SET PASSIVE 17 MULTI LEVEL REPORT NOT BUILT DUE TO ODR SIZE LIMIT ADJUST NAMELIST 0 events 1 2 3 4 5 6 ToB 9 10 11 12 13 14 15 16 17 0 SYNOP SYNOP Manual Land 0 0 0 496 329 0 0508 54 O 125 0 0 0 0 0 0 SYNOP Automatic L 0 0 0 81 415 0 O x 2439 0 77 0 D o0 0 0 0 SHIP 0 0 0 30 6 0 0o o0 0 0 0 0 0 0 0 0 0 SHIP Automatic 0 0 0 15 7 0 0 o0 59 QU 18 0 0 0 0 0 0 0 AIREP CODAR 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 AIREP Aircraft 0 0 0 199 0 0 0 0 4 0 3 0 2 0 0 0 0 AMDAR 0 0 0 233 67 0 0 0 D 185 0 69 0 342 2359 38 ACARS o 0 0 144 50 0 0 0 3 0 109 40 394 0 148 1632 87 COLBA Const Lev B 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 SATOB SATOB 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 High res VIS Wind 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 AMV 0 D 0 0 0 8 0 0 0 0 0 0 0 0 0 0 0 SST as DRIBU 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 O DRIBU BATHY 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 DRIBU Drifting Bu 18 0 0 399 0 0 0 0 223 0 0 0 0 0 0 0 0 0 TEMP TEMP Land 0 0 0 116 0 0 0 0 71 0 7 0 0139 0 226 0 TEMP SHIP 0 0 18 4 2 0 o o0 3 90 0 0 0 0 0 6 0 0 PILOT PILOT Land 0 0 0 12 0 0 0 0 5 8 4
46. Figure 8 16 Example file YUSTATS warning messages related to specified array sizes Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 171 types of messages The first one relates to the arrays which store internally the observation reports themselves from the local sub domain and the second one to arrays containing the observation increments from the total model domain that are used at a certain timestep Both types of events the relation to certain NAMELIST parameters and considerations on the recommendations on how much to increase the values of these parameters are already decribed in detail in Section 8 2 5 8 2 7 YUVERIF Nudging Verification File VOF YUVERIF is called Verification Observation File VOF and lists all the active observations of which the observation time lies within a selected period between the beginning and the end of the model integration time It is important to note that under normal circumstances it never lists all the observations which are used actively in the nudging and which e g enter the statistics shown in file YUSTATS This is due to the use of a finite temporal weight function for the relaxation so that the nudging also uses data that are older than the beginning of the model integration time In a data assimilation cycle moreover it normally also uses data with observation time later than the end of the model run These data outsi
47. LOG Selection of analysed water content of snow FALSE lan_w_i LOG Selection of analysed interception water FALSE lan_w_so LOG Selection of analysed or external soil moisture FALSE lan_w_cl LOG Selection of analysed climatological soil water content FALSE lan_vio3 LOG Selection of analysed vertical integrated ozone FALSE lan hmo3 LOG Selection of analysed ozone maximum FALSE lan plcov LOG Selection of analysed plant cover FALSE lan lai LOG Selection of analysed leaf area index FALSE lan rootdp LOG Selection of analysed root depth FALSE lan rho snow LOG Selection of analysed density of snow FALSE The control of the parameters lan xxx is as follows TRUE Fields are used from external analyses or from an interpolated coarse grid field FALSE Fields are used from the continous data assimilation of the COSMO Model Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 14 GRIBOUT Controlling the Grib Output 144 7 14 GRIBOUT Controlling the Grib Output How to use the NAMELIST group gribout is explained in Section 8 4 in detail Basic parameters for controlling the output Name Type Definition Purpose Comments Default ydir CHAR Directory of the output file di yvarml CHAR List of model variables for output yvarpl CHAR List of variables for pressure levels plev REAL Pressure levels
48. PA oe Roy RH ee n 40 922 The NetCDE Data Format scsi 4 Boe Ae ee RO WU wes cnn 42 5 271 AGE Conventions 5 s s e Ok a poe a Ee eee Bee pee e i a 42 0 22 Namelist Ip p 23 4 m e e SoG p Re UR UR aa ee eee Y Ye 43 Da3 netCDP Header Section so 2 6 ba ea kk se RR Re aa a Eo a 43 5 2 4 Useful Post Processing Utilities aoao a 46 6 Input Files for the COSMO Model AT 6 1 File for Namelist Input ev coc gom o moon RR Pee s 4T 6 2 Conventions tor File Names s esci a Be eee o Oe o y wes 49 6 9 Imtialand Boundary Data ee o ommo be Xo oo OE RR e mon 51 6 4 Observation Input Files si ecsedi Roe xo Ue Kec a ee 8 53 Part VII User s Guide 4 28 Contents Contents iii 6 4 1 Templates for observation types for which Table Driven Code Forms TDCF defined by WMO exist o o 55 6 4 2 Observation types with templates proposed or approved by WMO 64 6 4 3 Observation types without templates proposed by WMO 69 644 The blacklist ile occasion hee S 73 7 Namelist Input for COSMO Modell 75 7 1 LMGRID Specifying the Domain and the Model Grid 77 7 2 RUNCTL Parameters for the Model Run 78 7 3 DYNCTL Parameters for the Adiabatic Model 85 7 4 PHYCTL Parameters for the Diabatic Model 94 7 5 TUNING Parameters for tuning dynamics and physics 104 7 6 DIACTL Parameters for Diagnostic Out
49. Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files e single level AMDAR File name cdfin_amdar The template follows the proposed WMO descriptor TM 311010 and is described in the table below The use of the variables is defined as follows need COSMO asks stringently for this variable and will abort if variable is absent opt X variable exists and is read used but COSMO will not abort if it does not exist up variable exists but is not read by COSMO 4 descriptor exists only in the BUFR file not in the NetCDF file use WMO descriptor type mnemonics meaning 031021 MADDF Associated field significance 1 3 11 005 Standard AMDAR report need 0 01 008 char 8 YAIRN Aircraft identification 0 01 023 int NOSNO Observation sequence number e 3 01 021 Latitude and Longitude need 005001 float MLAH Latitude high accuracy degree need 006001 float MLOH Longitude high accuracy degree 3 01 011 Year month and day need 0 04 001 int MJJJ Year need 0 04 002 int MMM Month need 0 04 003 int MYY Day 3 01 013 Hour minute and second need 0 04 004 int MGG Hour need 0 04 005 int NGG Minute 0 04 006 int MSEC Second need 0 07 010 int NFLEV Flight level 2 opt 0 08 009 int NDEPF Detailed phase of flight 3 need 0 11 001 int NDNDN Wind direction need 0 11 002 float NFNFN Wind speed opt
50. Relative humidity 0 07 032 float MHOSENO Height of sensor above marine deck platform for wind measurement 007 033 float MHAWASO Height of sensor above w surf wind 0 08 082 int NACH2V Artificial correction of sensor height to another value 007 033 float MHAWASI Height of sensor above w surf wind opt 0 08 021 int MTISI1 Time significance 2 time averaged opt 0 04 025 int NGGTP Time period in minutes need 0 11 001 int NDNDN Wind direction need 0 11 002 float NFNFN Wind speed opt 0 04 025 int NGGTPO Time period in minutes opt 0 11 041 float NFXGU Maximum wind gust speed opt 0 04 024 int MGGTP Time period in hours opt 013 011 float MRRR Total precipitation kg m2 Table notes 1 Only one of the variables MHOBNN and MHOSNN are needed to exist MHOBNN is preferred to exist and to be used if both variables exist and have non missing values because it should provide the precise height of the barometer for the pressure measure ment which is the observation with the most critical dependency on sensor or station height 2 n in the variable type definition means that this variable has an additional dimension i e several values may be present in one report If the corresponding replication factor MDREP is zero for all reports in the NetCDF file then these multi dimensional variables do not need to exist and probably will not exist in the NetCDF file Part V
51. VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 3 Horizontal and Vertical Grid Structure 14 k 1 2 j 1 2 i 1 2 i Figure 3 1 A grid box volume AV Ac AA Ay showing the Arakawa C Lorenz staggering of the dependent model variables transformation to a specific terrain following system where in principle any unique function of geometrical height z can be used As a default either a generalized sigma type coordinate 7 based on base state pressure or a generalized Gal Chen coordinate u based on height can be chosen In a second step this vertical coordinate is mapped onto the computational coordinate C with discrete coordinate values Cj k and an equidistant grid spacing of A 1 The latter mapping is by a table which relates specific values of the terrain following coordinate 7 or u to the Nc 1 values of the half level values Cy 1 2 In this way a user defined vertical grid stretching can be easily applied Details on the set up of the vertical grid are provided in Part I of the Documentation Dynamics and Numerics To render the model code independent on 77 or u all metric terms involving the three compo nents 3 4 of the Jacobi matrix are evaluated numerically on the computational grid These terms are rewritten in the form J 1 po Jo 1 Opo VG Via VG Vay ae 1 VG d ys goo where y y Opo O denotes the change of base state pressure with C In discretized form we
52. Version 4 12 and has been re placed by reduction factors for the interior and the boundary zone see below l diff Smag LOG to switch on off the Smagorinsky diffusion TRUE hd corr u REAL eliminated in 4 12 replaced by hd corr u in bd hd corr u bd REAL Reduction factor for the horizontal diffusion flux for u 0 25 v and w in the boundary zone of the domain hd corr u in REAL Reduction factor for the horizontal diffusion flux for u 0 25 v and v in the interior zone of the domain hd corr t REAL eliminated in 4 12 replaced by hd corr t in bd Also there are separate factors now for t and pp hd corr t bd REAL Reduction factor for the horizontal diffusion flux for t 0 0 in the boundary zone of the domain hd corr t in REAL Reduction factor for the horizontal diffusion flux for t 0 0 in the interior zone of the domain hd_corr_p_bd REAL Reduction factor for the horizontal diffusion flux for pp 0 0 in the boundary zone of the domain hd_corr_p_in REAL Reduction factor for the horizontal diffusion flux for pp 0 0 in the interior zone of the domain hd_corr_q REAL eliminated in 4 12 replaced by hd_corr_q_ in bd hd corr q in bd REAL Renamed in 4 26 to hd corr trcr in bd hd corr trcr bd REAL Reduction factor for the horizontal diffusion flux for 0 0 tracers in the boundary zone of the domain hd corr trcr in REAL Reduction factor for the horizontal diffusion flux for 0 0 tracers in the interio
53. WMO descriptor 0 33002 0 data ok 1 data suspect 3 missing info set for aircraft wind profiler RASS data 2 blacklisted gross error 4 not in valid height range surface level data height or height distance to orography too large upper air humidity above 300 hPa level 5 bad reporting practice SYNOP pressure bad reporting practice uv aircraft height not measured derived from p using std atmosphere upper air height without temperature obs buoy wind zero wind speed radar VAD wind small wind speed aircraft wind bad roll angle quality dew point temperature not active mixing ratio temperature or pressure not active relative humidity temperature not active if needed generally sensor not at appropriate height or measurement duration not appropriate 6 gross error for multi level temperature or wind found in special check temperature lapse rate check horiz wind wind speed shear or directional shear check rel humidity gt 96 and bias corrected to saturation upper air pressure derived from reported height using model atmosphere bit 28 of main flag word flag indicating observation level is below surface Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 176 bit numbers for observation status flag entry 7 and QC flag entry 8 0 horizontal wind
54. YUCAUTN is created in the operational suite or even in an experiment by the experimentation system NUMEX If an unknown observation type is found the reason has to be investigated The event may indicate that there is a problem with the observation data base or with a pre processing step before the data are read by the COSMO model If an array size is too small the value s of some NAMELIST parameter s have to be increased appropriately this requires testing that there is enough memory available on the processors of the computer File YUCAUTN also provides recommendations on how much to increase the value of which NAMELIST parameter Even though the recommended increase is often sufficient to obtain appropriate array sizes everywhere this is not always the case particularly in situations where several arrays are too small for several reasons and the surplus data add to each other Note that similar types of messages and recommendations are also given in file YUSTATS see Section 8 2 6 Figure 8 10 shows excerpts of an example file Each message contains the model timestep after t 27 at which the message was issued While in a real YUCAUTN file the messages are ordered according to the timestep they have been ordered thematically in the example shown here for convenience Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 165 CAUTION t 0 1923 LOCAL S
55. a rather complex task both for the experienced and even more for the non experienced user This User s Guide serves in a first instance as a complete reference for all the different NAMELIST groups and variables with which the execution of the model can be controlled It also includes a description on how to install the package and gives additional necessary information e g on the Grib format used for I O Knowing the meaning of all NAMELIST variables normally is not enough to find the way through the possible configurations of the model Therefore a description would be desirable that explains how the variables can be put together to give a meaningful setup or which variable settings contradict each other or simply are not possible We apologize that such a description is not yet available but it will be developed in the future It will explain how the different components of the model see Fig 2 1 can be selected and which configurations are possible Up to then Part VII of the model documentation is organized as follows First an overview on the model formulation and the data assimilation is given In Section 4 the installation of the package is explained The necessary input files of the model are listed in Section 6 and Section 5 1 gives a short description of the GRIB code used for input and output of the meteorological fields Section 7 then is the complete reference for all NAMELIST variables Sections 8 1 and 8 2 finally describe
56. and file names radar in CHAR directory in which the radar data files reside PR 100 incl the blacklist amp beam height map files blacklist CHAR file name of blacklist for radar data blacklist_dx grib1 100 height_file CHAR file name of radar beam height maps height_dx gribl1 100 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 10 EPSCTL Controlling the Ensemble Prediction Mode 136 7 10 EPSCTL Controlling the Ensemble Prediction Mode labelepsmode Basic control parameters Name Type Definition Purpose Comments Default iepsmem INT ID of the member in the ensemble ID gt 0 1 As a local solution for Grib 1 iepsmem is coded in product definition section 52 ipds 52 iepstot INT Total number of ensemble members gt 0 ed As a local solution for Grib 1 iepstot is coded in product definition section 51 ipds 51 iepstyp INT ID of the ensemble generation type ID gt 0 1 As a local solution for Grib 1 iepstyp is coded in product definition section 50 ipds 50 Parameters for changing input fields Name Type Definition Purpose Comments Default fac_plcov REAL Modification factor for PLCOV 1 0 rmin plcov REAL Lower limit for values of PLCOV 0 0 rmax plcov REAL Upper limit for values of PLCOV 1 0 fac lai REAL Modification factor for LAI 1 0 rmin lai REAL Lower limit for val
57. and octets 11 and 12 indicate the value of this level Table 5 5 shows the code figures used for the COSMO Model For reserved values or if not defined octets 11 and 12 shall contain zero All 3 D variables except the vertical velocity are defined on terrain following main levels In GRIB these main levels are coded as level type 110 hybrid layers between two adjacent hybrid levels which are the half levels in the COSMO Model i e the layer interfaces In this case octet 11 contains the level index of the upper half level and octet 12 contains the level index of the lower half level The vertical velocity and the height of the half levels are coded as level type 109 hybrid levels i e the model half levels In this case octet 11 contains zero and octet 12 contains the level index of the model half level Pressure levels ipds 8 100 and height levels ipds 8 105 are used when the interpolation from model to specified p or z surfaces is switched on for model output Table 5 3 GRIB tables for parameter element indicator number Version number of Comment GRIB table ipds 2 2 WMO table of indicator parameters 201 205 national tables of DWD for internal use Octets 13 17 contain the reference time of the data the start of a forecast the time for which an analysis is valid or the start of an averaging or accumulation period The year of Part VII User s Guide 4 28 Section 5 Data Formats for I O
58. and read only if the compile option DNUDGING is used for the production of the COSMO binary NUDGING contains the variables that control all the processes which require meteorological observations except for the use of satellite radiances in order to produce synthethic satellite images see NAMELIST group SATCTL Figure 7 2 shows an example for it as an excerpt of the script that is used to run the COSMO Model Note that the whole of the namelist group NUDGING has no effect at all as long as the NAMELIST variable luseobs of NAMELIST group RUNCTL is set to FALSE This means that if a free model forecast is to be performed solely it is sufficient to set luseobs FALSE and the group NUDGING does not have to be concerned with At present there a four main processes requiring observations data assimilation based on the nudging technique for atmospheric variables verification which means here simply the writing of a NetCDF feedobs file see first comments in Section 8 and or a verification observation file VOF see Section 8 2 7 for the purpose of observation input for the LETKF analysis scheme or input for verification tools latent heat nudging LHN for the assimilation of radar derived surface precipitation rates production of 2 dimensional 2D surface level analyses based on synoptic observa tions these analyses can be used for validation purposes or as input for the variational
59. array i e at each local node on a distributed memory computing platform only the data related to the respective model sub domain are stored The size of the ODR arrays depends on the NAMELIST parameters maxsgo for single level reports maxmlo for multi level reports and maxgpo for GPS ZTD or IWV reports The second message in Figure 8 10 simply means that the ODR size for multi level reports is too small in order to accommodate all multi level aircraft re ports which could have been created from the input single level reports In this case the aircraft reports do not have to be omitted unless maxsgo is also too small but they are assimilated as single level reports rather than multi level reports as preferred The second group of 3 messages relates to the arrays containing the observation increments from the total model domain used at a certain timestep There are four types of sets of increments in the scheme and hence four types of families of arrays in the code multi level upper air single level surface level and surface pressure increments The length of the re spective arrays correspond to the number of stations in case of temporal linear interpolation or reports otherwise with active observation increments or reports to be written to the YUVERIF or NetCDF feedobs files These lengths are given by the NAMELIST parameters maxmlo maxgpo 2 1 for multi level maxuso for upper air single level maxsgo f
60. at the observation point the value of the spread ing parameter usually height at the current model level the maximum spatial weight at any grid point on the previous for surface level data only and current model level within the local sub domain and the 2 observation increments The entry vCS fac in the lines airmy denote the reduction factor of the scale of the vertical weight function below and above a single level aircraft observation The lines zuwi or ztwi zqwi provide the station identity the sum of weighted zonal wind resp temperature or humidity increments the sum of weights and the sum of squared weights from the previously processed and the present observations at the lowest but fifth model level at grid point ionl jon1 The same applies to lines omu see Figure 8 21 resp omt omq and lines ntstep ztwi without squared weights except that the sum is over all observations The lines nudge horiz wind or Tqnudge deliver the local coordinates of grid point ion12 jon12 the model level and the analysis increments of the wind components resp of temperature of specific humidity from nudging temperature data and of humidity from nudging humidity data at that point The lines itera q11 indicate the iteration in solving the Poisson equa tion to derive geostophic surface pressure increments from wind increments based on 10 m wind data From these pressure increments geostr
61. average u momentum flux surface average v momentum flux surface average sensible heat flux surface average latent heat flux surface surface water drainage sum over forecast soil water drainage sum over forecast maximal windspeed in 10m geopotential height vertical velocity p dot in pressure coordinate system relative humidity total precipitation Part VII User s Guide 4 28 Section 8 Model Output REFERENCES 193 References Davies H and R Turner 1977 Updating prediction models by dynamical relaxation An examination of the technique Quart J R Met Soc 103 225 245 Dudhia J 1993 A nonhydrostatic version of the Penn State NCAR mesoscale model Validation tests and simulation of an Atlantic cyclone and cold front Mon Wea Hev 121 1493 1513 Hess R 2001 Assimilation of screen level observations by variational soil moisture analysis Meteorol Atmos Phys 77 155 166 Jacobsen I and E Heise 1982 A new economic method for the computation of the surface temperature in numerical models Contr Atmos Phys 55 128 141 Kain J S and J M Fritsch 1993 Convective Parameterization for Mesoscale Mod els The Kain Fritsch Scheme In The Representation of Cumulus Convection in Nu merical Models Meteorological Monographs No 46 American Meteorol Soc 165 170 Kessler E 1969 On the distribution and continuity of water substance in atmospheric circulation Mete
62. beginning of each section Bit positions within octets are referred to as bit 1 to 8 where bit 1 is the most significant bit and bit 8 is the least significant bit Thus an octet with only bit 8 set to 1 would have the integer value 1 Part VII User s Guide 4 28 Section 5 Data Formats for I O 5 1 The GRIB Binary Data Format 36 Table 5 1 Form of GRIB code Section number Name Contents 0 Indicator Section GRIB length of record GRIB edition number 1 Product Definition Section Length of section identification of the coded analysis forecast field 2 Grid Description Section Length of section optional grid geometry as necessary 3 Bit map Section Length of section the bit per optional grid point placed in suitable sequence Binary Data Section Length of section data values 5 End Section TTTT 5 1 2 Indicator and End Section The Indicator Section has a fixed length of 8 octets The first four octets shall always be character coded as GRIB according to the CCITT International Alphabet No 5 The remainder of the section shall contain the length of the entire GRIB record including the Indicator Section expressed in binary form over the left most 3 octets i e 24 bits in octet 5 7 followed by the GRIB edition number currently 1 in binary in the remaining octet 8 The End Section has a fixed length of 4 octets These octets are character coded as 7777 according to the Intern
63. budget surface mean value over forecast APAB S m average photosynthetic active radiation surface mean value over fcast ASOB T m average solar radiation budget model top mean value over forecast ATHB_T m average thermal radiation budget model top mean value over fore cast VIO3 vertical integrated ozone content HMO3 ozone maximum ZO m roughness length g FR_LAND m part of land in the grid cell Part VII User s Guide 4 28 Section 8 Model Output 8 4 Output of Forecast Fields 192 SOILTYP PLCOV LAI ROOTDP ALB RAD RLAT RLON FC RAIN GSP SNOW GSP RAIN CON SNOW CON U_10M V_10M T_2M TD_2M TMIN_2M TMAX_2M PRR_CON PRS_CON PRR_GSP PRS_GSP AUMFL_S AVMFL_S ASHFL_S ALHFL_S RUNOFF_S RUNOFF_G VMAX_10M FI OMEGA RELHUM TOT_PREC B B BBBBBBBBBRB TT BBBBBE SB B g N soil type of the land degree of plant covering leaf area index root depth albedo of the ground geographical latitude geographical longitude Coriolisparameter rain precipitation sum over forecast snow precipitation sum over forecast rain convective sum over forecast snow convective sum over forecast zonal wind in 10m meridional wind in 10m temperature in 2m dew point in 2m minimum temperature in 2m maximum temperature in 2m rate of precipitation convective rain rate of precipitation convective snow rate of precipitation scale rain rate of precipitation scale snow
64. change 0 005 0 031 0 072 Q nudging p incr T corr no T nudge 0 007 0 042 0 080 0 nudging p incr T nudge no evapo 1 694 7 566 1 024 0 nudging p incr T nudge complete 0 025 0 112 0 030 0 nudging T incr T corr no T nudge 0 000 0 000 0 001 0 nudging T incr T nudge no evapo 0 023 0 006 0 166 0 nudging T incr T nudge complete 0 000 0 000 0 002 0 nudging q incr RH nudge no evapo 0 000 0 000 sadat p nudging q incr T RH nudge complete 0 000 0 000 0 019 0 nudging u incr geostrophic 0 000 0 000 0 000 0 nudging v incr geostrophic 0 000 0 000 0 000 0 181 126 ztwi omyt 0 00 0 00000 000 0 0000 0 4132 002 002 r V gr pt 2 75 u gr pt 0 59 1 0 002 67 100053 101361 I12 0 131 0 133 030 0 124 0 133 020 0 113 0 133 085 0 128 0 133 002 0 002 0 002 002 0 002 0 002 001 0 001 0 001 000 0 000 6 580 000 0 000 0 206 002 0 001 0 000 000 0 000 0 000 Figure 8 21 Example file YUPRINT forth part related to weighted increments and balancing sure levels at grid point ionl jonl Mostly self explantory information on the geostrophic wind correction is given by the lines containing geost incr not shown Last but perhaps most essentially nearly self explanatory lists of the coordinates of the grid points at which the maximum absolute values for the analysis increments occur on each model level are provided together with the values of the increments themselves Figu
65. considered this parameter has to be changed and the program has to be recompiled With nOgp and nincgp the first output and the interval of outputs in time steps can be controlled alternatively hOgp and hincgp for specifying these values in hours Figure 8 1 shows an example of a file M stationname with the short form of the grid point output This short form contains the following information for every grid point A header specifying the initial date of the forecast and the and j indices of the model domain the model orography m the fraction of land 96 within the grid cell the geo graphical latitude i and the geographical longitude A For every time step the following quantities are listed in one line HH date PS surface pressure reduced to sea level DF10M wind direction and speed kn where 1 m s 2 kn at 10m above surface DF500M wind direction and speed kn at 500m above surface DF850 wind direction and speed kn at 850 hPa DF700 wind direction and speed kn at 700 hPa DF500 wind direction and speed kn at 500 hPa TG surface temperature C T2M temperature C 2m above surface TD2M dew point C 2m above surface T30M temperature C 30m above surface T850 temperature C at 850 hPa T700 temperature C at 700 hPa T500 temperature C at 500 hPa HML cloud cover high medium low range 0 8 ground fog range 0 8 HBAS base height of convective cloud abov
66. coordinate Name Type Definition Purpose Comments Default ydirini CHAR Directory path of the initial data file interpolated data ES from a driving model or analysis from a continuous COSMO Model data assimilation lchkini LOG Checking the initial data for min max values FALSE yvarini CHAR has been eliminated in Version 4 12 lana qi LOG If TRUE use the cloud ice field contained in the initial FALSE data file as initial condition for cloud ice otherwise initial conditions for cloud ice are set in the model lana qr qs LOG If TRUE values for rain and snow are read from the initial FALSE conditions Otherwise values are set in the model lana gg LOG If TRUE values for the graupel scheme are provided in the FALSE initial data lana rho snow LOG If TRUE values for the density of snow are provided in the FALSE initial data nlgw ini INT Number of prognostic soil water levels in initial data This 2 parameter is only used for the old soil model ydirhhl CHAR directory that contains the HHL file which is necessary for no reading writing GRIB2 general vertical coordinate ynamhhl CHAR name of the file that contains the HHL file which is neces Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 13 GRIBIN Controlling the Grib Input 142 Control parameters for boundary data Name Type Definition Purpose
67. fixed name blklsttmp containing a blacklist and a whitelist is read by COSMO The existence of this file is mandatory if itype_obfile 2 and any of the namelist parameters Inudge lverif or lsurfa is true All these input files must reside in the directory given by the namelist parameter ycdfdir e Content of the NetCDF observation input files The NetCDF observation input files cdfin are usually created by direct conversion of BUFR files using the bufrx2netcdf program This means that they contain the same variables as the input BUFR files Hence if possible all BUFR reports in a file should use exactly the same template in order to allow for a complete conversion into NetCDF bufrx2netcdf will only convert those observation reports in a BUFR file which have exactly the same template as the first report converted and it will neglect the other reports Therefore if the BUFR reports do not have an identical template several calls of bufrx2netcdf using the x option to skip the previous reports with Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files 54 other templates to a BUFR file are necessary to create several NetCDF files for the same observation file type A BUFR file without any observations will result in a file cdfin with zero length Such a file must be deleted before starting running COSMO COSMO expects the NetCDF
68. for the COSMO Model 6 4 Observation Input Files 55 The ifxy attribute for any variable in the NetCDF file is equal to the BUFR descriptor for that variable These BUFR descriptors are described in http www wmo int pages prog www W MOCodes TDCFtables htmlZz TDCFtables link BUFR CREX Table B Classification of elements and on the same page the link Code and Flag Tables associated with BUFR CREX Table B provides the details of the code tables Observational reports are rejected unless they contain appropriate non missing values for observation time year month day hour minute location latitude longitude and for some observation types station altitude and some form of station identity Note that for some observation file types the BUFR and the NetCDF files produced at DWD contain an additional variable for each or most of the existing variables of section 3 This variable contains a quality flag related to the value of the original variable The variable name of this quality flag variable is equal to the name of the original variable plus a suffix Q However all the quality flag variables of this type are obsolete i e they are never needed never used or read by COSMO they always contain missing values only and therefore they are never described hereafter In the following sub sections of this section the sequences of variables templates related to BUFR Section 3 are detailed for the different ob
69. for the period of validity of minimum and max 6 0 imum values of the 2m temperature TMIN 2M and TMAX 2M i e every hincmxt hours the corresponding arrays are reset to default values hincmxu REAL Interval in hours for the period of validity of minimum and 1 0 maximum values of the 10 m wind gusts VMAX 10M i e ev ery hincmxu hours the corresponding array is reset to default values Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 2 RUNCTL Parameters for the Model Run 84 Control parameters for debug purposes Additional informations on the status of the model run can be obtained by increasing the debug level However this reduces the efficiency of the model run and blow up the standard ascii output enormously Therefore these options are recommended to be used by model developers and or in case of problems with the execution of a model run Name Type Definition Purpose Comments Default idbg level INT Selects the verbosity of ASCII output during a model run 2 The higher the value the more debug output is written to standard output To switch on off debug output for special components the additional flags 1debug xxx have to be set to TRUE FALSE ldump ascii LOG Tosave the ASCII Files written during the model run to disk TRUE in every time step This may cause a performance slow down on some machines and can be switched off Then the files are only
70. from several coarse grid models can be processed GME IFS COSMO Model Option for periodic boundary conditions Top Boundary Conditions Options for rigid lid condition and Rayleigh damping layer Initialization Digital filter initialization of unbalanced initial states Lynch et al 1997 with options for adiabatic and diabatic initialization Physical Parameterizations Subgrid Scale Turbulence Prognostic turbulent kinetic energy closure at level 2 5 including effects from subgrid scale condensation and from thermal circulations Option for a diagnostic second order K closure of hierarchy level 2 for vertical turbulent fluxes Preliminary option for calculation of horizontal turbulent diffusion in terrain following coordinates 3D Turbulence Surface Layer Parameterization A Surface layer scheme based on turbulent kinetic energy including a laminar turbulent roughness layer Option for a stability dependent drag law formulation of momentum heat and moisture fluxes according to similarity theory Louis 1979 Grid Scale Clouds and Precipitation Cloud water condensation and evaporation by sat uration adjustment Precipitation formation by a bulk microphysics parameterization including water vapour cloud water cloud ice rain and snow with 3D transport for the precipitating phases Option for a new bulk scheme including graupel Option for a simpler column equilib rium scheme Subgrid Scale Clouds Subgrid scal
71. from the same station which are less than tipolmx hours apart from each other gt at most 2 reports per station used at each timestep linear interpolation in time of pairs of surface level reports less than tipmxsu hours apart from each other max time span in hours between 2 upper air reports to allow for temporal linear interpolation max time span in hours between 2 surface level reports to allow for temporal linear interpolation TRUE TRUE wtukrsa wtukrse wtukara wtukare wtuksua wtuksue REAL REAL REAL REAL REAL REAL temporal weights for single reports temporal radii of influence relative to observation time tops if reports from the same station are assimilated independently from each other using saw tooth shaped temporal weights radius towards the past for radiosonde data radius towards the future for radiosonde data radius towards the past for aircraft data radius towards the future for aircraft data radius towards the past for surface level data radius towards the future for surface level data c c oc co o Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 121 Spatial weights basic mode of spreading Name Type Definition Purpose Comments Default msprpar INT switch specifying the surfaces along which upper air 1 observation increments are primarily
72. grid mapping rotated pole U coordinates slonu slatu float V time level srlat rlon V standard name grid northward wind V long name V component of wind V units m s 1 V grid mapping rotated pole V coordinates slonv slatv float TOT PREC time rlat rlon TOT PREC standard name precipitation amount TOT PREC long name total precipitation amount TOT PREC units kg m 2 TOT PREC grid mapping rotated pole TOT PREC coordinates lon lat TOT PREC cell methods time sum float ASOB_S time rlat rlon ASOB S standard name surface net downward shortwave flux ASOB S long name surface net downward shortwave radiation ASOB S units W m 2 ASOB S grid mapping rotated pole ASOB S coordinates lon lat ASOB S cell methods time mean float VMAX_10M time height 10m rlat rlon VMAX 10M standard name wind speed Part VII User s Guide 4 28 Section 5 Data Formats for I O 5 2 The NetCDF Data Format 45 VMAX 10M long name maximum 10m wind speed VMAX 10M units m s 1 VMAX 10M grid mapping rotated pole VMAX 10M coordinates lon lat VMAX 10M cell methods time maximum global attributes title Simulation in the EU project ENSEMBLES institution GKSS source CCLM4 project id ENSEMBLES experiment id ERA40 50km realization 1 Conventions CF 1 0 conventionsURL http www
73. gt 0 necessary for transformation from one ro tated grid to another rotated grid dlat REAL Meridional rotated lat direction grid spacing in degrees 0 008 dlon REAL Zonal rotated lon direction grid spacing in degrees 0 008 startlat_tot REAL Latitude of the lower left grid point of the total domain 7 972 in degrees north gt 0 rotated coordinates startlon_tot REAL Longitude of the lower left grid point of the total domain 1 252 in degrees east gt 0 rotated coordinates ie tot INT Number of gridpoints of the total domain in west east di 51 rection of the rotated coordinates je tot INT Number of gridpoints of the total domain in south north 51 direction of the rotated coordinates ke tot INT Number of gridpoints of the total domain in vertical direction 20 i e number of layers or main levels there are ke_tot 1 half levels The specifications for the model domain and the grid size are compared to the values from the headers of the data files the Grid Description Section GDS of the Grib files or the headers of NetCDF files for the initial and boundary fields If they do not correspond the program will print an error message and abort Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 2 RUNCTL Parameters for the Model Run 78 7 2 RUNCTL Parameters for the Model Run The namelist group RUNCTL contains the paramet
74. have Vh Apo k Po k 1 2 Po r 1 2 gt 3 10 po k 5 00 rsaja Jaya for y y and the base state pressure po on model main levels Additionally the height of model half levels z 41 2 resulting from the coordinate transformation is stored as a 3 D array The Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 3 Horizontal and Vertical Grid Structure 15 iS Waz 770 1 2 model top u T Pos o Y k 1 MEN o A LEES k 1 2 E A DE k 1 2 4 u T Pos Po Y L kzNr gea A w z TET Es N 1 2 surface iz i i 1 2 Figure 3 2 Vertical staggering of variables and metric terms in a grid box column with N layers Dashed lines are the model half levels separating the main levels full lines base state density on main levels then results from the discretized hydrostatic relation ven A 9 Zk 1 2 k4 1 2 and the main level base state temperature results from the equation of state Fig 3 2 illus trates the vertical staggering of model variables as well as base state variables and metric terms used in the discretization In order to implement boundary conditions and to apply the domain decomposition strategy for code parallelization in a convenient way the horizontal extent of the computational domain is chosen to be smaller than the total domain size The lateral physical boundaries are positioned with a spatial offs
75. hpa h 4 2 K Pmain T OV QC QI REL HUM CLC CLC CON U V SPEED HML hPa Grd C g kg mg kg g 5 m s m 1 34 19 76 61 0 004 0 000 0 000 15 01 0 00 0 00 42 73 1 29 42 75 22184 5 2 51 27 78 10 0 004 0 000 0 000 25 42 0 00 0 00 30 44 13 05 33 12 19807 4 3 69 20 74 72 0 004 0 000 0 000 20 32 0 00 0 00 22 42 14 96 26 95 18062 0 4 88 30 73 62 0 004 0 000 0 000 21 87 0 00 0 00 13 30 14 65 19 79 16634 0 37 1010 77 3 46 1 696 0 000 0 000 35 13 0 00 0 00 14173 1437 2 21 226 7 38 1016 66 3 24 1 722 0 000 0 000 36 43 0 00 0 00 1433 0 99 1 66 179 6 39 1020 97 2 95 1 740 0 000 0 000 37 73 0 00 0 00 1 06 0 76 14 30 145 4 40 1024 03 2 71 1 759 0 000 0 000 38 93 0 00 0 00 0 85 0 60 1 04 121 2 K Phalf W TKVM TKVH HHL hPa cm s m 2 s m 1 19 91 0 000 23588 5 2 42 73 0 741 1 000 1 000 20780 5 3 60 24 1 3599 1 000 1 000 18834 3 4 78 75 1 979 1 000 1 000 17289 8 38 1013 71 0 338 1 000 1 000 199 7 39 1018 82 0 269 1 000 1 000 159 6 40 1022 50 0 186 1 000 1 000 131 1 41 1025 29 0 168 111 2 Surface variables TCM 0 00000 m s TCH 0 00000 m Z0 0 75001 w m2 SHFL 0 007 N m2 UMFL 0 000 LHFL 0 002 N m2 VMFL 0 001 g kg OV S 2 125 kg m2 RUNOFF_S 0 000 RUNOFF_G 0 000 Plants LAI 1 060 Ozone VIOS3 0 078 PLCOV 0 670 HMO3 5477 051 ROOTDP 0 120 Soil temperatures T_SNOW 0 460 Soil moistures Snow W SNOW 0 000 dgr C T S 0 460 mm H20 WI B 0 000 TG 0 460 FRESHSNW 1 000 kg m3 RHO_SNOW 50 000 m H_SNOW 0 000 T_SO 0 0 460 T_SO 1
76. in the source code of the COSMO Model and also in the INT2LM To satisfy the calls from mpe io to the data base system an additional file dummy db f90 is provided Part VII User s Guide 4 28 Section 4 Installation of the COSMO Model 4 2 Working with the VCS 31 4 1 5 libRTTOV7 a Since Version 3 7 the COSMO Model contains an interface to the RTTOV7 library Radia tive Transfer Model This interface has been developed at the DLR Institute for Atmospheric Physics in Oberpfaffenhofen Together with the RT TOV7 library it is possible to compute synthetic satellite images brightness temperatures and radiances derived from model vari ables for Meteosat5 7 and Meteosat Second Generation Since Version 4 18 also the use of newer RTTOV libraries namely RTTOV9 is possible The RTTOV model has been developed by UKMO et al in the framework of the ESA NWP SAF To use any version of the RT TOV model a license is necessary For getting this license please contact nwpsaf metoffice gov uk Usage of the RT TOV libraries can be controlled by conditional compilation and setting the macros RTTOV7 and or RTTOV9 Note that RTTOV7 has been modified at DWD to be used in parallel programs For the usage or RTTOV9 and also RTTOV10 a special interface mo rttov ifc f90 is necessary which can also be obtained from DWD If the license and hence the RTTOV libraries is not available the corresponding macros must not be set The computatio
77. increment dT mean corresponding to the thickness increment within the given vertical range Finally the entries V mult T mult q mult and z mult indicate that the corre sponding observations are set passive in the given pressure range due to the multi level check While rejection of temperature implies rejection of humidity rejection of height does not imply rejection of temperature here Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 164 8 2 5 YUCAUTN Nudging Warning Messages on Insufficient Array Sizes The size of arrays which are used to store non gridded observational information are a func tion of several NAMELIST parameters This means that the values of these NAMELIST param eters determine the size of these arrays If the values are too small there are various places in the program where such arrays may fail to accommodate all the available data In such a case the program will not crash or stop but it will simply omit the surplus data and issue warning messages which always contain the label CAUTION This allows to grep for it yet there may be also other types of messages containing the word CAUTION Messages on short array sizes related to individual observational reports are written to the files YUREJCT see section 8 2 3 YUPRINT section 8 2 8 and or to the standard output The messages on YUPRINT are writt
78. is described in Section 6 1 The namelist groups the variables their meanings and possible values are described in Chapter 7 GRIB Code or NetCDF files for the initial and boundary values These files are de scribed in Section 6 2 and in Section 6 3 respectively NetCDF observation input files or alternatively an AOF file which contain the ob servational information for data assimilation nudging and for producing a NetCDF feedobs file The purpose of the feedobs file see first remarks in Section 8 is to serve as input to a LETKF analysis scheme or to verification tools The NetCDF observation input files including a blacklist file are described in Section 6 4 6 1 File for Namelist Input The COSMO Model uses NAMELIST input to specify runtime parameters The parameters are splitted into several groups which are distributed to the components Table 6 1 lists the components the groups and the corresponding INPUT_ files The last group of component Input Output GRIBOUT can occur several times Every group can determine a different list of variables for output and also different output steps The program provides default values for all parameters To change a default value an appro priate NAMELIST statement has to appear in the corresponding ASCII file INPUT_ The form of a NAMELIST statement depends on the specific platform you are using but is always similar to the following refer to the Language Reference Manual of
79. it has been adapted also to the coarser resolutions Now all applications can be run using the Runge Kutta scheme and the Leapfrog scheme will no longer be supported Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 3 DYNCTL Parameters for the Adiabatic Model 86 Name Type Definition Purpose Comments Default 12tls LOG The default time integration scheme is a 2 time level Runge TRUE Kutta scheme with time split treatment of acoustic and grav ity waves 12tls TRUE Alternatively a 3 time level Leapfrog time split scheme may be used 12t1s FALSE lsemi imp LOG Instead of the standard Leapfrog time split scheme a Leapfrog FALSE scheme with semi implicit treatment of acoustic and gravity waves may be used This option works only with 12tls FALSE Caution the semi implicit scheme is not yet fully evaluated lcond LOG Toinclude TRUE or exclude FALSE cloud water conden TRUE sation and evaporation during the forecast lcori LOG Run with Coriolis force TRUE has been moved from group RUNCTL to here in Version 4 8 lmetr LOG Run with metric terms TRUE If FALSE the spherical geometry of the grid is neglected and the Coriolis force is assumed to be constant on the model domain f plane approximation has been moved from group RUNCTL to here in Version 4 8 lcori deep LOG To take coriolis terms cos q into account
80. missing value 2 Flight level is defined relative to the ICAO standard sea level pressure and is readily converted to static air pressure using standard formulae i e using the ICAO standard atmosphere Hence flight level is not the geometrical height In the above file types the variables MIAA and NHHH also denote the same type of flight level as variable NFLEV and are therefore not geometrical height either Once converted the original resolution either 100ft or 10ft in the BUFR report is lost hence it is desirable to disseminate the element in the received form 3 This phase of flight table is expanded to indicate wind quality from roll angle or roll and pitch combined and also to indicate the method of ascent and descent observation interval selection either by time or pressure increments 4 For temperature it is in fact only required that at least one of the variables MTDBT or MTN exists in any of the NetCDF file s but it does not really matter which one The unified ACARS format unif produced at DWD file type cdfin_acars contains ad ditionally the following variables which do not occur in either of the ACARS formats _uk and _us obtained via GTS except for variable MPTT but which are part of the AMDAR template proposed by WMO unif descript type mnemonics meaning 1 02 000 Delayed replication of 2 descriptors 0 31 001 int
