Home

JULES v2.0 User Guide - NERC Open Research Archive

1. Flag value Notes b Backward time average ending at given time Will be interpolated with time C Centred time average centred on given time Will be interpolated with time f Forward time average starting at given time Will be interpolated with time ae Instantaneous value at the given time Will be linearly interpolated with time nb Backward time average ending at given time Value will be held constant with time nc Centred time average centred on given time Value will be held constant with time nf Forward time average starting at given time Value will be held constant with time Depending upon the time interpolation flags driving data may need to be supplied for one or two times that fall before or after the times for the integration The interpolation scheme implemented in JULES for flags b c and f is a simplified version of the Sheng and Zwiers 1998 method that conserves the period means of the driving data file In order to ensure conservation of the average these flags can be used only if the data period is an even multiple of the model timestep i e if driveDataPer 2 n timestep n 1 2 3 In these cases the curve fitting process tends to produce occasional values near turning points that fall outside the range of the input values Note that for centred data flags c and nc the time of the data should be given as that at the start of the averaging period rather than the centre e g
2. Variable name Type and Notes permitted values readFile logical Switch controlling location of data TRUE read from an external file FALSE read from the run control file filename character The name of the file to be read Only used if readFile TRUE Note For many applications the best approach may be to read the parameters from the files provided with the JULES code via readFile TRUE since this removes the risk that values can be changed by an accidental edit to the run control file nnvgInFile integer The number of non vegetation surface types for which gt nnvg parameters are available in the input file charact er Name of each surface type This list must include the non vegetation surface types used in this run as defined in INIT OPTS variable nvgName see Section 6 2 Special cases soil this surface type must always be present water this is used to indicate open water such as lakes ice this is used to indicate land ice such as glaciers Each special type must be represented by not more than one type e g we cannot have two soil types Page 52 of 100 albsnc_nvg real Snow covered albedo Only used if 1 spec _albedo FALSE See HCTN30 Table 1 albsnf nvg real Snow free albedo See HCTN30 Table 1 Only used if 1 spec albedo FALSE catch nvg real Capacity for water kg m See HCTN30 p7 gs nvg real Surfa
3. 2 timeRun 2 spinTol spinTol Table 8 Description of the INIT_TIME section Variable name Typeand Notes permitted values timestep integer Timestep length seconds gt A typical timestep is 30 to 60 minutes If the timestep is too long the model becomes numerically unstable dateMainRun 1 2 integer These specify the start and end times for the timeRun 1 2 array integration Each run of JULES consists of an optional character spin up period and the main run that follows the spin 8 array up See below for more about the specification of the spin up For simplicity the same times of day are used for both the spin up and main periods The main run starts at timeRun 1 on dateMainRun 1 and ends at timeRun 2 on dateMainRun 2 Dates should be given in format yyyymmdd All dates must be gt 0 Times should be given in format hh mm ss All times are in Greenwich Mean Time UTC but see 6 3 1 dateSpin 1 2 integer The dates for the spin up period in the format array yyyymmdd Elements 1 and 2 are the start and end dates respectively The spin up phase of the integration must be over times that either immediately precede the main run In this case the spin up phase is from timeRun 1 on dateSpin 1 to timeRun 1 on dateSpin 2 where dateSpin 2 equals dateMainRun 1 Page 23 of 100 OR are the same as those for the main run In this case the spin up pha