81. more Most often this relates to multi level aircraft reports which may have been created only after the observation time This can happen when new observations with the same aircraft identifier station ID are read Then all multi level reports are deleted first and multi level reports are derived again based on all the single level reports from the same aircraft even if some of those reports are older than the current model time Normally the aircraft observations missing on the feedobs file s should already have been written as a single level report report or as part of another multi level report Lines of the type considered can also relate to other observation types Typically this occurs for redundant reports from stations which issue very frequent reports of observation types where temporal linear interpolation is applied between active reports The lines starting with nudge render the last timestep of the model run the index of the lowest middle or top model level and temperature and specific humidity Tq resp pressure and zonal wind pu at grid point ionl jonl in unformatted form This often allows to diagnose tiny differences between runs done with slightly different model versions coarse grid size for Poisson solver 63 63 STEP 0 dt dt2 dtdeh within the nudging 40 80 0 0111 STEP 0 dt dt2 dtdeh outside the nudging 20 40 0 0056 Surface analysis GRIB output file lansfc opened t2m GR
82. msprpsu 0 gnudg 0 0006 0 0012 0 0006 0 0006 gnudgsu 0 0006 0 0012 0 0000 0 0006 gnudgar 0 0006 0 0000 0 0006 0 0000 gnudggp 0 0003 vcorls 333 333 04 04 vcutof 0a 1oy DAT O y he z vcorlsu 013 013 002 00001 vcutosu 0 75 0 75 4 0 0 001 vcsnisu 2500 2500 9 ps p rhvfac 1 0 0 0 0 83 0 83 rhinfl 0 70 0 Ow rhtface 1 3 1 43 1 3 y 1 3 rhiflsu TUN TU DOG ny 7O rhtfsu 1 0 1 43 1 0 1 0 fnondiv 0 8 cnondiv 0 1 r QUuEOT r 34 5 e Bud y 3 9 4 345 4 tnondiv 1 1 cutofsu 2 0 3 5 y 2 0 2 0 topobs 849 1099 799 699 botmod 1099 1099 1099 899 lscadj TRUE TRUE TRUE FALSE dtqc 720 qcvf 6 0 y ig 210 0 1 0 qcc Ox SOO Dias BP qccsu IZ SOO 12 ay a lp mqcorr92 2 lsynop TRUE laircf TRUE ldribu TRUE ltemp TRUE lpilot TRUE lsatem FALSE lscatt TRUE lcd122 TRUE lcd123 TRUE lgps TRUE igpscen 30 23 26 24 29 33 34 37 32 0 21 35 25 lcd132 TRUE lcd133 TRUE lcd137 TRUE maxmlo 800 maxsgo 6000 maxuso 5000 maxgpo 7000 nolbc 5 altopsu 100 5000 5000 5000 thairh 20 exnlat 90 exslat 90 exwlon 180 exelon 180 lsurfa TRUE lt2m TRUE ht2a 0 ht2i 1 lrh2m TRUE hh2a 0 hh2i 1 1ff10m TRUE hffa 0 hffi 1 lprecp FALSE hprc 0 raintp 12 lpraof FALSE lprodr TRU
83. neglected if f Az gt doromx qcfpst REAL maximum enhancement factor of the quality weight for 1 5 surface pressure observations due to observed 3 hourly surface pressure tendency 0 ps qcfpst is applied if 3 p gt 25 hPa the enhance ment decreases linearly with decreasing p to 1 for 0 ps 3hPa Use of observations and reports Name Type Definition Purpose Comments Default altopsu 4 REAL surface level observations above height altopsu in m 100 are neglected 5000 if altopsu 0 all surface level observations 5000 assigned to land grid points are neglected 5000 thairh REAL for contructing multi level reports piecewise profiles 20 from single level aircraft reports maximum horizontal distance in km between the resulting multi level report and the original location of any single level report included in this multi level report lgpsbias LOG seasonal daytime dependent bias correction applied to GPS FALSE derived IWV integrated water vapour data mqcorr92 INT switch for bias correction for Vaisala RS92 radiosonde humidity 0 0 no correction for humidity 1 correction of solar radiation bias only 2 correction of total bias incl nighttime bias nolbc INT number of grid rows at the lateral boundaries of the COSMO 5 Model domain where all reports are neglected Part VII User s Guide 4 28 Section 7
84. of NetCDF can only be avoided if also the Nudging is switched off If Nudging the data assimilation shall be used a NetCDF library has to be available because the observation processing is done via NetCDF 4 1 3 libmisc a For special applications in the data assimilation the nudging a library libmisc a is used For external users these routines have been provided in the grib library up to version DWD libgrib1_061107 from 6th November 2007 Newer versions of the grib library do not contain these routines anymore and the misc library has to be used instead The use of 1ibmisc a can be controlled together with the assimilation component with the macro NUDGING which turns on off the whole assimilation component If DNUDGING is specified for compilation the assimilation component is compiled and linked and the library libmisc a has to be available If it is not specified the assimilation component is not compiled and linked and libmisc a is not necessary then 4 1 4 libcsobank a libsupplement a The COSMO Model and INT2LM use a tool for parallel asynchronous I O from or to files or a data base system only for Grib The routines for that tool are grouped together in a module mpe_io f90 In the VCS of DWD mpe_io f90 is provided as an external module hence it is not in the source code of the model library mpe_io 90 uses the two libraries libcsobank a and libsupplement a For users outside DWD mpe io f90 has been included
85. of regular time boxes in h in which analysis increments are computed once and then used to update the model variables at all timesteps within the time box this length corresponds to NAMELIST variable tconbox in s AI box denotes the time box interval in h and mean the middle of the interval ie mean declares the time in h for which the temporal weights used to compute the analysis increments are exactly valid next AI denotes the time in h at which new analysis increments are to be computed for the next time and obs process the next time at which new observations must be read again from the AOF If the observations are read from NetCDF observation input files the obs process entry does not have any meaning and is simply equal to zero Next the path name of the BLACKLIST WHITELIST file is given This is followed by a list of NetCDF observation input files which do not exist due to missing data but which would be read if they existed Then for each of the existing NetCDF observation input files the number of reports are indicated and the time interval from which all reports have to be read currently This last type of lines is written again whenever new observations are read which is typically once every hour Analysis Increments AI held constant during time boxes of ca 0 067 hours hour AI box 0 000 0 056 mean 0 028 next AI 0 07 obs process 0 00 open and read BLA
86. of sensor above water surface for temperature opt 0 04 024 int MGGTP2 opt 0 04 024 irit MGGTP1 Start of time period in hours opt 0 04 024 int MGGTP3 i a opt 0 04 024 iit MGGTP2 End of time period in hours opt 012111 float MTXTXH Maximum temperature over period specified T A a d MOOD Start of time period in hours RA r o a uM End of time period in hours opt 012112 float MINTNH Minimum temperature over period specified Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files 59 use L S WMO descriptor type mnemonics meaning dg E ue a Wind data 007 032 float MHOSEN5 Height of sensor a ground deck opt 0 07 033 float MHAWAS2 Height of sensor a water surface F ats p Ae 0 02 002 int NIW Type of instrumentation for wind measurement opt 0 08 021 int MTISI Time significance 2 time averaged opt 004025 int NGGTP Time period 10 min or since a significant wind change need 0 11 001 int NDNDN Wind direction need 011002 float NFNFN Wind speed 008021 int MTISIO Time significance missing value 1 03 002 Replicate next 3 descript twice opt 0 04 025 int 2 NGGTPO Time period in minutes 011043 int 2 NMWGD Maximum wind gust direction opt 0 11 041 float 2 NFXGU Maximum wind gust speed
87. pressure 4 0 if nsprpar 2 In p is replaced by vpblsu 4 REAL Gaussian vertical radius of influence in potential 99 0 temperature differences between the observation 99 0 increment point Po and the target model level 99 0 at the horizontal observation location 99 0 gt this defines an additional Gaussian weight that makes the total vertical weight depend on the near surface stability even if nsprpar 1 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 123 Horizontal weights for upper air observations Name Type Definition Purpose Comments Default isotropic horizontal correlation function wy Ar s wey 1 Ar s e for scalar model variables rhinfl 4 REAL constant part se of the correlation scale s 0 0 s se fu Su Sy 70 0 se ferm 1 f 1 100 0 0 0 0 rhvfac 4 REAL multiplication factor f to the vertically varying part 1 0 Sy of the correlation scale s 0 0 s is given in km as a function of pressure p hPa 0 83 p 1000 850 700 500 400 300 250 200 150 50 0 53 s 70 80 90 100 100 110 115 120 125 125 sI s 70 80 90 100 100 100 100 110 120 120 rhtfac 4 REAL temporal factor f scaling the correlation scale s 1 3 at the beginning 1 43 and the end of the 1 3 nudging period for 1 3 an individual ob servation relative to s
88. rates into the convective scale model COSMO DE at DWD Quart J R Meteor Soc 134 1315 1326 Thomas S C Girard G Doms and U Sch ttler 2000 Semi implicit scheme for the DWD Lokal Modell Meteor Atmos Phys 75 105 125 Tiedtke M 1989 A comprehensive mass flux scheme for cumulus parameterization in large scale models Mon Wea Rev 117 1779 1799 Wicker L and W Skamarock 1998 A time splitting scheme for the elastic equa tions incorporating second order Runge Kutta time differencing Mon Wea Rev 126 1992 1999 Wicker L and W Skamarock 2002 Time splitting methods for elastic models using forward time schemes Mon Wea Rev 130 2088 2097 Part VII User s Guide 4 28 REFERENCES
89. reports see below The next three lines inform about an observation level being redundant within a multi level report Pressure level observation error and observed value for height resp quality flags for other variables are provided both from the active and the redundant level The following information is given for each active or passive single level report stars always indicate missing values Observation type and station identity status indicator active P passive M used as part of a multi level report zonal and meridional wind components u v m s temperature T K and relative humidity rh pressure level p hPa height z at observation level m and station height m Observation errors assigned to wind temperature humidity and height data station altitude minus height of model orography coordinates of the grid point to which the report is assigned longitude and latitude code type and observation type see Figure 8 12 reported observation time h relative to the initial model time AIREP EU5331 M 0 4 5 1 265 5 0 66 840 38 1550 2 5 1 0 Q lQ exxwee 313 297 12 27 48 50 244 2 2 51 AIREP EU5331 M 0 2 24 1 266 5 0 62 848 67 1470 2 5 1 0 0 10 313 297 12 25 48 47 244 2 2 51 AIREP EU5331 M 0 1 3 1 267 0 0 63 858 08 1380 2 5 1 0 0 1O0 313 297 12 23 48 45 244 2 2 51 AIREP EU5331 M 0 7 3 5 268 0 0 56 869 69 1270
90. saved at the end of the run lprintdeb all LOG In most cases the debug output is only written from one pro FALSE cessor with ID 0 With 1printdeb_a11 TRUE all proces sors will print the debug output ldebug dyn LOG To switch on off the debug output for the dynamics FALSE ldebug gsp LOG To switch on off the debug output for the microphysics FALSE ldebug rad LOG To switch on off the debug output for the radiation FALSE ldebug tur LOG To switch on off the debug output for the turbulence FALSE ldebug con LOG To switch on off the debug output for the convection FALSE ldebug soi LOG To switch on off the debug output for the soil processes FALSE ldebug io LOG To switch on off the debug output for the input and output FALSE ldebug mpe LOG To switch on off the debug output for the asynchronous I O FALSE module mpe io2 f90 introduced in Version 4 27 ldebug dia LOG To switch on off the debug output for the diagnostics FALSE ldebug ass LOG To switch on off the debug output for the assimilation FALSE ldebug lhn LOG Toswitch on off the debug output for thelatent heat nudging FALSE ldebug art LOG To switch on off the debug output for the ART module in FALSE troduced in Version 4 9 linit fields LOG To switch on off initialization of local variables introduced FALSE in Version 4 9 Part VII User s Guide 4 28 Section 7 Namelist Input for
91. temp meas 012102 float MTFNH Wet bulb temperature need 012103 float MTDNH Dew point temperature opt 0 13 003 int MUUU Relative humidity 96 dag ES ae Visibility data 0 07 032 float MHOSENO Height of sensor a ground vis opt 007033 float MHAWASO Height of sensor above water surface for visibility opt 0 20 001 float MVV Horizontal visibility 3 02 034 Precipitation past 24 hours 0 07 032 float MHOSENI Height of sensor a gr precip opt 013 023 float MRR24 Total precipitation past 24 hours 3 02 004 Cloud data opt 020010 int MN Cloud cover total 96 opt 0 08 002 int MVTSU Vertical significance opt 0 20 011 int MNH Cloud amount of low or middle clouds opt 020013 float NH Cloud base height above surface opt 020012 int MCC Cloud type low clouds Cz opt 0 20 012 int MCCO Cloud type middle clouds C3 opt 0 20 012 int MCC1 Cloud type high clouds Cy opt 0 31 001 int MDREP Delayed descriptor replication 3 02 005 Individual cloud layers of masses opt 0 08 002 int n MVTSUO Vertical significance 2 opt 0 20 011 int n MNHO Cloud amount 2 opt 0 20 012 int n MCC2 Cloud type C 2 opt 0 20 013 float n NHO Height of base of cloud 2 Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model
92. temperature humidity pressure surface level obs height upper air obs IWV vertically integrated water vapour ZTD zenith total delay of GPS signal D owwnnre level identity pressure code entry 10 level identity for non SYNOP observations pressure code for SYNOP observations bit meaning code meaning 0 maximum wind level 0 sea level 1 tropopause 1 station level pressure 2 TEMP D part 2 850hPa level geopotential 3 TEMP C part 3 700hPa level geopotential 4 TEMP B part 4 500gpm level pressure 5 TEMP A part 5 1000gpm level pressure 6 surface level 6 2000gpm level pressure 7 significant wind level 7 3000 gpm level pressure 8 significant temperature level 8 8000gpm level pressure several bits can be set at the same time 9 900hPa level geopotential 10 1000 hPa level geopotential 11 500hPa level geopotential combined cloud and weather group entry 14 bits meaning 0 3 type of high cloud VUB WMO Code table 0509 4 7 type of middle cloud VUB WMO Code table 0515 8 11 cloud base height VUB WMO Code table 1600 12 15 type of low cloud VUB WMO Code table 0513 16 19 cover of low cloud if gt 0 else of middle cloud VUB WMO Code table 2700 20 23 total cloud cover VUB WMO Code table 2700 24 30 present weather VUB WMO Code table 4677 combined weather and ground group entry 15 for SYNOP obs bits meaning 0 8 present weather WMO descr
93. the ASCII output of the COSMO Model and Section 8 4 provides information on the output model fields Part VII User s Guide 4 28 Section 2 Introduction Figure 2 1 Schematic view of the different COSMO Model components Dynamics 3 time level split explicit Leapfrog 2 time level split explicit Runge Kutta several vari ants 3 time level semi implicit Leapfrog Physics Assimilation Diagnostics Observation Near surface Grid Scale Clouds and Precipitation with or without 3D transport of precipitating particles processing Surface analysis Nudging of atmo spheric variables and surface pres sure weather parameters warm rain scheme category ice scheme 2 category ice scheme 3 category ice scheme Subgrid Scale Turbu lence Closure 1 D diag closure 1 D TKE based diagnostic closure 3 D TKE based prognostic closure Mean values Meteographs Volume and Area Integrals Latent Heat Nudging Synthetic Sa tellite Pictures Parameterization of Surface Fluxes Standard bulk trans fer scheme TKE based surface scheme Initialization Digital filtering I O Moist Convection Tiedtke mass flux Kain Fritsch scheme Shallow convection Subgrid sc orography Soil and Surface 2 layer soil Grib NetCDF Restart multi layer soil lake model sea ice scheme Part VII User s Guide 4 28
94. the different types of communications perform surely depends on the computer used Therefore it would require some testing to find the optimal settings Name Type Definition Purpose Comments Default nprocx INT Number of processors in the x direction of the grid 1 nprocy INT Number of processors in the y direction of the grid 1 nprocio INT Number of processors for GRIB I O 0 num_asynio_comm INT To choose the number of asynchronous I O communi 0 cators for NetCDF With several communicators it is possible to parallelize the output over the files to be written the GRIBOUT namelists num iope percomm INT To choose the number of asynchronous I O processes 0 per communicator for NetCDF I O With several pro cesses per communicator it is possible to do a parallel writing of single files This is only possible if the par allel NetCDF library is available and the code has been compiled with the preprocessor directive DPNETCDF nboundlines INT Number of boundary lines at each side of a subdomain 2 specifies the size of the halo around each subdomain where variables are updated by neighboring processors Along nboundlines at the lateral boundaries of the to tal domain the variables are set to boundary data from the driving host model A minimum of 2 boundlines is required ltime proc LOG Eliminated replaced by itype timing in Version 4 4 ltime mean LOG Eliminated replaced by itype timing i
95. the extrapolation of pressure obs observed values ke obs interpolated to level ke inc obs increments at ke Sta height surf pressure 10m horizontal wind 2m temperature 2m rel humid t id diff obs ke inc 10m obs at ke inc obs ke inc obs ke inc EU1312 0 227 9 k xk kk kk x x 28 2 13 7 3 2 4 1 221 4 2 4 kk kk e A x x 0 00 10946 s 53 3937 6 931 4 0 4 1 9 0 7 3 2 1 8 0 0 0 0 273 9 272 6 0 0 0 89 0 94 0 07 Vertical profiles of observation increments from TEMP PILOT AIRCRAFT station 06610 255 271 56 levels mbotlv mtoplv 6 2 8 4 1 12 u incr v incr T incr Od incr RH incr pressure height pot T 06610 Bee EAR AR IR A EK 1 578 0 00099 0 219 95800 535 278 36 EKKKKKKKKKKKAKK k 1 553 kk kk kk kk kx KKK 95659 547 278 36 kck ck ck ockckckck ck ck ckck ko ko kk 1 455 X k kk k kk k ke k ke k k k 95113 593 278 41 cock ck ck ck ck ckckckckckck kck ck k Li BQ GR KKK KKK KK KK KKK kx kx kk 94394 654 278 69 kckckckckckck ck ck ck ckck ko ko kk 1 156 kk kk kk k kx KKK 93483 32 279 08 aee cipe de siege cese Ke ER 1 009 0 00075 0 180 92700 799 279 32 1 055 20 039 0 916 0 00074 0 178 92500 817 279 39 1 027 0 045 Rs eoe apos 92367 828 279 43 station ZIMM BKG 260 271 24 levels mbotlv mtoplv 0 0 5 0 0 11 quality weight Od incr RH incr pressure height pot T ZIMM BKG_ 0 48531 0 00050 0 128 90514 990 279 89 0 57270 0 00048 0 127 89441 1085 280
96. the vertical mass flux at the convective cloud base in terms of the grid scale variables For shallow and penetrative convection it is assumed that this mass flux is proportional to the vertically integrated moisture convergence between the surface and the cloud base In case of Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 5 Physical Parameterizations 22 midlevel convection the mass flux is simply set proportional to the grid scale vertical velocity Given the mass flux at cloud base the vertical redistribution of heat moisture and momentum as well as the formation of precipitation is then calculated by integrating a simple stationary cloud model for both updrafts and downdrafts This finally allows to compute the convective tendencies i e the feedback of the subgrid vertical circulation onto the resolved flow The downdrafts are assumed to originate at the level of free sinking As an additional closure condition the downdraft mass flux in this level is set proportional to the updraft mass flux at cloud base via a coefficient yg which is a disposable parameter In the present version of the scheme yg is set to a constant value of 0 3 In subsaturated regions below cloud base the precipitation in the downdrafts may evaporate with a parameterized rate Depending on the temperature of the lowest model layer the precipitation is interpreted as convective snow or rain The parameterization sche
97. the whole forecast This is needed for climate simulations where the forecast is splitted into several runs on a computer and hstop only indicates the stop of a single run The format is as above i e either yyyymmddhh or gt yyyymmddhhmise If omitted ydate_end is not set and the end of the whole forecast is specified by hstop or nstop resp 2 2 lyear 360 LOG To use a climatological year with 360 days Eliminated replaced by itype calendar in Version 4 4 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 2 RUNCTL Parameters for the Model Run 79 Name Type Definition Purpose Comments Default itype_calendar INT To specify which calendar is used during the forecast 0 introduced in Version 4 4 0 gregorian calendar 1 every year has 360 days 2 every year has 365 days hstart REAL Number of the first hour relative to the initial time given 0 0 by ydate_ini For a regular forecast hstart 0 0 If hstart gt 0 0 the model is in restart mode and it requires restart data for Leapfrog from 2 consecutive time levels for Runge Kutta only from 1 time level to resume a forecast run hstop REAL Duration of the forecast in hours 0 0 nstop INT Number of time steps to be performed determines the fore 0 cast range nstop dt relative to the initial time can be specified alternatively to hstop dt REAL Time step in seconds
98. valid at the observation time rhfgps REAL additional scaling factor f to the horizontal correla 0 45 tion scale s which is applied only for humidity profiles derived from GPS IWV Sape Tope 8 cutofr 4 REAL cut off in multiples of correlation scales s 3 5 of the horizontal correlation function Wy 3 5 8 5 3 5 non isotropic correction w to isotropic function Wey vesni 4 REAL square of Gaussian radius of influence 2500 in potential temperature O diff if nsprpar lt 1 2500 in log pressure differences if nsprpar 2 2500 between the target grid point and the point at the 2500 horizontal observation location on the surface along which observation increments are spread laterally see figure for vcsnisu below Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 124 Horizontal weights for surface level observations Name Type Definition Purpose Comments Default isotropic horizontal correlation function w Way 1 Ar s e for scalar model variables rhiflsu 4 REAL constant part se of the correlation scale s in km 70 0 Se St Se 1E fe 1 1 w ios w temporal weight TU wi p ght 70 0 rhtfsu 4 REAL temporal factor f scaling the correlation scale s 1 0 at the beginning and the end of the nudging period 1 43 for an individual obs
99. which is defined as a character variable 3 Global attributes The global attributes contain general information about the data The attributes conventions conventionsURL and creation date are set within the model itself The other global attributes can be set by the user via namelist IOCTL see above Part VII User s Guide 4 28 Section 5 Data Formats for I O 5 2 The NetCDF Data Format 46 5 2 4 Useful Post Processing Utilities e ncdump Freeware Shows information about the contents of a netCDF file This pro gram is part of the netCDF standard software package e ncview freeware Visual browser for netCDF http meteora ucsd edu pierce ncview home page html e NCO freeware Software package including several programs to manipulate netCDF data http nco sourceforge net CDO freeware Software package including several programs to manipulate grib and netCDF data This is the successor of the PINGO package at the German Climate Research Centre DKRZ http www mpimet mpg de cdo e Others An extensive listing of software that uses netCDF is available from http www unidata ucar edu packages netcdf software html Part VII User s Guide 4 28 Section 5 Data Formats for I O AT Section 6 Input Files for the COSMO Model The COSMO Model requires several input files ASCII files called INPUT_ see below for the exact filenames that contain the namelist variables The form of these files
100. 0 11 031 int MB Degree of turbulence opt 0 11 036 float NMDEWX Derived equivalent vertical gust speed need 0 12 101 float MTDBT Temperature dry bulb T 0 33 025 int MAIV ACARS interpolated values need 0 08 004 int MPHAI Phase of flight need 0 02 064 int MQARA Wind quality roll angle opt 0 13 003 int MUUU Relative humidity need 0 12 103 float MTDNH Dew point temperature opt 0 13 002 float MMIXR Mixing ratio 1 02 000 Delayed replication of 2 descriptors 0 31 001 int MDREP Delayed descriptor replication factor 011075 float n MMPI Mean turbulent intensity EDR 4 011076 float n MPTI Peak turbulent intensity EDR 4 0 11 037 int MTUIN Turbulence index EDR 0 11 039 int NTIED Extended time of occur of peak EDR 0 11 077 int NRED EDR reporting interval 0 20 042 int NAICE Ice no ice 0 20 043 float NPLWC Peak liquid water content 0 20 044 float NALWC Average liquid water content 0 20 045 int NSLD Supercooled water droplet conditions Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files 68 Table notes 1 Variable MADDF does not exist in the WMO descriptor TM 311010 but is added at 2 DWD by the decoding software Flight level is defined relative to the ICAO standard sea level pressure and is readily converted to static air pressure using standard formulae i e using the ICAO standar
101. 0 127 0 097 22 3119 0 386 697 697 0 366 40 272 23 5616 0 659 504 504 0 190 0 149 23 2749 0 419 730 730 0 433 0 334 24 5327 0 767 524 524 0 259 0 209 24 2408 0 448 762 762 0 501 0 399 25 5058 0 8967 543 543 0 335 0 276 25 2093 0 473 792 792 0 568 0 465 26 4809 0 951 561 561 0 411 0 348 26 1806 0 492 820 820 0 631 0 530 27 4580 1 021 578 578 0 487 0 423 27 1544 0 507 847 847 0 690 0 594 28 4371 1 077 594 594 0 560 0 498 28 1308 0 518 871 871 0 745 0 656 29 4181 1 120 609 609 0 631 0 571 29 1097 0 525 893 893 0 795 0 713 30 4011 1 150 622 622 0 694 0 641 30 911 0 529 913 913 04 839 0 7696 31 3860 1 169 635 634 0 753 0 706 31 748 0 530 931 931 0 878 0 813 32 3728 1 182 645 645 0 805 0 765 32 608 0 528 946 946 0 911 0 855 33 3614 1 187 655 655 0 850 0 818 33 489 0 526 959 959 0 938 0 891 34 3517 1 188 663 663 0 890 0 863 34 391 0 522 970 970 0 958 0 922 35 3437 1 185 670 670 0 921 0 902 35 31i 0 517 979 9179 0 973 0 947 36 3372 1 181 675 675 0 948 0 933 36 249 0 512 986 987 0 984 0 966 37 3321 1 176 680 680 0 967 0 958 37 202 0 508 992 992 0 991 0 981 38 3283 1 171 683 683 0 982 0 976 38 168 0 504 996 996 0 996 0 992 39 3255 1 167 685 685 0 993 0 990 39 144 0 502 999 999 1 000 1 000 40 3235 1 164 687 687 1 000 1 000 40 Figure 8 24 Second part of example messages on nudging written to standard output
102. 0 1840 ztwi om t 06610 0 0086 0 9657 0 3401 zqwi om q 06610 0 00344256 0 5075 0 1302 mul infl km 277 EU4591 352 477 z obs v cut 2082 3508 1391 4567 k range 34 20 air infl km 492 SAS903 234 517 20647 zu vwi omyu 0 025 0 042 0 7700 sfc infl km 140 06610 167 103 94994 zqwi omyq 2 0 198833 6 1970 3 8493 thresh uv k 31 31 260 452 01427 24 mx omy k 1 k 0 642 0 490 uv incr 1 9 0 1 thrair T k 27 31 27 358 478 EU7654 1607 1721 max omy 0 792 T incr 0 0 1 9 air cor EU1337 113 167 103 408 spr par ob cutof 10949 7722 14176 at k 7368 640 43 thrair uv k 13 13 6 167 103 EU1337 10949 7368 max omy 0 240 uv incr 1 0 2 1 Figure 8 20 Example file YUPRINT third part related to spreading of observation increments The lines sfc kl sfc cor sfc infl and omyk for surfac level data resp mul kl air kl ras cor air cor mul infl and air infl for upper air data provide a variety of information on the spreading of observation increments Figure 8 20 The lines thresh and thrair also relate to the spreading of surface level resp upper air single level data and consist of the observation type uv for wind T for tem perature RH for humidity the current vertical model level the lowest for upper air data only and uppermost model level influenced by the observation the grid point coordinates of the station the station identity the height
103. 00 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 19 02 2008 19 02 2008 rrn rsn mm 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 15 21 47 15 21 47 Example file YUPRHUMI Figure 8 4 Model Output Section 8 Part VII User s Guide 4 28 8 1 ASCII Output for the Forecast Model 155 prsk precipitation rate for convective snow rrn sum of precipitation rain since start of the forecast rsn sum of precipitation snow since start of the forecast The file YUPRHUMI is always written With the NAMELIST parameters nOmeanval and ninc meanval of diactl the first output and the interval of the outputs in time steps can be controlled With ldump ascii TRUE FALSE the flushing of YUPRHUMI to disk in every time step can be switched on off An example of YUPRHUMI is shown in Figure 8 4 Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 156 8 2 ASCII Output Related to the Use of Observations The ASCII ou
104. 001 a snow analysis and a sea surface temperature SST analysis including an analysis of sea ice cover These analysis schemes are not integrated into the COSMO code but are programs on their own Therefore they are not covered by this User s Guide even though a scientific descrip tion or outline of them is included in the COSMO Documentation Part III on data assimilation Diagnostic surface analyses Basic Namelist setting lsurfa TRUE In contrast the data assimilation code of the COSMO model includes a module with a Cressman type successive correction analysis scheme which can be used to compute a set of other 2 dimensional surface level analyses This set comprises of a 2 m tem perature 2 m relative humidity 10 m wind speed and surface precipitation analysis While the precipitation analysis is purely based on rain gauge surface synoptic data the other analyses use the corresponding model field as a first guess purely to help defining the small scale details that are not resolved by the surface observations All these analyses are used only for diagnostic purposes As an exception the daytime 2 m temperature and optionally 2 m humidity analyses are used in the variational soil moisture analysis Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 29 Section 4 Installation of the COSMO Model This chapter explains the steps necessary to compile and run the model Section 4 1 lis
105. 04 5 768 5 125 5 192 5 144 5 809 5 159 5 811 5 169 5 804 5 177 5 803 5 185 5 813 5 191 5 828 15 21 47 15 21 47 Example file YUPRMASS Figure 8 3 Model Output Section 8 Part VII User s Guide 4 28 154 8 1 ASCII Output for the Forecast Model ntst 0 1OYU0O14 COO r2 OO i000 JACA CO PO ES CO CO CO CO CO CO CO CO hO PO PO PO PO NO P2 PO PO PO OY Ui 4 C IN P2 O xo O0 OY Ui 4 C No PP O to Experiment Elapsed time for providing initial and boundary values Number Initial mean values for nstart cloud water 0 000 Experiment ep qc 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 cloud ice 1 SUN 17 0 rain Snow 0 000 0 000 0 000 Number t SUN 17 0 qi qr qs kg kg 1E 5 0 085 0 000 0 386 0 076 0 000 0 385 0 072 0 000 0 383 0 069 0 000 0 382 0 067 0 000 0 380 0 065 0 000 0 379 0 064 0 000 0 377 0 062 0 000 0 376 0 061 0 000 0 374 0 060 0 000 0 373 0 059 0 000 0 371 0 059 0 000 0 370 0 058 0 000 0 369 0 057 0 000 0 367 0 057 0 000 0 366 0 056 0 000 0 365 0 056 0 000 0 364 0 055 0 000 0 362 0 055 0 000 0 361 0 055 0 000 0 360 0 054 0 000 0 359 0 054 0 000 0 2358 0 054 0 000 0 357 0 054 0 000 0 357 0 054 0 000 0 356 0 054 0 000 0 355 0 0
106. 1 allowed values are in the range 0 5 0 1 Runge Kutta like method following iadv order 2 3 Second order centered differences 4 5 Fourth order centered differences 0 2 4 Include linear extrapolation of horizontal wind to surface 1 3 5 No extrapolation of horizontal wind to surface For itype fast waves 2 the settings 102 104 112 114 122 and 124 are possible The meaning of the three digits led are 1 To make the value bigger than 100 to indicate the settings for itype fast waves 2 e 0 No extrapolation of u and v to the bottom e Linear extrapolation of u and v to the bottom e 2 Quadratic extrapolation of u and v to the bottom d 2 Centered differences of 2nd order to discretize dh dx and dh dy d 4 Centered differences of 4th order to discretize dh dx and dh dy Include diabatic forcing due to latent heat in the Runge Kutta scheme has been eliminated in Version 4 18 This is now set by the new parameter y scalar advect Switch to use semi Lagrangian advection of the moisture variables in the Runge Kutta scheme TRUE or not has been eliminated in Version 4 18 This is now set by the new parameter y scalar advect type of Euler forward advection of the moisture variables vanLeer PPM Bott 2 Bott_4 114 TRUE Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 3 DYNCTL Parameters for the Adiabatic Model 89 Name Type
107. 13 VEST NGAA 999 00 1 08 0 00 50 50 0 00 0 30 0 00 0 30 ps QC 60195 11 Obs Mod Thresh 853 60 849 22 4 30 Time Obs Mod 0 0 0 0 uv QC 08495 0 Obs Mod Thr 24 7 5 2 16 0 4 3 27 9 Time O M 2 4 0 0 P 10 IWV sc QC 60571 0 Obs Mod Thr 5 56 10 79 5 13 T O M 1 8 0 0 bias w 0 0 0 0 V mult QC 08019 137 Time Obs Mod 0 0 0 0 P 916 368 Figure 8 18 Example file YUPRINT first part related to observation statistics and quality control Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 179 The lines mladm sgadm and gpadm provide the local number of currently active multi level single level resp GPS reports or stations and list the following items of them station indices for 2 reports i e 1 report in the past and 1 report in the future grid point co ordinates station identity 2 observation times in h i e for the 2 reports 999 denotes missing report lengths of the decreasing part selected increasing part and minimum pos sible increasing part of the temporal weight function the 2 temporal weights according to the temporal function selected the 2 temporal weights assuming that temporal linear inter polation cannot be applied A statement containing QC is made whenever the quality status of an observation changes due to the threshold quality control QC i e when an observation does not pass the QC for the first time or w
108. 134 72246 6 136 47912 6 134 74341 6 136 47945 6 134 10169 6 137 WHITELIST 6 136 10204 6 137 72246 6 136 10384 6 137 74341 6 136 WHITELIST 6 137 10169 6 137 10204 6 137 10384 6 137 In the standard format the first line is fixed and the following lines are the entries in the whitelist with the following 3 columns 1 station identity 2 observation type currently only 6 PILOT which includes profilers 3 observation code type 132 European wind profiler 133 European RASS SODAR 134 Japan wind profiler RASS 136 US wind profiler RASS 137 Radar VAD wind profiles The standard format assumes a whitelist exactly for those observation code types for which there are entries on whitelist In the blacklist file used at DWD currently no RASS profiler station is on the whitelist which implies that COSMO will assume that no whitelist exists for RASS Hence all RASS reports would be used actively unless they are on the blacklist This is indeed the case i e the temperature profiles from all the known RASS stations are put on the blacklist see e g station 10266 in the example above There is an alternative format for the whitelist which can be used but is not yet as thor oughly tested A line containing WHITELIST and observation type and code type precedes the whitelist for each code type even if the whitelist for that code type is empty Only this alternative format allows to use empty whitelists and
109. 150 C SURFACE LEV OBS 90 C OR gt 40 C HUMIDITY ABOVE 300 HPA LEVEL HUMIDITY EXCEEDING ALLOWED VALUE 120 HUMIDITY FORCED TO BE SATURATED T gt 0 HUMIDITY FORCED TO BE SATURATED T O HUMIDITY FORCED TO BE lt 100 T gt 0 HUMIDITY FORCED TO BE lt 100 T 0 HUMIDITY AT 2M HEIGHT OR HEIGHT DISTANCE TO OROGRAPHY TOO LARGE WIND DIRECTION MISSING WIND SPEED MISSING WIND DIRECTION FLAGGED OR ABSOLUTE VALUE 360 DEGREES WIND SPEED FLAGGED WIND SPEED 0 DRIBU lt 0 OR gt 150 M S P gt 700HPA gt 90 M S WIND AT 10M HEIGHT OR HEIGHT DISTANCE TO OROGRAPHY TOO LARGE WIND SPEED SHEAR TOO LARGE WIND DIRECTION SHEAR TOO LARGE PRECIPITATION AMOUNT EXCEEDING THRESHOLD LIMIT ZENITH PATH DELAY MISSING OR TOO SMALL 19 2021 2223 24 2526 27 28 29 30 31 32 33 3435 36 37 38 67 10 0 16 178 490 0200 47 58 0 0 027390 0 0 0 1g 0 0190 102 1 1 11435573559 0 0 018510 0 O 0 13 00 0 0 3 00 0 0 0 0 0 0 0 30 0 0 0 04301214 0 23 69 0 0 0 95 95 24 24 0 991 0 0 0 0 00 0 0 0 0 0 0 0 0 0 0 0 0 00 0 0 71 SYNOP Manual Lan SYNOP Automatic Figure 8 14 Example file YUSTATS forth part incomplete Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 170 1 O HOURLY MAX TOTAL NUMBER OF STATIONS WITH ACTIVE OBSERVATION INCREMENTS only for MPP applications S multi uppe
110. 2 5 1 0 0 L0 313 296 12 21 48 43 244 2 2 53 AIREP EU5331 0 1 2 0 268 5 0 65 882 51 1150 2 5 T 0 0 10696 0312298 12 15 48 41 244 2 2 53 AIREP EU5331 P 1 1 4 5 268 8 0 50 879 29 31180 42 5 1 0 O0 10 k xw 312 296 12 18 48 42 244 2 2 53 AIREP EU5331 P 0 2 2 1 268 8 0 63 892 51 1150 2 5 1 0 0 10 312 296 12 12 48 40 244 2 2 54 AIREP EU5331 M 0 3 2 6 269 0 0 62 881 43 31160 2 5 1 0 O lQ t xk 311 296 12 10 48 40 244 2 2 54 AIREP EU5331 M 0 6 3 5 2069 3 0 58 882 51 1150 2 5 1 0 0 109 317 295 12 06 48 38 244 2 2 54 AIREP EU5331 P 0 2 3 1 269 3 0 55 0882 51 1150 2 5 T 0 Q 10sskexeoexie 311 295 12 03 48 38 244 2 2 56 AIREP EU5331 M 0 1 3 1 269 5 0 54 882 51 1150 2 5 1 0 0 109 w 310 295 12 01 48 37 244 2 2 56 AIREP EU5331 M 0 4 2 6 270 0 0 52 890 05 1080 2 5 1 1 0 1O 310 295 11 98 48 37 244 2 2 56 AIREP EU5331 M 0 6 2 5 270 8 0 53 902 01 DIO Seer FD Ad Qu pQsceexaoee C3T0 295 11 95 48 37 244 2 2 58 AIREP EU5331 M 0 3 2 1 270 8 0 62 2913 00 BO eRe F5 1l 010r ERRE RR CSTOLZO5 11 93 48 37 244 2 2 58 AIREP EU5331 M 0 5 2 5 272 5 0 58 922 98 TU 2 5 3 1 0 TOS 9 309 295 11 91 48 37 244 2 2 58 AIREP EU5331 M 0 4 2 6 272 8 0 60 930 81 TEQ cee 02 5 3 1 QTQssekekdeese 309 295 11 89 48 37 244 2 2 60 AIREP EU5331 M 0 3 2 6 273 2 0 56 938 69 GAD RHEE Ib 1 Og Osee ee 309 295 11 87 48 37 244 2 2 60 AIREP EU5331 M 0
111. 2 5 YUCAUTN Nudging Warning Messages on Insufficient Array Sizes 164 8 2 6 YUSTATS Nudging Statistics on Observation Processing 166 8 2 7 YUVERIF Nudging Verification File VOF 171 8 2 8 YUPRINT Nudging Other Aspects a asoa saco ad taraua 178 8 2 9 Standard Output Basic Monitoring of Nudging 182 8 2 10 YUSURF 2 D Surface Analyses 0 0200004 186 So NetCDF Peedobs Mile o sog seose RUE ee ee S WU ee no 8 188 8 4 Output of Forecast Fields i s cos ono o x oko exon m Re 9o ROUES X e 189 References 193 Part VII User s Guide 4 28 Contents Section 1 Overview on the Model System 1 1 General Remarks The COSMO Model is a nonhydrostatic limited area atmospheric prediction model It has been designed for both operational numerical weather prediction NWP and various scien tific applications on the meso 8 and meso scale The COSMO Model is based on the prim itive thermo hydrodynamical equations describing compressible flow in a moist atmosphere The model equations are formulated in rotated geographical coordinates and a generalized terrain following height coordinate A variety of physical processes are taken into account by parameterization schemes Besides the forecast model itself a number of additional components such as data assimi lation interpolation of boundary conditions from a driving host model and postprocessing utilitie
112. 21 0 370 8 029 22 1 250 7 658 23 0 370 7 442 24 0 360 7 288 25 0 360 7 210 26 260 6 887 2 0 370 6 880 28 0 360 6 796 29 0 390 6 489 30 1 280 5 791 31 0 360 5 203 32 0 360 4 570 33 1 290 4 018 34 0 360 3 653 Number for nstart a dry static energy Number ps hPa 967 195 967 324 967 265 967 289 967 284 967 289 967 299 967 294 967 303 961 297 967 301 967 297 967 298 967 295 967 294 967 291 967 288 967 284 967 281 967 277 967 272 967 267 967 262 967 257 9672252 967 246 967 240 967 233 967 226 967 220 967 214 967 208 967 200 967 190 967 180 T SUN 17 02 2008 18 UTC Elapsed time for providing initial and boundary values REAL 0 for several variables 3139 1 17 93 J kg moist static e 314962 08 SUN 17 02 2008 18 UTC dse J kg E 3 0 005 0 001 0 000 0 003 0 005 0 006 0 008 0 009 0 010 0 011 0 013 0 014 0 015 0 016 0 017 0 019 0 020 0 021 0 022 0 024 0 025 0 026 0 027 0 028 0 029 0 030 0 031 0 032 0 033 0 034 0 035 0 036 0 037 0 037 0 038 mse J kg E 3 0 002 0 012 0 014 0 020 0 023 0 027 0 030 0 032 0 035 0 038 0 040 0 042 0 045 0 047 0 049 0 051 0 053 0 055 0 057 0 060 0 062 0 064 0 065 0 067 0 069 0 071 0 072 0 074 0 076 0 077 0 079 0 080 0 082 0 083 0 084 ke J kg 0 621 0 924 1 220 1 463 1 679 1 872 2 044 2 202 2 338 2 464 20913 2 671 2 159 24828 2 889 29 93 2 975
113. 25 will contain the integer value 20 The time or time interval for which the data are valid with respect to the reference time is coded in octets 18 21 ipds 16 ipds 19 Octets 19 and 20 contain two periods of time P1 and P2 The units of the values of P1 and P2 are defined in octet 18 Currently we use hours as the time unit but other values may be more appropriate for special applications of the model as the maximum integer number in an octet is 256 Thus for long term climate runs or short term cloud simulations other time units must be chosen The WMO code table for the unit of time in P1 and P2 is given in Table 5 6 The meaning of the time period P1 in octet 19 ipds 17 and of the time period P2 in octet 20 ipds 18 given in the units coded in octet 18 depends on the time range indicator which is contained in octet 21 ipds 19 The WMO code table allows for a large number of indicators including averages and accumulation over a number of forecasts and analyses For the COSMO system we use only a few standard indicators as shown in Table 5 7 Part VII User s Guide 4 28 Section 5 Data Formats for I O 5 1 The GRIB Binary Data Format 40 Table 5 6 Code table for unit of time ipds 16 Meaning ipds 16 Meaning ipds 16 Meaning 0 Minute 5 Decade 11 6 hours 1 Hour 6 Normal 12 12 hours 2 Day 7 Century 13 253 Reserved 3 Month 8 9 Reserved 254 Second 4 Year 10 3 hours Table 5 7 Ti
114. 3 02 044 Evaporation data 3 02 045 Radiation data 3 02 046 Temperature change data Table notes 1 Only one of the variables MHOBNN and MHOSNN is strictly needed to exist MHOBNN is preferred to exist and to be used if both variables exist and have non missing values because it should provide the precise height of the barometer for the pressure measure ment which is the observation with the most critical dependency on sensor or station height The use of n in the variable type definition means that this variable has an additional dimension i e several values may be present in one report If the corresponding replica tion factors MEDRE or MDREP are zero for all reports in the NetCDF file then the corresponding multi dimensional variables do not need to exist and probably will not exist in the NetCDF file File names for fixed land stations for sea stations e TEMP TEMP MOBIL and TEMP SHIP TEMP for mobile land stations TEMP MOBIL TEMP SHIP cdfin temp cdfin temp cdfin tempship The template given by the common sequence descriptor TM 3 09 052 is used for the observation types TEMP fixed land stations TEMP MOBIL mobile land stations as well as TEMP SHIP sea stations The table below lists all the variables of this template and their use in COSMO Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observatio
115. 33 i e square of Gaussian vertical radius of influence 0 333 in log pressure differences if msprpar lt 1 0 04 in potential temperature O diff if msprpar 2 0 04 reasonable values in latter case 275 275 33 33 vcutof 4 REAL cut off for Gaussian vertical correlation function w 0 75 In pcus poos 0 75 vcutof 5 such that w e lade se 1 0 peut resulting cut off pressure 1 0 if nsprpar 2 In p is replaced by lsvcorl LOG for aircraft observations only TRUE decrease of the vertical correlation scales as given by vcorls such that the resulting correlation func tions dotted lines are adjusted individually for each observation and take the value of 0 5 halfway between the present observation and nearest observation above resp below if they are nearly colocated in the horizontal and are reported by the same aircraft vertical weights for surface level observations vcorlsu 4 REAL square of the vertical correlation scale 0 013 i e square of Gaussian vertical radius of influence 0 013 in log pressure differences if nsprpar lt 1 0 002 i e default e folding decay height for T f 300m 0 002 in potential temperature O diff if nsprpar 2 reasonable values 11 1 11 1 1 33 1 33 vcutosu 4 REAL cut off for Gaussian vertical correlation function w 0 75 In peut Pobs 0 75 A E t y such that w e S REF QS 4 0 peut resulting cut off
116. 4 int MGG Hour need 0 04 005 int NGG Minute 0 04 006 int MSEC Second 3 01 114 Horiz vert coord of launch site 3 01 021 Latitude and Longitude need 005001 float MLAH Latitude high accuracy degree need 006001 float MLOH Longitude high accuracy degree need 0 07 030 float MHOSNN Height of station above MSL 2 need 0 07 031 float MHOBNN Height of barometer above MSL 2 opt 0 07 007 int MH Height of release of sonde above MSL opt 0 33 024 int MSEQM Station elevation quality mark 2 3 02 049 Cloud info reported with vert soundings opt 0 08 002 int MVTSU Vertical significance opt 0 20 011 int MNH Cloud amount of low or middle clouds opt 0 20 013 float NH Cloud base height above surface opt 0 20 012 int MCC Cloud type low clouds Cz opt 0 20 012 int MCCO Cloud type middle clouds Cm opt 0 20 012 int MCCI Cloud type high clouds Cy 0 22 043 float MTNOO Sea water temperature Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files 61 use WMO descriptor type mnemonics meaning need 0 31 002 int MEDRE Extended delayed descriptor replicat fac 3 03 054 Temperature dewpoint wind data at a pressure level with sonde position opt 0 04 086 int n NLTPD Time displacement since launch time s need 0 08 042 int n MEVSS Extended vertical sounding significance need 0 07 004 float n MPN Pressure ve
117. 48 0 10 60 594 11 2 540 82 9 0 42 319 0 00 0 0 0 00 0 0 13 0 11 292 111 0 15 105 508 8 3 292 111 9 0 337 64 0 00 0 0 0 00 0 0 14 0 16 113 40 0 14 114 40 9 4 175 25 9 4 55 16 0 00 193 109 0 00 649 26 33 0 17 420 64 0 22 23 249 9 6 641 218 10 1 621 200 0 13 16 238 0 22 23 249 39 0 20 410 55 0 23 344 648 9 1 642 220 7 3 255 88 0 01 16 238 0 02 23 249 40 0 20 410 55 0 23 344 648 9 2 642 220 7 4 258 84 0 00 16 238 0 01 23 249 Figure 8 22 Example file YUPRINT fifth part related to analysis increments Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 182 8 2 9 Standard Output Basic Monitoring of Nudging As already mentioned e g in Section 8 2 5 warning messages containing the label CAUTION are also issued to the standard output if the values of the NAMELIST variables maxmlo maxuso maxsgo or maxgpo are not large enough to hold in memory and process all observations In such a case the program will not crash or stop see Section 8 2 5 In the standard output such a message is written for each individual multi level observation report which cannot be used due to insufficient array size Furthermore summary messages indicate how many reports of which type had to be discarded for that reason Figure 8 23 lists the types of regular messages that are also written to the standard output by the nudging The first line specifies the length
118. 5 0 20 256 13 6 22 1 222 9 0 09 300 0 8880 3 7 0 5 0 20 10 7 32 14 7 23 9 218 3 0 06 259 0 Wkwew 3 5 Rips eee 2 13 6 22 1 218 3 0 05 250 0 10060 3 5 O 5 11 8 32 12 1 24 x x k k 206 0 3 GARA RRA RRA RARA 128 12 1 24 1 217 9 0 02 200 0 11480 3 5 0O0 6 13 2 32 oekekdeeeeeeee 219 6 0 01 155 09 000000k TRE RAKE 256 7 3 11 9 218 6 0 01 150 0 13320 3 4 O 7 15 2 32 0 3 11 0 215 6 0 01 L0Q Q0 w x 3 3 Q Bgskseew 394 2 4 11 8 214 3 0 01 70 0 t8140 3 2 0 8 19 5 8 z346 3 4 214 1 0 01 50 0 20250 3 2 0 9 22 5 8 0 6 1 9 212 4 0 01 30 0 23430 3 3 0 9 25 0 8 2 9 0 7 212 1 0 01 20 0 25940 3 6 1 0 32 0 8 11 0 0 3 220 6 0 01 LD OeAeAERE A D qp 129 PILOT 10266 OBS CODE TYPE 6 132 STA MOD HEIGHT 57 56 LON 11 84 LAT 53 31 G P 306 374 HOUR 0 5 u v t rh p z w err t er rh er z er lev z 4 lt 2 SER KERR RRR 958 9 525 DLR KKK RRR RRR k ERK 0 0 8 L2 3K KK k k kk k k k 941 6 669 2 1 AK HR k RRR RRR REE 0 1 0 Q pxdoxke ek 924 5 814 2 244 x k eek 0 0 6 D Bee eek 907 9 958 2 24 RRR RR RHR RE 0 0 8 3 4 x eek 891 4 1103 2 99 d kkk k 0 PILOT 10266 P OBS CODE TYPE 6 133 STA MOD HEIGHT 57 56 LON 11 84 LAT 53 31 G P 306 374 HOUR 0 5 u v E rh p z v err t er rh er z er lev RRRKKKKAKKRAKK 270 ape 956 9 542 BAKER ERR RR RRR RK 0 kkkkkkkkkkkkkk 269 Toe 940 4 679 IR ae 0 kkkkkkkkkkkk kk 269 Qe 924 3 816 o ok ek RR RR k 0 kkkkkkkkkkkkkk 269 Doe 908 3 954 AERA RARA RRA RR R
119. 5 1 The GRIB Binary Data Format 39 Table 5 4 Process identification numbers process id number ipds 4 Comment 131 Analyses from data assimilation cycle for former model domain 132 Forecasts and initialized analyses for former model domain 134 Analyses from data assimilation cycle for COSMO EU 135 Forecasts and initialized analyses for COSMO_EU 137 Analyses from data assimilation cycle for COSMO DE 138 Forecasts and initialized analyses for COSMO DE Table 5 5 Types of fixed levels or layers used by the COSMO Model level type Meaning ipds 9 ipds 10 ipds 8 1 Ground or water surface 0 0 2 Cloud base level 0 0 3 Level of cloud tops 0 0 4 Level of 0 C isotherm 0 0 8 Top of atmosphere 0 0 100 Pressure isobaric level 0 Pressure in hPa 102 Mean sea level 0 0 103 Specified height above 0 Height in m mean sea level 105 Specified height level 0 Height in m above ground 109 Hybrid level half levels 0 Level number k 110 Hybrid layer main level Level number Level number of between two hybrid levels of top k bottom k4 1 111 Depth below land surface 0 Depth in cm 112 Layer between two depths Depth of upper Depth of lower below land surface surface in cm surface in cm the century is coded in octet 13 and the century 100 years in octet 25 For a reference time within the year 2000 octet 13 will contain the integer value 100 and octet
120. 54 0 000 0 354 0 054 0 000 0 353 0 054 0 000 0 353 0 054 0 000 0 352 0 054 0 000 0 352 0 054 0 000 0 351 0 054 0 000 0 350 0 054 0 000 0 350 0 054 0 000 0 349 0 054 0 000 0 349 0 054 0 000 0 348 2 2008 18 UTC REAL S 1 12 graupel all g kg 0 000 2 2008 18 UTC 0H SUN 17 02 2008 qg prrs prss prrk mm D 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0H SUN 17 02 2008 18 UTC 18 UTC prsk 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 0
121. 7 9 NUDGING Controlling the Data Assimilation 131 Directory and file names related to nudging Name Type Definition Purpose Comments Default itype obfile INT Type of observation input file s 1 1 AOF 2 NetCDF yaofpath CHAR path directory unless working directory and filename of aof 70 the input observation file AOF ycdfdir CHAR directory where NetCDF observation input files and the P 70 blacklist file reside Diagnostic output Name Type Definition Purpose Comments Default lprodr LOG diagnostic print about observation data record ODR TRUE on file YUOBSDR see section 8 2 2 ionl INT grid point coordinates for control output 167 jonl INT on file YUPRINT see section 8 2 8 103 ionl2 INT grid point coordinates for further control output 167 jon12 INT on file YUPRINT see section 8 2 8 103 diagnostic print of observational input from AOF files only lpraof LOG diagnostic print of analysis observation file AOF FALSE on file YUAOFEX see section 8 2 1 dinlat REAL latitude of northern boundary of diagnostic print area deg 0 dislat REAL latitude of southern boundary of diagnostic print area deg O diwlon REAL longitude of western boundary of diagnostic print area deg O dielon REAL longitude of eastern boundary of diagnostic print area deg O noctrq INT observation or code type of reports to be printed 9 if noctrq 9 then print all report types
122. 7 Aircraft ascent descent profile data for 1 level with lat lon indicated need 0 07 010 int n NFLEV Flight level 2 3 3 01 021 Latitude and Longitude opt 0 05 001 float n MLAHO Latitude high accuracy degree 3 opt 0 06 001 float n MLOHO Longitude high accuracy degree 3 need 0 11 001 int n NDNDN Wind direction 3 need 0 11 002 float n NFNFN Wind speed 3 need 0 02 064 int n MQARA Wind quality roll angle 3 need 0 12 101 float n MTDBT Temperature dry bulb temperature 3 need 0 12 103 float n MTDNH Dew point temperature 3 0 13 002 float n MMIXR Mixing ratio 1 3 Table notes 1 The Variables MMIOGC MMIOGS NOSNO and MMIXR do not exist in the WMO BUFR descriptor TM 311009 but are added at DWD by the data base decoding software 2 Flight level is defined relative to the ICAO standard sea level pressure and is readily converted to static air pressure using standard formulae i e using the ICAO standard atmosphere Hence flight level is not the geometrical height Once converted the original resolution either 100ft or 10ft in the BUFR report is lost hence it is desirable to disseminate the element in the received form 3 The use of n in the variable type definition means that this variable has an additional dimension here for vertical levels If the corresponding replication factor MDREP is zero for all reports in the NetCDF file then these multi
123. 774 9 9 9 o 2 9 9 90 9 692 27 48 13 0 9 9 9999 9999 9999 9999 9999 1 Q671 948 4839 20 750 749 1 14 0 00 0 284 295 7 19 2726 1000 9 750 6 0 20 T 329 59 z9 IPEETPDETNROS X7622 77115 9 174 211 40 15 9999 9 9 9 9 3 EU5331 1215 4842 52 999 495 2 244 140000000 00 0 312 296 1 26 2685 655 88251 50 7 0 400140600 384 9 56 91 18 118 3 EU5331 1207 4839 53 999 481 2 244 140000001 100001 0 311 295 6 35 2693 585 88251 50 7 0 400140600 384 9 58 18 69 240 6 EU5331 1207 4839 53 999 481 2 244 140000000 00 4 311 295 6 35 2693 585 88251 50 a 0 400140600 384 3 26 2690 629 88144 60 E 0 400140600 384 7 35 2680 570 86970 270 7 4 400140600 384 1 31 2670 039 85808 380 7 4 400140600 384 2 41 2665 624 84868 470 E 4 400140600 384 4 51 2655 667 84038 550 7 4 400140600 384 58 18 69 240 9999 29 75 45 199 9999 91 20 22 291 9999 35 63 6 257 9999 40 6 18 295 9999 20 78 29 270 9999 9999 3 EU5331 1210 4840 53 999 477 2 244 140000001 100001 0 311 296 3 26 2690 629 88144 1160 7 0 400140600 384 9 51 75 47 178 3 EU5331 1212 4840 53 999 495 2 244 140000102 20002 0 312 296 2 21 2688 640 88251 1150 Fi O 400140600 384 9 73 142 12 133 3 EU5331 1202 4837 54 999 458 2 244 140000001 100001 0 310 295 1 31 2695 543 88251 1150 7 0 400140600 384 9 56 16 95 317 3 EU5331 1204 4839 54 999 481 2 244 140000102 20002 0 311 295 2 31 2693 552 88251 1150 7 0 400140600 384 9 122 28 69 274 9 xXxxxx 0 0 0 0 Q 0 0 0 00 0 0 0
124. 8 9 5 7 0 70 RH 0 05 0 80 0 76 RH t DIALS 35 2 4 560 0 58 9 5 7 0 51 RH 0 22 0 76 q mult EU5331 244 2 4 618 0 48 9 12 5 p top 538 0 q mult EU5331 244 2 5 707 3 48 7 12 4 p top 639 2 z mult 17220 35 2 5 1000 0 38 4 27 2 p top 50 0 z 17220 39 2 5 10 0 38 4 27 2 312 1 p top 10 0 zz 324 2 RH 60571 35 360 500 0 SS 262 0 6007 RH 0 11 0 88 IWV sc 60571 35 3 0 400 0 31 5 2 2 5 2 IWV Disi 12 0 bias w2 0 0 0 0 p TEMP 02527 35 3 0 948 4 57 7 12 5 5 0 ps 948 4 955 0 ps 02527 11 3 0 953 0 37 7 12 5 35 0 ps 953 0 959 5 ps 02527 35 340 953 0 S7 X25 5 0 ps 953 0 959 6 z mult 02527 35 340 953 0 257 7 1245 p top 500 0 I 26038 35 3 0 250 0 59 4 24 6 5 0 E 207 1 212 4 T 26038 35 3 0 245 0 59 4 24 6 5 1 T 206 7 212 5 T mult 26038 35 3 0 250 0 59 4 24 6 p top 150 0 dz 26038 35 3 0 250 0 59 4 24 6 96 9 p top 100 0 dz 119 9 dT mean 4 5 ps scc 14427 11 3 0 1011 8 43 8 15 2 3 4 ps 1011 8 1015 3 bias w2 Od 120 ps scc 16584 14 3 0 1011 9 41 1 16 8 3 6 ps 1011 9 1008 3 bias w2 0 3 Sa Figure 8 9 Example file YUQUCTL Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 163 The entries for each line denote the type of rejection The entries uv uv 10m T T 2m RH and RH 2m indicate that an upper air resp surface level wind temperature resp humidity observation is rejected by the individual threshold quality c
125. 992 This scheme is based on a two stream version of the general equation for radiative transfer and considers three shortwave solar and five longwave ther mal spectral intervals Clouds aerosol water vapour and other gaseous tracers are treated as optically active constituents of the atmosphere which modify the radiative fluxes by absorption emission and scattering As an extension to the original scheme a new treatment of the optical properties of ice par ticles has been introduced which allows a direct cloud radiative feedback with the predicted ice water content when using the cloud ice scheme for the parameterization of cloud and precipitation Numerically the parameterization scheme is very cost intensive Thus it is called only at hourly intervals during an operational forecast on the meso scale The resulting short and longwave heating rates are then stored and remain fixed for the following time interval In case of high resolution simulations the calling frequency of the radiation scheme can be increased to allow for a better representation of the interaction with the cloud field The radiation can also be computed on a coarser grid to save computation time 3 5 2 Grid scale Precipitation Basic Namelist settings lphys TRUE lgsp TRUE The basic parameterization scheme for the formation of grid scale clouds and precipitation is an adapted version of the DM scheme It is based on a Kessler type bulk formulation an
126. CKLIST WHITELIST from blklsttmp NOTE NO FILE cdfin tempdrop nc NOTE NO FILE cdfin amdar ml nc NOTE NO FILE cdfin amdar vp nc NOTE NO FILE cdfin acars uk nc NOTE NO FILE cdfin acars us nc NOTE NO FILE cdfin synop mob n Qo NOTE NO FILE cdfin metar nc NOTE NO FILE cdfin satob nc processing 415 reports from 58 241 min from file cdfin temp processing 14 reports from 58 241 min from file cdfin tempship processing 7 reports from 58 241 min from file cdfin pilot processing 5 reports from 58 241 min from file cdfin pilot p processing 2829 reports from 28 151 min from file cdfin amdar processing 3709 reports from 28 151 min from file cdfin acars processing 118 reports from 28 91 min from file cdfin wprof processing 20 reports from 28 91 min from file cdfin rass processing 556 reports from 28 91 min from file cdfin radar vad processing 8662 reports from 58 151 min from file cdfin synop processing 335 reports from 58 151 min from file cdfin ship processing 246 reports from 58 151 min from file cdfin buoy processing 16369 reports from 58 151 min from file cdfin gps zenith processing 705 reports from 58 151 min from file cdfin ascat processing 0 reports from 58 151 min from file cdfin_qscat Figure 8 23 First part of example messages on nudging written to standard output Part VII
127. COSMO Modell 7 3 DYNCTL Parameters for the Adiabatic Model 85 7 3 DYNCTL Parameters for the Adiabatic Model The namelist group DYNCTL contains control parameters for the numerical methods used to solve the thermo hydro dynamic model equations the adiabatic part of the COSMO Model and to specify the boundary conditions and the numerical filtering for the COSMO solution in particular close to the boundaries The specifications for the parameters in DYNCTL are included in the file INPUT DYN The namelist parameters of this group are described in the subsections main switches for the time integration parameters for the semi implicit time integration scheme parameters for the Runge Kutta scheme parameters for the lateral boundary conditions horizontal diffusion lower and upper boundary condition additional numerical filters and spectral Nudging Main switches for the time integration The main switch for the dynamics in the COSMO Model is 12t1s Several options for the choice of numerical schemes are available with one of the schemes only The more recent developments are implemented for the Runge Kutta scheme 12t1s TRUE only In the first years of the COSMO Model the Leapfrog scheme was the only available dy namical core 12tls FALSE But development for a 2 timelevel Runge Kutta scheme started early First this scheme has been developed for very high resolutions below 4 5 km But later
128. CS type make help for a list of available entries Outside the VCS the following entries are available Note that the special entry has to correspond to the settings of the macros purexe A pure binary without nudging and synthetic satellite images DNUDGING DRTTOVx and DNETCDF must not be set allexe A binary with nudging and synthetic satellite images if DNUDGING DRTTOVx and DNETCDF are set nudexe A binary with nudging but without synthetic satellite images DNUDGING and DNETCDF must be set and DRTTOVx must not be set satexe A binary without nudging but with synthetic satellite images DNUDGING must not be set and DRTTOVx must be set More entries can be added on your own Part VII User s Guide 4 28 Section 4 Installation of the COSMO Model 4 5 Running the Code 34 4 5 Running the Code To run the code several ASCII files INPUT_ have to be provided that contain values for the NAMELIST variables The form of these INPUT files is described in Chapters 6 and 7 They are created by the provided run scripts See the manual for your system on how to invoke the binary created in the last step Part VII User s Guide 4 28 Section 4 Installation of the COSMO Model 35 Section 5 Data Formats for I O All input and output fields of the COSMO Model and the preprocessor program providing interpolated initial and boundary conditions can be stored in GRIB or in NetCDF format Restart files
129. Consortium for Small Scale Modelling as FOR SMALL SCALE MODELING A Description of the Nonhydrostatic Regional COSMO Model Part VII User s Guide U Sch ttler G Doms and C Schraff COSMO V4 29 October 2013 OD O www cosmo model org DEA Arp a Er perla Protezione Ambientale Printed at Deutscher Wetterdienst P O Box 100465 63004 Offenbach Germany Contents Contents 1 Overview on the Model System 1 Li General Remarks a r x ice o9 o ek R A a OX GRE as 1 1 2 Basic Model Design and Features 2l 3 1 3 Organization of the Documentation snis se ad woni a ae eae a ioa 6 2 Introduction 8 3 Model Formulation and Data Assimilation 10 3 1 Basic State and Coordinate System llle 10 3 2 Differential Form of Thermodynamic Equations 11 3 3 Horizontal and Vertical Grid Structure a a ea een 12 3 4 Numerical Integration ll 16 3 4 1 Runge Kutta 2 timelevel HE VI Integration 17 3 4 2 Leapfrog 3 timelevel HE VI Integration 17 3 4 3 Leapfrog 3 timelevel Semi Implicit Integration 18 3 5 Physical Parameterizations a sa ce so osa aon moa a i e aa Eaa E E a 18 DID Radiation ss saosa awo ae e SO a ae a EUR a a mh Re En 19 3 0 2 Grid scal e Precipitation lt lt o cs ea sooo Be a amp Pa D e 19 3 09 Moist WOnVeCtiGn s s miere die E Rea e ba d OUR a T s 21 3 5 4 Vertical Turbulent Diffu
130. D SUM W Number of scans 3 Scan 1 radius 200000 m Scan 2 radius 100000 m Scan 3 radius 50000 m Figure 8 27 Example file YUSURF first part on description of methods Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 187 After a description of the grid point itself station identity latitude longitude height ob servation increment and total weight of the increment at the grid point are also provided for these observations For precipitation observations assigned to the grid point itself this is complemented by the vertical distance zvdis and vertical weight zwi in the lines starting with zmodor Finally the resulting analysis increment of the scan and some statistics on the analysis are provided Total number of surface observations extracted from ODR nobtot 1049 zdistm 189364 zrormx 0 947 wwa 0 055 zdistm 192541 zrormx 0 963 wwa 0 038 Diagnostic 2M TEMPERATURE analysis scan 1 grid point ix 167 iy 103 lat 47 54 lon 8 74 height 462 Observations influencing this grid point stat id lat lon height increment weight 06771 45 84 8 93 352 714 21 0 08 06770 46 00 8 96 301 0 02 0 16 10733 49 30 8 91 240 20 33 0 02 Resulting t2m analysis increment 0 15 Resulting t2m analysis increment 0 04 STATISTICS ON ANALYSIS OF 2M TEMPERATURE UNIT DEGREES SUM OF CHANGES 60422 7556 NO OF ANAL POINTS 194081 A
131. E ldiasa TRUE ionl 255 jonl 271 ionl2 255 jonl2 271 end input ass Figure 7 2 Excerpt of run script for COSMO Model to create the INPUT ASS file related to the NAMELIST group NUDGING Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 116 General variables controlling the nudging Name Type Definition Purpose Comments Default nudgsta INT start of nudging period in timesteps 0 nudgend INT end of nudging period in timesteps 0 or hnudgsta REAL start of nudging period in hours 0 0 hnudgend REAL end of nudging period in hours 0 0 relative to initial time of model run tconbox REAL time box in s for which analysis increments are computed 3 dt once and then held constant dt model timestep nwtyp INT if gt 1 then compute net weighted observation increments i e 1 preliminary analysis increments Av in Section 3 2 of Part IIT for nwtyp different sets of observing systems separately niwtyp 20 INT for each of the nwtyp sets of observing systems 1 0 0 number of observation and or code types which belong to and define that set of observing systems for obs code types see below use of observation and code types nwtyp number of non zero elements in niwtyp iwtyp 50 INT for each of the nwtyp sets of observing systems 0 0 0 obser
132. E 0 kkkkkkkkkkkkxkk 270 QA AAA 892 8 1091 xxi RRR k 0 SYNOP Q671 QT EIS 212 6 3 00 7505 750 1 0 Q lQ tee 1 284 295 9 48 48 38 141 2 00 SYNOP 07497 p LU 0 2 278 5 0 52 917 90 868 B86g eenecesceeiobeeesiex 787 253 252 6 76 45 61 111 2 00 SYNOP 02868 P 5 1 7 4 262 0 0 92 1002 60 0 488 ienek 156 417 596 29 14 66 16 141 2 00 SYNOP 02876 4 3 5 5 271 1 0 80 1003 40 55 5 3 6 1 0 0 10 7 4 2 396 571 25 39 65 00 141 2 00 SYNOP 63108 6 0 3 0 280 0 0 94 1017 50 0 0 3 6 1 0 0 10 14 5 0 223 497 1 70 60 79 241 2 00 SYNOP DBFC 2 0 2 2 273 9 0 73 1024 30 0 0 3 6 1 0 0 10 14 5 0 320 402 13 39 55 00 211 2 00 DRIBU 63551 ERNE LEE OR OE Howe LOA O 0 QE OE RE eee S 0 230 605 0 84 67 53 165 4 0 00 DRIBU 64071 20 5 55 0 265 3 1 00 4 2 0 QU 54 Lee eee ee 0 171 632 9 25 68 46 165 4 2 00 Scatt 004 4 7 USARA ICE EERE 0 U TL REA AER 0 227 395 3 36 54 46 123 9 1 63 Scatt 004 4 3 cU qoe eee eee e edes ede 0 Qu Qi eee dee eee 0 235 395 4 23 54 48 123 9 1 63 Figure 8 6 Example file YUOBSDR first part Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 159 way as the finite observation error indicator of the VOF see section 8 2 7 It reveals which observations of the active report have been missing and were replaced by data from the rejected report Rejected multi level reports are written to file YUOBSDR in a format very similar to that for the active
133. Feedback File Definition As the file name starts with fof the feedobs file is also sometimes called fof file Considering the purpose of the NetCDF feedobs file it should contain all the observations that are potentially used for grid point verification This means that it contains many more observation types than actively used in data assimilation i e e g in a LETKF analysis However it does not include all types of exotic observations present e g in an original surface synoptic report and it also fails to contain all the header and complementary information e g on the instrumentation from the originial observation reports Thus the feedobs file cannot be considered a complete surrogate or extension of the original BUFR observation files resp NetCDF observation input files Part VII User s Guide 4 28 Section 8 Model Output 8 4 Output of Forecast Fields 189 8 4 Output of Forecast Fields The results of the model forecast can be written to Grib Edition 1 or NetCDF files The Grib Code is explained in Section 5 1 in more detail The files to which the forecast fields are written obey to the File Name Conventions explained in Section 6 2 Depending on the type of data the filenames get a certain extension p Forecast fields on pressure levels z Forecast fields on geometric z levels s Synthetic satellite images All fields on model levels soil and surface fields are written to a file without
134. G LOG INT LOG LOG INT LOG INT REAL Main switch to include soil processes by running a soil model Main switch to activate the sea ice scheme Introduced in Version 4 10 Main switch to include lake processes by running the lake model FLake Switch to run the new multi layer soil model TERRA ML If Imulti layer FALSE the standard soil model TERRA based on the two layer EFR method is used Switch to run the multi layer snow model If 1multi snow FALSE the standard snow model within TERRA ML is used Introduced in Version 4 11 Number of prognostic soil water levels in the standard soil model TERRA i e only for lmulti layer FALSE At present only nlgw 2 and nlgw 3 are implemented Caution the number of soil water levels must be identical to the number of levels provided in the initial and boundary conditions Switch to include melting processes within the soil when run ning the multi layer soil model TERRA_ML Sub option for the 1melt parameter IF lmelt_var TRUE the soil freezing temperature is treated to be dependent on the water content otherwise a constant freezing temperature is used Number of active soil layers in the new multi layer soil model TERRA ML The total number of layers is ke_soil 1 since a climatological layer with time independent temperature and soil moisture values is added below the lowest active layer Switch to use a minimum stomata resistance map for
135. IB field written on file r2m GRIB field written on file fff GRIB field written on file fff GRIB field written on file NOTE m 1 report EU5185 2 05 hrs 99892 Pa maybe not written to YUVERIF NOTE GPS report ZIM2 LPTR 2 53 hrs Pa maybe not written to YUVERIF NOTE m 1 report EU3421 1 95 hrs 65764 Pa maybe not written to YUVERIF nudge Tq 270 40 276 9854634975543 4 096394061700189E 003 nudge pu 270 40 1775 909540306221 2 396582502917388 nudge Tq 270 20 256 6670491482951 1 044913113842654E 003 nudge pu 270 20 96 66768992558980 19 39489036339958 nudge Tq 270 1 211 5194561088256 3 947515866446530E 006 nudge pu 270 1 766 6450024472723 1 518195750920522 Figure 8 25 Third part of example messages on nudging written to standard output Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 185 or configurations and helps to identify the existence of errors which lead to violations of bit identical reproducibility Note that once such a difference or error is present at a grid point to which observations are assigned then the nudging process will usually spread it very quickly to the whole model domain including point ionl jon1 It is further mentioned again that CAUTION messages related to insufficient array sizes are also written to the standard output see also Section 8 2 5 This includes messages on individual multi level repor
136. IGHT FLAGGED 1 3 HRS STATION 11164 NO SURFACE LEVEL REPORT DERIVED FROM TEMP PILOT STATION 06620 CODE TYPE 132 NOT ON WHITELIST 0 0 HRS STATION 07145 CODE TYPE 137 NOT ON WHITELIST 0 5 HRS FLIGHT TRACK THINNING EU4611 3 173 393 TOO CLOSE TO OBS TIME 3 72 FLIGHT TRACK THINNING EU4611 3 80 330 TOO CLOSE TO OBS TIME 3 77 FLIGHT TRACK THINNING EU4611 3 83 330 TOO CLOSE TO OBS TIME 3 77 FLIGHT TRACK CHECK EU6363 0 80 216 HORIZONTAL CONFIDENCES 83 9 57 9 FLIGHT TRACK CHECK EU3311 0 07 344 HORIZONTAL CONFIDENCES 101 0 58 5 FLIGHT TRACK CHECK EU1234 3 52 238 HORIZONTAL CONFIDENCES 59 2 101 0 EXAGGERATED HORIZONTAL COLOCATION EU0350 El 28 REPORTS FROM 1 68 TO 3 70 FLIGHT TRACK CHECK LHEU0456 4 80 780 HORIZONTAL CONFIDENCES 9 0 0 0 FLIGHT TRACK CHECK EU3268 2 70 376 VERTICAL CONFIDENCES 29 0 81 2 FLIGHT TRACK CHECK EU9145 3 88 290 LON SIGN FOREWARD CONFIDENCE 54 89 FLIGHT TRACK CHECK EU8969 1 61 376 LON SIGN FOREWARD CONFIDENCE 29 89 STA 62023 OBTYP 1 BLACKLISTED Z V T Q 0 o 1100 0 0 0 0 0 STA 16115 OBTYP 1 BLACKLISTED Z V T Q 1100 0 0 0 0 0 0 0 STA EU8742 OBTYP 2 BLACKLISTED Z V T Q 0 0 0 o 1100 0 0 0 STA KLM791 OBTYP 2 BLACKLISTED Z V T Q 0 o 1100 0 0 0 0 0 STA RCH7440 OBTYP 2 BLACKLISTED Z V T Q 0 o 1100 o 1100 0 0 0 STA 62337 OBTYP 5 BLACKLISTED Z2 V T Q 30 o 1100 0 30 0 0 0 STA 34247 OBTYP 5 BLACKLISTED Z2 V T Q 150 0 0 0 150 0 0 0 STA 22522 OBTYP 5 BLACKLISTED
137. II User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files 66 e GPS GNSS zenith total delay and water vapour File name cdfin gps zenith GPS or GNSS Global Navigation Satellite System GPS GLONASS GALILEI reports on zenith total path delay ZTD are obtained via GTS from the UK Met Office as BUFR reports in a template which has been approved by WMO and has the Table D descriptor TM 3 07 022 These reports can be directly converted into NetCDF The following table describes this template and a description can also be found at http egvap dmi dk support formats egvap_bufr_v10 pdf The formats consist of the descriptors in the following table where need COSMO asks stringently for this variable and will abort if variable is absent opt X variable exists and is read used but COSMO will not abort if it does not exist up variable exists but is not read by COSMO GPS descript type mnemonics meaning need 0 01 015 char 20 YSOSN Station or site name need 0 04 001 int MJJJ Year need 0 04 002 int MMM Month need 0 04 003 int MYY Day need 0 04 004 int MGG Hour need 0 04 005 int NGG Minute need 0 05 001 float MLAH Latitude high accuracy degree need 0 06 001 float MLOH Longitude high accuracy deg need 0 07 001 int MHP Height of station m 0 08 021 int MTISI Time signi
138. INGLE LEVEL OBS BEYOND maxsgl 5573 gt INCREASE NAMELIST VARIABLE maxsgo BY AT LEAST 1381 CAUTION t 0 88 LOC MULTI LEV AIRCRAFT REPORTS BEYOND ARRAY SIZE CAUTION t 0 203 LOCAL MULTI LEVEL OBS BEYOND maxmll 244 INCREASE NAMELIST VARIABLE maxmlo BY AT LEAST 292 CAUTION t 0 980 LOCAL GPS IWV OBS BEYOND maxgpl 3344 gt INCREASE NAMELIST VARIABLE maxgpo BY AT LEAST 587 CAUTION t 90 2617 UPPER AIR SINGLE LEVEL OBS INCR ARRAY SIZE 2500 gt INCREASE NAMELIST VARIABLE maxuso BY AT LEAST 117 CAUTION t 162 4372 SURFACE PRESSURE OBS INCREMENTS ARRAY SIZE 4350 gt INCREASE NAMELIST VARIABLE maxsgo BY AT LEAST 22 CAUTION t 270 1374 MULTI LEVEL STATIONS OF OBS INCR ARRAY SIZE 1351 INCREASE NAMELIST VARIABLE maxmlo maxgpo OR maxtvo BY AT LEAST 23 CAUTION t 252 2452 IWV INCREMENTS FOR HUMIDITY CHECK ARRAY SIZE 2350 gt INCREASE NAMELIST VARIABLE maxmlo OR maxgpo BY AT LEAST 102 CAUTION total number of reports 35088 FOF size max rep 25056 gt INCREASE SUM OF NAMELIST VARIABLES maxmlo maxsgo 2 maxgpo 2 maxtvo BY AT LEAST 3344 FOR NetCDF FEEDOBS FILE fof Figure 8 10 Example file YUCAUTN The first group of 4 messages relates to the observation reports themselves After reading the observations to be used at a certain timestep or later on are stored in arrays called observation data record ODR The ODR is a local
139. Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 126 Observation increments from surface level reports Name Type Definition Purpose Comments Default loiqv2m LOG TRUE 2 m humidity observation increments as differences FALSE of specific humidity instead of relative humidity lqfqv2m LOG TRUE quality weight factor for 2 m humidity observations FALSE dependent on 2 m temperature observation increments Threshold quality control Name Type Definition Purpose Comments Default dtqc REAL timestep in s for the threshold quality control QC 720 0 in addition QC is applied to an observation when it is used for the first time QC thresholds v for upper air data qcc 4 REAL constant part v of the thresholds at observation time 0 0 0 0 YIr QT fo p 0 units for I wind v m s surface pressure hPa 0 7 temperature T K relative humidity qcvf 4 REAL multiplication factor f to the vertically varying part wt 5 0 of the QC thresholds f 1 0 the 2 element relates to height thickness checks for 10 0 multi level temperature rather than to surf pressure 0 0 wth is given as a function of pressure p hPa for radiosonde v v Tih and aircraft vtr T wind and temperature p 000 850 700 500 400 300 250 200 150 100 50 vihr 23 23 25 3 0 3 5 3 7 3 5 3 5 34 83 33 vihr 25 25 30
140. Lorenz vertical grid staggering Spatial Discretization Second order finite differences For the two time level scheme also 1st and 3rd to 6th order horizontal advection default 5th order Option for explicit higher order vertical advection Time Integration Two time level 2nd and 3rd order Runge Kutta split explicit scheme after Wicker and Skamarock 2002 and a TVD variant Total Variation Diminishing of a 3rd order Runge Kutta split explicit scheme Option for a second order leapfrog HE VI horizontally explicit vertically implicit time split integration scheme including extensions proposed by Skamarock and Klemp 1992 Option for a three time level 3 d semi implicit scheme Thomas et al 2000 based on the leapfrog scheme Part VII User s Guide 4 28 Section 1 Overview on the Model System 1 2 Basic Model Design and Features 5 Numerical Smoothing 4th order linear horizontal diffusion with option for a monotonic ver sion including an orographic limiter Rayleigh damping in upper layers 2 d divergence damping and off centering in the vertical in split time steps Initial and Boundary Conditions Initial Conditions Interpolated initial data from various coarse grid driving models GME ECMWF COSMO Model or from the continuous data assimilation stream see below Option for user specified idealized initial fields Lateral Boundary Conditions 1 way nesting by Davies type lateral boundary formulation Data
141. MDREP Delayed descriptor replication factor 011075 float n MMPI Mean turbulent intensity EDR 1 011076 float n MPTI Peak turbulent intensity EDR 1 0 11 037 int MTUIN Turbulence index EDR 0 11 039 int NTIED Extended time of occurrence of peak EDR 0 11 077 int NRED EDR reporting interval 0 20 042 int NAICE Ice no ice 0 20 043 float NPLWC Peak liquid water content 0 20 044 float NALWC Average liquid water content 0 20 045 int NSLD Supercooled water droplet conditions Table note 1 The use of n in the variable type definition means that this variable has an additional dimension i e several values may be present in one report If the corresponding replica tion factor MDREP is zero for all reports in the NetCDF file then the variables MMPI and MPTT do not exist in the NetCDF file e Wind Profiler RASS temperature profiler Radar VAD wind profiles File names cdfin_wprof cdfin_rass resp cdfin radar vad Wind Profiler reports WP Radio Acoustic Sounding System temperature profile reports RASS and Radar Velocity Azimuth Display wind profile reports VAD are usually obtained as BUFR reports with a variety of templates At DWD BUFR reports with a unified template for each of the 3 data types are produced from the original reports for subsequent conversion into NetCDF These unified templates are described below Part VII User s Guide 4 28 Section 6 Inpu
142. MLOLO Longitude coarse accur deg need 0 07 030 float MHOSNN Height of station above MSL 1 need 0 07 031 float MHOBNN Height of barometer a MSL 1 opt m 0 33 024 int MSEQM Station elevation quality mark Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files 57 use L S WMO descriptor type mnemonics meaning 3 02 031 Pressure data 3 02 001 Pressure group need 010004 float MPPP Pressure opt 0 10 051 float MPPPP Pressure reduced to MSL opt 0 10 061 float NPPP 3 hour pressure change 0 10 063 int NA Characteristic of press tendency 0 10 062 float NP24 24 hour pressure change opt 0 07 004 float MPN Pressure standard level opt 0 10 009 int NHHHN Geopotential height gpm of the standard level 3 02 035 Basic synoptic instanteous data 3 02 054 SHIP instanteous data s 7 de 2 ES Do Temperature 4 humidity data opt 0 07 032 float MHOSEN Height of sensor above local ground marine deck platform for temp humidity meas opt 007033 float MHAWAS Height of sensor above water surface temp hum meas need 012101 float MTDBT Temperature dry bulb temperat 0 02 039 int MMOWTM Method of wet bulb
143. MMIOGS Identific of orig gen sub centre need need need 0 01 008 char 8 YAIRN Aircraft identification 0 01 006 char 8 YXXNN Aircraft identification F x 0 01 023 int NOSNO Observation sequence number opt 0 02 001 int NIX Type of station 0 02 002 int NIW Instrument type for wind measure 0 02 061 int NS1 Aircraft navigational system need need need 005 001 float MLAH Latitude high accuracy degree need need need 006001 float MLOH Longitude high accuracy deg need need need 0 04 001 int MJJJ Year need need need 0 04 002 int MMM Month need need need 0 04 003 int MYY Day need need need 0 04 004 int MGG Hour need need need 0 04 005 int NGG Minute 0 04 006 int MSEC Second ned 007002 float MHHH Height or altit vert level 1 2 need opt 007004 float MPN Pressure vertical level 1 0 07 007 int MH Height need 0 07 010 int NFLEV Flight level 1 2 need 0 10 070 int MIAA Indicated aircraft altitude 1 2 opt 008 009 int NDEPF Detailed phase of flight 3 need need need 0 11 001 int NDNDN Wind direction need need need 0 11 002 float NFNFN Wind speed opt opt 0 11 031 int MB Degree of turbulence 0 11 032 float MHBT Height of base of turbulence E 011033 float MHTT Height of top of turbulence opt opt 0 11 036 float NMDEWX Derived equiv
144. N coincides with the surface level of the sounding Therefore for the three T EMP types and unlike for other observation types MHOSNN is preferred to exist and be used if both variables have non missing values n n in the variable type definition means that this variable has an additional dimension used here for the vertical levels and hence several values may be present in one report If the corresponding replication factors MEDRE or MDREP are zero for all reports in the NetCDF file then the corresponding multi dimensional variables do not need to exist and probably will not exist in the NetCDF file e PILOT either with height or pressure as vertical coordinate File names for PILOT with height as vertical coordinate edfin pilot for PILOT with pressure as vertical coordinate cdfin_pilot_p Irrespective of the vertical coordinate used the templates for PILOT from fixed land stations PILOT MOBIL from mobile land stations as well as PILOT SHIP from sea stations are identical The common sequence descriptors TM 3 09 050 and TM 3 09 051 are used for PILOTS with pressure resp height as vertical coordintate These templates are identical to each other except for the vertical coordinate TM 3 09 050 contains pressure but lacks height as a variable whereas for TM 3 09 051 the variable pressure does not exist and the vertical level is expressed as height Part VII User s Guide 4 28 Sec
145. Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 128 Use of observations and reports continued Name Type Definition Purpose Comments Default observation area all reports outside of it are neglected obnlat REAL latitude of northern boundary of observation area deg 90 obslat REAL latitude of southern boundary of observation area deg 90 obwlon REAL longitude of western boundary of observation area deg 180 obelon REAL longitude of eastern boundary of observation area deg 180 Use of observation types and code types Name Type Definition Purpose Comments Default exclusion area reports inside of it are neglected if their observation or code type is set to FALSE exnlat REAL latitude of northern boundary of exclusion area deg 90 exslat REAL latitude of southern boundary of exclusion area deg 90 exwlon REAL longitude of western boundary of exclusion area deg 180 exelon REAL longitude of eastern boundary of exclusion area deg 180 observation type reports inside the exclusion area are set passive if their type is set to FALSE lsynop LOG observation type SYNOP TRUE laircf LOG observation type AIREP aircraft TRUE lsatob LOG observation type SATOB FALSE ldribu LOG observation type DRIBU drifting buoy TRUE ltemp LOG observation type TEMP TRUE lpilot LOG observat
146. OG To switch on a diagnostic initialization of rain and snow in FALSE case that no initial data are given Up to COSMO Model Version 3 6 the parameter itype_gscp was used to switch on off a prognostic treatment of rain and snow but only for the 2 timelevel Runge Kutta scheme for irunge kutta 0 Now only the specific kind of parameterization scheme can be chosen with itype gscp and there are two additional parameters to control the prognostic precipitation Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 4 PHYCTL Parameters for the Diabatic Model 96 Radiation These parameters control the Ritter amp Geleyn radiation parameterization scheme of the COSMO Model Here among others the frequency of the calculation the surface albedo the back ground aerosols and the greenhouse gas concentration scenario can be specified Name Type Definition Purpose Comments Default lrad LOG Main switch for including radiation If TRUE the model TRUE is run with the radiation scheme which computes heating rates in the atmosphere solar and thermal and the energy balance solar and thermal at the ground To save comput ing time the radiation scheme will be called at certain time intervalls defined by nincrad or hincrad Between two con secutive calls of the scheme the radiative heating rates are kept constant nincrad INT Interval in time steps between two cal
147. Q 90Qd dock e BOQESERE 1 090 Qua esee 0 335 322 14 44 50 02 35 5 2 25 SYNOP 11520 P 2 1 2 2 273 4 0 91 2983 00 303 303 1 0 0 10 7 5 1 335 322 14 44 50 02 111 0 00 SYNOP 11520 2 1 2 2 273 4 0 91 983 00 303 303 1 0 0 10 7 5 1 335 322 14 44 50 02 111 0 00 SYNOP 11520 2 1 2 2 273 6 0 92 983 70 303 303 1 0 0 10 7 5 1 335 322 14 44 50 02 111 1 00 SYNOP 11520 1 2 2 8 274 0 0 90 984 40 303 303 1 0 0 10 7 5 1 335 322 14 44 50 02 111 2 00 TEMP 11520 OBS CODE TYPE 5 35 STA MOD HEIGHT 302 301 LON 14 44 LAT 50 02 G P 335 322 HOUR 2 3 v E rh p z v err t er rh er z er lev 0 9 2 8 274 0 0 90 985 0 302 xde e xe 4 3 64 2 8 6 4 d xd eek 958 0 2 uec 128 RRA 271 0 1 00 945 pce 1 1 0 10 256 1 2 7 9 269 9 1 00 925 0 803 2 2 1 1 0 10 4 3 32 20 9 8 266 3 0 93 850 0 1468 2 4 1 0 0 10 4 4 32 P g 4kxdo eek 800 0 2 4kkkkkkkkkkkkkkk 32 ET A RE 165 0 2 GARA RRA RARA 128 KRRKKEKREREEER 258 5 0 90 738 Ok eKRKARKEEEE 0 8 0 10 256 RRRKKKKREKEEEE 259 9 0 89 732 Oe eRe KARKEEEE 0 8 0 10 256 5 4 10 7 257 3 0 59 700 0 2956 2 5 0 7 0 10 5 2 32 10 0 11 2 xxx 600 09 2 9kkkkkkkkkkkkkkk 32 ocekeekeeeeee 247 3 0 25 568 0Xxx 0 5 0 10 256 14 3 19 3 241 4 0 14 500 0 5410 3 4 0 4 0 15 8 4 32 Xocekeekeeeee 240 9 0 09 AT3 OR RRRKARKEEER 0 4 0 15 256 ocexdekee eee 241 1 0 11 469 0 0 4 O 15 256 17 8 29 0 233 4 0 12 400 0 6970 3 5 0 4 0 15 9 8 33 oce eee e 226 3 0 20 344 0 0
148. SMO Model Sensor Satellite Channel Central Wavelength MVIRI METEOSAT 7 1 WV6 4 MVIRI METEOSAT 7 2 IR11 5 SEVIRI MSG 1 2 4 IR3 9 SEVIRI MSG 1 2 5 WV6 2 SEVIRI MSG 1 2 6 WV7 3 SEVIRI MSG 1 2 7 IR8 7 SEVIRI MSG 1 2 8 IR9 7 SEVIRI MSG 1 2 9 IR10 8 SEVIRI MSG 1 2 10 IR12 1 SEVIRI MSG 1 2 11 IR13 4 Two fields have been implemented into the COSMO Model to take care of the output of the synthetic satellite images These fields can be specified by the shortnames SYNME7 output of the products for MVIRI METEOSATT SYNMSG output of the products for SEVIRI MSG1 or MSG2 Although special channel and products can be chosen via Namelist the implementation is such that all channels and all products of a special instrument are computed Computation of the SynSat products can be controlled by the following namelist parameters Name Type Definition Purpose Comments Default itype rttov INT To specify which RTTOV Version shall be used 7 Possible values 7 9 num sensors INT Number of sensors used during the calculation 0 lcon clw LOG To specify whether convective liquid water shall be used in the FALSE computations lsynsat LOG To activate computation of synthetic satellite images default FALSE behaviour from former versions before 4 26 lobsrad LOG To activate satellite observation processing This can only be FALSE done if the model is compi
149. Section 2 Introduction 10 Section 3 Model Formulation and Data Assimilation 3 1 Basic State and Coordinate System The COSMO Model is based on the primitive hydro thermodynamical equations describing compressible nonhydrostatic flow in a moist atmosphere without any scale approximations A basic state is subtracted from the equations to reduce numerical errors associated with the calculation of the pressure gradient force in case of sloping coordinate surfaces The basic state represents a time independent dry atmosphere at rest which is prescribed to be horizontally homogeneous vertically stratified and in hydrostatic balance By introducing the basic state the thermodynamic variables temperature T pressure p and density p can be formally written as the sum of a height dependent reference value and a space and time dependent deviation T T z T p polz p p pb 2 tP 3 1 where To z po z and po z are related by the hydrostatic equation Opo gpo 3 2 De gpo RaT 3 2 and the equation of state po poRaTo Ra is the gas constant of dry air In principle the vertical profile To z of temperature can be specified arbitrary since we do not linearize the model equations with respect to the reference state For practical reasons we prescribe a constant rate 3 for the temperature increase with the logarithm of pressure as proposed by Dudhia 1993 0T0 0 ln po B The integration of the hydrostatic equation
150. U P 6 5 42 Hee REE seisesioeeee 922865 0 247 1 4 0 00 46 279 253 9 09 45 80 80012 0 50 GPS COMO METO P 6 4 41 2 xoc 2285 1 247 1 77 0 00 46 279 253 9 09 45 80 80012 0 75 GPS COMO ROB PRIRRERERAIAARERAEERARERRARARES 2279 8 247 1 7 FERRER 46 279 253 9 09 45 80 83712 0 50 GPS COMO ROB postcode FRE RR 0278 1 24 1 28 FRESE 46 279 253 9 09 45 80 83712 0 75 GPS COMO ASIL PL EREARRS FIRES E SER oboe EE RE On qa cg ZAT 2 4 SEERE 46 279 253 9 09 45 80 82112 0 50 GPS COMO ASI PRIREAERRERA RACER AEESARERAAERRES 0261 2 247 2 6 FERRE 46 279 253 9 09 45 80 82112 0 75 GPS COMO SGON PREFEAERRA AER CER RERRA CERRAR 2275 0 24 2 2 RARE 46 279 253 9 09 45 80 82912 0 50 GPS COMO SENL_ xdexoee d oboe RAICES FR ERR 02734 247 25 KERR 46 279 253 9 09 45 80 82912 0 75 GPS COMO SIGNI Poe Ae eR EER Ee sop RE RHE DD To ZAT 2 3 Nee 46 279 253 9 09 45 80 92912 0 50 GPS COMO SCNI Posse seater ea koe se ree O I E 022712 6 247 2 8 FIERA 46 279 253 9 09 45 80 92912 0 75 GPS GOMO BKG FFAECEREEREAEAEREEERERERRAERRESR 0273 6 247 1 0 Lad kd 46 279 253 9 09 45 80 83012 0 50 Figure 8 7 Example file YUOBSDR second part Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 160 Figure 8 6 provides examples of a TEMP radiosonde a wind profiler and a RASS multi level report For multi level reports a self explanatory header line is followed by lines each providin
151. V specific water vapor content in g kg QC specific cloud water content in g kg QI specific cloud ice content in g kg REL HUM relative humidity in CLC cover with grid scale clouds in CLC_CON cover with convective clouds in U zonal wind component in m s V meridional wind component in m s SPEED wind speed in m s HML height of model main levels in m For the half levels the following parameters are printed Phalf pressure in hPa of the half levels W vertical velocity TKVM turbulent diffusion coefficients for momentum TKVH turbulent diffusion coefficients for heat and moisture HHL height of model half levels in m In addition the values of several near surface and soil parameters are printed 8 1 2 YUSPECIF NAMELIST parameters YUSPECIF contains all NAMELIST variables their default and their actual values At the end of this file the vertical coordinate parameters sigma k and the values of the reference atmosphere are printed YUSPECIF is always written 8 1 3 YUCHKDAT Checking the Grib input output data YUCHKDAT contains information about fields that are read from or written to GRIB files For every field the maximum value the minimum value together with the corresponding indices and the mean value are written This output can be controlled with the NAMELIST parameters e lchkini in gribin check the initial data e lchkbd in gribin check the boundary data e lcheck in gribout check
152. VERAGRE CHANGE 0 3113 STANDARD DEVIATION 0 8121 MAX CHANGE x 3 9370 MAX POSITIVE CHANGE 3 9370 AT MODEL LAT LON 3 3500 0 4500 MAX NEGATIVE CHANGE 1 9850 AT MODEL LAT LON 4 7750 3 8250 zdistm 14959 zrormx 0 374 wwa 0 623 zdistm 26371 zrormx 0 659 wwa 0 217 zmodor 462 rpaght 540 zvdis 78 zvdmax 462 zwi 0 931 wwa 0 511 zmodor 462 rpaght 445 zvdis 17 zvdmax 462 zwi 0 997 wwa 0 210 Diagnostic PRECIPITATION analysis scan 1 grid point ix 167 iy 103 lat 47 54 lon 8 74 height 462 Observations influencing this grid point stat_id lat lon height increment weight 06679 47 48 8 90 540 11 00 0 51 Q942 47 77 8 82 445 2 00 0 21 Resulting prc analysis increment 7 61 STATISTICS ON ANALYSIS OF PRECIPITATION AMOUNT UNIT mm SUM OF CHANGES 497251 8071 NO OF ANAL POINTS 194081 AVERAGRE CHANGE 2 5621 STANDARD DEVIATION 8 7602 MAX CHANGE 137 0000 MAX POSITIVE CHANGE 137 0000 AT MODEL LAT LON 5 3250 0 7250 MAX NEGATIVE CHANGE 0 0000 AT MODEL LAT LON 5 0000 2 8000 Figure 8 28 Example file YUSURF second part on current analyses Part VII User s Guide 4 28 Section 8 Model Output 8 3 NetCDF Feedobs File 188 8 3 NetCDF Feedobs File For data assimilation or verification purposes a special NetCDF feedobs file sometimes also mis called feedback file can be written This file contains the observations themselves from a specifiable perio
153. Y ux veos ez tex where is the contra variant vertical velocity in the C system Ja TE w JG acosy a D is the three dimensional wind divergence which is calculated from 1 E JA Ou haces deos dio Ax 1 Ow acosplOA vaot dy n VG 8C VG OC In deriving the prognostic equation for the perturbation pressure from the continuity equa tion a source term due to diabatic heating has been neglected For most meteorological applications this source term is much smaller than the forcing by divergence This approxi mation is also used in many other nonhydrostatic simulation models 3 3 Horizontal and Vertical Grid Structure The model equations 3 5 are solved numerically using the traditional finite difference method In this technique spatial differential operators are simply replaced by suitable finite Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 3 Horizontal and Vertical Grid Structure 13 difference operators The time integration is also by discrete stepping using a fixed timestep At The terrain following coordinate system with the generalized vertical coordinate allows to map the irregular grid associated with the terrain following system in physical space onto a rectangular and regular computational grid Thus constant increments AA grid spacing in longitudinal direction Ap grid spacing in latitudinal direction AC grid spacing in directio
154. a dynamical bottom boundary condition FALSE Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 3 DYNCTL Parameters for the Adiabatic Model 93 Additional numerical filters Name Type Definition Purpose Comments Default epsass REAL Value of the filter coefficient e in the Asselin time filter to damp 0 15 the computational mode in the 3 time level Leapfrog scheme used only for 12t1s FALSE Caution both too small e lt 0 01 and too large values e gt 0 25 may cause instabilities alphaass REAL Weight for Williams 2009 modification to the Asselin time filter 1 0 0 5 alphaass lt 1 betasw REAL Value of the P parameter used for time weighting of the future 0 40 values in the vertically implicit treatment of acoustic sound waves 3 0 gives a time centred average with no damping DB 1 results in a fully implicit vertical scheme with strong damping of acoustic and gravity wave modes Slight positive off centering is recommended to damp vertically propagating sound waves betagw REAL Same as betasw but used for gravity waves 0 40 beta2sw REAL Same as betasw but used in the p 7 dynamics for sound 0 40 waves beta2gw REAL Same as betagw but used in the p T dynamics for gravity 0 40 waves xkd REAL Coefficient for divergence damping 0 1 Spectral Nudging Name Type Definition Purpose Comments Default lsp
155. al applications in climate simulations Mon Wea Rev 120 303 325 Sch r C D Leuenberger O Fuhrer D L thi and C Girard 2002 A new terrain following vertical coordinate formulation for atmospheric prediction models Mon Wea Rev 130 2459 2480 Schraff C 1996 Data assimilation and mesoscale weather prediction A study with a forecast model for the Alpine region Publication 56 Swiss Meteorological Institute Z rich Schraff C 1997 Mesoscale data assimilation and prediction of low stratus in the Alpine region Meteorol Atmos Phys 64 21 50 Skamarock W C and J B Klemp 1992 The stability of time split numerical methods for the hydrostatic and the nonhydrostatic elastic equations Mon Wea Hev 120 2109 2127 Skamarock W C P K Smolarkiewicz and J B Klemp 1997 Preconditioned conjugate residual solvers for Helmholtz equations in nonhydrostatic models Mon Wea Rev 125 587 599 Sommeria G and J W Deardorff 1977 Subgrid scale condensation in models of non precipitating clouds J Atmos Sci 34 344 355 Stauffer D and N Seaman 1990 Use of four dimensional data assimilation in a limited area mesoscale model Part I Experiments with synoptic scale data Mon Wea Rev 118 1250 1277 Stauffer D and N Seaman 1994 Multiscale four dimensional data assimilation J Appl Meteor 33 416 434 Stephan K S Klink and C Schraff 2008 Assimilation of radar derived rain
156. albedo 1 This parameter has been introduced in Version 4 23 1 surface albedo is a function of soiltype method up to now and still default 2 surface albedo is determined by two external fields for dry and for saturated soil 3 A background albedo is prescribed by external fields which give average values for every month 4 The vegetation albedo is modified by forest fraction Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 4 PHYCTL Parameters for the Diabatic Model 98 Moist Convection These parameters specify the convection parameterization used In particular at resolutions below 3 km the deep convection parameterization should be switched off since this process is mainly a grid scale process for such high horizontal resolutions Name Type Definition Purpose Comments Default lconv ltiedtke lkainfri lshallow itype_conv nincconv lconf avg lcape lctke lconv inst LOG LOG LOG LOG INT INT LOG LOG LOG LOG Main switch for including subgrid scale convection If TRUE the model is run with a moist convection parameterization which computes the effect of moist convection on temperature water vapour and horizontal wind in the atmosphere and the precipitation rates of rain and snow at the ground To save computing time the convection scheme may not be called ev ery time step but at certain intervals define