4. conFrac real gt 0 Convective precipitation covers the fraction conFrac of the gridbox io rad type integer 1 2or3 Flag indicating what radiation fluxes are input 1 Downward fluxes of short and longwave radiation are input Normally this is the preferred option 2 Downward shortwave and net all wavelengths Page 63 of 100 downward radiation are input The modelled albedo and surface temperature are used to calculate the downward longwave flux 3 Net downward fluxes of short and longwave radiation are input The modelled albedo and surface temperature are used to calculate the downward fluxes of shortwave and longwave radiation iowindSpeed logical Switch indicating how wind data are input TRUE the wind speed is input FALSE the two components of the horizontal wind e g the southerly and westerly components are input zl uv real The height m at which the wind data are valid This gt 0 0 height is relative to the zero plane not the ground zl tg real The height m at which the temperature and humidity gt 0 0 data are valid This height is relative to the zero plane not the ground gt ASCBIN If fileFormatLatLon asc Nigam Ore jejo E nfieldFile integer Number of fields in each file nHeaderFile integer The number of headers at the start of each file see gt 0 Section 5 2 nHeaderTime integer The number of headers at
5. v1 Gridbox southerly wind component m s vlOm Gridbox southerly wind component at 10m height m s7 Page 96 of 100 Table 36 A list of output variables that have a single value at each land gridpoint Name Description canopy Gridbox canopy water content kg m cs Gridbox soil carbon kg C m CV Gridbox mean vegetation carbon kg C m depthFrozen Gridbox depth of frozen ground at surface m depthUnfrozen Gridbox depth of unfrozen ground at surface m gpp Gridbox gross primary productivity kg C m s gs Gridbox surface conductance to evaporation m s1 hfSnowMelt Gridbox snowmelt heat flux W m landIndex Index gridbox number of land points liceIndex Index gridbox number of land ice points 1itCMn Gridbox mean carbon litter kg C m 360days lwUp Gridbox surface upward LW radiation of land points W m assuming an emissivity of 1 npp Gridbox net primary productivity kg C ms radnet Surface net radiation of land points W m respP Gridbox plant respiration kg C m s resps Gridbox soil respiration kg C m s respSDrout Gridbox mean soil respiration for driving TRIFFID kg C m 360days runoff Gridbox runoff rate kg m s smcAvailTop Gridbox available moisture in top 0 0m of soil kg m smcAvailTot Gridbox available moisture in soil column kg m smcTot Gridbo
6. 360days gLeafPhenP PFT mean leaf turnover rate over phenology period 360days gstomP PFT bulk canopy stomatal conductance for water vapour m s1 gppP PFT gross primary productivity kg C m s laiP PFT leaf area index laiPhenP PFT leaf area index after phenology LitcP PFT carbon litter kg C m 360days nppDrOutP PFT mean NPP for driving TRIFFID kg C m 360days nppP PFT net primary productivity kg C m s rdcP Canopy dark respiration without soil water dependence mol CO m s respPP PFT plant respiration kg C m s respWDrOutP PFT mean wood respiration for driving TRIFFID kg C m 360days D Page 98 of 100 Table 38 A list of output variables that have a single value for each tile at each land gridpoint Name Desciption alb1T Tile land albedo waveband 1 direct beam visible alb2T Tile land albedo waveband 2 diffuse visible alb3T Tile land albedo waveband 3 direct beam NIR alb4T Tile land albedo waveband 4 diffuse visible canopyT Tile surface canopy water for snow free land tiles kg m catchT Tile surface canopy water capacity of snow free land tiles kg m ecanT se evaporation from canopy surface store for snow free land tiles kg m s eiT Tile sublimation from lying snow for land tiles kg ms esoilT Tile surface evapotranspiration from soil moisture store for snow free land tile kg m s fqwT Tile surface moisture flu
7. and is given by ntype npft nnvg In the standard setup JULES models 5 vegetation types and 4 non vegetation types npft 5 nnvg 4 However the model domain need not contain all 9 types e g the domain could consist of a single point with 100 grass The amount of each type in the domain is set in the section INIT FRAC Section 6 5 Page 19 of 100 ntiles integer The number of tiles for each gridbox npft nnvg This is the number of surface energy balances that are solved or 1 for each gridbox ntiles ntype a separate energy balance is calculated for each surface type ntiles 1 aggregate parameter values are used to solve a single energy balance per gridbox Generally ntiles ntype is preferred but ntiles 1 is used by the Met Office forecast model where computational cost must be kept to a minimum When using TRIFFID ntiles must equal nt ype 9 pftName 1 npf character Names of PFTs When JULES looks for parameter values T array for the PFTs it looks for these names nvgName 1l nnv_ character Names of non vegetation surface types When JULES looks g array for parameter values for the surface types it looks for these Must include names soil nxIn integer The number of columns of data in the input grid see further gt discussion of the grid in Section 6 4 nyIn integer The number of rows of data in the input grid gt The total number of point