157. alues are used at the boundary zone 1 A zero gradient condition is used for qr qs qg 2 qr qs qg are set to 0 0 at the boundary zone 3 No presetting is done in the whole the boundary zone must not be chosen for Leapfrog applications lradlbc LOG To run with radiative lateral boundary conditions if TRUE FALSE or with Davies conditions if FALSE In Version 4 5 this switch has been moved from the group RUNCTL to DYNCTL relax fac REAL reduction factor for strength of lateral boundary relaxation 0 01 only if radiative lateral boundary conditions are used Switch has been introduced in Version 4 5 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 3 DYNCTL Parameters for the Adiabatic Model 91 Horizontal diffusion m between adjacent gridpoints in the orographic flux limiter With increasing height difference the diffu sive fluxes are gradually decreased to become zero at hd dhmax Active only for itype hdiff 2 Caution The parameter hd dhmax should be adjusted when changing the grid spacing Name Type Definition Purpose Comments Default lhordiff LOG Main switch to include horizontal diffusion TRUE itype hdiff INT Parameter to select a scheme for horizontal diffusion 2 1 Regular 4th order linear scheme 2 Monotonic 4th order linear scheme with orographic limiter lhdiff mask LOG has been eliminated in
158. are always written in pure binary format The basic namelist settings to control the format are yform read string and yform write string where string can be either gt grbi or ncdf 5 1 The GRIB Binary Data Format GRIB means gridded binary and is designed for the international exchange of processed data in the form of grid point values expressed in binary form The GRIB code is part of the FM system of binary codes of the World Meteorological Organization WMO Currently we use Edition 1 of the GRIB code with number FM 92 VIII For coding details see the Manual on Codes International Codes Volume 1 2 of WMO WMO Publication No 306 1995 In this section we describe only the basic features of the GRIB code which are relevant for the I O of the COSMO system 5 1 1 Code Form Each GRIB coded record analysis or forecast field consists of a continuous bit stream which is made up of a sequence of octets 1 octet 8 bits The representation of data by means of series of bits is independent of any particular machine representation The octets of a GRIB messsage are grouped in sections see Table 5 1 where the length of the record and the length of the sections are expressed in octets Section 0 has a fixed length of 8 octets and Section 5 has a fixed length of 4 octets Sections 1 2 3 and 4 have a variable length which is included in the first three octets of each section Octets are numbered 1 2 3 etc starting at the
159. are assigned their default values offered by FLake Since no tile approach is used in the COSMO model i e each COSMO model grid box is characterised by a single land cover type only the grid boxes with the lake fraction in excess of 0 5 are treated as lakes Each lake is characterised by its mean depth Deep lakes are currently treated with the false bottom That is an artificial lake bottom is set at a depth of 50 m The use of such expedient is justified since strictly speaking FLake is not suitable for deep lakes because of the assumption that the thermocline extends down to the lake bottom However as the deep abyssal zones typically experience no appreciable temperature changes using the false bottom produces satisfactory results A Global Land Cover Characterization GLCC data set http edcdaac usgs gov glcc with 30 arc sec resolution that is about 1 km at the equator is used to generate the lake fraction filed The filed of lake depth is generated on the basis of a data set developed at DWD that contains mean depths of a number of European lakes and of major lakes of the other parts of the world Notice that unless tile approach is used to compute the surface fluxes only the lake depth external parameter filed is actually required to use FLake within the COSMO model Setting the lake depth to its actual value for the COSO model grid boxes with the lake fraction in excess of 0 5 and to a negative value say 1m otherwise unambi
160. ata used value ok 3 only passive data used value not ok 4 no data at all usable 15 x resp y coordinate of model grid point to which report is assigned The station characteristics is given by the following bit pattern the description related to a bit number is true only if that bit takes the value of 1 bit numbers of station characteristics 0 NOoR WN HP 9 13 20 21 22 26 single level report set passive because it is used as part of a multi level report report set passive because at least 1 flag at positions 2 6 or 20 21 is set flag station location outside of user specified area flag distance between model orography and station altitude too large flag suspicious aircraft identity flag observation or code type excluded at station location user specified flag redundant report report located at sea grid point station correction indicator 10 station suspicion indicator important station indicator 19 instrument specification word only for obs reports read from AOF file flight track error flag flight thinning flag 25 indicator for phase of flight aircraft code table WMO descriptor 0 08 004 2 unsteady 3 4 level flight 5 ascending 6 descending 7 missing value 27 aircraft roll angle WMO descriptor 0 02064 not used for report status Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations
161. ation of the COSMO Model 4 3 Preparing the Code 32 The following files and subdirectories are created CompilerFlags To specify which module is compiled with which set of compiler options FileNames To define the names of binaries and or libraries LinkLibs To define the libraries for the link step Makefile Link to a makefile for compiling and linking Options To set the compiler and linker options Parallel To set the number of parallel tasks for compiling edid Script to edit the SCCS decks mk batch Script to submit a batch job optional obj Directory containing object files of the files in src src Directory containing modified source files work Directory containing files you are working on Normally correct defaults are set by your administrator You can change Options Parallel and LinkLibs according to your needs see also the part for the Source Code Administrator User without VCS If the VCS is not available you have got a tar file cosmo_yymmdd_x y where yymmdd de scribes the date in the form Year Month Day and x y gives the version number as in the DWD Version Control System VCS By de taring a directory is created with the following contents DOCS Contains a short documentation of the changes in version x edid Script to edit files in src and store them in work Fopts Definition of the compiler options and also directories of libraries LOCAL Contains several examples of Fo
162. ational Alphabet No 5 Thus the beginning and the end of a GRIB record can be identified by the character coded words GRIB and 7717 All other octets included in the code represent data in binary form Each input or output array defined on the rotated lat lon grid of the COSMO model e g the surface pressure or the temperature at a specified model level is coded as a GRIB record Various records can be combined in a single GRIB file At DWD still a modified Grib file format is used because of historical reasons In Grib 0 there was no information about the length of a record Therefore DWD added additional 4 bytes before the GRIB and after the 7777 the so called controlwords Also 4 bytes were added at the beginning and at the end of the file DWD plans to remove these additional bytes in the future but at the moment a correct processing of Grib files is possible only with these controlwords To ensure this an environment variable has to be set before running programs linked with the DWD Grib library export LIBDWD_FORCE_CONTROLWORDS 1 Part VII User s Guide 4 28 Section 5 Data Formats for I O 5 1 The GRIB Binary Data Format 37 Table 5 2 Contents of the Product Definition Section array Octet Contents of PDS ipds i number Value Remarks 1 1 3 54 Length of the PDS in octets 2 4 2 Version number of the GRIB indicator table see Table 5 3 3 5 78 Identification of originating gen
163. c Namelist settings 12t1s FALSE lsemi imp FALSE This method is a variant of the scheme which is based on a Leapfrog integration for the slow modes from time level n 1 to time level n 1 using an integration interval of 2At The slow mode tendencies are evaluated at time level n for horizontal advection using standard second order centered differences and at time level n 1 for most physical forcings Vertical advection and vertical diffusion are calculated by a quasi implicit scheme The integration step is then subdivided into a number N of small time steps Ar according to 2At N AT and the prognostic equations 3 11 are stepped forward according to rt uar SAT foAT 3 12 Figure 3 4 illustrates the basic idea of the time splitting scheme In the integration of 3 12 sound waves are treated explicitly for horizontal directions using the forward backward method while implicitly for the vertical direction HE VI Thus the small time step Ar is limited by the CFL stability criterion for horizontal but not for vertical sound wave prop agation This makes the HE VI scheme numerically very efficient for large grid aspect ratios ie Ax Az gt 1 which are typically used in meso and meso y applications Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 5 Physical Parameterizations 18 Figure 3 4 The time splitting algorithm An additional 3 D divergence damping as well as a sli
164. contains control parameters for the DFI initialization These parameters are only in effect if the main switch 1dfi in RUNCTL is set to TRUE Name Type Definition Purpose Comments Default ndfi INT Indicator for kind of filtering 0 0 No filtering is done 1 Foreward stage filtering launching using a diabatic foreward integration 2 Two stage filtering using an adiabatic backward integration followed by a diabatic foreward integration nfilt INT Indicator for method of filtering 1 1 Dolph Chebyshev filter is used no other filter is implemented tspan REAL Time span in seconds for the adiabatic and diabatic stages of 3600 0 the initialization Caution The time span has to be less or equal to the time the first boundary data set is provided dtbak REAL Time step s for the backcast filtering stage 90 0 It is recommended to set dtbak dt dtfwd REAL Time step s for the forecast filtering stage 90 0 It is recommended to set dtfwd dt taus REAL Cuttoff time period in seconds of the filter High frequency 3600 0 components with periods less than taus are filtered It is recommended to set taus to a value smaller or equal to tspan Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 114 7 9 NUDGING Controlling the Data Assimilation The NAMELIST group NUDGING is required
165. convective mm RS snow amount grid scale and convective mm WS water content of snow m GRID POINT Frankfurt Flughafen Ts 23 J 70 Initial Date SUN 17 02 2008 18 UTC HSURF es 111 208 FR LAND 100 000 LAT dgr 50 055 LON dgr 8 637 SOIL TYPE sandy loam 4 HH PMSL DF10M DF500M DF850 DF700 DF500 TG T2M h hpa dgr m s dgr m s dgr m s dgr m s dgr m s 18 0 1039 60 234 1 271 4 342 5 344 7 343 13 0 5 dd 19 0 1039 67 244 1 271 4 332 6 341 7 348 14 1 0 0 5 20 0 1039 35 241 2 276 4 332 6 338 8 349 14 1 4 1 1 21 0 1039 12 237 2 280 4 327 6 346 9 349 15 2 0 1 8 SUN 17 02 2008 18 UTC TD2M T30M T850 degree centrigrade 9 6 8 9 8 3 7 6 CONN O1 00 OU oo0oo 3 23 al oak T700 Sel 7 0 0 6 64D T500 HML octas 21 9 000 22 0 100 21 8 100 22 0 100 OOo HBAS HTOP 10 m OOOO oo0oo RR 0 00 0 00 0 00 0 00 mm RS 0 00 0 00 0 00 0 00 WS 0 000 0 000 0 000 0 000 Example file M stationname with short grid point output Figure 8 1 Section 8 Model Output Part VII User s Guide 4 28 8 1 ASCII Output for the Forecast Model 150 GRID POINT Frankfurt Flughafen I 23 J 70 Initial Date SUN 17 02 2008 18 UTC HSURF m 111 208 FR LAND 100 000 LAT dgr 50 055 LON dgr 8 637 FC 1E4 s 1 119 SOIL TYPE sandy loam 4 Actual date SUN 17 02 2008 18 UTC 0 0 time step 0 PS hpa 1025 3 DPSDT
166. ctl the first output and the interval of the outputs in time steps can be controlled With ldump_ascii TRUE FALSE the flushing of YUPRMASS to disk in every time step can be switched on off An example of YUPRMASS is shown in Figure 8 3 8 1 5 YUPRHUMI Protocolling the forecast with humidity variables YUPRHUMI contains meanvalues of model variables related to humidity and rain rates First the initial values of the humidity variables are written In the next lines the following values are written ntstep actual time step Real elapsed time since the last line was printed qc area mean value of cloud water qi area mean value of cloud ice qr area mean value of rain qs area mean value of snow qg area mean value of graupel prrs precipitation rate for grid scale rain prss precipitation rate for grid scale snow prrk precipitation rate for convective rain Part VII User s Guide 4 28 Section 8 Model Output 153 8 1 ASCII Output for the Forecast Model Experiment LM Initial mean values surface pressure hP 967 278 Experiment LM ntstep Real dpsdt S Pa s E 2 0 3 680 22 809 1 200 25 425 2 0 360 17 142 3 0 370 14 223 4 250 1 841 5 0 370 12 827 6 0 360 11 485 y 0 360 1 888 8 1 260 L0 9 17 9 0 400 11 277 0 0 360 0 802 11 1 270 10 548 2 0 360 10 328 3 0 360 0 335 14 0 360 9 545 5 260 8 093 16 0 360 9217 L7 0 360 8 472 8 0 360 8 484 19 1 360 8 104 20 0 360 7 968
167. d atmosphere Hence flight level is not the geometrical height Once converted the original resolution either 100ft or 10ft in the BUFR report is lost hence it is desirable to disseminate the element in the received form This phase of flight table is expanded to indicate wind quality from roll angle or roll and pitch combined and also to indicate the method of ascent and descent observation interval selection either by time or pressure increments The use of n in the variable type definition means that this variable has an additional dimension i e several values may be present in one report If the corresponding replica tion factor MDREP is zero for all reports in the NetCDF file then the variables MMPT and MPTT do not exist in the NetCDF file e multi level AMDAR File name cdfin amdar ml Caution The use of this observation file type in COSMO has not been tested thoroughly yet mainly due to a lack of data of this type at least over Europe The use of it is at the user s own risk The template follows the proposed WMO descriptor TM 3 11 009 and is described in the table below which is split into two parts for convenience There is also a proposed descriptor TM 3 11 008 for aircraft profiles without latitude and longitude reported at each level Descriptor TM 311008 equals TM 311009 except that 311007 is replaced by 3 11006 which in turn is the same as 3 11007 except that 301021 is
168. d the integration is done for variables in the interior compu tational domain i 3 N 2 and j 3 Ny 2 3 4 Numerical Integration Because the governing nonhydrostatic equations describe a compressible model atmosphere meteorologically unimportant sound waves are also part of the solution As acoustic waves are very fast their presence severely limits the time step of explicit time integration schemes In order to improve the numerical efficiency the prognostic equations are separated into terms which are directly related to acoustic and gravity wave modes and into terms which refer to comparatively slowly varying modes of motion This mode splitting can formally be written in the symbolic form Ov Bt Sy Ius 3 11 where v denotes a prognostic model variable fy the forcing terms due to the slow modes and sy the source terms related to the acoustic and gravity wave modes sy is made up Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 4 Numerical Integration 17 of the pressure gradient terms in the momentum equations the temperature and pressure contributions to the buoyancy term in the w equation and the divergence term in the pres sure and the temperature equation The subset of equations containing the s terms is then integrated with a special numerical scheme The COSMO Model provides four different in tegration methods 3 4 4 Runge Kutta 2 timelevel HE VI Integra
169. d uses a specific grouping of various cloud and precipitation particles into broad categories of water substance The particles in these categories interact by various microphysical processes which in turn have feedbacks with the overall thermodynamics Microphysical processes are parameterized by corresponding mass transfer rates between the categories and are formu lated in terms of the mixing ratios as the dependent model variables Besides water vapour in the gaseous phase three categories of water are considered by the default scheme cloud water is in the form of small suspended liquid phase drops Cloud droplets are smaller than about 50 um in radius and thus have no appreciable terminal fall speed relative to the airflow rain water is in the form of liquid phase spherical drops which are large enough to have a non negligible fall velocity An exponential Marshall Palmer size distribution is assumed for the raindrops and a drop terminal velocity depending only on drop diameter is prescribed Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 5 Physical Parameterizations 20 Snow is made up of large rimed ice particles and rimed aggregates which are treated as thin plates with a specific size mass relation Particles in this category have a non negligible terminal velocity which is prescribed to depend only on particle size n exponential Gunn Marshall size distribution is assumed The b
170. d REAL end of verification period in hours 0 0 mveripr INT type of verification observation file s written 3 0 no file written equivalent to lverif false 1 NetCDF feedobs feedback file for EnKF or verif 2 ASCII file VOF YUVERIF 3 both NetCDF feedobs and ASCII VOF files mruntyp INT type of current model run used for increments written to the 1 NetCDF feedobs and or ASCII VOF verification files no increments written to VOF 2 increments from current assimilation run 40 increments from current forecast run lverpas LOG on off switch for writing also passive reports to the verification TRUE file s Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 118 Corrections to balance the analysis increments Name Type Definition Purpose Comments Default ntpscor ptpstop INT REAL temperature correction density pressure balancing switch for hydrostatic temperature correction to balance the surface pressure analysis increments Ap such that the implied upper level pressure increments Ap decrease with increasing height according to the functions given below and become zero for p lt Prop 0 no hydrostatic temperature correction 3 xe Ap ays _ _P Pto 1 Ap 1 gt m a gt P 2 Ptop l p P Pto Ap 3m n gt V Ee vo
171. d by nincconv Between two consecutive calls of the scheme the convective tendencies and precipitation fluxes are kept constant Eliminated replaced by itype conv in Version 4 4 Eliminated replaced by itype conv in Version 4 4 Eliminated replaced by itype conv in Version 4 4 To specify the type of convection parameterization 0 Tiedtke scheme 1 This option has been eliminated 2 This option has been eliminated 3 Shallow convection based on Tiedtke scheme Interval in time steps between two calls of the convection scheme Switch to apply a horizontal smoothing of the convective forc ings moisture convergence surface moisture flux and vertical velocity prior to calling the convection scheme Enables a CAPE type closure within the Tiedtke convection scheme not fully tested yet Enables a turbulent kinetic energy closure within the Tiedtke scheme not fully tested yet Switch to write instantaneous TRUE or aggregate FALSE values of top con and bas_con to model output FALSE TRUE FALSE FALSE FALSE Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 4 PHYCTL Parameters for the Diabatic Model 99 Vertical turbulent diffusion There is one main switch to include turbulent diffusive fluxes 1tur which activates both the parameterization of turbulent diffusion in the atmosphere and of surface layer fluxes A further selection of scheme
172. d definition DWD uses bit maps to send only those data from the global model GME to the national weather services running the COSMO Model that are needed for the corresponding domain 5 1 6 Binary Data Section This section contains all values of the defined grid usually in a packed format At DWD typically 16 bits are used to store a packed value Part VII User s Guide 4 28 Section 5 Data Formats for I O 5 1 The GRIB Binary Data Format A1 Table 5 8 Contents of the Grid Description Section array Octet Contents of GDS igds i number Value Meaning 1 1 3 202 Length of GDS in octets including the vertical coordinate parameters here for ke 35 layers i e ke 1 36 half levels 2 4 40 NV Number of vertical coordinate parameters four base state parameters 4 ke 4 1 values of the vertical coordinates of the half levels 3 5 43 PV Location octet number of the list of vertical coordinate parameters 4 6 10 Data representation type according to WMO code table 6 10 assigns a rotated latitude longitude grid 5 1 8 325 Number of gridpoints in zonal direction 6 9 10 325 Number of gridpoints in meridional direction T 11 13 17000 Rotated latitude of the first gridpoint in millidegrees 8 14 16 12500 Rotated longitude of the first gridpoint in millidegrees 9 17 0 Resolution flag according to WMO code table 7 0 means that the grid spacing is not
173. d in Section 6 4 In practical applications the nudging term usually remains smaller than the largest term of the dynamics so that the dynamic balance of the model is not strongly dis turbed The coupling between the mass and wind field innovations is primarily induced implicitly by the model dynamics If the assimilation process is successful the model fields will be close to dynamic balance at the beginning of the forecast and an initial ization step is not required Latent Heat Nudging Basic Namelist setting llhn TRUE Radar derived precipitation rates can be assimilated by an extra Latent Heat Nudging scheme Stephan et al 2008 It computes additional temperature and humidity increments at each model column independently from each other It is tuned and should be used only for convection permitting model configurations with horizontal mesh widths of lt 3km The observation input is gridded precipitation rates read in the form of extra Grib files Further Grib files can be read optionally containing a blacklist and radar beam height maps utilised for bright band detection Analysis of surface and soil fields outside COSMO code In addition to the nudging type assimilation schemes for the atmosphere a set of 2 dimensional intermittent analysis schemes can be applied for some of the surface and soil fields in a full data assimilation cycle for the COSMO model This comprises of a variational soil moisture analysis Hess 2
174. d used without changing the COSMO code Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files 73 6 4 4 The blacklist file The blacklist file contains both a blacklist and whitelist The blacklist contains the stations with the variables and vertical ranges that are known to have a bad quality statistically Observations of that variables and in that vertical ranges are then excluded permanently from active use in the assimilation The whitelist contains all stations related to certain observation types which are known to issue observations of good quality The difference in concept between the blacklist and the whitelist becomes important when a new station sends observations of a certain type via GTS and these data have unknown quality These observations will be excluded from active use only if a whitelist exists for that observation type Whitelists are useful for types where there are not too many stations and the risk that a new station will deliver data of minor quality is considered rather high This often applies to remote sensing observation types A station related to a certain observation type can appear both on the whitelist and several times on the blacklist The whitelist activates the station as a whole and in the blacklist certain vertical ranges for certain variables can be excluded then from active use The blacklist file is a formatted ASCII file and has a fo
175. d within the model integration period see Namelist parameter hversta etc Moreover it also accommodates estimated observation errors and other pieces of infor mation which depend on the model state of the COSMO run itself This includes the quality control flags bias corretions as well as the simulated observations derived from the model state exactly at the respective observation times The purpose of the feedobs file is to serve as input for a Local Ensemble Transform Kalman Filter LETKF analysis or for a utility which computes the simulated observations from various model runs and writes them into an extended NetCDF feedback file This feedback file then contains all the input information required for grid point verification Its format and in principle its contents are identical to that of the feedobs file except that such feedback files can contain the simulated observations from more than just one model run The format is described in a separate documentation Feedback File Definition which can also be found on the COSMO web site www cosmo model org The file name of the feedobs file has the following form fof yyyymmddhhttss nc with yyyy year mm month dd day hh hour tt minute ss second yyyy means year in 4 digits while the other quantities are expressed in 2 digits each The date and time refers to verification_ref_date and verification_ref_time as described in the
176. d035 LOG TEMP code 35 land radiosonde TRUE 1cd036 LOG TEMP code 36 ship radiosonde TRUE 1cd037 LOG TEMP code 37 mobile radiosonde TRUE 1cd135 LOG TEMP code 135 drop sonde TRUE 1cd039 LOG TEMP code 39 land rocket TRUE 1cd040 LOG TEMP code 40 ship rocket TRUE 1cd032 LOG PILOT code 32 land PILOT TRUE 1cd033 LOG PILOT code 33 ship PILOT TRUE 1cd038 LOG PILOT code 38 mobile PILOT TRUE 1cd132 LOG PILOT code 132 European wind profiler TRUE 1cd133 LOG PILOT code 133 European SODAR RASS TRUE 1cd136 LOG PILOT code 136 US wind profiler RASS FALSE 1cd137 LOG PILOT code 137 radar VAD FALSE 1cd122 LOG SCATT code 122 OSCAT TRUE 1cd123 LOG SCATT code 123 ASCAT TRUE 1cd096 LOG GPS code 96 ground based GPS this code type is used TRUE only for GPS data read from a special ASCII file in COST 716 format igpscen 20 INT array of processing centres of GPS reports used actively 1 1 the order of centres determines the preference in the redundancy check 1 means no active centre the numbers N indicating the processing centres are given in the WMO Common Code Table C 12 for BUFR CREX data WMO descriptor 001 034 the code types for GPS data if read from NetCDF files are set to X Np where X 800 or X 900 these code type have to be used e g for setting iwtyp Part VII User s Guide 4 28 Section 7 Na
177. de the model integration time are never written to YUVERIF YUVERIF is written only if the NAMELIST variables 1verif TRUE and mveripr gt 2 Passive reports and the deviations of the observed values from the model values i e the observation increments are optionally included Furthermore YUVERIF can be post processed by means of additional programs to include also the deviations e g from different forecast runs and to derive statistical quantities Thus it can be used for monitoring validation and verification purposes Figure 8 17 shows a short example for YUVERIF The file begins with a header part which is mostly self explanatory Note that the initial date and hour of the run entry in the table at the end of the file header relates to the formal initial time of the forecast which is set to the final model integration time for an assimilating run Then the file body with the list of reports follows In the current example it consists of a subset of the reports already shown in Figures 8 6 and 8 7 for YUOBSDR namely 3 GPS ZTD IWV 1 scatterometer 1 buoy 2 Synop 6 aircraft single level and 1 aircraft multi level report The last line of the file has always the same form in order to indicate the end of file Each report consists of a report header a regular report body and an optional report body extension which contains the devations of the observed values from the model values In the report header und regular bod
178. dex c 1nd c 1 10 c_sea REAL Surface area index of gridpoints over sea 1 5 c_sea 1 10 c_soil REAL Surface area index of the evaporating fraction of gridpoints 1 0 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 5 TUNING Parameters for tuning dynamics and physics 105 Reasonable values are rain n0 factor 0 0 1 0 Introduced in Version 4 14 Name Type Definition Purpose Comments Default e surf REAL Exponent to get the effective surface area 1 5 e surf c 0 1 10 rlam heat REAL Scaling factor for the thickness of the laminar boundary layer 1 0 for heat rlam heat 0 1 10 rlam mom REAL Scaling factor for the thickness of the laminar boundary layer 0 0 for momentum rlam_mom 0 1 a_heat REAL Factor for turbulent heat transport 0 74 a_heat 0 01 100 a_mom REAL Factor for turbulent momentum transport 0 92 a_mom 0 01 100 a hshr REAL Length scale factor for separate horizontal shear production 0 2 of TKE Introduced in Version 4 10 a stab REAL Length scale factor for the stability correction 0 0 Introduced in Version 4 10 d heat REAL Factor for turbulent heat dissipation 10 1 d heat c 0 01 100 d mom REAL Factor for turbulent momentum dissipation 16 6 d mom 0 01 100 c diff REAL Factor for turbulent diffusion of TKE 0 2 c diff 0 10 clc diag REAL
179. dimensional variables do not need to exist and probably will not exist in the NetCDF file 6 4 3 Observation types without templates proposed by WMO e ACARS File names cdfin_acars cdfin_acars_uk resp cdfin_acars_us As a standard way implemented at DWD ACARS can be read by COSMO in 2 different ways 1 A file type cdfin_acars_us with BUFR obtained via GTS from ARINC Center 56 USA denoted as _us in the table below plus another file type cdfin_acars_uk with BUFR obtained via GTS from UK Met Office and Canada denoted as _uk or 2 A single file type cdfin_acars in a unified format defined by DWD which contains the reports from the two other files denoted as unif The formats consist of the descriptors in the following table where need COSMO asks stringently for this variable and will abort if variable is absent opt variable exists and is read used but COSMO will not abort if it does not exist P variable exists but is not read by COSMO gt variable does not exist in this format Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files 70 unif uk us descript type mnemonics meaning opt 0 01 033 int MMIOGC Identific of orig generat centre 0 01 034 int
180. discarded Note analysis field HMO3 with time range indicator 13 is discarded Note analysis field VIOS3 with time range indicator 13 is discarded Note analysis field RHO SNOW with time range indicator 13 is discarded te analysis field T ICE with time range indicator 13 is discarded te analysis field H ICE with time range indicator 13 is discarded te analysis field H ICE with time range indicator 0 is used zz te analysis field CE with time range indicator 0 is used te analysis field RHO_SNOW with time range indicator 0 is used te analysis field W_SNOW with time range indicator 0 is used 0 0 analysis field T_SNOW with time range indicator is used zz O OO O0O0O00O0000000O00000000000000 00 OQ ct oO te analysis field W_I with time range indicator is used te analysis field T_SNOW with time range indicator 13 is discarded te analysis field W_SNOW with time range indicator 13 is discarded Note analysis field W_I with time range indicator 13 is discarded Note analysis field T_SO with time range indicator 13 is discarded te analysis field T_SO with time range indicator 13 is used te analysis field T_SO with time range indicator 13 is used te analysis field T_SO with time range indicator 13 is used Note analysis field T_SO with time range indicator 13 is used Note analysis field T_SO with time range indicator 13 is used te analysis field T_SO with time range indicator 13 is used te analysis fie
181. e 0 5 the difference betw the scaled and the original latent heating profile is the LHN temperature increment lhn logscale LOG use of logarithmic scaling factors a 1 log a TRUE so as to unbias the absolute size of positive and negative LHN increments this also scales Amar and Qmin limits restrictions reduced weighting of LHN increments lhn limit LOG absolute limits for LHN temperature increments TRUE abs lhn lim REAL upper limit in K s for the absolute value of the LHN 50 3600 temperature increments applied if 1hn limit lhn incloud LOG restriction of LHN increments to cloudy layers only TRUE kbot lhn INT index of lowest model level at which ke_tot ktop_lhn INT index of uppermost LHN increments are applied 1 ktop_temp REAL temperature of uppermost model level at which LHN incre 999 9 ments are applied as alternative to ktop 1hn lhn wweight LOG additional weighting of LHN increments weight decreases FALSE with increasing mean horizontal wind speed if gt 20 m s 1hn_coef REAL overall scaling factor lt 1 for the LHN temperature incre 1 0 ments i e a kind of nudging coefficient lhn hum adj LOG application of humidity adjustment TRUE Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 135 Observation and file input for LHN Name Type Definition Purpose Comments Default spat
182. e 133 RASS SODAR Eur 40 4 36 0 code type 136 Wind Prof RASS US 0 0 0 0 code type 137 RADAR VAD Wind Prof 1089 488 573 242 0 observation type SATEM 0 0 0 0 0 observation type Scatterometer 1322 1145 0 177 code type 123 ASCAT scatterometer 1322 1145 0 177 code type 122 QuickScat scatterom 0 0 0 0 0 observation typel2 GPS 27791 14736 11349 4389 code type 800 GPS by METO 4270 2127 2097 347 code type 900 GPS by MET_ 0 0 0 0 code type 821 GPS by ASI_ 939 152 727 162 code type 823 GPS by GFZ_ 0 0 0 0 code type 824 GPS by GOP_ 352 153 199 46 code type 924 GPS by GOPE 0 0 0 0 code type 825 GPS by IEEC 0 0 0 0 code type 826 GPS by LPT_ 802 406 188 308 code type 926 GPS by LPTR 1800 722 607 833 code type 829 GPS by SGN_ 3581 2483 906 519 code type 929 GPS by SGN1 3335 600 2567 377 code type 830 GPS by BKG_ 564 548 4 143 code type 930 GPS by BKGH 0 0 0 0 code type 832 GPS by ROB_ 0 0 0 0 code type 833 GPS by KNMI 1122 683 439 228 code type 933 GPS by KNM1 621 376 245 106 code type 834 GPS by NGAA 4602 4094 400 184 code type 934 GPS by NGA_ 0 0 0 0 code type 835 GPS by IGE_ 2424 1228 917 579 code type 837 GPS by ROB_ 3379 1164 2053 557 code type 899 GPS by XXX_ 0 0 0 0 0 Notes on the table above Rejected passive means that the whole report is rejected set passive Partly rejected and partly passive reports are labeled active A report can be labeled active even if part of its data is black listed Figure 8 12 Exa
183. e 4 28 Section 7 Namelist Input for COSMO Modell 7 4 PHYCTL Parameters for the Diabatic Model 101 Surface layer fluxes These parameters control the calculation of the turbulent fluxes of latent and sensible heat Name Type Definition Purpose Comments Default itype tran INT Main prameter to select a specific surface layer parameteri 1 zation 1 Standard Louis type scheme 2 New TKE based scheme including a laminar sub layer lfreeslip_sfc LOG In Version 4 17 this functionality has been moved to the Namelist group ARTIFCTL but with a new name lnosurffluxes m h To switch on off the surface momentum fluxes even if a tur bulence scheme is used The following parameters apply only for the new surface layer scheme i e only in case of itype tran 2 imode tran INT Type of surface atmosphere transfer 1 1 Based on diagnostic TKE in the surface layer 2 Based on prognostic T KE in the surface layer itype_wcld INT Type of cloud water diagnosis 1 1 Diagnosis based on a relative humidity scheme 2 Diagnosis based on a statistical scheme icldm_tran INT Mode of cloud representation to take into account subgrid 0 scale condensation within the new surface layer parame terization itype tran 2 Options 0 1 and 2 as for icldm turb i e 0 No clouds are considered 1 Only grid scale clouds are considered 2 Grid and sub grid scale water clouds are considered cl
184. e cloudiness is interpreted by an empirical function depending on relative humidity and height A corresponding cloud water content is also inter preted Option for a statistical subgrid scale cloud diagnostic for turbulence Moist Convection Tiedtke 1989 mass flux convection scheme with equilibrium closure based on moisture convergence Option for the Kain Fritsch Kain and Fritsch 1993 convec tion scheme with non equilibrium CAPE type closure Shallow Convection Reduced Tiedtke scheme for shallow convection only Radiation two stream radiation scheme after Ritter and Geleyn 1992 short and longwave fluxes employing eight spectral intervals full cloud radiation feedback Soil Model Multi layer version of the former two layer soil model after Jacobsen and Heise 1982 based on the direct numerical solution of the heat conduction equation Snow and interception storage are included Option for the old two layer soil model employing the extended force restore method still included Fresh Water Lake Parameterization Two layer bulk model after Mironov 2008 to predict the vertical temperature structure and mixing conditions in fresh water lakes of various depths Sea Ice Scheme Parameterization of thermodynamic processes without rheology after Mironov and B 2004 The scheme basically computes the energy balance at the ices surface using one layer of sea ice Part VII User s Guide 4 28 Section 1 Ov
185. e msl HTOP top height of convective cloud above msl RR rain grid scale and convection RS snow grid scale and convection WS water content of snow Figure 8 2 shows an example of a file M_stationname with the long form of the grid point output This long form contains the following information for every grid point HSURF height of model orography in m FR LAND land fraction LAT geographical latitude in of the grid point Part VII User s Guide 4 28 Section 8 Model Output 149 8 1 ASCII Output for the Forecast Model Model LM KKK Start of the forecast Short meteograph of the LM at selected grid points Meaning of the parameters HH forecast hour hours PS mean sea level pressure hpa DF10M wind direction and speed at 10m degree knots DF500M wind direction and speed at 950 hPa degree knots DF850 wind direction and speed at 850 hPa degree knots DF700 wind direction and speed at 700 hPa degree knots DF500 wind direction and speed at 500 hPa degree knots TG ground temperature C T2M temperature at 2m C TD2M dew point temperature at 2m C T30M temperature at 30m lowest level C T850 temperature at 850 hpa C T700 temperature at 700 hpa C T500 temperature at 500 hpa C HML cloud cover high medium low 0 8 ground fog 0 8 HBAS base height of con cloud above msl hpa HTOP top height of con cloud above msl hpa RR rain amount grid scale and
186. e tot 92 For a complete reference of all NAMELIST parameters see Chapter 7 Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 2 Conventions for File Names 49 6 2 Conventions for File Names The initial and boundary fields needed for the model are provided either in Grib or in NetCDF format Also for the output files one can choose between Grib or NetCDF Restart files are written in binary format with full precision There is one file for the initial fields and also for every set of boundary fields The following conventions apply for the filenames A file name for the COSMO Model or the INT2LM has the general form yheader ydate yextension for Grib files Or yheader ydate yextension nc for NetCDF files where yheader ydate and yextension have the following meaning yheader File header 3 characters first character specifies the model GME global model COSMO Model ECMWF model IFS Integrated Forecast System A general global climate model Q oO F OF second character analysis file uninitialized analysis file initialized boundary file forecast files H Hh o n p restart files third character specifies the region covered by the data f full model domain s subdomain ydate There are two forms of specifying the date either with the full date or relative to the start date In the name of analysis files second character in the
187. eccessary e g to include a constant pressure gradient in terms of a geostrophic wind Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 2 RUNCTL Parameters for the Model Run 82 Note To generate artificial initial and boundary conditions for idealized model runs the routines gen ini data to specify initial data and gen bound data to specify boundary data con tained in the source file src artifdata f90 are used The configuration of such idealized model runs can be done with the help of a large number of namelist parameters in the namelist ARTIFCTL in the file INPUT IDEAL which offers a wide variety of choices for orogra phy atmosphere and soil profiles for initial and boundary conditions and convection triggers However there is no unique way for this specification as it depends on the problem to be considered and therefore it is possible to extend src artifdata f90 according to special needs The documentation of the namelist group ARTIFCTL would go beyond the scope of this User Guide therefore an extra document will be provided for running idealized cases Parameters for parallel and sequential execution This section contains parameters for the decomposition parallelization of the model domain and for generating additional information about timings for the different parts of the model run Furthermore the type of communication can be specified see parameters 1datatypes ncomm type How
188. ecnudge LOG to switch on off spectral nudging FALSE yvarsn CHAR list of fields for spectral nudging must be a subset of yvarbd U V pp sn REAL lowest pressure level for spectral nudging 850 0 alpha_sn REAL amplification factor for spectral nudging 0 05 isc sn INT spectral nudging in i direction 2 jsc sn INT spectral nudging in j direction 2 nincsn INT to define a time increment for calling the Spectral Nudging 1 Note The spectral nudging has nothing to do with the Nudging used as assimilation scheme in the COSMO Model Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 4 PHYCTL Parameters for the Diabatic Model 94 7 4 PHYCTL Parameters for the Diabatic Model The namelist group PHYCTL contains parameters controlling the physical parameterizations All parameters are only active if the parameter lphys in Namelist RUNCTL is set to TRUE in order to enable physical parameterizations There is one main switch for each physical process to turn on off this process and to activate additional parameters and sub options for the corresponding parameterisation The specifications for the parameters in PHYCTL are included in the file INPUT_PHY The namelist parameters of this group are described in the subsections Grid Scale Precipitation Radiation Moist Convection Vertical turbulent diffusion Surface layer fluxes Soil Processes Subgrid Scale Orog
189. ed geographical coordinates e Specifying the time unit of the forecast range With the variable ytunit the time unit of the forecast range form of the output file name can be specified Specifying the system where the variables are written to There are two possibilities ysystem FILE The results are written to grib files in the directory ydir ysystem BASE This possibility is not implemented yet e Specifying the packing rate for the Grib code nrbit gives the number of bits that will be used for storing one real variable of the model Possible values are 4 8 16 24 32 and 48 e Specifying control output If lcheck TRUE minimum maximum and mean values of the fields are calculated and written to file YUCHKDAT Table 8 1 Basic output fields for the COSMO Model Field defaults Meaning U mpz zonal wind speed V mpz meridional wind speed W mz vertical wind speed T mpz temperature QV specific water vapor content m m specific cloud water content QI m specific cloud ice content m specific rain content m specific snow content QG specific snow content Part VII User s Guide 4 28 Section 8 Model Output 8 4 Output of Forecast Fields 191 PP deviation from reference pressure P m z pressure PS m surface pressure PMSL m surface pressure on mean sealevel HHL m geometrical height of half levels HSURF m height of surface topography T S m tempe
190. ed to the Use of Observations 167 1 0 DISTRIBUTION OF PROCESSED ACTIVE PASSIVE REJECTED REPORTS FOR ASSIMILATION 0 processed active passive rejected 0 0 total number of reports 59556 38808 14041 11437 reports with unknown obs code type 0 0 0 0 0 observation type 1 SYNOP 16011 11860 1025 4959 code type 11 SYNOP Manual Land 5955 4951 457 1025 code type 14 SYNOP Automatic Land 9430 6418 537 3830 code type 21 SHIP 126 90 6 30 code type 22 SHIP Abbreviated 0 0 0 0 code type 23 SHIP Reduced SHRED 0 0 0 0 code type 24 SHIP Automatic 500 401 25 74 code type 140 METAR 0 0 0 0 0 observation type AIREP 11716 10214 926 576 code type 41 CODAR 0 0 0 0 code type 141 AIREP Aircraft 386 178 9 199 code type 144 AMDAR 4721 4167 321 233 code type 244 ACARS 6609 5869 596 144 code type 241 COLBA Const Lev Ball 0 0 0 0 0 observation type SATOB 0 0 0 0 0 observation type DRIBU 778 138 0 640 code type 63 BATHY 0 0 0 0 code type 165 DRIBU Drifting Buoy 778 138 0 640 0 observation type TEMP 549 114 1 434 code type 35 TEMP Land 531 111 0 420 code type 36 TEMP SHIP 18 3 1 14 code type 37 TEMP Mobile 0 0 0 0 code type 39 ROCOB Land 0 0 0 0 code type 40 ROCOB SHIP 0 0 0 0 code type 135 TEMP DROP 0 0 0 0 0 observation type PILOT 1389 601 740 262 code type 32 PILOT Land 25 2 3 20 code type 33 PILOT SHIP 0 0 0 0 code type 38 PILOT Mobile 0 0 0 0 code type 132 Wind Profiler Eur 235 107 128 0 code typ
191. el identity bit pattern or pressure code SYNOP as decimal number see below 3 hourly pressure tendency GPS zenith wet delay mm 10 Pa 3h 4 500 low cloud cover octas horizontal visibility 10 m combined cloud and weather group code tables as octal number see below SYNOP combined weather amp ground group code tables as octal number see below single level aircraft degree of turbulence WMO decriptor 0 11031 as octal number precipitation amount rr over 12hrs 1 denotes 0 lt rr lt 0 1 mm 1 10 mm minimum temperature at 2m during past 12 hrs 1 10 K maximum temperature at 2m during past 12 hrs 1 10 K only for AOF read min ground temperature at 5cm in past 12hrs 1 10 K maximum wind speed of gusts over 1 hour m s maximum wind speed of gusts over 6 hrs m s only for AOF read global radiation sum over 1 hour 10 kJ m2 Missing values are denoted by 9 for quantities which do not take negative values by 999 for height and by 9999 otherwise Entries 7 8 9 10 14 and 15 are detailed by the following lists main flag word entry 9 main flag groups on horiz wind temperature humidity pressure or geopot bit position for flag groups 0 7 14 21 bit position within flag groups of 7 bits each a 1 bit flag is set if the bit set to 1 0 aircraft wind roll angle code table WMO descriptor 0 02 064 otherwise 2 bit dataset quality flag
192. elf explaining i e no additional tables are needed The contents of an output file can be listed with the program ncdump which installs automatically along with the netCDF library The netCDF library also defines a machine independent format for representing scientific data Together the interface library and format support the creation access and sharing of scientific data The netCDF software was developed at the Unidata Program Center in Boulder Colorado The freely available source can be obtained as a compressed tar file or a zip file from Unidata or from other mirror sites http www unidata ucar edu packages netcdf index html Information on the F90 implementation can also be obtained from http www unidata ucar edu packages netcdf f90 index htm In the current implementation of netCDF I O data are in 32bit accuracy For GRIB format an additional packing can be done usually using 16 bit accuracy But this packing will lose information NetCDF I O can be turned on via the yform read and the yform write parameters in the namelist IOCTL The parameters can be chosen independently e g it is also possible to have GRIB as input and netCDF as output format and vice versa Name Type Definition Purpose Comments Default yform read CHAR Format of input data grbi1 grb1 GRIB formatted input ncdf netCDF formatted input yform_write CHAR Format of output data grbi1 grb1 GRIB formatted output ncdf netCDF
193. en by one processing unit and often take into account only one sub domain on distributed memory machines In contrast the whole model domain is considered by the messages on YUREJCT and the standard output In addition there are summary CAUTION messages related to insufficient array size occurring in any of the sub domains written to the files YUSTATS section 8 2 6 and YUCAUTN In an operational setting it is important that the model does not crash due to insufficient array size Otherwise any simple increase from one day to the next of the number of obser vations that are input to the data assimilation scheme could potentially cause a crash of the operational suite Omitting the presumably rather small number of surplus observations will usually lead only to a minor degradation of the analysis if at all On the other hand it should be made sure that the data assimilation does not run for weeks or months with too small array sizes The file YUCAUTN serves this purpose It is not created at all except if one of the following two events occurs Either there is an insufficient array size related to observational infor mation or an observation with unknown observation type has been read Thus if the file YUCAUTN is produced by the COSMO model this implies that action needs to be taken to ensure an optimal use of the data in subsequent data assimilation cylces At DWD e mails are sent automatically to responsible persons if the file
194. ency check However if the entry follows immediately the entry ps for the same report e g report 16080 in Figure 8 9 it implies that the spatial consistency check accepts the observation by cancelling the rejection suggested by the individual threshold quality control The ps scc lines are always complemented with the value for the bias correction applied to the model value used in the check and with the total weight of the observations used to determine this bias The same as for ps scc applies to entry IWV sc if it is related to a ground based GPS observation IWV sc can also occur for a radiosonde report in this case this means that the whole humidity profile report is rejected In the case of the entry z the height observation increment as derived hydrostatically from multi level temperature increments and possibly a surface pressure increment ex ceeds the threshold at the given pressure level This implies that all temperature humidity and geopotential data from that level up to the top level p top of the multi level profile are set passive Similarly the entry dz means that these types of data are rejected due to the hydrostatic thickness check between the given pressure level and the pressure level p top which in this case does not coincide with the top level of the report in general The specification of the threshold and the thickness increment is complemented here by the mean temperature
195. ent atmosphere The Charnock formula to estimate the surface fluxes over sea is also reformulated using T KE COSMO EU and COSMO DE use the TKE based surface transfer scheme 3 5 6 A subgrid scale orography scheme Basic Namelist settings lphys TRUE 1sso TRUE 3 5 7 Soil Processes Basic Namelist settings lphys TRUE lsoil TRUE The calculation of the surface fluxes requires the knowledge of the temperature and the specific humidity at the ground The task of the soil model is to predict these quantities by the simultaneous solution of a separate set of equations which describes various thermal and hydrological processes within the soil If vegetation is considered explicitly additional exchange processes between plants ground and air have to be taken into account a The soil model TERRA Basic Namelist settings Imulti layer FALSE For land surfaces the soil model TERRA provides the surface temperature and the specific humidity at the ground The ground temperature is calculated by the equation of heat conduction which is solved in an optimized two layer model using the extended force restore method Jacobsen and Heise 1982 The soil water content is predicted for two or three layers by the Richards equation Evaporation from bare land surfaces as well as transpiration by plants are derived as functions of the water content and only for transpiration of radiation and ambient temperature Most parameters of the soil
196. er w_g2 water content of the medium soil layer w_g3 water content of the lower soil layer Ww cl climatological water content of the lowest soil layer Initial field required for cloud ice rain snow and graupel if appropriate NAMELIST parameters are set Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 3 Initial and Boundary Data 52 qi specific cloud ice content qr specific rain content qs specific snow content qg specific graupel content Next all possible fields necessary for the boundary files are listed This again depends on Namelist switches All boundary fields are provided for the full forecast domain Boundary fields required for the COSMO Model u bd zonal wind speed v bd meridional wind speed pp bd deviation from reference pressure t bd temperature t snow bd temperature of snow surface qv bd specific water vapor content qc bd specific cloud water content qv s bd specific water vapor content on the surface w snow bd water content of snow Boundary fields required for 2 layer soil model t s bd temperature at the boundary soil atmosphere t m bd temperature between upper and medium soil layer w gi bd water content of the upper soil layer w g2 bd water content of the medium soil layer w g3 bd water content of the lower soil layer For the multi layer soil model no boundary fields are necessary If the model should run with cloud ice also the following field is needed
197. erating centre DWD has WMO number 78 4 6 132 Generating process identification number allocated by originating centre see Table 5 4 5 7 255 Number of grid used from catalogue defined by the originating centre Octet 7 set to 255 indicates a non cataloged grid in which case the grid is defined in the grid description section 6 8 128 Block flag the value 128 indicates that the grid description section is included 7 9 71 Indicator of parameter element number from GRIB table in ipds 2 8 10 1 Indicator of type of level see Table 5 5 9 10 11 12 0 Value of level height pressure etc for which the data are included see Table 5 5 11 13 98 Year start time of forecast analysis time 12 14 10 Month start time of forecast analysis time 13 15 28 Day start time of forecast analysis time 14 16 0 Hour start time of forecast analysis time 15 17 0 Minute start time of forecast analysis time 16 18 1 Indicator of unit of time range see Table 5 6 17 19 11 P1 period of time number of time units time units given by octet 18 ipds 16 18 20 0 P2 period of time number of time units time units given by octet 18 ipds 16 19 21 0 time range indicator see Table 5 7 20 22 23 0 Number of forecasts included in average when octet 21 ipds 19 indicates an average or accumulation of forecasts or analyses otherwise set to zero 21 24 0 Number of forecast
198. erface and the structure of the database system All users outside DWD have to work with file based IO Most Namelist input parameters from DATABASE are not relevant in this case except nout sockets and nin sockets which both have to be set to 0 default in order to enable file based IO Name Type Definition Purpose Comments Default yinit order CHAR String to initialize the data base interface program csodaban ak nix for GRIB IO including specifications to write GRIB output to the data base system yana tab CHAR Specifications to retrieve initial data from the data base PRU system ybd tab CHAR Specifications to retrieve boundary data from the data base system nout sockets INT Number of sockets for database output per PE 0 0 means File IO nin sockets INT Number of sockets for database input per PE 0 0 means File IO has to be lt 1 iretry INT Number of seconds to retry on database failure 0 ibackup size INT Size of incore backup space in bytes by a database failure 1 ybackup_dir CHAR Directory path for outcore backup idbg level INT Debug level for mpe_io 0 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 13 GRIBIN Controlling the Grib Input 141 7 13 GRIBIN Controlling the Grib Input Control parameters for initial data sary for reading writing GRIB2 general vertical
199. ers that control the basic configuration of a model run The specifications for the parameters in RUNCTL are included in the file INPUT ORG The namelist parameters of this group are described in the subsections initial time forecast range time step and calendar used basic control parameters parameters related to artificial initial and boundary conditions parameters for parallel and sequential execution parameters for diagnostic min max model output and parameters for debug purposes Initial time forecast range time step and calendar used Name Type Definition Purpose Comments Default ydate ini ydate bd ydate end CHAR CHAR CHAR Initial date and time of the forecast Since Version 4 24 there are two different formats e ydate ini yyyymmddhh or e ydate ini yyyymmddhhmise where yyyy year mm month dd day hh hour mi minutes and se seconds If the date is specified in the old format still the default all file names are also treated as usual If the date is given with minutes and seconds then all relevant files are also written with minutes and seconds in the file names This will be necessary if a sub hourly analysis cycle will be used Start date and time of the forecast from which the bound ary fields are used specified by year month day hour T he format is yyyymmddhh as above If omitted ydate bd is set to ydate ini End date for
200. ervation relative to s valid at the 1 0 observation time see figure for rhtfac above 1 0 cutofsu 4 REAL cut off in multiples of correlation scales s 2 0 of the horizontal correlation function way 3 5 2 0 2 0 non isotropic correction w to isotropic function Wyy vcsnisu 4 REAL square of Gaussian radius of influence 2500 in potential temperature differences 2500 AED 9 ie w e PUE if msprpsu lt 1 9 in log pressure differences if nsprpsu 2 between the target grid point and the point at the horizontal observation location on the surface along which observation increments are spread laterally spreading of surface level observation increments along model levels nsprpsu 0 rhfpsdd REAL for scaling horizontal correlation scale for surface pressure data 1 0 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 125 Geometry for lateral spreading of horizontal wind Name Type Definition Purpose Comments Default 2 dimensional wind correlation functions correlation functions for longitudinal and transverse wind Ar s components w e ris f uL swn 0 wr se a Arjs e 777 cnondiv REAL constant part ye of the non divergence correction factor Yn 0 1 Yn Ye fo Ww me Ye fo Ww gt 1 1 wi fnondiv REAL multiplication factor f to the vertically varyi
201. erview on the Model System 1 3 Organization of the Documentation 6 Data Code 1 3 Terrain and Surface Data All external parameters of the model are available at various resolutions for a pre defined region covering Europe For other regions or grid spacings the external parameter file can be generated by a preprocessor program using high resolution global data sets Assimilation Basic Method Continuous four dimensional data assimilation based on observation nudg ing Schraff 1996 Schraff 1997 with lateral spreading of upper air observation increments along horizontal surfaces Explicit balancing by a hydrostatic temperature correction for sur face pressure updates a geostrophic wind correction and a hydrostatic upper air pressure correction Assimilated Atmospheric Observations Radiosonde wind temperature humidity air craft wind temperature wind profiler wind and surface level data SYNOP SHIP BUOY pressure wind humidity Optionally RASS temperature radar VAD wind and ground based GPS integrated water vapour data Surface level temperature is used for the soil moisture analysis only Radar derived rain rates Assimilation of near surface rain rates based on latent heat nudging Stephan et al 2008 It locally adjusts the three dimensional thermodynamical field of the model in such a way that the modelled precipitation rates should resemble the observed ones Surface and Soil Fie
202. es for those common sequences which contain only variables that are not used by COSMO For convenience the table is split into a report header and three report body parts Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files 56 The use of the variables first column of table is defined as follows need opt The existence COSMO asks stringently for this variable and will abort if variable is absent but will not abort if values are equal to missing value variable exists and is read used but COSMO will not abort if it does not exist used means here that it is e g written to the feedobs file but it does not imply active use in the data assimilation variable exists but is not read by COSMO descriptor a BUFR common sequence or a BUFR data description operator exists only in the BUFR file not in the NetCDF file descriptor itself exists only in the BUFR file not in the NetCDF file however the descriptor indicates a common sequence of variables which are present in the NetCDF file but which are not used by COSMO and hence are not detailed here column L for land stations 3 7 column S for SHIP sea stations of the variables or descriptors for the different observation file types according to the above mentioned BUFR common sequence descriptors is defined as follows ME exists for both typ
203. es of land stations resp for sea stations EE exists for fixed land stations but not for mobile land stations m exists for mobile land stations but not for fixed land stations DO does not exist use L S WMO descriptor type mnemonics meaning f 3 01 090 Fixed surface station ID time horiz 4 vertical coordinates m 3 01 092 Mobile surface station ID time horiz vertical coordinates 3 01 093 Ship ID movement date time horiz vertical coordinates 3 01 036 Ship ID movement time lat lon need m 0 01 011 char 9 YDDDD Ship or mobile land sta identifier 0 01 012 int MDS Platform motion direction 0 01 013 int NVS Platform motion speed m 0 01 003 int MA WMO region number f 3 01 004 Surface station identification need f 0 01 001 int MII WMO block number need f 0 01 002 int NIII WMO station number opt f 0 01 015 char 20 YSOSN Station or site name need 0 02 001 int NIX Type of station 3 01 011 Year month and day need 0 04 001 int MJJJ Year need 0 04 002 int MMM Month need 0 04 003 int MYY Day 3 01 012 Hour minute need 0 04 004 int MGG Hour need 0 04 005 int NGG Minute e 3 01 021 Latitude and Longitude need 005001 float MLAH Latitude high accuracy degree need 006 001 float MLOH Longitude high accuracy deg need 005002 float MLALA Latitude coarse accuracy deg need 0 06 002 float
204. esponding coarse grid model run In this case the interpolation program writes so called ready files which simply indicate that a lateral boundary file for a certain forecast time has been success fully and completely written During program execution the model checks for the existence of these files in certain time intervals and in this way waits until the required lateral boundary data set is available for the next READ operation on the corresponding input file In a simi lar way the COSMO Model writes ready files for each output file that has been successfully written This allows to run postprocessing utilities in parallel to the model execution Basic control parameters Name Type Definition Purpose Comments Default lgen LOG Moved to group RUNCTL and renamed with lartif data lasync io LOG If TRUE the model runs with extra processors for asyn FALSE chronous IO lprefetch io LOG Enables reading of boundary files ahead of time i e when FALSE the forthcoming I O operation will be reading a GRIB file then this can be done simultaneously with the preceding compute steps Prefetching can only be enabled with true asynchronous I O lasync_io TRUE itype_gather INT To choose the type of gathering output fields 1 1 default gather the fields by an extra communica tion per level 2 gather fields by one communication for nproc levels with MPLALLTOALLV ngribout INT To specify
205. et from the outer boundaries to the interior T his offset is NogAA AA A in A direction and NogAqo Av A in y direction where Nog nboundlines as program variable denotes the number of grid intervals used to define the position of the physical boundaries By default Nog is set to 2 larger but not smaller numbers for Nog may be specified by the user Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 4 Numerical Integration 16 All grid points interior to the physical boundary constitute the computational or model interior domain where the model equations are integrated numerically These are points with subscripts i j running from i Nog 1 NA Nog and j Nogtl1 No Nog The extra points outside the interior domain constitute the computational boundaries At these points all model variables are defined and set to specified boundary values but no dynamical computations are done For No 2 we have two extra lines of grid points adjacent to each physical boundary see Fig 3 3 v T ie v u id T ee A A v uTu TuTuTu T j v AN T 2222222222222 222 E Y E Thanananmsamuaanssamnuaumamanasnaumanasanananananlana MH d 3 i 2 3 A A N 2 N 1 Ny Figure 3 3 Horizontal model domain for N x N grid points and an offset of N g 2 for the position of the physical boundaries dotted The computational boundaries are shade
206. extension All fields can be written for the full domain or a subdomain if the NAMELIST parameters ydomain s and slon slat elon elat in the group gribout are specified To distinguish NetCDF from Grib files the NetCDF files contain the suffix nc The output of the forecast fields is controlled by the NAMELIST group gribout the name of this group comes from the time when only Grib was implemented but it it also valid for NetCDF output It is possible to specify several instances of this group The NAMELIST parameter ngribout in the group IOCTL has to be set accordingly For every instance you have to define a list of variables for output and a description of the special kind of output For that purpose you have to set the NAMELIST variables contained in gribout see Section 7 14 properly e Specifying the list of variables for output yvarml Variables on model levels e g yvarml U V HSURF yvarpl Variables on pressure levels plev A list of pressure levels to which the model variables are interpolated yvarzl Variables on z levels zlev A list of z levels to which the model variables are interpolated yvarsl Variables that contain images for channels of selected satellites By specifying yvarxl default x m p z s a predefined list of variables is written Table 8 1 gives a list of basic model variables that can be specified for output The seco
207. ficance need 0 04 025 int NGGTP Time period or displacement min need 0 10 004 float MPPP Pressure Pa need 0 12 001 float MTN Temperature dry bulb temperature K need 0 13 003 int MUUU Relative humidity 96 need 0 33 038 int NQFGD Quality flags for ground based GNSS data 008 022 int MTOTN Total number accumulation average 0 02 020 int n MSACL Satellite classification 0 01 050 int n MPTID Platform transmitter ID number 005021 float n MDA Bearing or azimuth degree_true need 0 07 021 float n MDE Elevation degree need 015 031 float n NADES Atmospheric path delay in sat signal m need 015 032 float n NEERR Estimated error in atmospheric path delay m 0 08 060 int MSSMS Sample scanning mode significance 015 033 float NDPDL Diff in path delays for Limb views m 0 15 034 float NEEPDD Estimated error in path delay difference m 0 08 060 int MSSMSO Sample scanning mode significance 015 033 float NDPDLO Diff in path delays for Limb views m 0 15 034 float NEEPDDO Estimated error in path delay difference m need 0 15 035 float NCZWV Component of zenith path delay due to water vapour m need 0 13 016 float NWLN Precipitable water kg m 2 0 15 011 float MLIED LOG 10 of integrated electron density log 1 m 2 Table note n in the variable type definition means that this variable has an additional dimension Currently this is set to a fixed value of 25
208. files to contain specific variables Some of the variables are mandatory while others are optional This implies that the BUFR file prior to con version into NetCDF must also contain these variables at least the mandatory ones Any BUFR files lacking these variables or containing the quantities in a different form with e g different variable names must first be converted into a BUFR file according to the specifications below prior to conversion into NetCDF The descriptions of the templates below applies to the NetCDF input files However the templates in particular of BUFR Section 3 are based upon those of the BUFR files Therefore reference is made to descriptions of BUFR templates where available by WMO and additional common sequence BUFR descriptors are included in the description even though they are absent in the NetCDF file The specified variable types int float char relate only to the NetCDF file after conversion from BUFR From BUFR Section 1 the following elements are mandatory for all observation types i e COSMO will abort if any of these are absent in the NetCDF file descriptor type variable name in NetCDF file meaning int edition_number BUFR edition number usually 4 C 11 int sectionl centre originating generating data centre C 12 int sectionl subcentre e g processing centre for GPS reports int sectionl update sequence nr upd seq number station correction C 13 int secti
209. for the ice thickness The result is a computationally efficient bulk model that incorporates much of the essential physics Importantly FLake does not require re tuning i e empirical constants and parameters of FLake should not be re evaluated when the model is ap plied to a particular lake There are of course lake specific external parameters such as depth to the bottom and optical properties of water but these are not part of the model physics Using the integral approach the problem of solving partial differential equations in depth and time for the temperature and turbulence quantities is reduced to solving ordinary differential equations for the time dependent quantities that specify the tem perature profile FLake carries the equations for the mean temperature of the water column for the mixed layer temperature and its depth for the temperature at the lake bottom and for the shape factor with respect to the temperature profile in the lake Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 5 Physical Parameterizations 26 thermocline a stably stratified layer between the bottom of the mixed layer and the lake bottom In case the lake is cover by ice additional equations are carried for the ice depth and for the ice surface temperature The lake surface temperature i e the quantity that communicates information between the lake and the atmosphere is equal to either the mixed layer te
210. formatted output 5 2 1 CF Conventions The basic conventions for netCDF Output in COSMO are the Climate and Forecast CF conventions These define standards on the naming and structuring of the netCDF output The latest description of the CF conventions can be found on the Lawrence Livermore WEB page http cf pcmdi llnl gov The values for units and standard name have fixed values and are defined by the CF conventions Values for long name and the name of the parameter field can be freely chosen by the user The long name is often used by graphic programs in creating the legend of figures The FillValue attribute holds the value set for missing data The name of the parameter is used to extract a certain field from the output file Part VII User s Guide 4 28 Section 5 Data Formats for I O 5 2 The NetCDF Data Format 43 5 2 2 Namelist Input The netCDF I O can be controlled via the Namelist IOCTL In addition to the parameters yform read and the yform write described above global attributes can be defined to describe the model simulation yncglob institution CHAR originating center name aid yncglob title CHAR title string for the output d yncglob source CHAR program name and version Ec yncglob contact CHAR identification of the project of the simulation Mes yncglob project id CHAR identification of the experiment of the simulation yncglob experiment id CHAR contact e g
211. g the following information for one observation level horizontal wind components temperature relative humidity pressure height observation errors for wind temperature humidity and height and level identity T he latter is a bit pattern as specified for the VOF see section 8 2 7 Figure 8 6 also includes examples for drifting buoy scatterometer wind Synop ship and Synop surface land reports Two of the latter reports are set passive since all their observations are passive due to large differences between observation level and model orography Figure 8 7 shows an example of an descending aircraft before landing The process of deriving a multi level report from the original single level reports is interrupted at a level where the aircraft had to stay on hold before it was allowed for the final landing As a result two instead of one multi level reports are created and at the holding altitude of 882 hPa several single level reports remain all of which except for one are set passive due to redundancy Figure 8 7 is completed by examples of ground based GPS reports Compared to other single level reports the values for wind are replaced here by integrated water vapour and zenith wet delay values in mm height by zenith total delay ZTD in mm and the observation errors by the ZTD error and the bias correction in mm All reports shown here are from one station COMO but the raw data have been processed by different processi
212. ght time off centering in the vertical implicit formulation is applied to damp acoustic modes On the big time step the Asselin time filter and a 4th order horizontal diffusion are used for numerical smoothing While this 3 timelevel HE VI integration was the default time scheme of the COSMO Model in the beginning it has now been replaced by the 2 timelevel Runge Kutta schemes 3 4 3 Leapfrog 3 timelevel Semi Implicit Integration Basic Namelist settings 12t1s FALSE lsemi imp TRUE Because the HE VI scheme integrates the horizontal momentum equations explicitly steep orography may provoke instabilities in small scale applications Full 3D semi implicit schemes can avoid such stability problems by treating all pressure gradient and divergence terms implicitly both vertically and horizontally HI VI scheme thus a small time step is not used Moreover 3D semi implicit schemes have the potential to become more cost effective than split explicit schemes at higher resolution where the grid aspect ratio is more isotropic and where the number of small time steps increases with the sound speed Courant number for low Mach number flows The derivation of the scheme is based on the 3 timelevel Leapfrog integration and uses the time tendency formulation to minimize cancellation errors An elliptic equation for the pressure perturbation tendency L 5 p qp is obtained by forming the divergence of the momentum equations and eliminating the b
213. given 10 18 20 3250 Rotated latitude of the last gridpoint in millidegrees 11 21 23 7750 Rotated longitude of the last gridpoint in millidegrees 12 24 25 0 Longitudinal direction increment grid spacing in A direction not given 13 26 27 0 Meridional direction increment grid spacing in direction not given 14 28 64 Scanning mode flag according to WMO code table 8 64 means that points scan in i and j direction and adjacent points in i direction are consecutive 15 19 29 32 0 Reserved set to zero 20 33 35 32500 Geographical latitude of rotated southern pole in millidegrees 21 36 38 10000 Geographical longitude of rotated southern pole in millidegrees 22 39 42 0 Angle of rotation 26 65 43 202 List of vertical coordinate parameters each packed on 4 octets length 4 x NV octets first the three parameters defining the base state igds 26 p0sl igds 27 t0sl igds 28 dt01p then the parameter igds 29 vcflat of the hybrid coordinate system and finally the ke 4 1 values of the vertical coordinate n k of the model half levels for k 1 ke 1 in igds 30 igds 65 Part VII User s Guide 4 28 Section 5 Data Formats for I O 5 2 The NetCDF Data Format 42 5 2 The NetCDF Data Format netCDF network Common Data Form is an interface for array oriented data access and a library that provides an implementation of the interface The netCDF of the COSMO Model is s
214. guously specifies the grid boxes for which the lake surface temperature should be computed A sea ice scheme Basic Namelist settings lsoil TRUE lseaice TRUE The presence of sea ice on the oceans surface has a significant impact on the air sea interactions Compared to an open water surface the sea ice completely changes the surface characteristics in terms of albedo and roughness and therefore substantially changes the surface radiative balance and the turbulent exchange of momentum heat and moisture between air and sea In order to deal with these processes the COSMO model includes a sea ice scheme Mironov 2008 Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 6 Data Assimilation 27 COSMO EU and COSMO DE use the multi layer soil model and the FLake Model The sea ice scheme is only used in COSMO EU 3 6 Data Assimilation Basic Namelist setting luseobs TRUE The requirements for the data assimilation system for the operational model are mainly determined by the very high resolution of the model and by the task to employ it also for very short range forecasting Hence detailed high resolution analyses of the atmosphere have to be able to be produced frequently and this requires a thorough use of asynoptic and high frequency observations such as aircraft data and remote sensing data Note that the synoptic scales are largely determined by the lateral boundary conditions provided by t
215. h drainage and diffusion czbot w so REAL to specify depth of bottom of last hydrological active soil layer 2 5 Subgrid Scale Orography These parameters control the parameterization of the effect of unresolved orography on the resolved scales of motion The main effect is an additional surface drag over mountains Note Additional external parameter fields describing the subgrid scale orography are needed as input fields They can be provided using an appropriate version of INT2LM Name Type Definition Purpose Comments Default lsso LOG Main switch to include subgrid scale orography processes TRUE nincsso INT Interval in time steps between two calls of the SSO scheme 5 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 5 TUNING Parameters for tuning dynamics and physics 104 7 5 TUNING Parameters for tuning dynamics and physics The namelist group TUNING contains parameters that can be used to tune special components and packages of the parameterizations and dynamics This namelist group is intended to be used mainly by the EXPERTS The parameters can be used to adapt the behaviour of the model to special regions applications and resolutions The specifications for the parameters in TUNING are included in the file INPUT_ORG In the following table some limitations and ranges for meaningful values of the different parameters are given in the form
216. hafen 135 Lindenberg Obs 900 Hohenpeissenberg lmulti layer TRUE lmelt TRUE lmelt var TRUE ke soil 7 czml soil 0 005 0 02 0 06 0 18 0 54 1 62 4 86 14 58 lconv TRUE itype conv 0 lcape FALSE lsso TRUE end input phy cat gt INPUT DIA lt lt end input dia amp DIACTL nOmeanval 0 nincmeanval 1 lgplong TRUE lgpshort FALSE lgpspec FALSE nO0gp 0 hincgp 1 0 Figure 7 1 Excerpt of run script from the COSMO Model to create INPUT files Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 1 LMGRID Specifying the Domain and the Model Grid 77 7 1 LMGRID Specifying the Domain and the Model Grid The namelist group LMGRID contains parameters that specify the lat lon coordinates of the pole of the rotated grid the position of the model domain within the rotated grid the size and resolution of the model grid The specifications of the parameters for this group are included in the file INPUT_ORG Name Type Definition Purpose Comments Default pollat REAL Geographical latitude of the rotated north pole in degrees 32 5 north gt 0 for a non rotated lat lon grid set pollat 90 pollon REAL Geographical longitude of the rotated north pole in degrees 170 0 east gt 0 for a non rotated lat lon grid set pollon 180 polgam REAL Angle between the north poles of two rotated grids in de 0 0 grees east
217. he driving model and the main purpose of the assimilation scheme is to analyze the meso scales By design 3 dimensional analysis methods tend to be less appropriate for this purpose They do not allow to account for the exact observation time of asynoptic data and they make it necessary to neglect most of the high frequent data unless the analysis scheme is applied very frequently at significant computational costs Moreover the geostrophic approximation a key ingredient of some of these schemes as used e g for the GME is of limited validity on the meso scale Therefore 4 dimensional methods offer potential advantages since they include the model dynamics in the assimilation process directly However the 4 dimensional variational 4DVAR method has been too expensive in the past for operational application considering the small amount of time available to produce the analyses and forecasts a Observation Nudging Basic Namelist setting lnudge TRUE Therefore a scheme based on the observation nudging technique has been developed to define the atmospheric fields It is based on an experimental nudging assimilation scheme which had been developed for the former hydrostatic model DM and its Swiss version SM Schraff 1996 Schraff 1997 and which compared favorably with the at that time operational Optimum Interpolation analysis of DM in a number of test cases he scheme for COSMO has then been adapted to the nonhydrostatic modelling
218. he earth Cpa and Cya are the specific heat of dry air at constant pressure and constant volume g is the gravity acceleration f is the Coriolis parameter Ry and Rg are the gas constants for water vapour and dry air p is the density of moist air which is calculated as a diagnostic variable from the equation of state p p Ra 1 R Ra Dd d aP T 3 6 q is the specific humidity q represents the specific water content of a category of liquid water cloud or rain water and q represents the specific water content of a category of frozen water cloud ice snow or graupel The corresponding precipitation fluxes are denoted by P and Py The terms My denote contributions from subgrid scale processes as e g turbulence and convection and Qr summarizes the diabatic heating rate due to this processes The various sources and sinks in the equations for the humidity variables due to microphysical processes of cloud and precipitation formation are denoted by S and Sf The calculation of all these terms related to subgrid scale processes is done by physical parameterization schemes An overview of the schemes used in the COSMO Model is given in Section 3 5 The term B in the equation for the vertical velocity is the buoyant acceleration given by po T To pTo vi ot B lilq q q 3 7 d p T po T Ra 9 The advection operator in terrain following coordinates is defined as 1 0 o 0 o rdi I e M COS
219. he passive reports are also listed if 1verif TRUE and lverpas TRUE The criteria to distinguish between passive and fully rejected i e omitted reports are given in Figure 8 13 In addition a message is also written if a report or a single observation level within a multi level report is redundant The first three lines in Figure 8 6 relate to redundant surface level reports At first the station identity and code type see Figure 8 12 of the redundant report are given This is followed by a list of properties of the resulting active report consisting of the rotated wind components temperature relative humidity pressure station identity code type longitude latitude order index station correction flag replacement indicator and report time relative to the initial time of the model run The replacement indicator is organized in the same REDUNDANT 60571 35 ACTIVE 1 3 21 6 290 2 0 399 6057 1 35 22 3 31 501 EF 0 2 0 REDUNDANT 11520 14 ACTIVE 2 X 2 2 273 4 0 91 98300 11520 11 14 5 50 02 T 0 0 0 gt REDUNDANT 11520 35 ACTIVE 0 2 9 274 0 0 9 ee 01520 35 14 5 50 01F 32 2 3 11520 Z ACT REJ P ZERR Z 39900 40000 QeQeekeeee 56970 11520 T ACT REJ P T FLAG 39900 40000 0 400 11520 O ACT REJ P Q FLAG 39900 40000 0 200 TEMP 11520 0 9 S258 274 0 0 00 cekeeckkesdeedek 309 ke 122 A eee 0 335 322 14 44 50 02 35 5 2 25 TEMP 11520 P 0 9 2 8 274 0
220. he same element number but different code table numbers for various fields The element numbers and code tables used by the COSMO Model are described below The program grbin1 of the supplementary GRIB library griblib can be used to decode GRIB binary code Besides the decoded data set this program does also retrieve the contents of the octets of the PDS in an integer array ipds To illustrate the structure of the PDS Table 5 2 shows the contents of the product definition section of a binary coded output array the total cloud cover CLCT The GRIB record for this field is valid for 28 10 1998 00 UTC 11 h and was created at 28 10 1998 7 04 UTC by a forecast of the COSMO Model Octet 4 ipds 2 assigns a table number to the parameter indicator number given in octet 9 Currently we use three additional code tables besides the WMO table see Table 5 3 A full list of variables defined by these tables is available from DWD Octet 6 ipds 4 indicates the process identification number which is allocated by the origi nating centre Currently we use only eight different process numbers for forecasts or analyses see Table 5 4 At DWD this number is strongly connected to the data base system because it also specifies the different application and whether it is a forecast or an analysis The level or layer for which the data are included in the GRIB record is coded in octets 10 12 ipds 8 ipds 9 where octet 10 indicates the type of level
221. header a or i the full date is specified ydate yyyymmddhh with yyyy year mm month dd day hh hour Example laf1992072100 COSMO Model uninitialized analysis for full model domain from July 21st 1992 In forecast boundary or restart files ydate consists of a single character the time unit of forecast range ytunit followed by a string ydate ytunit string Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 2 Conventions for File Names 50 Depending on ytunit the string has the following meaning t timestep mode forecast range given in timesteps f forecast mode the forecast range is given in the form ddhhmmss where dd day hh hour mm minute ss second c climate mode the forecast range is given in the form yyydddhh where yyy year ddd day of the year hh hour d day mode the full date is given in the form yyyymmddhh where yyyy year mm month dd day hh hour yextension 1 character optional Extension e g data interpolated from model to pressure levels Examples 1bff00050000 COSMO Model file with boundary values for hour 5 1fff01233000 COSMO Model forecast at day 1 23 hours and 30 minutes 1r f01000000 COSMO Model restart file for day 1 Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 3 Initial and Boundary Data 51 6 3 Initial and Boundary Data To start a forecast the files containing
222. heat and moisture fluxes due to condensation and evaporation of cloud water is not taken into account 1 D TKE based diagnostic closure Basic Namelist settings itype turb 3 13dturb FALSE For the COSMO Model a new scheme has been developed which is based on a prog nostic equation for turbulent kinetic energy TKE that is a level 2 5 closure scheme The new scheme includes the transition of turbulence which contributes mainly to the fluxes diffusive turbulence to very small scale dissipative turbulence by the action of small scale roughness elements and the handling of non local vertical diffusion due to the boundary layer scale turbulence Most important seems to be the introduc tion of a parameterization of the pressure transport term in the TKE equation that accounts for TKE production by subgrid thermal circulations The whole scheme is formulated in conservative thermodynamic variables together with a statistical cloud scheme according to Sommeria and Deardorff 1977 in order to consider subgrid scale condensation effects 3 D closure Basic Namelist settings itype turb 5 7 13dturb TRUE The parameterization of subgrid scale turbulent processes also called a subgrid scale SGS model is of particular meaning for highly resolved LES like model simulations For resolutions reaching to the kilometer scale a more adequate turbulence param eterization scheme should be used For both versions described above there is the po
223. hen it is accepted after it has failed the QC in the past After some information e g on extrapolating surface pressure and on processing of aircraft data and multi level data YUPRINT provides the following self explanatory lists Figure 8 19 surface level observations interpolated to the lowest model level multi level observation in crements from TEMP PILOT or AIRCRAFT reports or in a different format as derived from GPS ZTD reports The lines ps spread Figure 8 20 relate to the spreading of pressure observation incre ments at the lowest model level They consist of the station identity the local part of the nudging weights i e temporal and quality weight of the 2 reports from the station the total weight at grid point ion12 jon12 from all the reports processed previously including the present report the 2 observation increments in Pa the total sum of weighted increments at that grid point the distance in km of the station to that point the indication whether that point is within the area of influence and the orography dependent correction of the lat eral weights at that point The line ps ana incr provides the local coordinates of point ion12 jon12 and the final surface pressure analysis increment at that point Interpolation of surface level observations to the lowest model level ke upper air single level reports are also listed here station height differences with the flag s are scaled as for
224. how many NAMELIST groups GRIBOUT are 1 contained for the model run Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 11 IOCTL Controlling the Grib I O 138 Name Type Definition Purpose Comments Default nvers INT The version number of a model run for documenting pur 1 poses nvers is coded in the PDS of GRIB output files and is the only parameter to distinguish GRIB output files for the same case but coming from different model versions ncenter INT To specify the WMO identification of generating center 78 nsubcenter INT To specify the WMO identification of the originating sub 255 center to set GRIB2 metadata nlocaldefnr INT To specify the local definition number for GRIB local sec 1 tions The default value of 1 means that no local section is present yform_read CHAR Specifies the format of input files The following formats grbi are implemented grbi default read GRIB1 data with DWD s GRIB library apix read GRIB 1 or 2 data with ECMWF s grib api ncdf read NetCDF data yform write CHAR In Version 4 18 this parameter has been moved to GRIBOUT This gives the possibility that it can be spec ified for every output group Specifies the format of output files can be grb1 or ncdf ymode read CHAR Specifies the mode how files are opened for reading v E ymode write CHAR Specifies the mode how files are opened for writing
225. ht_2m units m height_2m positive up float height_10m height_10m height_10m standard_name height height_10m long_name height above the surface height 10m units m height 10m positive up float soili soili soili standard name depth Soili long name depth of soil layers Soili units m soili positive down soili bounds soili_bnds float soili bnds soili sbnds Soili bnds long name boundaries of soil layers double time time time standard name time time long name time time units seconds since 1979 01 01 00 00 00 time calendar proleptic gregorian time bounds time bnds double time bnds time tbnds time bnds long name time bounds time bnds units seconds since 1979 01 01 00 00 00 float P time level rlat rlon P standard name air pressure P long name pressure P units Pa P grid mapping rotated pole P coordinates lon lat float PS time rlat rlon PS standard name surface air pressure PS long name surface pressure PS units Pa PS grid mapping rotated pole PS coordinates lon lat float T time level rlat rlon T standard name air temperature T long name temperature T units K T grid mapping rotated pole T coordinates lon lat float U time level rlat srlon U standard name grid eastward wind U long name U component of wind U units m s 1 U
226. ial filtering lhn filt LOG vertical filtering of LHN temperature increments TRUE lhn relax LOG horizontal filtering both of the TRUE observed and simulated precipitation fields input LHN temperature increments output nlhn relax INT number of iterations of 2 grid point scale horiz filtering 2 thresholds and weighting of observation input thres lhn REAL threshold in mm s below which the rain rate is consid 0 1 3600 ered to be zero in many aspects within the LHN scheme rad wobs lhn REAL maximum range from radar site with full observation weight 100 in km lhn spqual LOG use of a spatial quality function of the input rain rates FALSE observation and auxilliary file input nradar INT maximum number of radar stations in obs input data 33 lhn dt obs REAL time step of the observation input data in minutes 5 0 noobs date 36 CHAR dates without radar data ps 12 as YYYYMMDDHHTT where YYYY year MM month DD day HH hour TT minute lhn black LOG use of a blacklist for radar data TRUE this requires an additional Grib file see blacklist lhn bright LOG bright band detection and rejection of flagged data TRUE this requires radar beam height maps lhn height LOG use of radar beam height maps TRUE this requires an additional Grib file see height file Name Type Definition Purpose Comments Default directory
227. idual nudging time window relative to observation time in h time in h for which analysis increments are to be computed currently end of period for which these analysis increments are valid temperature observation error in K temperature nudging coefficient for aircrafts in s observation level pressure in Pa Time is always specified relative to the initial model time except where indicated differently NUMBER OF SINGLE AND MULTI LEVEL AND GPS DATA TO BE PRINTED NODE CART ID NTOTSGO NTOTSG NTOTMLO NTOTML NTOTGPO NTOTGP 0 0 128 0 38 0 152 0 864 0 163 0 3023 14 0 78 0 6 0 11 MAX LOCAL NUMBER 0 5810 0 430 0 4046 LOCAL ARRAY SIZE 8360 557 11704 GLOBAL NUMBER 0 14748 0 1150 0 15611 airep 0 13 0 630 0 011 EU3080 0 5 0 5 Les 0 0 0 03 0 8 0 0006 72990 mladm 45 multi level reports mladm 1 0 1 41 55 02591 999 00 2 67 0 00 3 00 3 00 0 00 0 12 0 00 0 12 mladm 45 0 128 35 87 EU7654 999 00 0 33 0 00 50 50 0 00 0 80 0 00 0 80 sgadm 429 surface level and 354 upper air single level reports sgadm 1 843 0 217 119 22892 0 00 999 00 0 50 0 00 0 00 0 94 0 00 0 94 0 00 sgadm 429 0 1750 19 101 LF4B 999 00 1 17 0 00 50 17 0 00 0 24 0 00 0 24 sgadm 430 13 0 31 84 EU3080 0 47 999 00 0 50 0 00 0 00 0 01 0 00 0 01 0 00 sgadm 783 0 785 71 70 EUO301 999 00 1 52 0 00 50 50 0 00 0 01 0 00 0 01 gpadm 181 GPS reports gpadm 1 1 701 91 22 BUDP BKG_ 0 50 0 50 1 00 00 00 0 47 0 53 0 00 0 69 gpadm 181 0 1327 14
228. ikely to be of some value for monitoring the model run A small part of the information written to YUPRINT is also written to the standard output or to file YUCAUTN These parts are described in the respective sections 8 2 9 and 8 2 5 Much of the information is local i e it relates only to a certain sub domain which accommodates the grid point with coordinates given by the NAMELIST variables ionl jonl and which coincides with the area processed by one node on an MPP massively parallel platform i e with distributed memory computing environment Also most types of information are given only at the first timestep or once per hour After some self explanatory header information about the run some details about the pro cessing of aircraft data and cloud observations the numbers of single level reports NSGOB of multi level reports NMLOB and of GPS reports NGPOB currently stored on each sub domain related to nodes CART_ID are provided Figure 8 18 NTOTSGO NTOTMLO and NTOTGPO denote the number of reports which have been read at previous timesteps GLOBAL NUMBER relates to the number of reports on the total model domain The lines starting with airep provide the following information for an aircraft report number of active reports internal report index passive report flag end of nudging period in timesteps timestep in h station identity observation time beginning and end of indiv