8. ndim character array Name of each of the dimensions in the netCDF file name character See above under gt ASCBIN SDFName character The name of the variable as used in a SDF See discussion of SDF in Section 5 2 2 varNameFile character See above under gt ASCBIN interpFlag character See above under gt ASCBIN See Table 33 The meteorological variables required by a run of JULES are determined by the choice of flags such as ioPrecipType The variables that are listed must then match this expectation Table 25 Names of meteorological driving variables Name Description Comments lw_down Downward longwave radiation W m Used with rad _type 1 lw net Net downward longwave radiation W m Used with rad type 2 sw_down Downward shortwave radiation W m Used with rad type 1 sw_net Net downward shortwave radiation W m Used with rad type 2 rad net Net all wavelength downward radiation W m Used with rad_type 3 precip Precipitation rate kg m s Used with precipType 1 precipCR Convective rainfall rate kg m s Used with precipType 3 and 4 precipCs Convective snowfall rate kg m s Used with precipType 4 precipLR Large scale rainfall rate kg m s Used with precipType 3 and 4 precipLs Large scale snowfall rate kg m s Used with precipType 3 nd 4 precipTR Rainfall rate kg m s Use
9. start the iterative calculation of gs for the first timestep only frac land pts The fraction of land area Always required but can be read ntype of each gridbox that is at INIT FRAC This variable has covered by each surface to be set either in this section of type the run control file or in the section tagged INIT FRAC see Section 6 5 If TRIFFID is being used see Section 6 2 frac must be set here as part of the initial condition e g from a model dump If TRIFFID is not being used i e the fraction of each type is static the fraction may be set here as part of the initial condition or in INIT FRAC The switch readFracIc described in that section is important in this case rgrain land_pts Snow grain size um on Only used if 1 spec albedo ntiles each tile TRUE 7 J snow _grnd land_pts The amount of snow on Only used if can _model 4 A ntiles the ground beneath the value should be given for all tiles canopy kg m on each but it is only updated for tiles that tile refer to PFTs that have snowCanPFT 1 see Section 6 8 lai land_pts Leaf area index of each Only set here if TRIFFID is npft PFT switched on in INIT OPTS see Section 6 2 If TRIFFID is not used LAI is not a prognostic variable and it is initialised in either INIT VEG PFT or INIT VEG VARY canht_ft land_pts Height m ofeach PFT Only set here if TRIFFID is used npft see comments for lai
10. zl_tq gt ASCBIN 1 nfieldDriveFile 0 0 0 ndriveHeaderFile ndriveHeaderTime ndriveHeaderField T noNewLineDrive gt VARS pstar 1 nf psfc name field flag name t 1 nf temp q 1 nf humid u 1 nf uwind v 1 nf vwind 1lw_down 1 nf long sw_down 1 nf solar precipTR 1 nf liqp precipTS 1 nf solp gt ENDVARS ndriveFileTime 162 indicates the number of files for each variable readList TRUE indicates that the names and times of each file are read from the file file list txt The first few lines of this file are shown in Figure 5 List of meteorological data files Columns are file name start date yyyymmdd start time hh mm ss met _data svv_data vv198207 dat 19820701 03 00 00 met_data tvv_data vv198208 dat 19820801 03 00 00 met_data vv_data vv198209 dat 19820901 03 00 00 Er A Gx rest of file not shown Figure 5 Example list of driving data files using file name templating The presence of Svv in each file name shows that we are using variable name templating see Section 6 18 The dates show that we in fact have monthly files but note that we cannot use time templating for these files because the start time of 03H does not conform to the requirements Page 67 of 100 described in Table 31 Furthermore files for each variable are stored in separate directories For example skipping ahead to after gt VARS we see that the humidity variable is held
11. 97 Table 38 A list of output variables that have a single value for each tile at each land gridpoint eee eeeeseeeeeeeerees 98 Table 39 A list of output variables that have a single value for each tile type at each land gridpoint eee 98 Table 40 A list of output variables that have a single value for each soil level at each land gridpoint 0 eee 99
12. E 5 W 3 W To do this we use the following entries in the run control file Only the lines in bold are relevant and irrelevant sections have been omitted gt INIT GRID F F pointsList landOnly TT subArea subAreaLatLon 355 0 357 0 55 0 57 0 xcoord 1 2 ycoord 1 2 1 npoints F readFilePoints points dat fileNamePoints gt INIT LAND T readFileLand bin fileFormatLand grid gra fileNameLand Page 34 of 100 gt ASCBIN 0 0 nheaderFileLand nheaderFieldLand 1 fieldLand gt INIT LATLON T regLatLon 55 5 353 5 regLatl regLonl 1 0 1 0 regDlat regDlon F readFileLatLon prn fileFormatLatLon grid gra fileNameLatLon pointsList F indicates that the model grid will be determined by the land fraction field and also latitude and longitude in this case landOn1ly F indicates that both sea and land points will be selected subArea T indicates a sub section of the input grid is requested subAreaLatLon T indicates that the sub section will be specified by a range of latitude and longitude shown by xcoord and ycoord to be 355 E to 357 E 55 0 N to 57 0 N note we could enter the longitude range as 5 to 3 npoints is irrelevant because the number of points will be determined as the number of points the model finds within the given lat lon range readFileLand T indicates that the land fraction field is read from the binary fi
13. JULES variables that define the model state prognostic variables ccccssessecsseceeeeseeeseeeeeeseeeeeeeseenseeerees 70 Table 28 Description of options in the INIT OUT section 0 0 cece ceeeseeecscscsseseseeecececssesesesesecsesssessesesecsesesenecsesaenes 75 Table 29 Description of options for each output profile eeececcecceseceseceseceeecseecseeeaeeeaeeeeseesseenseeseceaeeaeceeeeeeneeses 78 Table 30 Valid substitution strings for substitution templates cccescescesseesseeseceecseeeseeeeeeceeeeeeeseceseeeaeeaeeaeessecseeens 85 Table 31 Requirements for the time of first data in time templated files 0 00 ee eecseeeecneeeeceseeeeeseceeesecnecaeeseeneeeeeeaees 86 Table 32 Examples of the use of file name templating ceececceeseceseceseeseeceecseecseecaeeeaeeeaeeseceeesseeeseeeaeenaeenaeenaecneeens 86 Table 33 Time interpolation Matses cok a E E E RH ae eee a ee 88 Table 34 Key procedures for reading input data cccecccsccsseesseeseeeseeeeceseeeecnsecnsecaecsaeesecsaecsaecnaecaaecaaecaeeeaeeeneenseereeas 91 Table 35 A list of output variables that have a single value at each gridpoint 0 0 0 0 cccccceesceeseeesceeeceeeeeceeeceeesseeneeees 95 Table 36 A list of output variables that have a single value at each land gridpoint cceecceeseesceeeeeeeceseeeeceeeeseeseenes 96 Table 37 A list of output variables that have a single value for each PFT at each land gridpoint eee eeeeeeeenees
14. Only used if readFile TRUE nheaderFile integer The number of headers at the start of the file not used if gt 0 readFile FALSE See Section 5 2 nheaderField integer The number of headers before each field not used if gt 0 readFile FALSE See Section 5 2 name character The name of a soil hydraulic variable These names must be chosen from the list in Table 15 below At present all 9 variables must be provided fieldNumber integer The field number of the first level of data in the input file that is to be used for a variable See discussion of fields in Section 5 1 Note that if readFile FALSE this is interpreted slightly differently it is the variable number not field number nsoilDim integer Data in SDF is held on array of nsoilDim dimensions gt dimName 1 ns character array Names of dimensions in file oilDim name character See under gt ASCBIN above SDFname character The name of a variable containing data as it appears in the SDF multH real Multiplier for matric suction at saturation sathh to convert to absolute suction m The input values of sathh are multiplied by multH The suction at saturation is generally less than zero but JULES uses the absolute value hence it is often necessary to set multH 1 Only used if readFile TRUE multCon real Multiplier for saturated hydraulic conductivity satcon to convert to units of kg m s which is equiv