229. ion type PILOT TRUE lsatem LOG observation type SATEM FALSE lgps LOG observation type GPS FALSE lscatt LOG observation type SCATT scatterometer TRUE code type reports inside the exclusion area are set passive if their code type is set to FALSE 1cd011 LOG SYNOP code 11 manual land TRUE lcd014 LOG SYNOP code 14 autom land TRUE 1cd021 LOG SYNOP code 21 manual ship TRUE 1cd022 LOG SYNOP code 22 abbrev ship TRUE 1cd023 LOG SYNOP code 23 reduced ship TRUE 1cd024 LOG SYNOP code 24 autom ship TRUE 1cdi40 LOG SYNOP code 140 METAR TRUE Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 129 Use of observation types and code types continued Name Type Definition Purpose Comments Default code type continued reports inside the exclusion area are set passive if their code type is set to FALSE 1cd041 LOG AIREP code 41 CODAR TRUE 1cd141 LOG AIREP code 141 AIREP TRUE 1cd241 LOG AIREP code 241 constant level balloon TRUE 1cdi44 LOG AIREP code 144 AMDAR TRUE 1cd244 LOG AIREP code 244 ACAR TRUE 1cd088 LOG SATOB code 88 SATOB TRUE 1cd188 LOG SATOB code 188 SST TRUE 1cd063 LOG DRIBU code 63 bathy sphere TRUE 1cd064 LOG DRIBU code 64 TESAC TRUE 1cd165 LOG DRIBU code 165 drifting buoy TRUE 1c
230. iptor 0 02 003 9 13 past weather WMO descriptor 0 02 004 14 19 time period of past weather VUB WMO Code table 4019 keys 0 7 1 Dh 2 12h 3 18h de 24h 5 1h 6 2h 7 3h 20 21 accuracy flag for low cloud cover 0 high accuracy 1 low accuracy 22 2T state of ground WMO descriptor 0 02 062 28 30 precip obs duration 2 12 hrs 0 otherwise Some VUB WMO Code tables relate to WMO descriptors in the following way VUB WMO table 0509 code figures 10 19 of WMO decriptor 0 20012 10 VUB WMO table 0513 code figures 20 29 of WMO decriptor 0 20012 20 VUB WMO table 0515 code figures 30 39 of WMO decriptor 0 20012 30 VUB WMO table 4677 code figures 00 99 of WMO decriptor 0 20 003 e e 92 2 Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 177 In VUB WMO Code table 2700 code figures 0 8 indicate the cloud cover in octas and code figure 9 indicates sky or clouds invisible VUB WMO Code table 1600 is defined as follows 0 50m 1 100m 2 200m 3 300m 4 600m 5 1000m 6 lt 1500m 7 lt 2000m 8 2500m 9 gt 2500 m or cloud free 15 undefined Report Body Extension for Increments The regular report body is followed by an optional report body extension if the NAMELIST variable mruntyp gt 0 It consists of observation increments or full model value
231. it simulation of deep convective clouds This is available only as a prognostic scheme This scheme is used in the COSMO DE The use of itype gscp 1 2 is not recommended for real cases Also the diagnostic schemes lprogprec FALSE are no longer used by DWD Therefore they are no longer tested and evaluated 3 5 3 Moist Convection Basic Namelist settings lphys TRUE lconv TRUE For model applications on the meso a and meso f scales down to grid spacings of 5 10 km cumulus convection is a subgrid scale process which requires a parameterized representation And even on the meso y scale it turned out that a parameterization of shallow convection still is necessary The COSMO Model offers three options a Mass flux Tiedtke scheme Basic Namelist settings ltiedtke TRUE lkainfri FALSE lshallow FALSE The mass flux scheme of Tiedtke 1989 which is used in the COSMO EU has been implemented for the meso o and meso f scale This parameterization discriminates three types of moist convection shallow convection penetrative convection and mi dlevel convection which are treated by different closure conditions Both shallow and penetrative convection have their roots in the atmospheric boundary layer but they differ in vertical extent Midlevel convection on the other hand has its roots not in the boundary layer but originates at levels within the free atmosphere As a closure condition the Tiedtke scheme requires a formulation of
232. k kk OF 2147483647 13552 103651382 32 17777777777 32360 613314066 40 0 131073280 256 0 764002400 400 2147483647 109580320 17777777777 642010040 133757695 776175377 2147483647 18867 20020109 16166 0 177777717717 44663 114275615 37446 0 2692 2678 64 T2 0 5204 5166 100 177 0 ck ckckckck ck ck ck k ck ck ck ckck ck ck ck kckck ck k ck ck ck ck ck ck k ck kckckckckck ck ck ck ck ck ck ckck ck ck ck kc Kk kk A OF 0 27 4 2 144 1100 1158 2147483647 1032192 33 4 2 220 2114 2206 17777777777 3740000 0 2381 212 0 0 4515 324 0 0 0 35 4 1 11 1200 1200 1169 65 43 4 1 13 2260 2260 2221 101 0 10160 999 502 2147483647 0 0 23660 1747 766 17777777777 0 0 0 332 4 5 35 1200 1200 1450 768 514 4 5 43 2260 2260 2652 1400 0 9830 150 1453 507904 23146 226 2655 1740000 2147483647 14057 103553716 127 17777777777 33351 613015264 177 1 2147483647 17777777777 117968928 702010040 Figure 8 5 Example file YUAOFEX 2147483647 19038 20020109 56 0 IUVPUPDPUTTTI 45136 114275615 70 0 2679 2676 5167 5164 REPORT t r REPOR REPORT Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 158 8 2 2 YUOBSDR Nudging Active and Passive Reports File YUOBSDR lists the active reports It is written only if the NAMELIST parameter 1prodr is set to TRUE T
233. ld T_SO with time range indicator 13 is used Note analysis field T_SO with time range indicator 13 is used Note analysis field T_SO with time range indicator 0 is used te analysis field W_SO with add element number 20 is used Note analysis field W_SO with add element number 20 is used Note analysis field W_SO with add element number 0 is discarded Note analysis field W_SO with add element number 0 is discarded Figure 8 26 Example messages on the originating processes for 2 dimensional grib fields which are part of the initial condition Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 186 8 2 10 YUSURF 2 D Surface Analyses YUSURF relates to the determination of two dimensional analyses which are based mainly on observations and can be used for verification purposes and as input for separate analy sis schemes such as the variational soil moisture analysis There are four types of analyses namely the 2 m temperature analysis the 2 m relative humidity analysis the 10 m wind speed analysis and the precipitation analysis While the latter is based purely on observa tions the first three include the corresponding fields of the model run as first guess fields The basic functions and parameters of the successive correction schemes are first outlined in YUSURF Figure 8 27 Then some basic information is provided for the analysis at grid point ionl jonl
234. lds Additional two dimensional intermittent analysis Soil Moisture Analysis Daily adjustment of soil moisture by a variational method Hess 2001 in order to improve 2 m temperature forecasts use of a Kalman Filter like background weighting Sea Surface Temperature Analysis Daily Cressman type correction and blending with global analysis Use of external sea ice cover analysis Snow Depth Analysis 6 hourly analysis by weighted averaging of snow depth obser vations and use of snowfall data and predicted snow depth and Parallelization Code Structure Modular code structure using standard Fortran constructs Parallelization The parallelization is done by horizontal domain decomposition using a soft coded gridline halo 2 lines for Leapfrog 3 for the Runge Kutta scheme The Message Passing Interface software MPI is used for message passing on distributed memory machines Compilation of the Code The compilation of all programs is performed by a Unix shell script invoking the Unix make command All dependencies of the routines are automatically taken into account by the script Portability The model can be easily ported to various platforms current applications are on conventional scalar machines UNIX workstations LINUX and Windows NT PCs on vector computers NEC SX series and MPP machines CRAY IBM SGI and others Model Geometry 3 d 2 d and 1 d model configurations Metrical terms can be adjusted
235. led with DRTTOV10 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell of ice particles e 1 Ou and Liou 1995 Atmos Res 35 127 138 e 2 Wyser et al see McFarquhar et al 2003 e 3 Boudala et al 2002 Int J Climatol 22 1267 1284 e 4 McFarquhar et al 2003 NOTE Only scheme 4 has been tested extensively 7 7 SATCTL Controlling the Synthetic Satellite Images 112 Name Type Definition Purpose Comments Default sat input 01 TYPE Structure to specify characteristics for first instrument and to specify the products that shall be generated CHAR Name of the satellite YY YY YY yy INT Satellite identification 0 CHAR Name of the sensor YY YY YY Y YY yyy INT Number of channels used 0 LOG To generate clear sky radiance FALSE LOG To generate cloudy sky radiance FALSE LOG To generate clear sky brightness temperature FALSE LOG To generate cloudy sky brightness temperature FALSE sat input 02 TYPE Structure to specify characteristics for second instrument and to specify the products that shall be generated same as above nchan input O01 INT input channel list for first sensor all 0 nchan input 02 INT input channel list for second sensor all 0 emiss input O1 REAL To read emissivities for all channels of first instrument all 0 0 emiss input 02 REAL To read emissivities for all channels of second instrument all 0 0
236. level reports see also report header 11 for upper air single level or very short surface level reports 10 for each observation level of multi level reports 11 for ground based GPS reports on ZTD IWV integ water vapour For multi level reports the regular report body consists of as many lines with 10 entries each as there are observation levels The following list declares the complete set of 22 entries regular report body 1 zonal wind component 1 10 m s for GPS reports derived IWV ice to water saturat bias adjusted 1 100 mm 2 meridional wind component GPS reported IWV 1 100 mm 1 10 m s 3 temperature 1 10 K 4 relative humidity ice to water saturation adjusted 1 10 96 5 pressure Pa 6 height GPS zenith total delay ZTD mm m 7 observation status flag bit pattern as decimal number see below if bit is set to 1 then the obs is active unless the QC flag is set or status Z 0 in report header 8 QC flag bit pattern as decimal number see below if bit is set to 1 then the obs is rejected by the threshold quality control bit 3 for height is also set if upper air obs is below model orography Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 175 9 10 11 12 13 14 15 16 17 18 19 20 21 22 main flag word for wind temperat humidity pressure as octal number see below lev
237. loud water ex ists The calculation of the fractional cloud cover o in each model layer is calculated based on a traditional scheme which has been used in the former operational hydrostatic models EM DM o is determined by an empirical function depending on the relative humidity the height of the model layer and the convective activity In addition to the EM DM scheme the contribution of convection to o is assumed to depend on the vertical extent of the con vection cell by prescribing a heuristic function Also a check for temperature inversions at the convective cloud tops is done to take anvils by an increase of c in case of inversions into account Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 5 Physical Parameterizations 23 3 5 4 Vertical Turbulent Diffusion Basic Namelist settings lphys TRUE ltur TRUE For vertical turbulent diffusion several schemes are available a 1 D diagnostic closure Basic Namelist settings itype turb 1 13dturb FALSE In the original EM DM scheme the vertical diffusion due to turbulent transport in the atmosphere is parameterized by a second order closure scheme at hierarchy level 2 0 Mellor and Yamada 1974 M ller 1981 This results in a diagnostic closure where the turbulent diffusion coefficients are calculated in terms of the stability of the thermal stratification and the vertical wind shear The impact of subgrid scale effects on the
238. ls of the radiation 360 scheme hincrad REAL As nincrad but time interval in hours In general an inter 0 0 val of 0 5 1 hour yields sufficient accuracy Can be specified alternatively to nincrad icldm_rad INT Parameter to select the mode of cloud representation i e 4 cloud cover and cloud water and ice content as input to the radiation parameterization 0 No clouds are considered 1 Only grid scale clouds are considered 2 Grid and sub grid scale water clouds are considered cloud cover and water content are calculated according to a relative humidity criterion itype wcld 1 or a statistical closure itype wcld 2 3 Grid and sub grid scale including convective water and ice clouds are considered cloud cover water con tent and ice content are calculated by the default di agnostic scheme 4 at the moment same as 3 forest LOG Switch to choose the external parameter fields describing ev FALSE ergreen and deciduous forest nradcoarse INT If nradcoarse gt 1 the radiation is computed on a coarser 1 grid to save computation time nradcoarse grid points in every horizontal dimension are combined Maximal possible value is nboundlines lradf avg LOG To average the radiative forcings when running on a coarser FALSE grid nradcoarse 1 lradtopo LOG To use topographic corrections for the radiation FALSE nhori INT Number of sectors for the horizont used by the topographic 24
239. mary messages are also written to YUSTATS As shown in Figure 8 16 there are two CAUTION CAUTION CAUTION CAUTION CAUTION RRR RR RRR RRR RN NR RR RR NR A N A A A SEI A 0 WARNING array size for multi level observations is too small to accommodate all observations Increase MAXMLO namelist by 312 usually ok for local obs array possibly still insufficient for the global obs increment array 0 WARNING array size for GPS observations is too small to accommodate all observations Increase MAXGPO namelist by 1190 usually ok for local obs array possibly still insufficient for the global obs increment array 1 0 CAUTION CAUTION CAUTION CAUTION CAUTION RRR RRR RRR RRR RRR WARNING array size for multi level obs increments is too small to accommodate all obs increments Increase MAXMLO namelist from 350 to at least 373 or increase MAXGPO namelist from2000 to at least2046 or increase MAXTVO namelist from 1 to at least 24 t WARNING array size for upper air single level obs increments is too small to accommodate all obs increments Increase MAXUSO namelist from 2500 to at least 2617 t WARNING array size for surface level obs increments is too small to accommodate all obs increments Increase MAXSGO namelist from 4000 to at least 4089
240. me is numerically very expensive Thus a timestep number increment can be specified for which the convection scheme is called The convective tendencies are then stored and remain fixed for the following time steps b Kain Fritsch scheme Basic Namelist settings ltiedtke FALSE lkainfri TRUE lshallow FALSE This scheme has lately been implemented for testing Up to now there is no detailed evaluation available c A scheme for shallow convection Basic Namelist settings ltiedtke FALSE lkainfri FALSE lshallow TRUE This scheme has been extracted from the Tiedtke scheme and can be used for the convection permitting scales It is applied for the COSMO DE Fractional Cloud Cover In the parameterization schemes for grid scale clouds and precipitation the condensation rate for cloud water is based on saturation equilibrium with respect to water Consequently a grid element is either fully filled with clouds at water saturation where q gt 0 relative humidity 100 or it is cloud free at water subsaturation where q 0 relative humidity lt 100 The area fraction of a grid element covered with grid scale clouds is thus a bivalued parameter which is either 1 or 0 However with respect to the calculation of radiative transfer but also for weather interpre tation in postprocessing routines it is useful to define a fractional cloud cover also for those grid boxes where the relative humidity is less than 100 and no grid scale c
241. me range indicators used by the COSMO Model ipds 19 Meaning 0 Forecast product valid for reference time P1 if P1 gt 0 or uninitialized analysis product valid for reference time P1 0 1 initialized analysis product valid for reference time P1 0 Product with a valid time ranging between reference time 4 P1 and reference time 4 P2 3 Average from reference time 4 P1 to reference time 4 P2 4 Accumulation from reference time P1 to reference time P2 product valid for reference time P2 5 1 4 Grid Description Section Section 2 of a GRIB record the grid description section GDS contains all information about the geometry of the grid on which the data are defined For all input and output files of the model this section is coded completely for every field contained in the file The program grbint of the supplementary GRIB library griblib retrieves the contents of the GDS in an integer array igds The contents of the grid description section of a model GRIB record is illustrated in Table 5 8 for the model domain used operationally at DWD The octets corresponding to the integer array igds are numbered relative to this section 5 1 5 Bit map Section This section is optional and provides the possibility to include only some grid points of the grid defined in the Grid Description Section The bit map is a sequence of bits with a bit to grid point correspondence ordered as defined in the gri
242. mean level pO k Ivi wind dir T p FI qv qc hPa m s deg K hPa m2 s2 g kg mg kg 1 30 00 0 174 14 29 0 728 0 13 211 7 0 000 0 000 2 50 18 0 074 73 22 1 228 0 08 88 7 0 000 0 000 3 70 87 0 025 124 75 0 064 0 04 32 0 000 0 000 4 92 55 0 241 46 38 0 176 0 06 41 5 0 000 0 000 5 114 84 0 E50 5 74 0 186 0 07 41 1 0 000 0 000 6 138 49 0 096 8 00 0 267 0 06 28 1 0 000 0 000 7 163 43 0 247 5 58 0 695 0 01 349 0 000 0 000 8 189 73 0 377 23 58 0 051 0 04 13 45 0 000 0 000 9 217 42 0 302 7 52 0 379 0 02 76 9 0 000 0 000 10 246 48 0 227 10 69 0 273 0 02 4 5 0 000 0 000 11 276 87 0 144 8 65 20 221 0 06 12 4 0 001 0 000 12 308 52 0 109 6 54 0 232 0 10 19 8 0 002 0 002 29 867 95 0 017 37 67 0 046 0 18 16 2 0 067 4 879 30 890 50 0 021 44 06 0 045 0 18 1549 0 084 7 029 31 910 85 0 021 38 69 0 109 0 18 15 5 0 122 3 819 32 928 96 0 031 69 08 0 146 0 18 14 9 0 109 1 592 33 944 79 0 092 76 08 0 138 0 17 14 3 0 101 0 094 34 958 36 0 207 50 72 0 138 0 17 13 9 0 083 0 338 35 969 73 0 301 20 85 0 175 0 16 1345 0 075 0 007 36 979 00 0 339 1 08 0 216 0 16 13 0 0 069 0 544 37 986 31 50 353 19 18 0 249 0 15 12 6 0 070 0 494 38 991 87 0 366 37 67 0 255 0 15 12 2 0 088 0 447 39 995 93 0 375 51 05 0 251 0 15 12 0 0 092 0 369 40 998 82 0 377 63 93 0 251 0 10 8 0 0 095 0 121 Figure 8 11 Example file YUSTATS first part incomplete Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Relat
243. melist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 130 The previous blocks of NAMELIST variables except for the verification block determine the meteorologically relevant contents of the nudging scheme i e they determine the direct explicit influence of the various observations on the model fields Once the values for these variables have been set in a quasi operational environment they usually do not need to be adjusted according to the day to day variation of the observational supply except when the observation system changes so dramatically that it modifies the fundamental behaviour of the system This is in some contrast to the following block of NAMELIST variables and in particular to the first 4 variables They determine the size of the run time arrays which contain the observations and observation increments They should be large enough to accommodate all these data but at the same time they must not be too large because these arrays contribute to the processor memory required Thus if the number of available observations e g aircraft data increases significantly it may occur that the arrays are filled up completely and some of the good data have to be neglected In such a case the program does not stop or crash but it will issue messages with the label CAUTION both to standard output and additional ASCII output files This is decribed in sections 8 2 6 and above all 8 2 5 The messages also
244. mer m gt P 2 Ptop upper boundary Prop of temperature correction in hPa 400 0 luvgcor qgeo qgeotop LOG REAL REAL geostrophic wind correction wind mass field balancing on off switch for geostrophic wind correction fraction of the full geostrophic wind increments that is added to the model wind fields at 1000 hPa fraction of the full geostrophic wind increments that is added to the model wind fields at level ptpstop in between this fraction is linearly interpolated TRUE 0 3 0 5 mpsgcor qgeops INT REAL geostrophic pressure correction wind mass field balancing switch for geostrophic pressure correction 0 no pressure correction correction balacing wind increments from scatterometer 2 corr balacing scatterometer 4 in situ 10 m wind increm fraction of the full geostrophic pressure increments that is added to the pressure field at the lowest model level 0 9 khumbal INT humidity temperature balancing radius in number of mesh widths of the area around a con vectively precipitating grid point in which specific humidity instead of relative humidity is preserved when temperature is nudged Special cases 1 relative humidity preserved everywhere gt 99 specific humidity preserved everywhere gt 100 specific humidity preserved additionally for incre ments from hydrostatic temperature correction 100
245. missing The use of the various variables is defined as follows need COSMO asks stringently for this variable and will abort if variable is absent opt X variable exists and is read used but COSMO will not abort if it does not exist P variable exists but is not read by COSMO descriptor exists only in the BUFR file not in the NetCDF file use WMO descriptor type mnemonics meaning opt 0 01 033 int MMIOGC Identific of orig generat centre 1 001 034 int MMIOGS Identific of orig gen sub centre 1 001 023 int NOSNO Observation sequence number 1 need 0 01 008 char 8 YAIRN Aircraft identification 3 01 011 Year month and day need 0 04 001 int MJJJ Year need 0 04 002 int MMM Month need 0 04 003 int MYY Day 3 01 013 Hour minute and second need 0 04 004 int MGG Hour need 0 04 005 int NGG Minute t 0 04 006 int MSEC Second 3 01 021 Latitude and Longitude need 0 05 001 float MLAH Latitude high accuracy degree need 0 06 001 float MLOH Longitude high accuracy degree need 0 08 004 int MPHAI Phase of flight Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files 69 use WMO descriptor type mnemonics meaning 1 01 000 Delayed replication of 1 descriptor need 0 31 001 int MDREP Delayed descriptor replication factor 311 00
246. model heat capacity water storage capacity etc strongly depend on soil texture Five different types are distinguished sand sandy loam loam loamy clay and clay Three special soil types are considered additionally ice rock and peat Hydrological processes in the ground are not considered for ice and rock Potential evaporation however is assumed to occur over ice where the soil water content remains unchanged Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 5 Physical Parameterizations 25 Z In the default configuration the thicknesses of the upper and lower thermal layers are taken to be 9 cm and 32 cm respectively and two layers of 10 cm and 90 cm depth are used for the hydrological calculations Below these soil layers climatological values for temperature and soil moisture are prescribed The soil model is run for all gridpoints with a land fraction larger or equal than 50 All other gridpoints are treated as sea points with an initial surface temperature which remains constant throughout a model run For operational applications the sea surface temperature is provided by an external analysis scheme The multi layer soil model TERRA ML Basic Namelist settings Imulti layer TRUE Recently the multi layer version TERRA ML of the soil model has been implemented in the COSMO Model as an option The main differences of this version in comparison to the older version TERRA are
247. mperature or in case the lake in question is covered by ice to the ice surface temperature In the present configuration a recommended choice the heat flux through the lake water bottom sediment interface is set to zero and a layer of snow over the lake ice is not considered explicitly T he effect of snow above the ice is accounted for parametrically through changes in the surface albedo with respect to solar radiation Optionally the bottom sediment module and the snow module can be switched on Then additional equations are carried for the snow surface tempera ture temperature at the air snow interface for the snow depth for the temperature at the bottom of the upper layer of bottom sediments penetrated by the thermal wave and for the depth of that layer Surface fluxes of momentum and of sensible and latent heat are computed with the operational COSMO model surface layer parameterization scheme Optionally a new surface layer scheme can be used that accounts for specific features of the surface air layer over lakes In order to be used within the COSMO model or within any other NWP or climate model FLake requires a number of two dimensional external parameter fields These are first of all the fields of lake fraction area fraction of a given numerical model grid box covered by lake water that must be compatible with the land sea mask used and of lake depth Other external parameters e g optical characteristics of the lake water
248. mple file YUSTATS second part Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 168 Reports may only be rejected or set passive for reasons given by report events 1 13 except for events 3 and 5 on station altitude for TEMPs and PILOTs Hence the number of these events must equal the number of rejected and passive reports In the verification mode i e if data are written to the VOF reports are rejected in case of report events 1 3 8 13 and event 4 if the re port is outside the model domain Otherwise the reports are set passive except that already passive reports are also rejected if they do not contain any data or if they are redundant and a subset of an active report Without verification reports are always rejected for events 1 to 13 0 1 REPORT EVENTS DEFINITIONS THEIR ORDER MATCHES THE ORDER OF THE CHECKS 1 DATA BASE FLAG ON LOCATION TIME ALTITUDE HIGH 2 OBSERVATION TIME OUT OF RANGE TOO OLD OR TIME MISSING 3 STATION ALTITUDE MISSING 4 STATION LOCATION OUT OF DOMAIN OR OUT OF USER SPECIFIED AREA 5 DISTANCE MODEL OROGRAPHY STATION ALTITUDE TOO LARGE 6 BLACKLISTED SHIP 7 OBSERVATION OR CODE TYPE EXCLUDED IN AREA AROUND STATION LOCATION 8 REPORT NUMBER EXCEEDING SIZE OF ODR gt ADJUST NAMELIST 9 NO ACCEPTED DATA IN REPORT 10 PRESSURE TOO SMALL lt 9 HPA OR MISSING WITH
249. n and orographically and thermally forced local wind systems Since April 2007 a meso y scale version is running operationally at DWD by employing a grid spacing of 2 8 km Applications with similar resolutions are now run by most COSMO partners We expect that this will allow for a direct simulation of severe weather events triggered by deep moist convection such as supercell thunderstorms intense mesoscale convective complexes prefrontal squall line storms and heavy snowfall from wintertime mesocyclones The requirements for the data assimilation system for the operational COSMO Model are mainly determined by the very high resolution of the model and by the task to employ it also for nowcasting purposes in the future Hence detailed high resolution analyses have to be able to be produced frequently and quickly and this requires a thorough use of asynoptic and high frequency observations such as aircraft data and remote sensing data Since both 3 dimensional and 4 dimensional variational methods tend to be less appropriate for this purpose a scheme based on the observation nudging technique has been chosen for data assimilation Besides the operational application the COSMO Model provides a nonhydrostatic mod elling framework for various scientific and technical purposes Examples are applications of the model to large eddy simulations cloud resolving simulations studies on orographic flow systems and storm dynamics development and valida
250. n optic data and remote sensing data use of pure Fortran constructs to render the code portable among a variety of com puter systems and application of the standard MPlI software for message passing on distributed memory machines to accommodate broad classes of parallel computers The development of the COSMO Model was organized along these basic guidelines How ever not all of the requirements are fully implemented and development work and further improvement is an ongoing task The main features and characteristics of the present release are summarized below Dynamics Model Equations Nonhydrostatic full compressible hydro thermodynamical equations in advection form Subtraction of a hydrostatic base state at rest Prognostic Variables Horizontal and vertical Cartesian wind components pressure per turbation temperature specific humidity cloud water content Optionally cloud ice content turbulent kinetic energy specific water content of rain snow and graupel Diagnostic Variables Total air density precipitation fluxes of rain and snow Coordinate System Generalized terrain following height coordinate with rotated geograph ical coordinates and user defined grid stretching in the vertical Options for i base state pressure based height coordinate ii Gal Chen height coordinate and iii exponential height coordinate SLEVE according to Sch r et al 2002 Numerics Grid Structure Arakawa C grid
251. n AG 1 of the independent variables are used to set up the computational grid To simplify the no tation we set the vertical grid spacing equal to one see below The discrete computational A v space is then represented by a finite number of grid points i j k where i corre sponds to the A direction j to the y direction and k to the direction The position of the grid points in the computational space is defined by Aj Ao 6 1 AA 1 Ny Ck k R 1 Ne N denotes the number of grid points in A direction N the number of points in the p direction and N the number of points in the G direction Ap and po define the south western corner of the model domain with respect to the rotated geographical coordinates A Thus 7 1 and i N correspond respectively to the western and the eastern boundaries of the domain Accordingly the southern and the northern borderlines are given by j 1 and j N The corresponding variables in the programs are dlon AA dlat Ay startlon tot Ao startlat tot qo ie tot N je tot N and ke tot Ne Every grid point i j k represents the centre of an elementary rectangular grid volume with side lengths AA Ay and AC The grid box faces are located halfway between the grid points in the corresponding directions i e at Aj41 2 Pj 1 2 and Ch 1 2 The model variables are staggered on an Arakawa C Lorenz grid with scalars temperature p
252. n Input Files 60 The use of the variables first column of table is defined as follows need COSMO asks stringently for this variable and will abort if variable is absent but will not abort if values are equal to missing value opt variable exists and is read used but COSMO will not abort if it does not exist used means here that it is e g written to the feedobs file but it does not imply active use in the data assimilation MP variable exists but is not read by COSMO ms descriptor a BUFR common sequence or a BUFR data description operator exists only in the BUFR file not in the NetCDF file use WMO descriptor type mnemonics meaning E 3 01 111 Identific of launch site instumentation 3 01 001 Station identification need 0 01 001 int MII WMO block number 1 need 0 01 002 int NIII WMO station number 1 need 0 01 011 char 9 YDDDD Ship or mobile land station identifier 1 need 0 02 011 int NRARA Radiosonde type opt 0 02 013 int NSR Solar and infrared radiation correction opt 0 02 014 int NSASA Tracking technique status of system opt 0 02 003 int NAA Type of measuring equipment 301113 Date time of launch opt 0 08 021 int MTISI Time significance 18 launch time x 3 01 011 Year month and day need 0 04 001 int MJJJ Year need 0 04 002 int MMM Month need 0 04 003 int MYY Day 3 01 013 Hour minute second need 0 04 00
253. n Version 4 4 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 2 RUNCTL Parameters for the Model Run 83 Name Type Definition Purpose Comments Default itype timing INT To specify how a timing of the program should be done 4 1 output hourly timings per task 2 output timings per task summed up over all hours 3 output hourly mean values for all tasks 4 output mean values for all tasks summed up over all hours lreproduce LOG Ensures the computation of reproducible results in par FALSE allel mode lreorder LOG If TRUE the numbering of the MPI processes can be TRUE reordered for the cartesian MPI communicator ldatatypes LOG Switch to choose between an explicit buffering FALSE FALSE or an implicit buffering TRUE using MPI datatypes for the boundary exchange ltime barrier LOG If TRUE some more barriers are called during the TRUE model run to separate communications from compu tations in a very clean way But without these barriers the model will run faster ncomm type INT To choose a certain type of communication 1 1 immediate send blocking receive and wait on the sender side 2 immediate receive blocking send and wait on the receiver side 3 using MPI SENDRECV Parameters for diagnostic min max model output Name Type Definition Purpose Comments Default hincmxt REAL Intervalin hours
254. n of the synthetic satellite images is not possible then 4 2 Working with the VCS The Version Control System is a programming environment tool based on the Concurrent Version System CVS The programming environment consists of several shell scripts or command procedures that are accessible from an administrator directory on DWD systems this directory is e rhome for0adm vcscmd you can refer to this directory with the shell variable ADM if it is set properly These command procedures serve to simplify tasks and contain safety features which may otherwise be easily forgotten External users having a collaboration with DWD can access the code of the COSMO Model and also of other models the necessary scripts for installing the programming environment tool and a description of that tool via ftp A list of all command procedures together with a short explanation can be obtained with ADM help 4 3 Preparing the Code Source Code Administrator for VCS As a source code administrator you have to provide the external code and libraries They have to be created on your system and put to a special directory They also have to be specified as EXTOBJ in LinkLibs in order to link them to the object files of the COSMO Model User with VCS If working with the VCS you have to create your own workbench within a special directory e g HOME model with the command ADM workbench 1m f90 Part VII User s Guide 4 28 Section 4 Install
255. nd column indicates whether a variable is in a special default list for output e Specifying the time steps when these variables shall be written There are two ways of specifying the output steps With a list of time steps ngrib or alternatively a list of hours hgrib e g ngrib 0 2 4 24 138 400 hgrib 0 0 0 5 1 0 1 75 4 3 Up to 100 different output steps can be specified Part VII User s Guide 4 28 Section 8 Model Output 8 4 Output of Forecast Fields 190 With a list of begin end and increment steps given in time steps ncomb or in hours hcomb The values have to be given in triples e g hcomb 0 0 12 0 1 0 12 0 24 0 0 5 24 0 48 0 2 0 With this specification the following output is performed From forecast hour 0 0 to 12 0 results are written every hour From forecast hour 12 0 to 24 0 results are written every 30 Minutes From forecast hour 24 0 to 48 0 results are written every 2 hours If nothing is specified for these variables results will be written every hour starting with the beginning of the forecast Specifying the domain for which these variables shall be written With the variable ydomain you can specify whether the variables are written for the full domain ydomain f default or for a subdomain ydomain s In case of a subdomain you also have to define the start and endpoints of this subdomain slon Slat elon and elat in rotat
256. ndard name grid longitude srlon long name staggered rotated longitude srlon units degrees float srlat srlat srlat standard name grid latitude srlat long name staggered rotated latitude srlat units degrees float lon rlat rlon lon standard name longitude lon long name longitude lon units degrees east float lat rlat rlon lat standard name latitude lat long name latitude lat units degrees north float slonu rlat srlon Part VII User s Guide 4 28 Section 5 Data Formats for I O 5 2 The NetCDF Data Format 44 slonu standard name longitude slonu long name staggered U wind longitude slonu units degrees east float slatu rlat srlon Slatu standard name latitude slatu long name staggered U wind latitude slatu units degrees north float slonv srlat rlon slonv standard name longitude slonv long name staggered V wind longitude slonv units degrees east float slatv srlat rlon Slatv standard name latitude slatv long name staggered V wind latitude slatv units degrees north float vcoord leveli vcoord long name terrain following coordinate vcoord units Pa vcoord p0sl 100000 vcoord t0sl 288 15 vcoord dtOlp 42 vcoord vcflat 0 22 float height_2m height_2m height_2m standard_name height height_2m long_name height above the surface heig
257. ng centres in order to obtain ZTD In such cases all except for one report have to be set to passive due to redundancy according to the preference given by NAMELIST parameter igpscen In the example the reports processed by BKG for observation time 0 5 and by SGN for time 0 75 are active 8 2 3 YUREJCT Nudging Rejected Reports YUREJCT see Figure 8 8 lists the station identities of the rejected or passive reports and indicates the reasons for their rejection For instance the station height and its difference to the model orography is provided if this difference is so large that every observed quantity from that report is excluded from use Most of the messages are self explanatory The FLIGHT TRACK messages all deliver station identity observation time relative to the model initial time and pressure of observation level They are complemented by position confidences in with original resp reversed sign for longitude if the line includes LON SIGN or position confidences for the foreward resp backward trajectory of the report sequence The BLACKLISTED messages indicate in general rejection of certain parts rather than com plete reports Observation types OBTYP are as specified in Figure 8 12 and the 4 pairs of numbers at the end of the lines indicate the lower and upper limit in hPa of the blacklisted vertical range for height geopotential or pressure wind temperature resp humidity
258. ng part y 0 8 of the non divergence correction factor Yn y is given as a function of pressure p hPa p 1000 850 700 500 400 300 250 200 150 50 0 4 05 0 5 0 5 0 5 0 6 0 65 0 7 0 75 0 75 tnondiv REAL temporal factor f scaling the correction factor Yn 1 1 at the beginning and the end of the nudging period for an individual observation relative to y valid at the observation time analogous to rhtfac above Observation increments from multi level reports Name Type Definition Purpose Comments Default computation of observation increments at model levels 1scadj 4 LOG vertical scale adjustment by vertical averaging the TRUE observed profile over each model layer instead of TRUE vertical interpolation as a method to convey the TRUE observational information to the model levels FALSE use of observation increments topobs 4 REAL at pressure p topobs in hPa only observation 849 increments at model levels are used 1099 i e increments at observation levels are not used T99 topobs is fixed at 1099 if msprpar 0 699 botmod 4 REAL at pressure p botmod in hPa only observation 1099 increments at observation levels are used 1099 i e increments at model levels are not computed 1099 botmod is fixed at 1099 if msprpar 0 899 botmod gt topobs must be satisfied Part VII User s Guide 4 28 Section 7 Namelist
259. nudging scheme are read from an AOF file see variable itype obfile in NAMELIST NUDGING and if NAMELIST variable lpraof is set to TRUE YUAOFEX then prints the complete observations that are read from the AOF file In this case the AOF file contains all the observations except for the GPS data that made available to the COSMO Model either for the purpose of nudging verification or production of two dimensional surface analyses Figure 8 5 shows an example of a file YUAOFEX which includes an aircraft report a SYNOP report and the beginning of a TEMP report For each report the first four lines contain the 19 items of the report header length of record for whole report length of preliminary array missing length of next record record number analysis box number observation type code type see Figure 8 12 station latitude in degree 100 9000 station longitude in degree 100 18000 c0 o0co0t SS d q a Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 157 10 11 12 13 14 15 16 17 18 19 date of observation synoptic exact time of observation station identity for characters 1 4 resp 5 8 ASCII allocating sequence data base key station altitude in yyyymmdd in m 1000 station characteristics instrument specifications yyyy year mm month dd day hh hour minutes in hh
260. o meet high resolution regional forecast requirements of weather services and to provide a flexible tool for various scientific applications on a broad range of spatial scales When starting with the development of the COSMO Model many NWP models operated on hydrostatic scales of motion with grid spacings down to about 10 km and thus lacked the spatial resolution required to explicitly capture small scale severe weather events The COSMO Model has been designed for meso f and meso y scales where nonhydrostatic effects begin to play an essential role in the evolution of atmospheric flows By employing 1 to 3 km grid spacing for operational forecasts over a large domain it is expected that deep moist convection and the associated feedback mechanisms to the larger scales of motion can be explicitly resolved Meso y scale NWP models thus have the princi ple potential to overcome the shortcomings resulting from the application of parameterized convection in current coarse grid hydrostatic models In addition the impact of topography on the organization of penetrative convection by e g channeling effects is represented much more realistically in high resolution nonhydrostatic forecast models In the beginning the operational application of the model within COSMO were mainly on the meso scale using a grid spacing of 7 km The key issue was an accurate numerical prediction of near surface weather conditions focusing on clouds fog frontal precipitatio
261. of the horizontal advection scheme Order of the explicit vertical advection scheme Option for using potential temperature as advected variable in the Runge Kutta scheme 0 Use Ty perturbation temperature for advection 1 Use 07 perturbation potential temperature 2 Use 0 full potential temperature Switch to use a limiter for temperature advection only for itheta adv 2 Maximal integer courant number in courant number inde pendent advection Switch to run with implicit vertical advection in the Runge Kutta scheme TRUE or not Parameter to select the treatment of fast waves 1 Old scheme from module fast waves rk f90 2 New scheme from module fast waves sc f90 with proper weightings for all vertical discretiza tions divergence operator in strong conservation form optionally isotropic fully 3D divergence damping optionally Mahrer 1984 discretization This is the new default from Version 4 27 on 1 FALSE TRUE Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 3 DYNCTL Parameters for the Adiabatic Model 88 Name Type Definition Purpose Comments Default itype bbc w ldiabf lh lsl adv qx yef adv qx INT LOG LOG CHAR Option for choosing bottom boundary condition for vertical wind before Version 4 12 this switch was named itype lbc w For itype fast waves
262. ogical switch if TRUE additional smoothing is done for FALSE mean sea level pressure and geopotential in mountainous areas lfi filter LOG Logical switch to filter the geopotential FI independent of FALSE the settings of 1 p filter Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 14 GRIBOUT Controlling the Grib Output 146 Name Type Definition Purpose Comments Default l pmsl filter LOG Logical switch to filter the mean sea level pressure PMSL TRUE independent of the settings of lp filter and or lz filter ysuffix CHAR To add an optional suffix to the file names 3 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 147 Section 8 Model Output The COSMO Model provides three kinds of output e The model fields resulting from the model integration can be written in GRIB or in NetCDF output more information on this can be found in Section 8 4 For a quick monitoring and diagnostic output several ASCII files are written These are described in Sections 8 1 and 8 2 e For data assimilation or verification purposes a special NetCDF feedobs file some times also mis called feedback file can be written Section 8 3 provides some infor mation on it A comprehensive description of the format of feedback files which are ex tended feedobs files is given in an extra doc
263. onl data category WMO data category C 13 int sectionl int data sub category internatl WMO data sub category int sectionl date year month day YY YYMMDD int sectionl time hour minute second HHMMSS The identifiers C denote BUFR CREX Common Code Tables which are detailed in http www wmo int pages prog www WMOCodes TDCFtables html TDCFtables link Common Code Tables to Binary and Alphanumeric Codes Observational reports which have missing values for sectionl data category or for section1 int data sub category will be rejected This also applies to GPS reports if sectionl centre or sectionl_subcentre have missing values From BUFR Section 2 the following elements are optional This means that they are read and stored internally for all observation types but it is not mandatory that they exist or contain non missing values i e COSMO will not abort if they do not exist descriptor type variable name in NetCDF file meaning int section2 ikz DWD internal data base ID int section2 decoding date year month day YY YYMMDD int section2_decoding_time hour minute second HHMMSS From BUFR Section 3 the variables read and used mandatorily or optionally are de tailed in the following sub sections for the different observation file types The variable names in the NetCDF files equal the BUFR mnemonics for these variables Part VII User s Guide 4 28 Section 6 Input Files
264. ontrol The same applies to the entries p TEMP and ps for pressure except that p TEMP refers to a single pressure datum at the lowest model level as derived from radiosonde geopotential data The entry IWV relates to an integrated water vapour value in mm which is derived from a ground based GPS report and possibly bias corrected and which is either rejected by the threshold quality control or is smaller than an absolute minimum threshold value 2mm For each entry the following properties are given station identity code type see Figure 8 12 observation time relative to the model initial time in h pressure in hPa at the observation level latitude and longitude in threshold value for the difference observed value model value and in some cases the difference between observed and model value For wind both the zonal and meridional value components are given complemented by the absolute value of the difference vector in m s For relative humidity in the temperature values in K are also provided if the humidity observation is rejected only due to the rejection of the corresponding temperature observation The same applies to aircraft temperature if it is rejected due to the rejection of the wind observations and vice versa The entry ps scc usually denotes a surface pressure report which after passing the individual threshold quality control is rejected by the spatial consist
265. ophic wind field can be derived again from which the maximum relative change in 0 1 96 in the iteration and the maximum mag nitude of the wind increments are also given for the subdomain with point ion12 jon12 Furthermore zdfi_max and max psaigeo 1 provide the largest positive and negative geopotential resp pressure increment on that subdomain The arrays of lines nudging list the analysis increments as a result of different processes as indicated at the given six pres Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations ntstep 0 k 35 T qv qc 276 84 0 004943 0 000000 18 Tgnudge T q ana incr 18 126 35 0 0008 0 000000 0 000 omu weight sqr weighted incr 35 0 0000 0 0000 0 0000 omu weight weighted incr 35 0 0001153 1 3699 geostroph_ps_corr uv ana incr 18 126 0 0000 0 0000 itera qll umax a 120 014440 0 164226 itera qll umax 2 64 963402 0 164223 itera qll umax 50 2 570915 0 175688 zdfi_max 8 0 2562859401204606 2 629577730880562E 002 max psaigeo 1 8 0 2927315768875584 3 061252112578843E max psaigeo 2 8 0 2634584191988025 2 755126901320959E nudge_horiz_wind uv ana incr 18 126 35 0 0138 0 0056 0 34 651 3 max unreduced geost incr on total domain 0 34 185 106 max unreduced geost incr on inner domain 0 34 224 134 0 83 reduced geost incr at i jonl 0 00 nudging pressure 3767 23293 55121 850 nudging p incr no T
266. or surface level and maxsgo maxmlo for surface pressure increment arrays Note that e g from the GPS ZTD observations profiles of humidity increments are derived Therefore the size of the multi level increment arrays also depends on maxgpo The next message relates to an array that is used purely for the spatial consistency quality check for integrated water vapour IWV derived from GPS ZTD and from radiosonde hu midity profile data T he final message in Figure 8 10 indicates that there is insufficient space in the NetCDF feedobs file to write all the observations onto that file Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 166 8 2 6 YUSTATS Nudging Statistics on Observation Processing YUSTATS provides statistics on the processed observations and on the analysis increments accumulated over time It consists of several parts In the first part Figure 8 11 the domain averged analysis increments integrated over time since the beginning of the model integration are provided once every hour the last hour is written almost at the end of the file For the 3 dimensional variables wind speed v wind direction temperature T pressure p geopotential FI specific water vapour content qv and specific cloud water content qc this information is given for each vertical model level separately and this results in vertical profile
267. orological Monographs No 32 American Meteorol Soc 84 pp Klemp J B and R B Wilhelmson 1978 The simulation of three dimensional convective storm dynamics J Atmos Sci 35 1070 1096 Liu X D S Osher and T Chan 1994 Weighted essentially non oscillatory schemes J Comput Phys 115 200 212 Louis J F 1979 A parametric model of vertical eddy fluxes in the atmosphere Bound Layer Meteor 17 187 202 Lynch P D Girard and V Ivanovici 1997 Improving the efficiency of a digital filtering scheme Mon Wea Rev 125 1976 1982 Mellor G and T Yamada 1974 A hierarchy of turbulence closure models for plan etary boundary layers J Atm Sc 31 1791 1806 Mironov D 2008 Parameterization of lakes in numerical weather prediction De scription of a lake model COSMO Technical Report No 11 Deutscher Wetterdienst Deutscher Wetterdienst D 63004 Offenbach Germany Mironov D and R B 2004 Testing the new ice model for the global NWP system GME of the German Weather Service Technical documentation 1220 WMO WMO Geneve Switzerland Miller E 1981 Turbulent flux parameterization in a regional scale model In ECMWF Workshop on planetary boundary layer parameterization pp 193 220 25 27 November 1981 Part VII User s Guide 4 28 REFERENCES REFERENCES 194 Ritter B and J F Geleyn 1992 A comprehensive radiation scheme for numerical weather prediction models with potenti
268. ot fully supported any more in the sense that it does not include new features introduced in the data assimilation part after COSMO V4 17 In particular writing NetCDF feedobs files is not working completely The choice between the two formats of observation input is made by setting the value for the namelist parameter itype obfile of COSMO A value of 1 indicates AOF file input and a value of 2 NetCDF files input In the following items general properties of the NetCDF file observation input are described e Required files and file names The NetCDF observation input files have fixed file names which are given in the fol lowing sub sections and begin with cdfin_ For the observation file type SYNOP for instance the file name would be cdfin_synop For each observation file type it is pos sible to have several input files with suffix 2 3 etc for the file names the suffix 1 is not used An additional suffix nc is optional Thus for a second file for SYNOP data the file name cdfin_synop 2 nc would be possible The existence of any of the NetCDF observation input files is optional If there are no observations available of a certain type then the corresponding NetCDF file should be missing in the input directory Any file with a corresponding file name must have the correct format as described further below and files with zero length are not allowed In addition a blacklist file with
269. oud cover and water content are calculated according to a relative humidity criterion itype wcld 1 or a statistical closure itype_wcld 2 itype synd INT Type of diagnosis of synoptic station values 1 1 Interpolation of screen level 2 m 10 m values using traditional similarity theory 2 Interpolation of screen level 2 m 10 m values based on profile relations used in the new surface layer scheme lprfcor LOG Using the profile values of the lowest main level instead of the FALSE mean value of the lowest layer for surface flux calculations Not tested should be set to FALSE Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 4 PHYCTL Parameters for the Diabatic Model 102 Soil Processes These parameters control the parameterization of soil and vegetation processes Mainly the configurations of the soil and vegetation model TERRA of the lake model FLake and of the snow model are specified The optimal configuration depends on the region investigated Note Additional external parameter fields are required by some of the methods used as additional input fields e g for FLake They can be provided using an appropriate version of INT2LM Name Type Definition Purpose Comments Default lsoil lseaice llake lmulti layer lmulti snow nlgw lmelt lmelt var ke soil lstomata ke snow czml soil LOG LOG LOG LO
270. output lgplong TRUE or the special grid point output for the physics lgspec TRUE The parameter nmaxgp 100 per default in the module src gridpoints f90 gives the maximum number of gridpoints for which calculations can be done For every grid point the output is stored for all time steps to a file M stationname where stationname is a name that can be specified by namelist input If no name is specified the geographical coordinates are used instead Name Type Definition Purpose Comments Default nOgp INT Time step of the first grid point calculation Alternatively 0 hOgp REAL Same as nOgp but time in hours 0 0 nincgp INT Time interval in time steps between two calls for grid point undefined output Alternatively hincgp REAL Same as nincgp but time in hours 0 0 station TYPE The list of stations for grid point output can be specified with undefined list tot a derived type declaration The components of the type are igp INT i index jgp INT j index rlatgp tot REAL geographical latitude rlongp tot REAL geographical longitude ystation name CHAR name of the station lgpshort LOG Calculate and print a short form of grid point output FALSE 1 line step lgplon LOG Calculate and print a long form of grid point output FALSE Bp ong 8 8 about 1 page step lgspec LOG Calculate and print a special form of grid point output for FALSE physics diagnostic
271. plants Introduced in Version 4 11 Number of layers in the multi layer snow model Introduced in Version 4 11 Array to specify the main levels of the ke soil 1 soil layers in meters The default specification is 0 005 0 02 0 06 0 18 0 54 1 62 4 86 14 58 The maximum number of soil layers is limited to 20 FALSE FALSE FALSE FALSE FALSE TRUE TRUE TRUE see left Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 4 PHYCTL Parameters for the Diabatic Model 103 Name Type Definition Purpose Comments Default itype trvg INT Parameter to select the type of parameterization for transpi 2 ration by vegetation 1 Bucket version 2 BATS version itype evsl INT Parameter to select the type of parameterization for evapo 2 ration of bare soil 1 Bucket version 2 BATS version itype root INT Parameter to select the type of root distribution 1 Introduced in Version 4 11 1 Uniform Default 2 Exponential following Arora amp Boer 2003 itype heatcond INT Parameter to select the type of soil heat conductivity 1 Introduced in Version 4 11 1 Use average soil moisture 2 Take into account soil moisture soil ice itype hydbound INT Parameter to select the type of hydraulic lower boundary 1 Introduced in Version 4 11 1 Allow for drainage but not diffusion 2 Rigid lid not yet implemented 3 Ground water wit
272. provide information on which of the NAMELIST variables need to be increased in order to allow the use of all the good data Observation dependent array sizes Name Type Definition Purpose Comments Default size of internal arrays to store observations or increments maximum number of reports or stations which are actively used at the same timestep in the total model domain maxmlo INT maximum number of multi level reports 300 maxsgo INT maximum number of surface level reports 3000 maxuso INT maximum number of upper air single level reports 900 maxgpo INT maximum number of ground based GPS reports 3000 Note maxmlo maxsgo maximum number of surface pressure reports maxmlo maxsgo and maxgpo also scale the size of the arrays which contain all the multi level single level upper air and surface level resp GPS reports on the local sub domain on distributed memory computers which have to be stored at a certain timestep to be used then or later on maxmlv INT maximum number of observation levels in multi level reports 100 mxfrep INT maximum number of reports in NetCDF feedobs file 1 mxfobs INT maximum number of observations in NetCDF feedobs file 1 Note if mxfrep nxfobs lt 0 then reasonable values are computed dynamically from maxmlo hverend etc Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell
273. pts files for different computers Makefile For compiling and linking the programs runxxx2yy Scripts to set the Namelist values for a configuration from model xxx to application yy and start the program src Subdirectory for the source code obj Subdirectory where the object files are written ObjDependencies Definition of the dependencies between the different source files Objfiles Definition of the object files work Subdirectory for intermediate files Here also the source code for mpe io f90 and the dummies for the external libraries are included in src dummy_db f 90 The directories obj and work are empty and can there fore get lost by the tar process If so you have to create them again In edid you have to adapt the pathnames if you want to work with it Before compiling and linking the program you should check and if necessary adapt the KIND type parameters listed below which are used for selecting the precision of REAL variables in the program and the precision of INTEGER variables of the grib library Part VII User s Guide 4 28 Section 4 Installation of the COSMO Model 4 4 Compiling and Linking 33 4 4 Compiling and Linking Before compiling check and adapt the necessary parameters see above All other input variables for the program can be determined before running the program with the NAMELIST input see Chapter 7 You have to choose the options for compiling the code in the file Options if wo
274. put 107 7 7 SATCTL Controlling the Synthetic Satellite Images 110 7 8 INICTL Parameters for the Model Initialization 113 7 9 NUDGING Controlling the Data Assimilation 114 7 10 EPSCTL Controlling the Ensemble Prediction Mode 136 CELL fOCTL Controlling the Gab T O sso ive hae dad 3 ky xe 137 7 12 DATABASE Specification of Database Job 140 Gla GRIBIN Controlling the Grib Input 4 2 24 cev a a 141 7 14 GRIBOUT Controlling the Grib Output 144 8 Model Output 147 8 1 ASCII Output for the Forecast Model rss 147 8 1 1 M stationname Grid point output ara ek ea ewe 148 8 1 2 YUSPECIF NAMELIST parameters gt 2229 o 9 o 9 Ros 151 8 1 3 YUCHKDAT Checking the Grib input output data 151 8 1 4 YUPRMASS Protocolling the forecast with mass variables 152 8 1 5 YUPRHUMI Protocolling the forecast with humidity variables 152 8 2 ASCII Output Related to the Use of Observations 4 156 8 2 1 YUAOFEX Nudging Observation Input AOF 156 8 2 2 YUOBSDR Nudging Active and Passive Reports 158 Part VII User s Guide 4 28 Contents Contents iv 8 2 3 YUREJCT Nudging Rejected Reports 160 8 2 4 YUQUCTL Nudging Data Rejected by Quality Control 162 8
275. quest and the agreement please contact the chairman of the COSMO Steering Committee In the case of a planned operational or commercial use of the COSMO Model package special regulations Part VII User s Guide 4 28 Section 1 Overview on the Model System 1 1 General Remarks 2 Table 1 1 COSMO Participating Meteorological Services DWD Deutscher Wetterdienst Offenbach Germany MeteoSwiss Meteo Schweiz Zurich Switzerland USAM Ufficio Generale Spazio Aero e Meteorologia Rome Italy HNMS Hellenic National Meteorological Service Athens Greece IMGW Institute of Meteorology and Water Management Warsaw Poland NMA National Meteorological Administration Bucharest Romania RosHydroMet Hydrometeorological Centre of Russia Moscow Russia ARPA SIMC Agenzia Regionale per la Protezione Ambientale dell Emilia Romagna Servizio Idro Meteo Clima Bologna Italy ARPA Piemonte Agenzia Regionale per la Protezione Ambientale Piemonte Turin Italy CIRA Centro Italiano Ricerche Aerospaziali Capua Italy AGeoBW Amt f r Geoinformationswesen der Bundeswehr Euskirchen Germany will apply The further development of the modelling system within COSMO is organized in Working Groups which cover the main research and development activities data assimilation nu merical aspects upper air physical aspects soil and surface physics aspects interpretation and applications verification and case studies refe
276. r air surface surface level single level level pressure 1 hour 1224 2608 3589 3960 2 hour T331 2617 3995 4372 3 hour 1293 2550 3995 4439 4 hour 1374 2531 3428 4115 total max 1374 2617 3995 4439 array bounds 1351 2500 4000 4350 Figure 8 15 Example file YUSTATS fifth part The fifth part Figure 8 15 finally delivers the hourly maximum total number of stations in case of temporal linear interpolation or reports otherwise with active observation increments There are four types of sets of increments in the scheme and hence four types of families of arrays in the code multi level upper air single level surface level and surface pressure increments The length of the corresponding arrays listed in the table as array bounds are a function of NAMELIST parameters maxmlo maxgpo 2 1 for multi level maxuso for upper air single level maxsgo for surface level and maxsgo maxmlo for surface pressure increment reports If the array lengths which depend on the selected values of these NAMELIST parameters are not sufficiently large there are various places in the program where arrays may fail to accommodate all the available data The program will not crash or stop in such a case but it will simply omit the surplus data and issue warning messages which always contain the label CAUTION For more information on this concept see Section 8 2 5 As already mentioned there sum
277. r of grid point rows and colums where lateral bound ary data are defined when using the 1bd frame option Number of vertical layer which separates frame and no frame boundary data For layers extending from the top of the model domain to ilevbotnoframe the boundary data are defined for the full horizontal domain this enables the application of a Rayleigh damping layer near the model top and for layers from ilevbotnoframe to the lowest layer the boundary data are defined on a frame FALSE FALSE FALSE FALSE FALSE FALSE 1 0 3 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 13 GRIBIN Controlling the Grib Input 143 Additional parameters for continuous data assimilation Name Type Definition Purpose Comments Default newbc INT Number of times that boundary update is analysis after 1 360 hour newbcdt INT Time step increment of boundary update from analysis 5o hnewbcdt REAL Hour increment of boundary update from analysis 0 0 nincboufac INT Factor to nincbound when new boundary update is analysis 2 lan t s LOG Selection of analysed tri 0 sea surface temperature FALSE lan_t_so0 LOG Selection of analysed tri 0 sea surface temperature for the FALSE multi layer soil model lan_t_snow LOG Selection of analysed snow temperature FALSE lan_t_cl LOG Selection of analysed climatological soil temperature FALSE lan_w_snow
278. r zone of the domain hd dhmax REAL Threshold value for the maximum height difference 250 0 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 3 DYNCTL Parameters for the Adiabatic Model 92 Lower and upper boundary condition Name Type Definition Purpose Comments Default lspubc LOG Option to use a sponge layer with Rayleigh damping in the TRUE upper levels of the model domain If lspubc FALSE a rigid lid upper boundary condition is used itype spubc INT To choose the type of Rayleigh damping in the upper levels 1 1 chooses the damping against boundary fields 2 chooses the damping against filtered forecast fields can be chosen if the boundary data is only available on frames 3 New sponge layer near upper model boundary accord ing to Klemp et al nfi_spubc2 INT Number of applications of smoother for the determination 10 of the large scale field used in the Rayleigh damping with itype_spubc 2 lrubc LOG Option to use a radiative upper boundary condition FALSE Not yet implemented rdheight REAL The bottom height m from which the Rayleigh sponge layer 11000 0 extends to the top of the model domain A cosine damping profile with maximum damping at the top and zero damping at rdheight is assumed nrdtau INT nrdtau dt is the e folding time for damping at the model top 5 Larger values of nrdtau result in weaker damping ldyn bbc LOG To choose
279. raphy Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 4 PHYCTL Parameters for the Diabatic Model 95 Grid Scale Precipitation These parameters control the parameterization of grid scale precipitation Note that the sub grid scale precipitation is controlled by the convection parameterization Name Type Definition Purpose Comments Default lgsp LOG Main switch for including grid scale precipitation If TRUE TRUE the model is run with a grid scale precipitation scheme which computes the effects of precipitation formation on temperature and the prognostic moisture variables in the atmosphere wa ter vapour cloud water optionally cloud ice rain snow and graupel as well as the precipitation fluxes of grid scale rain and snow at the ground itype gscp INT Control parameter to select a specific parameterization scheme 3 1 Kessler type warm rain parameterization scheme with out ice phase processes 2 Kessler type bulk formulation using cloud water rain and snow 3 Extension of the basic scheme with cloud water and cloud ice 4 Graupel scheme with prognostic cloud water cloud ice and graupel lprogprec LOG This switch has been eliminated in Version 4 23 All simulations are now done with prognostic precipitation ltrans prec LOG This switch has been eliminated in Version 4 23 All simulations are now done with prognostic precipitation ldiniprec L
280. rature of surface T SNOW m temperature of snow surface T G m temperature at the boundary soil atmosphere T M m temperature between upper and medium soil layer T S0 m temperature of multi layer soil levels Qv_s m specific water vapor content at the surface WG_1 m water content of the upper soil layer WG_2 m water content of the medium soil layer WG_3 water content of the lower soil layer W_SO m water content of multi layer soil levels TKVM turbulent diffusion coefficients for momentum in the atmosphere TKVH turbulent diffusion coefficients for heat and moisture in the atmosphere WI m water content of interception water W SNOW m water content of snow T CL m temperature between medium and lower soil layer W_CL m climatological water content of the lowest soil layer TCM m turbulent diffusion coefficients for momentum at the surface TCH m turbulent diffusion coefficients for heat and moisture at the surface CLCT m total cloud cover CLCH m cloud cover with high clouds CLCM m cloud cover with medium clouds CLCL m cloud cover with low clouds DPSDT tendency of the surface pressure BAS CON m index of lower boundary of convective clouds TOP CON m index of upper boundary of convective clouds HBAS CON m height of lower boundary of convective clouds HTOP CON m height of upper boundary of convective clouds ASOB S m average solar radiation budget surface mean value over forecast ATHB S m average thermal radiation
281. re 8 22 This is done at each timestep that new analysis increments are computed and it allows to monitor which variables at which levels are explicitly influenced by the nudging Analysis increments at timestep 12 max ABS of analysis increments of mass field prior to condensation evaporation except for pressure increments l T corr global T incr global qv incr due gl qv incr global RH incr global p incr global vel K coord K coord to T incr coord g kg coord coord Pa coord 1 0 000 0 0 0 04 634 91 0 0000 0 0 0 0000 0 0 0 00 0 0 0 689 635 92 2 0 000 0 0 0 08 640 87 0 0000 0 0 0 0000 0 0 0 00 0 0 0 740 273 189 9 0 000 0 0 0 04 326 582 0 0000 0 0 0 0000 0 0 0 00 0 0 1 600 268 196 10 0 000 0 0 0 05 325 185 0 0000 0 0 0 0001 113 320 0 11 113 320 1 743 533 518 38 0 067 195 648 0 07 114 40 0 0000 0 0 0 0326 300 225 0 62 300 226 13 929 195 648 39 0 067 195 648 0 07 114 40 0 0000 0 0 0 0329 300 225 0 62 301 225 14 097 195 648 40 0 067 195 648 0 07 114 40 0 0000 0 0 0 1434 520 84 1 47 649 645 14 215 195 648 max ABS of increments of horizontal wind field components in m s ana incr analysis increments net incr net observation increments geo corr geostrophic wind correction le u ana global v ana global u net global v net global u geo global v geo global vel incr coord incr coord incr coord incr coord corr coord corr coord 1 0 06 258 343 0 05 599 131 13 3 540 82 7 1 483 345 0 00 0 0 0 00 0 0 2 0 09 280 2
282. rence version and implementation and predictability and ensemble methods In 2005 the COSMO Steering Committee decided to define Priority Projects with the goal to focus the scientific activities of the COSMO com munity on some few key issues and support the permanent improvement of the model For contacting the Working Group Coordinators or members of the Working Groups or Priority Projects please refer to the COSMO web site The COSMO meteorological services are not equipped to provide extensive support to ex ternal users of the model If technical problems occur with the installation of the model system or with basic questions how to run the model questions could be directed via email to cosmo support cosmo model org If further problems occur please contact the members of an appropriate Working Group We try to assist you as well as possible The authors of this document recognize that typographical and other errors as well as dis crepancies in the code and deficiencies regarding the completeness may be present and your assistance in correcting them is appreciated All comments and suggestions for improvement or corrections of the documentation and the model code are welcome and may be directed Part VII User s Guide 4 28 Section 1 Overview on the Model System 1 2 Basic Model Design and Features 3 to the authors 1 2 Basic Model Design and Features The nonhydrostatic fully compressible COSMO Model has been developed t
283. ressure and humidity variables defined at the centre of a grid box and the normal velocity components defined on the corresponding box faces see Figure 3 1 For a given grid spacing this staggering allows for a more accurate representation of differential operators than in the A grid where all variables are defined at the same point In general we use second order centered finite difference operators i e the numerical discretization error is reduced by a factor of four when we increase the resolution by a factor of two For a detailed description of the numerical operators see Part I of the Documentation Dynamics and Numerics The grid box faces in vertical direction are usually referred to as the half levels These interfacial levels separate the model layers from each other The model layers labeled by integers k are also denoted as main levels Thus for a model configuration with N layers we have N 1 half levels The top boundary of the model domain is defined to be the half level C 1 2 above the uppermost model layer 1 At the lower boundary the coordinate surface becomes conformal to the terrain height The half level N 1 2 below the first model layer above the ground N defines the lower boundary of the model The discrete formulation of the model equations is independent on a specific choice for the vertical coordinate This is achieved by a two step transformation procedure First we apply a Part
284. rking within the VCS or in Fopts otherwise See the User Guide of your computer system for necessary and or desired options Before linking check that the Grib library necessary for the I O the external object files mpe_io o and dummy_db o and the necessary external libraries see 4 1 are available The COSMO Model is parallelized for distributed memory parallel computers using the domain decomposition technique and explicit message passing with the Message Passing Interface MPI Thus it can run on parallel platforms but also on sequential platforms where MPI is not available For this purpose an additional module dummy mpi f90 is provided which has to be linked with the model then sequential On single processor systems you can create a binary for se quential execution without using MPI To avoid warning messages by the linker a file dummy mpi f90 is provided to satisfy the MPI external references parallel On parallel computers with distributed memory you can cre ate a binary for parallel execution if MPI is available You can also create a sequential binary which can only run on one processor In the VCS environment the creation of one or more certain binaries is fixed Ask your administrator if you want to change the default Outside the VCS you can choose the binary by modifying Makefile You can invoke a make run by typing make entry On batch machines you can start a batch job for a make run with mk batch entry Within V
285. rmat as follow Format of Blacklist 211111111 T P G U P G O P W U P W O P T U P T O P D U P D O 01295 1 1100 0 0 0 0 0 0 0 0210 1 1100 0 0 0 0 0 0 0 924Y48JP 2 0 0 0 o 1100 0 0 0 ABX 2 0 o 1100 0 0 0 0 0 ABX 2 0 o 1100 0 0 0 0 0 ABX 2 0 o 1100 0 0 0 0 0 ABX 2 0 o 1100 0 0 0 0 0 17912 4 1100 0 0 0 0 0 0 0 21523 4 0 o 1100 0 0 0 0 0 07137 5 1100 O 1100 o 1100 0 0 0 10437 5 300 O 1100 0 0 0 0 0 10828 5 1100 o 1100 0 0 0 0 0 10204 6 0 O 500 0 0 0 0 0 10266 6 0 o 200 o 1100 0 0 0 10384 6 0 O 1100 850 0 0 0 0 The first line is fixed and the following lines are the entries in the blacklist with following 10 columns 1 station identity as wildcard allowed 2 observation type 1 surface level 2 aircraft 4 buoy 5 TEMP 6 PILOT 3 4 lower upper limit pressure in hPa of blacklisted vertical range for geopotential 5 6 blacklisted vertical range for horizontal wind 7 8 blacklisted vertical range for temperature 9 10 blacklisted vertical range for humidity Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files 74 Immediately after the last line of the blacklist the whitelist follows Standard format of Whitelist Alternative format of Whitelist WHITELIST WHITELIST 6 132 03019 6 132 03019 6 132 10135 6 132 10135 6 132 10266 6 132 10266 6 132 10394 6 132 10394 6 132 47912 6 134 WHITELIST 6 133 47945 6 134 WHITELIST 6
286. rticallocation Pa need 0 10 009 int n NHHHN Geopotential height gpm opt 0 05 015 float n MLADH Latitude displacement since launch site opt 0 06 015 float n MLODH Longitude displacement since launch site need 0 12 101 float n MTDBT Temperature dry bulb temperature need 0 12 103 float n MTDNH Dew point temperature need 0 11 001 int n NDNDN Wind direction degree true need 0 11 002 float n NFNFN Wind speed m s 0 31 001 int MDREP Delayed descriptor replication factor 3 03 051 Wind shear data at a pressure level 0 04 086 int n NLTPDO Time displacement since launch time s 0 08 042 int n MEVSSO Extended vertical sounding significance 0 07 004 float n MPNO Pressure vertical location Pa 0 05 015 float n MLADHO Latitude displacement since launch site 0 06 015 float n MLODHO Longitude displacement since launch site 0 11 061 float n NVBVB Absolute wind shear in 1 km layer below 0 11 062 float n NVAVA a wind shear in 1 km layer above m s Table notes 1 Only either the pair of variables MIT and NIIT or the single variable YDDDD is strictly needed to exist If both exist and have non missing values for a certain report then the values of the pair MIT and NIIT are used 2 Only one of the variables MHOBNN and MHOSNN is strictly needed to exist For radiosondes MHOBNN is found to be the pressure at the site where the ground check and calibration is done whereas MHOSN
287. s Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 6 DIACTL Parameters for Diagnostic Output 109 Computation of special diagnostics Since Version 4 8 it is possible to choose different variants for special diagnostic computations Name Type Definition Purpose Comments Default itype_diag t2m itype_diag_gusts INT To specify the method for computing the 2m temperature INT To specify the method for computing the maximal wind 1 Computation with an exponential canopy profile but with a diagnostic Prandtl layer interpolation even for scalars using an adopted canopy layer resistance 2 Computation with exponential canopy profile gusts 1 Dynamical gust derived from lowest model layer 19 Dynamical gust derived from 30 m ee Dynamical gust derived after Brasseur E Similar to 1 but here the gust factor depends weakly on the mean wind speed at 10 meters 1 Computation of surface and volume integrals Since Version 3 23 the COSMO Model offers the possibility to calculate volume integrals of arbitrary fields over an arbitrary cuboid defined in the numerical i e terrain following grid Also surface integrals of arbitrary vector fields fluxes over the surface of this cuboid can be computed With the following set of namelist variables the cuboid can be defined in terms of grid point indices of the total domain
288. s are required to run the model in NWP mode climate mode or for case studies The purpose of the Description of the Nonhydrostatic Regional COSMO Model is to provide a comprehensive documentation of all components of the system and to inform the user about code access and how to install compile configure and run the model The basic version of the COSMO Model formerly known as Lokal Modell LM has been developed at the Deutscher Wetterdienst DWD The COSMO Model and the triangular mesh global gridpoint model GME form together with the corresponding data assimila tion schemes the NWP system at DWD which is run operationally since end of 1999 The subsequent developments related to the model have been organized within COSMO the Consortium for Small Scale Modelling COSMO aims at the improvement maintenance and operational application of a non hydrostatic limited area modelling system which is now consequently called the COSMO Model The meteorological services participating to COSMO at present are listed in Table 1 1 For more information about COSMO we refer to the web site at www cosmo model org The COSMO Model is available free of charge for scientific and educational purposes es pecially for cooperational projects with COSMO members However all users are required to sign an agreement with a COSMO national meteorological service and to respect cer tain conditions and restrictions on code usage For questions concerning the re
289. s can be done by the itype turb and itype tran parameters Name Type Definition Purpose Comments Default ltur ninctura itype_turb imode_turb LOG INT INT INT Main switch for including turbulent diffusion If TRUE the model is run with a turbulence parameterization which calcu lates the transport coefficients for momentum Km and heat Kn also applied for water substance in the atmosphere and the transfer coefficients at the ground surface layer Over wa ter also the roughness length zo is computed To save comput ing time the exchange coefficients in the atmosphere may not be computed every time step but at certain time intervals de fined by ninctura Time step increment for recalculating the transport coefficients Km and Kj for vertical diffusion itype turb 1 or for re calculating the stability functions Sm and Sp itype_turb 3 which are used to diagnose Km and K from the predicted TKE When running with the Leapfrog time integration ninctura should be an odd mumber to avoid using the same time family all the time Parameter to select a specific vertical turbulent diffusion pa rameterization 1 Default diagnostic scheme 2 Not used 3 Prognostic TKE based scheme includes effects from subgrid scale condensation evaporation 4 Not used 5 7 If a threedimensional turbulence scheme will be used Mode of turbulent diffusion parametrization in case of itype
290. s missing from averages or accumulations 22 25 20 Century of reference time of data given by octets 13 17 23 26 255 Sub centre identification national use 24 27 28 0 Units decimal scale factor D 25 36 29 40 0 Reserved need not to be present 37 Al 254 Octets 41 54 are reserved for the originating centre The integer value 254 indicates that additional data follow We use this part as follows 38 42 0 not used 39 43 45 0 not used 40 46 0 not used 41 47 0 Additional indicator for a GRIB element number 42 48 98 Year of production of GRIB record 43 49 98 Month of production of GRIB record 44 50 11 Day of production of GRIB record 45 51 2 Hour of production of GRIB record 46 52 0 Minute of production of GRIB record 47 53 54 1 Version number currently 1 Part VII User s Guide 4 28 Section 5 Data Formats for I O 5 1 The GRIB Binary Data Format 38 5 1 3 Product Definition Section The Product Definition Section PDS contains the necessary information to identify the binary coded field contained in the GRIB record The most important octet in this section is the indicator of the meteorological parameter T he indicator relates a specific meteorological element to an integer number This indicator number is also referred to as GRIB number or element number and is defined in a separate code table More than one indicator code tables may be used in GRIB code Thus one can have t
291. s of the assimilating model run which writes YUVERIF This part can be extended by analogous increments from several other model runs e g forecasts by use of a post processing program lmstats which can read and write extended YUVERIF files Here the extended format of YUVERIF is described which can contain the increments from one or several runs Similarly to the regular report body the number of entries increments per model run depends on the basic type of reports Moreover the number of runs for which the increments of an observation level are written to the same line also depends on that type Both the number of entries and the number of runs on the same line are given in the following list increment body length basic type of report entries runs per line complete synoptic surface level report 14 1 short surface level report 9 2 very short surface level report 5 3 upper air single level report 4 4 multi level report each obs level 5 3 ground based GPS report 1 12 For multi level reports this means that the five increments from the first run followed by twice five increments from the next two runs are written to one line for the first observation level Analogously the increments from these three runs are then written for the other observation levels line by line Finally an extra line is added for multi level reports only with the three surface pressure increments in Pa from the three runs before the process i
292. s of the domain averaged analysis incre ments Note that the time integrated values are obtained by updating them with the analysis increments from the nudging without the latent heat nudging at each model timestep i e it is the sum over the nudging increments from every timestep This can be very different from the difference analysis minus first guess if first guess is e g a 3 hour free forecast The reason is that the model dynamics and physics can react to the changes of the model state from the nudging by producing modfied dynamics and physics tendencies in the subsequent timestep s The second part of YUSTATS Figure 8 12 shows the number of processed active passive and rejected reports for each observation type and code type The meaning of the type numbers is also specified The third part Figure 8 13 first mentions the conditions for a report to be set passive rather than rejected In contrast to rejected reports passive reports are processed further for being written to the VOF file YUVERIF and or the NetCDF feedobs file for verification purposes if NAMELIST variable lverpas is set to TRUE The subsequent REPORT EVENTS table declares the reasons in a statistical sense rather than 1 Diagnostics on time integrated analysis increments 2009022409 2 h horizontal mean of 2 hourly sums of analysis increments TOV g m2 TOC mg m2 Dry Static Energy J kg 389 070 7539 854 0 003 volumn
293. s repeated for the next one two or three runs and so on The following list declares the complete set of 14 increment entries report body extension on increments 1 zonal wind component for GPS reports IWV 1 100 mm 1 100 m s 2 meridional wind component 1 100 m s 3 temperature 1 100 K 4 relative humidity 1 10 96 5 pressure surface level reports geopotential upper air Pa m2 s2 6 total cloud cover full model value instead of increment octas 7 low cloud cover full model value instead of increment octas 8 horizontal visibility full model value instead of increment 10 m 9 precipitation amount full model value instead of increment 1 10 mm 10 minimum temperature at 2m during past 12 hrs 1 10 K 11 maximum temperature at 2m during past 12 hrs 1 10 K 12 minimum ground temperature at 5cm during past 12 hrs 1 10 K 13 max wind speed of gusts time range as in regular body m s 14 global radiation sum over 1 hour 10 kJ m2 Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 178 8 2 8 YUPRINT Nudging Other Aspects YUPRINT provides a large variety of information Much of it has been used mainly at the stage of developing and testing new pieces of code Instead of inflating this documentation by explaining all types of statements in detail only the most important ones are described here including those which are most l
294. servation types In the last sub section the blacklist file is described 6 4 1 Templates for observation types for which Table Driven Code Forms TDCF defined by WMO exist The BUFR templates which the NetCDF files described in this sub section are based on are described at http www wmo int pages prog www WMOCodes TemplateExamples html e SYNOP SYNOP MOBIL and SHIP File names for fixed land stations SYNOP cdfin_synop for mobile land stations SYNOP MOBIL cdfin synop mob for sea stations SHIP edfin ship The templates for observation type SYNOP and for SYNOP MOBIL follow the WMO common sequence descriptors TM 307 080 resp TM 307090 These sequences and also the use of the variables in COSMO are identical to each other except for the report header common sequences 3 01 090 esp 3 01092 The template for SHIP follows the WMO common sequence descriptor TM 3 08 009 Caution The use of observation file type SYNOP MOBIL in COSMO has not been tested yet due to a lack of testing opportunity because of a lack of data of this type Its use is at the user s own risk However as the template is almost the same as for fixed land SYNOP errors are unlikely If errors occur they should be reported to the author For all the 3 observation file types the table below lists all the variables which are used by COSMO plus some of the other variables but it does not detail the variabl
295. sion gt s lt a s salsaa ppa eaa nariai 23 3 5 5 Parameterization of Surface Fluxes s oci oo cecs ateua da la p oui 23 3 5 6 A subgrid scale orography scheme aooaa a a 24 Buh DO PTOCESSES nuu noe oh RO a a Kew ek Ep UTR Res ped 24 3 0 Data E Cni ocre eb ke ale aed ee Bate he ae ae is 27 Part VII User s Guide 4 28 Contents Contents ii 4 Installation of the COSMO Model 29 4 1 External Libraries for the COSMO Model aana 29 Aki Oera eae 45s Wed REIR D Ee RUE bea dh de ee A 29 A a yar ep er cae mo e oe a g x So ok Eom 9 e a RUNS R Ros da 30 Zu IAB onum oi Saas cage A bete Mn que e e a 30 4 1 4 libcsobank a libsupplement a o e 30 AMO LORITO 380 9k he eo ROREM ERR URS oe a ee eR 31 42 Working with the VOS succus a qu Re Ph eee EG E s 31 453 Preparing the Code esmerada eR 4 REGE A ee ee oe he eS 31 AA Complme and Linking ice rios A aana xx b ee xU m BON ee a hea 33 245 BRunnmpthe Code s sas due i a aeo ox Romo m XD GOR do eR A C En 34 5 Data Formats for I O 35 5 L Ehe GRIB Binary Data Format se cresci pa 05686544 yo RR bees 35 DLI Code Porm oc ey fay ek OHO eae Re bee ee See d EURO eee Y Y Re 35 5 1 2 Indicator and End Section o s sa e a e acate o ee 36 5 1 3 Product Definition Section adonde ewe ae Wc 38 5 144 Grid Description SeebloH s a oe a kom soe ee ox Rogo em a 40 Delay a 2d omo E Roo ROTADAE Yo Re UR OR Eve a Rogo ds os 40 5 106 Binary Data Section sooo 9er oq eee e n S
296. spread 0 spreading along model levels gt vertical weights depend approximately on differences in log pressure exactly on differ ences in scaled height between the point Por for which the observation increment is valid and the target level at the horizontal location of the observation spreading along horizontal surfaces gt vertical weights depend approximately on log pressure differences between point Por and the target grid point 2 spreading along isentropic surfaces gt vertical weights depend on potential tempera ture differences between point Por and the target grid point msprpsu INT switch specifying the surface along which surface level 0 observation increments are primarily spread 0 spreading along model levels spreading along horizontal surfaces 2 spreading along isentropic surfaces for msprpsu point Po is always a grid point at the lowest full model level A target grid point is any model grid point for which the analysis increments and hence the influence or weight of any observation is to be computed Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 122 Vertical weights Name Type Definition Purpose Comments Default vertical weights for upper air observations vcorls 4 REAL square of the vertical correlation scale 0 3
297. ssibility to use a 3 D closure scheme Up to now this has been implemented into the COSMO Model only for testing purposes COSMO EU and COSMO DE use the 1 D TKE based closure scheme 3 5 5 Parameterization of Surface Fluxes Basic Namelist settings lphys TRUE ltur TRUE Mesoscale numerical modelling is often very sensitive to surface fluxes of momentum heat and moisture These fluxes provide a coupling between the atmospheric part of the model and the soil model For both closure schemes described in Sec 3 5 4 a special surface layer scheme can be applied Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 5 Physical Parameterizations 24 a A bulk transfer scheme Basic Namelist settings itype tran 1 For the 1 D diagnostic turbulence scheme a stability and roughness length dependent surface flux formulation based on Louis 1979 is implemented b A TKE based surface transfer scheme Basic Namelist settings itype tran 2 In context with the TKE scheme a revised and consistent formulation for the trans port through the surface layer should be used This surface scheme extends the TKE equation to the constant flux layer and introduces an additional laminar layer just above the surface This makes it possible to discriminate between the values of the model variables at the rigid surface e g radiative surface temperatures and values at the roughness height zo lower boundary of the turbul
298. sts all the variables which are used by COSMO plus only a subset of the variables that are not used by COSMO For convenience the table is split into a report header and a report body part The use of the variables is defined as follows need COSMO asks stringently for this variable and will abort if variable is absent but will not abort if values are equal to missing value opt variable exists and is read used but COSMO will not abort if it does not exist used means here that it is e g written to the feedobs file but it does not imply active use in the data assimilation up variable exists but is not read by COSMO S descriptor for a BUFR common sequence exists only in the BUFR file not in the NetCDF file use WMO descriptor type mnemonics meaning need 0 01 005 int MABNN Buoy platform identifier need 0 02 001 int NIX Type of station 0 02 036 int NBOTY Buoy type opt 0 02 149 int MTODB Type of data buoy 3 01 011 Year month and day need 0 04 001 int MJJJ Year need 0 04 002 int MMM Month need 0 04 003 int MYY Day 3 01 012 Hour minute need 0 04 004 int MGG Hour need 0 04 005 int NGG Minute 008021 int MTISI Time significance 26 time of last known position E 3 01 011 Year month and day 3 01 012 Hour minute 3 01 021 Latitude and Longitude need 0 05 001 float MLAH Latitude high accuracy degree need 0 06 001 floa
299. t MLOH Longitude high accuracy degree need 0 07 030 float MHOSNN Height of station above MSL 1 0 01 012 int MDS Platform drift direction 0 01 014 float MDSDS Platform drift speed opt 0 33 023 int MQOBL Quality of buoy location Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files 65 use WMO descriptor type mnemonics meaning 3 02 021 Waves 306 004 Depths salinities temperatures 0 31 001 int MDREP Delayed descriptor replication factor 0 07 062 float n NZNZN Depth below sea water surface 2 0 22 043 float n MTNOO Sea water temperature 2 0 22 062 float n MSNSN Salinity 2 306 005 Depths directions speeds of currents 3 02 001 Pressure and pressure change need 0 07 031 float MHOBNN Height of barometer above MSL 1 need 0 10 004 float MPPP Pressure opt 0 10 051 float MPPPP Pressure reduced to MSL opt 0 10 061 float NPPP 3 hour pressure change 0 10 063 int NA Characteristic of pressure tendency opt 0 07 032 float MHOSEN Height of sensor above marine deck platform for temperature and humidity measurements 0 07 033 float MHAWAS Height of sensor above water surface for temperature and humidity need 0 12 101 float MTDBT Temperature dry bulb temperature need 0 12 103 float MTDNH Dew point temperature opt 0 13 003 int MUUU
300. t they will no longer be retained Parameters for control ouput For control output a number of variables are computed and written to the ASCII files YUPRMASS for mass variables and YUPRHUMI for humidity variables These variables are domain averages of quantities like surface pressure surface pressure tendency kinetic energy dry static energy moist static energy cloud water content absolute vertical velocity at certain levels precipitation rates and accumulated precipitation and domain maxima of absolute horizontal and vertical velocity This allows for a quick look control of the model run Name Type Definition Purpose Comments Default nOmeanval INT Number of time step for the first call of control output If 0 nOmeanval gt nstop no control output will be computed nincmeanval INT Interval in time steps between two calls of the control calcu 10 lations Ouput for COSMO Testsuite Name Type Definition Purpose Comments Default ltestsuite LOG To activate additional ASCII output which is used and evalu FALSE ated when running the COSMO Technical Testsuite Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 6 DIACTL Parameters for Diagnostic Output 108 Control parameters for grid point output Only one type of grid point output can be chosen either the short meteographs 1gpshort TRUE the long grid point
301. t Files for the COSMO Model 6 4 Observation Input Files 72 The formats consist of the descriptors in the following table where need COSMO asks stringently for this variable and will abort if variable is absent opt variable exists and is read used but COSMO will not abort if it does not exist P variable exists but is not read by COSMO variable does not exist in this format WP RASS VAD descript type mnemonics meaning opt opt opt 001 033 int MMIOGC Identif of orig generat centre 0 01 034 int MMIOGS Identif of orig gen sub centre need need need 0 01 001 int MII WMO block number need need need 0 01 002 int NIII WMO station number 0 02 001 int NIX Type of station need need need 0 04 001 int MJJJ Year need need need 0 04 002 int MMM Month need need need 0 04 003 int MYY Day need need need 0 04 004 int MGG Hour need need need 0 04 005 int NGG Minute need need need 005 002 float MLALA Latitude coarse accuracy deg need need need 006 002 float MLOLO Longitude coarse accur deg need need need 0 07 001 int MHP Height of station m opt opt 010018 char 5 YSSOSN Short station or site name opt opt opt 0 02 003 int NAA Type of measur equipment used opt opt 0 25 021 int MWCE Wind computation enhancement opt opt 0 08 021 int MTISI Time significance opt 0 04 026 in
302. t MSETP Time period or displacement s need need need 0 31 001 int MDREP Delayed descript replic factor opt opt 033002 int n MQINZ Quality information need need need 007 007 int n MH Height m need need 011001 int n NDNDN Wind direction degree need need 011 002 float n NFNFN Wind speed m s need 0 12 007 float n MTVIR Virtual temperature K opt opt opt 011006 float n MWMPS W component m s opt opt 021030 int n NSINOR Signal to noise ratio db opt opt 025034 int n NWPQ NOAA WP quality ctrl results opt opt 011050 float n NSTDFF Stand deviat wind speed m s 0 11 051 float n NSTDVF Std vertical wind speed m s Table notes If YSSOSN is present then MII and NIII are not strictly mandatory n in the variable type definition means that this variable has an additional dimension here for vertical levels If the corresponding replication factor MDREP is zero for all reports in the NetCDF file then these multi dimensional variables do not need to exist and probably will not exist in the NetCDF file Not all variables are listed here that are present in some of the templates but are not read and used by COSMO Some of the variables that are used but not needed e g in the wind profiler file but are not present e g in the VAD file could be added to the VAD file in the future an
303. te of Technology Running with zero vertical velocity as lower boundary condition in the fast waves solver FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE TRUE Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 2 RUNCTL Parameters for the Model Run 81 Parameters related to artificial initial and boundary conditions The following parameters are only in effect if the control parameter lartif data is set to TRUE This parameter controls the basic type of the simulation if it is set to TRUE a run with artificial or idealized initial and boundary data is performed If it is set to FALSE a real case simulation with initial and boundary data coming from a coarse grid driving model or from an assimilation run is done If lartif data TRUE is chosen there are parameters to control the type of lateral bound ary conditions relaxation periodic open and to include exclude metrical terms Also in this case another namelist ARTIFCTL will be read from the file INPUT_IDEAL and the namelist GRIBIN from file INPUT IO has no effect and will be skipped ARTIFCTL offers parameters for a wide variety of choices to configure idealized runs such as definition of orography ideal ized hills or pre user provided ASCII files specification of atmospheric and soil profiles for intitial and boundary conditions and artificial convection triggers warm b
304. th its assumption of column equilibrium for the precipitating particles we now solve the complete prognostic equations Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 5 Physical Parameterizations 21 for rain q and and snow q can be turned on off by Namelist switches This approach is therefore applicable to the meso y and smaller scales For this purpose a semi Lagrange advection scheme with a simple trilinear interpolation is implemented This scheme is cou pled by Marchuk splitting with the cloud microphysical sources and sinks and the implicit sedimentation flux calculation These options are available for the microphysics parameterization scheme All options but the graupel scheme are available as diagnostic scheme 1progprec FALSE or as prognostic scheme 1progprec TRUE a itype_gscp 1 A warm rain scheme which is similar to the original Kessler 1969 scheme all ice phase processes are ignored b itype_gscp 2 The basic scheme described above c itype_gscp 3 An extension of the basic scheme which includes cloud ice as an addi tional prognostic variable cloud ice scheme The scheme allows for an explicit repre sentation of ice clouds and a more complete simulation of precipitation formation in mixed phase clouds This scheme is used in the COSMO EU d itype_gscp 4 A graupel scheme in addition to cloud ice has been implemented re cently It allows for an explic
305. the BUFR file not in the NetCDF file D variable does not exist at all Z P WMO descriptor type mnemonics meaning 3 01 110 Identific of launch site instumentation 3 01 001 Station identification need 0 01 001 int MII WMO block number 1 need 0 01 002 int NIII WMO station number 1 need 0 01 011 char 9 YDDDD Ship or mobile land station identifier 1 need 0 02 011 int NRARA Radiosonde type opt 0 02 014 int NSASA Tracking technique status of system opt 0 02 003 int NAA Type of measuring equipment 3 01 113 Date time of launch opt 0 08 021 int MTISI Time significance 18 launch time 3 01 011 Year month and day need 0 04 001 int MJJJ Year need 0 04 002 int MMM Month need 0 04 003 int MYY Day 3 01 013 Hour minute second need 0 04 004 int MGG Hour need 0 04 005 int NGG Minute 0 04 006 int MSEC Second 3 01 114 Horiz vert coord of launch site 3 01 021 Latitude and Longitude need 005 001 float MLAH Latitude high accuracy degree need 0 06 001 float MLOH Longitude high accuracy degree need 0 07 030 float MHOSNN Height of station above MSL 2 need 0 07 031 float MHOBNN Height of barometer above MSL 2 opt 0 07 007 int MH Height of release of sonde above MSL opt 0 33 024 int MSEQM Station elevation quality mark Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files 63
306. the above example in this alternative format is also shown above Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 75 Section 7 Namelist Input for COSMO Modell The execution of the COSMO Model is controlled by 14 NAMELIST groups LMGRID specifying the domain and the size of the grid RUNCTL parameters for the model run DYNCTL parameters for the adiabatic model PHYCTL parameters for the diabatic model TUNING parameters for tuning dynamics and physics DIACTL parameters for the diagnostic calculations SATCTL parameters for the satellite images NUDGING controlling the data assimilation INICTL parameters for the initialization of model variables EPSCTL controlling the ensemble prediction mode IOCTL controlling the environment DATABASE specification of database job GRIBIN controlling the grib input GRIBOUT controlling the grib output These NAMELIST groups have to appear in the corresponding INPUT files See Tab 6 1 for the distribution of the NAMELIST groups to the INPUT_ files The INPUT_ files have to be in the directory from where the model is started Every group is read in a special subroutine called input groupname This subroutine also sets default values for all parameters and checks most parameters that have been changed for correctness and consistency The NAMELIST variables can be specified by the user in the run script for the model
307. the initial and boundary data have to provide all meteorological fields necessary for running the model It is checked whether all fields are present otherwise the run will be aborted Which fields are needed depends on the settings of special Namelist switches All possible initial fields required are listed below Initial fields required for the COSMO Model in all cases hsurf geopotential of the earth s surface gzo roughness length fr land part of land in the grid cell soiltyp soil type of the land plcov degree of plant covering lai leave area index rootdp root depth u zonal wind speed v meridional wind speed W vertical wind speed pp deviation from reference pressure t temperature t_snow temperature of snow surface t_s temperature at the boundary soil atmosphere qv specific water vapor content qc specific cloud water content qv_s specific water vapor content at the surface wi water content of interception water Ww Snow water content of snow vio3 vertical integrated ozone content hmo3 ozone maximum Initial fields required for multi layer soil model t_so soil temperature on the different layers Ww So soil moisture of the different layers rho snow density of snow freshsnw weighting function indicating freshness of snow Initial fields required for 2 layer soil model tm temperature between upper and medium soil layer t_cl temperature between medium and lower soil layer w_gl water content of the upper soil lay
308. the output data If none of these variables is set to TRUE YUCHKDAT is not written Part VII User s Guide 4 28 Section 8 Model Output 8 1 ASCII Output for the Forecast Model 152 8 1 4 YUPRMASS Protocolling the forecast with mass variables YUPRMASS contains meanvalues of model variables related to mass and deviations from initial mean values First the initial values of the following variables are written e area mean value of the surface pressure in hPa for the total model domain without boundary zone e volume mean values of dry static moist static and kinetic energy in J kg In the next lines the following values are written ntstep actual time step Real elapsed time since the last line was printed dpsdt area mean value of the tendency of the surface pressure This is a good measure for the noise in the model ps area mean value of the surface pressure dse deviation of the volume mean value of dry static energy from the initial value mse deviation of the volume mean value of moist static energy from the initial value ke deviation of the volume mean value of kinetic energy from the initial value vamx maximum of the horizontal velocity wamx maximum of the absolute vertical velocity waxxx area mean values of the absolute vertical velocity for three model layers in about 850 500 and 300 hPa The file YUPRMASS is always written With the NAMELIST parameters nOmeanval and ninc meanval of dia
309. tics on 2 dimensional surface analyses on file YUSURF FALSE see section 8 2 10 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 133 The remaining variables in this section are related to the latent heat nudging and are relevant only if 11hn TRUE Some basic variables on the LHN and its diagnosis Name Type Definition Purpose Comments Default Basic on off switches llhnverif LOG switch for verification FALSE skill scores against radar obs written to file YULHN lhn diag LOG switch for enhanced detailed diagnostic output TRUE written to file YULHN LHN period relative to the initial time of the model run nlhn start INT start of latent heat nudging LHN period in timesteps 0 nlhn end INT end of latent heat nudging LHN period in timesteps 0 or hlhn start REAL start of latent heat nudging LHN period in hours 0 0 hlhn end REAL end of latent heat nudging LHN period in hours 0 0 verification period nlhnverif start INT start of period of verification of LHN in timesteps 0 nlhnverif end INT end of period of verification of LHN in timesteps 0 or hlhnverif start REAL start of period of verification of LHN in hours 0 0 hlhnverif end REAL end of period of verification of LHN in hours 0 0 relative to the initial time of the model run
310. tion Basic Namelist settings 12tls TRUE lsemi_imp FALSE irunge kutta 1 2 This scheme with irunge kutta 1 is used for the COSMO DE and COSMO EU This scheme has been implemented into the COSMO Model as an alternative to the former default the Leapfrog scheme and can be combined with a forward backward scheme for integrating the high frequency modes of the elastic equations The first irunge kutta 1 variant is the normal 3rd order Runge Kutta scheme used by Wicker and Skamarock 2002 whereas the second one is a total variation diminishing TVD variant of 3rd order Liu et al 1994 irunge_kutta 2 Different horizontal advection upwind or centered differences schemes of 3rd to 6th order can be used the operators are formulated in advection form The vertical advection is nor mally treated in an implicit way using a Crank Nicolson scheme and centered differences in space Most slow tendencies such as vertical diffusion thermal solar heating parameterized convection and coriolis force are computed only once using values of the prognostic variables at time step n These tendencies are fixed during the individual Runge Kutta steps and contribute to the total slow mode tendencies which are integrated in several small time steps together with the fast mode tendencies in a time splitting sense In contradiction to this the whole 3D advection is computed in each Runge Kutta step 3 4 2 Leapfrog 3 timelevel HE VI Integration Basi
311. tion 6 Input Files for the COSMO Model 6 4 Observation Input Files 62 COSMO is coded such that from the file cdfin pilot p in addition to the mandatory pressure variable the height variable can also be read as an optional variable and from cdfin pilot in addtion to the mandatory height variable the pressure variable is also read if present This would allow for using PILOT reports which contain both pressure levels and height levels However since the current version of bufrr2netcdf is not able to produce such mixed PILOT NetCDF files this has never been tested The table below split into two parts lists all the variables and their use in COSMO The use of the variables first column of the header table columns Z and P of the body table for PILOT with height resp pressure as vertical coordinate is defined as follows need COSMO asks stringently for this variable and will abort if variable is absent but will not abort if values are equal to missing value opt variable exists and is read used but COSMO will not abort if it does not exist used means here that it is e g written to the feedobs file but it does not imply active use in the data assimilation opt variable does not exist in the template if it does exist nevertheless then it is read and used by COSMO Tapas variable exists but is not read by COSMO t descriptor for a BUFR common sequence exists only in
312. tion of large scale parameterization schemes by fine scale modelling and tests of computational strategies and numerical tech niques For these types of studies the model should be applicable to both real data cases and artificial cases using idealized test data Moreover the model has been adapted by other communities for applications in climate mode CCLM and or running an online coupled module for aerosols and reactive trace gases ART Part VII User s Guide 4 28 Section 1 Overview on the Model System 1 2 Basic Model Design and Features 4 Such a wide range of applications imposes a number of requirements for the physical nu merical and technical design of the model The main design requirements are i iii use of nonhydrostatic compressible dynamical equations to avoid restrictions on the spatial scales and the domain size and application of an efficient numerical method of solution provision of a comprehensive physics package to cover adequately the spatial scales of application and provision of high resolution data sets for all external parameters required by the parameterization schemes flexible choice of initial and boundary conditions to accommodate both real data cases and idealized initial states and use of a mesh refinement technique to focus on regions of interest and to handle multi scale phenomena use of a high resolution analysis method capable of assimilating high frequency asy
313. tion time window the messages do not imply automatically that all of these observations are never used It only means that these data are rejected at least at the individual observation time itself when the checks give the best estimation about the data quality and when the nudging weights would be largest List of Observations Rejected by the Threshold Quality Control at the Observation Time Obs Station ID Code Time Pressure Lat Lon Thresh Var Obs Model Var Obs Model Diff Op uv 08019 137 0 0 893 9 43 5 6 3 10 4 u 545 Shed ge Y m S 2238 1365 V mult 08019 137 0 0 915 8 343 5 6 3 p top 367 6 T 2m 62144 24 0 0 1000 1 53 4 Ligh 20 E 250 7 287 4 36 6 uv 10m 06012 14 0 0 995 7 62 3 6 3 12 0 Q 2 9 IL ov 1 2 10 0 16 7 RH EU6564 244 0 0 815 9 48 2 12 2 0 45 RH 0 46 0 96 RH EU6564 244 0 1 857 0 48 3 12 1 0 44 RH 0 48 0 95 RH EU6564 244 0 1 883 6 48 3 12 0 0 44 RH 0 46 0 94 RH EU6564 244 0 1 846 6 48 3 12 1 0 44 RH 0 48 0 96 RH EU6564 244 0 1 838 3 48 3 12 1 0 47 RH 0 50 0 97 q mult EU6564 244 0 1 846 6 48 3 2 1 p top 815 9 T mult EU9734 144 0 1 959 5 50 2 4 3 p top 729 9 T mult EU9734 144 0 2 694 2 50 3 4 1 p top 578 7 T mult EU9734 144 0 2 555 9 50 3 3 8 p top 483 4 T mult EU9734 144 0 3 465 6 50 2 Ja p top 406 5 uv RCH7440 14 0 3 206 5 60 0 20 3 106 1 u 21 8 18 0 p v T8 X328 Fak T RCH7440 14 0 3 206 5 60 0 20 3 6 9 T 233 1 223 4 uv EU526 244 0 7 507 1 51 2 11 6 14 0 u 19 8 18 7
314. tions 1 Domain used for verification rotated pole see above lower left corner 20 000 18 000 upper right 21 000 23 500 domain size 665 657 Types assigned to model runs in table below 0 the model run is one straightforward model integration 0 the model run comprises of x hourly periods from a series of cycled integrations starting at x hour intervals x 6 12 or 24 and the initial date and time relates to the latest of these integrations 2 the model run is a series of analyses forecast time horiz number model initial Jat the end of mesh of run date and hour verification width vertical description of no type of the run period hrs 1 deg levels the model run 1 02 2009022412 0 0000 16 00 40 analysis 4 COMO METO 910 4580 30 247 293 12 800 102 20002 0 279 253 603 650 9 9 9 2286 32 0 0 64 42 78 4 COMO SGN 910 4580 30 247 293 12 829 102 20002 0 279 253 433 9999 9 9 9 2275 32 0 0 64 9 92 4 COMO BKG 910 4580 30 247 293 12 830 0 00 0 279 253 411 9999 9 9 9 2274 32 0 0 64 9 114 1 004 337 5446 98 0 0 9 123 200 00 0 227 395 47 5 9 9 9 0 0 0 64 9 9 9 17777000707 17777777777 9 30 117 9999 9999 9999 9 9 9 9 1 63551 84 6746 20 0 0 4 165 200 00 0 230 604 9999 9999 9 9 101510 0 8 0 0 64 510 9 9 177011111177 177770711717 9 9999 9999 9999 9999 58 9 9 9 9 0 07497 677 4562 20 868 1649 1 1 412 402 2 253 252 10 2 2785 522 91790 868 0 0 203004020 1 501 3 2500 17714653400 1766252
315. tput described in this section is produced only if the compile option DNUDGING is used for the production of the COSMO binary and the NAMELIST variable luseobs is set to TRUE For some of the files additional prerequisites exist The ASCII files provide a helpful tool for a quick monitoring and diagnosis of the tasks related to the use of obser vations such as data assimilation input for verification and production of 2 dimensional analyses based on observations These files are e YUAOFEX Nudging observation input AOF e YUOBSDR Nudging list of active and passive reports e YUREJCT Nudging list of rejected reports e YUQUCTL Nudging list of data rejected by quality control YUSTATS Nudging statistics on observation processing e YUCAUTN Nudging warning messages indicating insufficient array sizes e YUVERIF Nudging verification file VOF e YUPRINT Nudging other various aspects e YUSURF 2 D surface analyses e YULHN Latent Heat Nudging description not yet available Note that the NAMELIST parameters related to the tasks which use observations are also written to file YUSPECIF This has already been described in Section 8 1 Furthermore the file YUTIMING is also extended with parts that consider the nudging or production of the YUVERIF and or NetCDF feedobs files and the latent heat nudging 8 2 1 YUAOFEX Nudging Observation Input AOF File YUAOFEX is written only if the observations for the
316. ts on multi level aircraft reports e g CAUTION for maxmlo nexceair O 117 as well as summary messages Finally there are messages Figure 8 26 which are not written by the nudging itself but which are related to the set up of the data assimilation as a whole It lists the originating process for a number of 2 dimensional grib fields which are used as part of the initial condition of the model integration Time range indicator 13 means that the corresponding field has been produced by an COSMO Model assimilation integration whereas a value of 0 indicates that the field originates from a snow analysis for T SNOW W SNOW and W_I a sea surface temperature analysis or another external data collection and interpolation process Additional element number 20 for W S0 implies finally that the initial soil moisture fields are produced by the variational soil moisture analysis instead of the nudging based COSMO Model assimilation run Note analysis field LAI with time range indicator 0 is used Note analysis field VIOS3 with time range indicator 0 is used te analysis field HMO3 with time range indicator 0 is used te analysis field PLCOV with time range indicator 0 is used Note analysis field ROOTDP with time range indicator 0 is used Note analysis field LAI with time range indicator 13 is discarded te analysis field ROOTDP with time range indicator 13 is discarded te analysis field PLCOV with time range indicator 13 is
317. ts the external libraries that are necessary to run different components of the model and what can be done if these libraries are not available Section 4 2 describes how to use the VCS Version Control System a programming environment tool developed at DWD for working with the model If the VCS is not available the source code together with a Makefile for compiling and linking and scripts for running the model are provided T he next sections give detailed informations on how to prepare compile link and run the COSMO Model 4 1 External Libraries for the COSMO Model For some components the COSMO Model uses external libraries Usage of most of these libraries can be controlled by conditional compilation To handle this the C preprocessor cpp must be called Most Fortran compilers activate the C preprocessor for files ending with a capital F in the suffix F or F90 INT2LM does not use capital letters in the suffix therefore a special compiler option has to be set to activate this preprocessor Take a look to the manual of your compiler to find out about this option At DWD a data base system can be used for this which needs special routines If these are not available dummy routines are provided to satisfy the external references 4 1 1 libgribl a This library has to be available and cannot be controlled by conditional compilation up to now The standard WMO data format the GRIB Grided Binary Version 1 can used for I O of
318. ts from the former version the following values have to be set itype gscp 3 and Leapfrog dynamics 15 0 itype gscp 3 and Runge Kutta dynamics 25 0 itype_gscp 4 and Runge Kutta dynamics 20 0 Introduced in Version 4 14 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 6 DIACTL Parameters for Diagnostic Output 107 7 6 DIACTL Parameters for Diagnostic Output DIACTL contains parameters to generate gridpoint and control ASCII output These param eters are only in effect if the main switch ldiagnos in RUNCTL is set to TRUE In this case some additional time integrated fields TDIV_HUM vertically integrated divergence of specific humidity and AEVAP S surface moisture flux are also calculated to allow for a mass budget calculation based on GRIB output Currently the following ASCII output can be generated Grid point output meteographs in a specific form results are written to files M stationname see Section 8 1 1 Control output for a quick look monitoring of the model run results are written to files YUPRMASS for mass variables and YUPRHUMI for humidity variables resp see Section 8 1 4 and 8 1 5 Note Up to model version 3 4 two addtional forms of ASCII output could be generated Di agnostics for various subdomains and differences between predicted and and boundary fields Since these diagnostics can be easily calculated from GRIB outpu
319. turb 3 0 Standard implicit treatment of vertical diffusion in the solver for slow processes slow tendencies f90 1 As option 0 but using Neumann boundary conditions for heat and moisture transport at the lower boundary specified fluxes instead of Dirichlet boundary condi tions specified values 2 Explicit treatment of vertical diffusion by calculating a process splitted explicit tendency 3 Alternative implicit treatment of vertical diffusion by calculating a process splitted implicit tendency TRUE Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 4 PHYCTL Parameters for the Diabatic Model 100 Name Type Definition Purpose Comments Default icldm turb 13dturb lprog_tke 13dturb_metr lexpcor ltmpcor lnonloc lcpfluc lturhor limpltkediff ltkesso ltkecon itype sher INT LOG LOG LOG LOG LOG LOG LOG LOG LOG LOG LOG INT Mode of cloud representation to take into account subgrid scale condensation within the turbulence parameterization in case of itype turb 3 Options 0 1 and 2 as for icldm rad i e 0 No clouds are considered 1 Only grid scale clouds are considered 2 Grid and sub grid scale water clouds are considered cloud cover and water content are calculated according to a relative humidity criterion itype wcld 1 ora statistical clos
320. ubbles soil hot spots Name Type Definition Purpose Comments Default lartif data LOG If TRUE the model runs with user defined initial and bound FALSE ary data All Namelist input from GRIBIN is skipped and the corresponding parameters are not in effect A Namelist group ARTIFCTL has to be defined instead If FALSE the model expects initial and boundary data in Grib or NetCDF format Control parameters for these files are contained in Namelist input group GRIBIN lartif data has been renamed named lgen before and moved from group GRIBIN to RUNCTL in Version 4 8 12dim LOG Run a 2 dimensional model version FALSE To run the model in 2 d mode the number of gridpoints in south north direction must be set to je tot 5 lperi LOG eliminated in Version 4 17 and replaced by lperi_x and lperi y see below lperi x LOG Use periodic lateral boundary conditions in x direction FALSE lperi y LOG Use periodic lateral boundary conditions in y direction FALSE lcori LOG moved to group DYNCTL in Version 4 8 lmetr LOG moved to group DYNCTL in Version 4 8 lradlbc LOG moved to group DYNCTL in Version 4 8 Also a 1 dimensional set up can be realized by setting the number of gridpoints in west east direction to ie tot 5 and choosing periodic lateral boundary conditions Depending on the problem at hand however some additional modification of the dynamics code might be n
321. udget equation for the specific water contents q of the various categories water vapour q cloud water q cloud ice q and graupel q9 depending on the scheme used take advective and turbulent transport into account and contain source and sink terms due to the micro physical processes of cloud and precipitation formation For rain water q and snow q only advective transport is considered The following mass transfer rates are considered by the scheme a condensation and evaporation of cloud water b the initial formation of rainwater by autoconversion and of snow by nucleation from the cloud water phase c the subsequent growth of the precipitation phases rain and snow by accretion riming deposition and shedding d evaporation of rainwater and sublimation of snow in subcloud layers and e melting of snow to form rain and freezing of rain to form snow The impact of the vertical motion of rain and snow relative to the airflow due to the sedi mentation of particles with their terminal velocities is also taken into account by the vertical divergence of the corresponding precipitation fluxes P and P4 Figure 3 5 illustrates the microphysical processes considered by this parameterization scheme Smelt Sfrz Snuc Srim rain Ss water Sev Sdep vapour P Fy Ps Figure 3 5 Hydrological cycle in the COSMO Model cloud and precipitation scheme In contrast to the former diagnostic precipitation scheme wi
322. ues of LAI 0 0 rmax lai REAL Upper limit for values of LAI 8 0 fac rootdp REAL Modification factor for ROOTDP 1 0 rmin rootdp REAL Lower limit for values of ROOTDP 0 0 rmax rootdp REAL Upper limit for values of ROOTDP 2 0 Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 11 IOCTL Controlling the Grib I O 137 7 11 IOCTL Controlling the Grib I O IOCTL contains some general parameters to control the Grib or NetCDF IO of a model run There is a main switch lgen to select a run with user defined artificial initial and boundary data you have to edit the module src_artifdata f90 for your purposes or a real case model run with initial and boundary data coming from a coarse grid driving model initial data can also be generated by a continuous data assimilation using the nudging technique Extensive IO such as reading and writing full domain files on a multi processor machine can slow down the model performance significantly For time critical applications such as oper ational NWP the COSMO Model has the capability for doing IO operations asynchronous with a certain pre or postponement to the model execution on dedicated extra processors At the moment the asynchronous IO is implemented only for Grib files In an operational environment the COSMO Model can also be run in parallel with the in terpolation program which generates lateral boundary conditions from the corr
323. ulation of the model the parameterization of physical processes and the four dimensional data assimilation The scientific documentation is independent of i e does not refer to the code itself Part IV will describe the particular implementation of the methods and algorithms as presented in Parts I III including information on the basic code design and on the strategy for parallelization using the MPI library for message passing on distributed memory machines not available yet The generation of initial and boundary conditions from coarse grid driving models is described in Part V This part is a description of the interpolation procedures and algorithms used not yet complete as well as a User s Guide for the interpolation program INT2LM Available postprocessing utilities will be described in the future in Part VI Finally the User s Guide of the COSMO Model provides information on code access and how to install compile configure and run the model The User s Guide contains also a detailed description of various control parameters in the model input file in NAMELIST format which allow for a flexible model set up for various applications All parts of the documentation are available at the COSMO web site http www cosmo model org content model documentation core default htm Part VII User s Guide 4 28 Section 1 Overview on the Model System Section 2 Introduction The usage of the program package for the COSMO Model is
324. umentation Feedback File Definition which can also be found on the COSMO web site www cosmo model org 8 1 ASCII Output for the Forecast Model For a quick forecast monitoring the model writes various control output to ASCII files These files are e YUSPECIF NAMELIST parameters e YUCHKDAT Checking the input output data from GRIB NetCDF YUPRMASS Protocolling the forecast with selected mean values for mass variables YUPRHUMI Protocolling the forecast with selected mean values for humidity variables YUDEBUG More detailed information for debugging purposes YUTIMING Timings for the different parts of the forecast In addition output for meteographs grid point output can also be done For every selected grid point a file M stationname is written Part VII User s Guide 4 28 Section 8 Model Output 8 1 ASCII Output for the Forecast Model 148 8 1 1 M stationname Grid point output The files M stationname provide a monitoring of model variables at single grid points There is a short form NAMELIST variable lgpshort TRUE or a long form 1gplong TRUE Only one of these two forms can be chosen If none of the two variables is set to TRUE the meteograph files are not printed The number of grid points considered are limited because of memory reasons Up to nmaxgp grid points can be chosen The parameter nmaxgp is contained in data runcontrol f90 nmaxgp 100 If more grid points should be
325. unidata ucar edu packages netcdf conventions html contact John Somebody somewhere de references http www ensembles eu org creation date 2005 09 16 20 50 38 The header is divided in three parts 1 Dimensions netCDF COSMO Description dimension dimension rlon ie number of grid points in rotated longitudinal direction mass points rlat je number of grid points in rotated latitudinal di rection mass points srlon ie number of grid points in rotated longitudinal direction flux points srlat je number of grid points in rotated latitudinal di rection flux points level ke number of vertical full levels levell kel number of vertical half levels height_2m single atmosphere level height 10m single atmosphere level soil ke soil number of soil layers soill ke soill number of soil layers 1 bnds B 2 bounds of variables soil bounds and time bounds time dimension for time series The time dimension is different from the other dimensions since it is declared as un limited This makes it possible to cat together several output files and construct time series and animations 2 Variables Variables can be divided in two categories coordinate variables and the meteorological quantities Coordinate variables have the same name as their dimension All variables are in 32bit i e float except of time which is in 64bit i e double and rotated pole
326. uoy ancy terms However the use of a nonorthogonal curvilinear coordinate system results in an elliptic operator containing cross derivative terms with variable coefficients A mini mal residual Krylov iterative solver GMRES was thus chosen to solve for the perturbation pressure tendency We found the convergence criterion proposed by Skamarock et al 1997 to be both sufficient and a robust predictor of when the RMS divergence of the flow has stabilized An efficient line Jacobi relaxation preconditioner was developed having the prop erty that the number of Krylov solver iterations grows slowly as the convergence parameter Ec decreases Once the solution for the pressure tencendy is known the other variables are updated by back substitution 3 5 Physical Parameterizations Some parts of the physics package of the COSMO Model are adapted from the former oper ational hydrostatic model DM Others have been widely rewritten or were replaced by new Part VII User s Guide 4 28 Section 3 Model Formulation and Data Assimilation 3 5 Physical Parameterizations 19 developments This section gives a short overview on the parameterization schemes used A detailed description is given in Part II of the Documentation Physical Parameterizations 3 5 1 Radiation Basic Namelist settings lphys TRUE lrad TRUE hincrad 1 0 To calculate the heating rate due to radiation we employ the parameterization scheme of Ritter and Geleyn 1
327. ure itype_wcld 2 Switch to choose a 3D turbulence scheme for itype turb 5 T Switch to choose a prognostic treatment of TKE for itype_turb 5 7 To switch on off the use of metric terms in the 3D turbulence Explicit corrections of implicitly calculated turbulent heat and moisture fluxes due to effects from subgrid scale condensation only if itype_turb 3 Should be set to TRUE to allow for a consistent treatment of diffusion coefficients and fluxes Consideration of thermal TKE source in enthalpy budget only if itype turb 3 Nonlocal calculation of vertical gradients used for turbulent diffusion only if itype turb 3 Consideration of fluctuations of the heat capacity of air only if itype turb 3 Switch to include horizontal turbulent diffusion It has been eliminated in Version 4 10 and replaced by limpltkediff Switch to include horizontal turbulent diffusion Implemented in Version 4 10 Switch to calculate SSO wake turbulence production for TKE Implemented in Version 4 10 Switch to consider convective buoyancy production for TKE Implemented in Version 4 20 Type of shear production for TKE Implemented in Version 4 10 1 Only vertical shear 2 Full isotropic 3D shear 3 Vertical shear and separted horizontal shear mode 1 FALSE FALSE TRUE FALSE FALSE FALSE FALSE FALSE TRUE FALSE FALSE Part VII User s Guid
328. v 15 6 1420 3 147 T EU526 244 0 7 507 1 51 2 11 6 5 1 T 267 0 260 8 dz EU526 244 0 6 984 7 51 4 2 2 26 1 p top 750 6 dz 45 9 dT mean 5 8 uv EU0350 244 0 7 197 0 50 1 96 16 1 u 215 9 12 7 v gt 18 2 437 0 19 1 T EU0350 244 0 7 197 0 50 1 8 6 7 1 T 221 7 213 6 uv RCH8125 14 0 8 196 8 057 5 17 6 16 1 u 2li Z9 op Y dos 12 8 3 4 T RCH8125 14 0 8 196 8 57 5 17 6 Fal T 229 1 218 8 ps scc SKEC 24 1 0 994 6 56 1 6560 3 3 ps 994 6 998 5 bias w2 0 1 8 5 z mult 40179 35 125 1000 0 32 0 34 8 p top 50 0 RH 60571 35 1 8 604 0 31 5 2 2 0 58 RH 0 02 0 66 RH 60571 35 148 500 0 31 5 2 2 0 56 RH 0 11 0 86 RH 60571 35 1 8 485 0 31 5 2 2 0 56 RH 0 10 0 82 IWV sc 60571 35 1 8 371 0 31 5 2 2 502 1l1WV 5 565 11 7 bias w2 0 0 0 0 uv 6080 35 149 10 0 45 4 943 225 17 u 19 4 S69 cy 15 5 3 6 23 5 T E 6080 35 A9 10 0 45 4 94 3 AZL T 220 9 211 1 z mult 17609 32 2 0 977 6 34 9 33 6 p top 783 1 z mult 17600 32 2 0 977 7 34 7 32 5 p top 844 1 ps 6080 35 2 0 985 0 45 4 93 bw ps 98520 990 0 ps scc 16080 35 2 0 985 0 45 4 9 3 5 5 ps 985 0 987 0 bias w2 3 0 17 5 IWV TRYN NGAA 834 2 2 61 4 12 4 2 82 IWV 6 87 9u 73 IWV NYKO NGAA 834 2 2 55 9 11 7 3 47 1WV 4 98 8 60 IWV OSV2 NGAA 834 2 2 60 2 17 2 3 05 IWV 5 24 9 08 IWV sc NYKO NGAA 834 2 2 55 9 11 7 3 9 IWV 5 0 8 6 bias w2 1 6 Dad IWV sc TRYN NGAA 834 2 2 61 4 12 4 3 0 IWV 6 9 9 8 bias w2 0 7 1 2 RH 2m 01415 35 2 4 1012 0 5
329. values 0 1 where stands for including and for excluding a value over land c_soil 0 c_Ind Name Type Definition Purpose Comments Default tkesmot REAL Time smoothing factor for TKE and diffusion coefficients 0 15 tkesmot c 0 2 wichfakt REAL Vertical smoothing factor for explicit diffusion coefficients 0 0 wichfakt 0 0 5 securi REAL Security factor for maximal diffusion coefficients 0 85 securi 0 1 tkhmin REAL Minimal diffusion coefficients for heat 0 4 tkhmin 0 2 tkmmin REAL Minimal diffusion coefficients for momentum 0 4 tkmmin c 0 2 zOm dia REAL Typical roughness length for a Synop station which is used 0 2 for the interpolation of screen level values of the 10 m wind instead of using the actual roughness length at the grid point zOm dia c 0 001 10 rat lam REAL Ratio of laminar boundary layer thickness for water wapour 1 0 and sensible heat rat lam 0 1 10 rat can REAL Scaling factor for the calculation of the canopy height 1 0 rat can c 0 10 rat sea REAL Ratio of laminar scaling factors for heat over sea 20 0 rat sea c 1 100 pat len REAL Length scale m of sub scale surface patterns over land 500 0 pat len c 0 10000 tur len REAL Maximal turbulent length scale m 500 0 tur len c 0 10000 c 1nd REAL Surface area index of gridpoints over land excluding leaf 2 0 area in
330. vation types as positive values and or code types as negative values by setting minus code type which define suc cessively one set after the other 0 denotes all the remaining observation and code types that are not specified to belong to another set of observing systems number of non zero elements in iwtyp sum over values in array niwtyp minus 1 kwtyp 22 INT mode of weighting W for multiple observations k specified for 1 1 1 each set of obseration systems m 1 gt Cw m 0 with wr _ Wein Cw m Wkm 2 gt Cum 1 25 Wy Cw m number of non zero elements in kwtyp E if nwtyp 1 nwtyp 2 if nwtyp gt 2 in this case the the first nwtyp entries are used for the nwtyp specified sets of observing systems the nwtyp 1 th entry is used combine the net obs increments from these sets for the final analysis increment and the nwtyp 2 th entry is for surface pressure Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 9 NUDGING Controlling the Data Assimilation 117 General variables controlling the verification without influence on the nudging Name Type Definition Purpose Comments Default nversta INT start of verification period in timesteps 0 nverend INT end of verification period in timesteps 0 or hversta REAL start of verification period in hours 0 0 hveren
331. ve e g s 3_mF or SL3_Mf are both recognized Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 3 DYNCTL Parameters for the Adiabatic Model 90 Parameters for the lateral boundary conditions Name Type Definition Purpose Comments Default lw freeslip LOG Free slip lateral boundary conditions are used for w by default TRUE in case of non periodic Davies type lateral boundaries Other wise w must be specified along the boundaries It is recommended to use free slip lateral boundary conditions for real data simulations lexpl lbc LOG Chooses an explicit formulation of the lateral boundary relax TRUE ation crltau REAL To set the time factor for the explicit formulation of lateral 1 0 boundary relaxation 7 crltau At rlwidth REAL If lexpl_lbc TRUE this parameter specifies the width of 85000 0 the relaxation layer in meters It should be 10 to 15 times the grid mesh size in meters but should not exceed 0 25 times the full domain size itype outflow qrsg INT To choose the type of relaxation treatment for qr qs qg 1 before Version 4 9 this switch was named itype lbcqx 1 qr qs qg are treated with the same lateral boundary relaxation as the other variables 2 norelaxation of qr qs qgis done at outflow boundary points itype lbc qrsg INT To choose the type of lateral boundary treatment for qr qs 1 qg i e which v
332. vertical gust speed need need 012001 float MTN Temperature dry bulb T 4 need 012101 float MTDBT Temperature dry bulb T 4 0 33 025 int MAIV ACARS interpolated values need need need 0 08 004 int MPHAI Phase of flight need need need 0 02 064 int MQARA Wind quality roll angle opt opt 013 003 int MUUU Relative humidity opt 0 12 103 float MTDNH Dew point temperature opt opt opt 0 13 002 float MMIXR Mixing ratio 011076 float MPTI Peak turbulent intensity EDR 0 20 041 int MAICI Airframe icing opt Opt opt 0 02 005 float MPOTO Precision of temperature observ 0 02 062 int NADRS Type of aircraft data relay system 0 02 070 int NOSLL Original specif of latit longit 0 02 065 char 5 YAGRS ACARS ground receiving system opt opt 0 33 026 int MMRQ Mixing ratio quality 0 04 015 int NGGTI Time increment E 0 04 032 int NGGTM Duration rel to following value 0 11 235 int 011235 unknown descriptor Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 4 Observation Input Files 71 Table notes 1 For the definition of the vertical level it is in fact only required that at least one of the variables NFLEV MIAA NHHH or MPN exists in any of the NetCDF file s but it does not really matter which one s And in order to use a report the corresponding vertical level must not be a
333. we num_gribtabs INT Specifies the number of different GRIB tables used in 6 COSMO Model variable table lst gribtabs LOG Identifications of the different GRIB tables see left Current table used are 2 201 202 203 204 205 lbdclim LOG If TRUE use additional boundary fields that are needed FALSE for long term simulations lbdsst LOG If TRUE use an additional boundary field to update the FALSE Sea Surface Temperature during NWP simulations ldwd grib use LOG If TRUE special DWD Grib settings are used TRUE l ke in gds LOG If TRUE write number of vertical levels as a vertical TRUE coordinate parameter to the GDS grib section Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 7 11 IOCTL Controlling the Grib I O 139 Writing and reading Restart Files nhour restart INT ydir restart CHAR ytunit restart CHAR Triplet to specify start stop and increment for writ 12 0 12 ing restart files Specifications are in full forecast hours Directory where to write the restart files 2d Time unit to specify the file name for restart files f Additional specifications for NetCDF IO yncglob institution yncglob title yncglob source yncglob contact yncglob project id yncglob experiment id yncglob references ncglob realization CHAR id CHAR r CHAR tes CHAR iid CHAR r CHAR tes CHAR ad INT 1
334. which then creates the INPUT_ files An excerpt of this run script is shown in Figure 7 1 for the forecast part and in Figure 7 2 for the nudging part Part VII User s Guide 4 28 Section 7 Namelist Input for COSMO Modell 76 HAH HH HE AE AE AE AE AE AE AE AE AE AE EE REAR FE FE FE FE FE FE FE FE FE FE FE FE cat together the INPUT files FEFE FEFE AE AE AE AE AE AE AE AE AE AE AE AE AE AE AE AE AE AE AE AE FE FE FE FE FE FE FE HE FE FE FE FE FE FE FE FE FE FE FE FE FE FE FE FE FE cat gt INPUT_ORG lt lt end_input_org amp LMGRID yvarzl default yvarpl default ydir gtmp routfor dat output end_input_io yvarsl default cat gt INPUT_DYN lt lt end_input_dyn amp DYNCTL lcpfluc FALSE ninctura 1 itype_turb 3 itype_wcld 2 icldm_rad 4 lsoil TRUE itype_evsl 2 itype_trvg 2 startlat_tot 20 0 startlon_tot 18 0 pollat 40 0 pollon 170 0 dlat 0 0625 dlon 0 0625 ie_tot 665 je_tot 657 ke_tot 40 amp RUNCTL hstart 0 0 hstop 48 0 dt 40 0 ydate_ini 2008021500 nprocx 8 nprocy 8 nprocio 0 lphys TRUE luse rttov QTRUE luseobs FALSE leps FALSE idbg level 2 amp TUNING clc diag 0 75 pat len 500 0 rlam heat 1 0 rlam mom 0 0 rat lam 1 0 rat_can 1 0 rat_sea 20 0 c 1nd 205 c_soil 1 0 c_sea 1 5 zOm_dia 0 2 crsmin 150 0 rat_sea 20 0 wichfakt 0 05 qeo 0 0 end_input_org cat gt INPUT IO lt lt
335. x intervals and is quite helpful for monitoring the assimilation Creation of NetCDF feedobs file fof OPENED e uwork fel2sra test 09022409 1me sx9 zlm4m zwtyp fof 20090224090000 nc feedobs file 4 newrep 4734 oldrep 0 newobs 38392 oldobs 0 STEP 0 NUMBER OF SINGLE OR PAIRS OF REPORTS WITH OBS INCREMENTS 8744 1259 multi level 106 TEMP 2 PILOT 16 WINDPROF 20 RADAR VAD 225 AIRCRAFT 1 RASS 889 GPS 0 RETRIEVALS 1999 sing lev aircraft 2878 in situ surface 2444 surf pressure 164 scatterometer Temperature correction for surface pressure nudging top at 400hPa in layer 14 relative to the surface pressure increment height T corr pressure z z dp level height T corr pressure z z dp level K hPa full hydro correl K hPa full hydro correl 7746 0 000 374 374 0 000 0 000 14 9030 0 000 312 312 0 000 0 000 14 7110 0 016 410 410 0 001 0 001 15 8587 0 000 332 332 0 000 0 000 15 6506 0 069 446 446 0 013 0 008 16 8157 0 000 354 354 0 000 0 000 16 5933 0 121 482 482 0 038 0 022 17 7742 0 000 375 375 0 000 0 000 17 5391 0 172 519 519 0 076 0 045 18 7344 0 010 397 397 0 000 0 000 18 4878 0 220 555 556 0 122 0 076 19 6962 0 129 419 419 0 007 0 007 19 4395 0 267 592 592 0 177 0 115 20 6598 0 272 441 441 0 033 0 025 20 3941 0 310 628 628 0 237 0 161 21 6252 0 408 462 462 0 073 0 055 21 3516 0 351 663 663 0 301 0 214 22 5924 0 538 483 483
336. y there are entries consisting of long bit patterns These entries are written to YUVERIF as octal numbers so that each digit consists of 3 bits This makes it easy to directly make out in the formatted ASCII file which bits are set The following description details the VOF file body for the case that the observations are read from NetCDF observation input files If the observations are read from an AOF file not all the details are exactly as described here particularly the flags Report Header For each type of report the header has the same format and consists of 15 entries in one line Part VII User s Guide 4 28 Section 8 Model Output 8 2 ASCII Output Related to the Use of Observations 172 VOF Verification Observation File Version 3 Verification period initial date and hour 2009022409 start 0 0000 end 3 0000 Set up of the reference model run used for threshold quality control QC LM grid pole 40 00 170 00 lower left corner 20 000 18 000 resolution 0 06250 0 06250 upper right 21 000 23 500 domain size 665 657 40 Initial date and time 2009022409 QC time step 720 s QC thresholds upper air vertical table 5 00 1 00 10 00 1 00 upper air constant part 0 00 500 00 0 00 0 00 upper air time factor 0 10 0 20 0 10 0 03 surface constant part 12 00 500 00 12 00 0 70 surface time factor 0 10 0 20 0 10 0 03 Number of model runs to compare with observa
337. your system 1 The ampersand amp character followed immediately by the name of the namelist group 2 A sequence of zero or more parameter value statements Part VII User s Guide 4 28 Section 6 Input Files for the COSMO Model 6 1 File for Namelist Input 48 Table 6 1 NAMELIST groups and INPUT_ files Component Description Group INPUT file Setup specifying the domain and the size of the grid LMGRID INPUT_ORG parameters for the model run RUNCTL parameters for tuning variables TUNING Dynamics parameters for the adiabatic model DYNCTL INPUT_DYN Physics parameters for the diabatic model PHYCTL INPUT_PHY Diagnostics parameters for the diagnostic calculations DIACTL INPUT_DIA parameters for the satellite images SATCTL Assimilation controlling the data assimilation NUDGING INPUT_ASS Additionals parameters for the initialization INICTL INPUT_INI parameters for controlling the EPS mode EPSCTL INPUT_EPS Input controlling the I O TOCTL INPUT IO Output parameters for using DWD s database system DATABASE controlling the grib input GRIBIN controlling the grib output GRIBOUT 3 to terminate the NAMELIST group Example In the following example new values are set for the parameters in the Namelist group 1mgrid amp lmgrid startlon tot 10 4 startlat tot 3 025 pollat 32 5 pollon 170 0 dlon 0 025 dlat 0 025 ie tot 72 j

COSMO Documentation Part VII: User's Guide

Contents

Download Pdf Manuals

Related Search

Related Contents