15. W m landAlbedol Gridbox albedo for waveband 1 direct beam visible landAlbedo2 Gridbox albedo for waveband 2 diffuse visible landAlbedo3 Gridbox albedo for waveband 3 direct beam NIR landAlbedo4 Gridbox albedo for waveband 4 diffuse NIR latentHeat Gridbox surface latent heat flux W m latitude Gridbox latitude longitude Gridbox longitude lsRain Gridbox large scale rainfall kg m s lsSnow Gridbox large scale snowfall kg m s1 lwDown Gridbox surface downward LW radiation W m precip Gridbox precipitation rate kg ms pstar Gridbox surface pressure Pa qipsm Gridbox specific humidity at 1 5m height kg kg qw1 Gridbox specific humidity total water content kg kg rain Gridbox rainfall rate kg m s snomltSurfHtf Gridbox heat flux used for surface melting of snow W m snowfall Gridbox snowfall rate kg m s1 snowMass Gridbox snowmass kg m surfHtFlux Gridbox net downward heat flux at surface over land and sea ice fraction of gridbox W m swDown Gridbox surface downward SW radiation W m tlp5m Gridbox temperature at 1 5m height K taux1 Gridbox westerly component of surface wind stress N m tauy1 Gridbox southerly component of surface wind stress N m tll Gridbox ice liquid water temperature K tstar Gridbox surface temperature K ul Gridbox westerly wind component m s ulOm Gridbox westerly wind component at 10 m height m s
16. a file or data period can be described by more than one time unit the longer unit is used For example a period of 60 minutes is described as 1 hour For example consider daily data held in one file per month This gives dataPerUnits day and filePerUnits 1 month Table 31 shows that the first data in each file must represent the 1 of the month as might be expected A file that started with data for the 2 of the month cannot be used with time templating even if a particular run does not require the data at that time Table 31 Requirements for the time of first data in time templated files dataPerUnits 1 year 1 month days hours minutes filePerUnits 1 year none Jan OlJan 00H OlJan 00H OlJan 1 month none 1 of 00H 1 of OOH 1 of month month month days none 00H 00H hours none 00H minutes none 6 18 2 Variable name templating Variable name templating is so called because it is expected to be used when related variables are stored in separate files with file names that are identical apart from a section that indicates what variable is in each file For example variable 1 could be in filel dat while variable 2 is in file2 dat Examples of the use of this type of templating are given in the next section If using variable name templating with non SDF formats the layout of each file must be similar the number of headers and the numb
17. all the output points This option requires that the model grid is rectilinear or is a subset of such a grid 4 the output grid will be the same as the input grid Depending upon the shapes of the input and model grids it may be possible to produce the same output grid via different combinations of the values of pointsFlag Similarly certain combinations will be less useful for particular grids outAreaLatLon logical Switch indicating how to interpret the coordinates outRangeX and Under some circumstances out DateStart 0 will also output the initial state of the model These circumstances are that the period of the output equals the timestep i e information for every timestep and that all output goes to a single file out FilePer 9 The timestamp information included with the output allows the user to determine whether this initial state has been output Page 80 of 100 outRangeY Only used if pointsFlag 1 1 TRUE co ordinates are longitude and latitude FALSE co ordinates are x and y indices column and row numbers outRangexX 1 2 real array x coordinates of the sub area to be output Depending on outAreaLatLon these are longitudes in range 180 to 360 or column numbers Only used if pointsFlag 1 1l Column numbers are those in the INPUT grid If values are column numbers the code uses the nearest integer to the input value outRangeyY 1 2 real array As outRangexX expect
18. but this is not important nx ny values will be read however they are presented Page 15 of 100 Table 4 Part of an example ASCII file that could be read by JULES There are 2 variabl Time level 1 Variable A 12 0 T546 Lie lt 1 5 0 2359 53 2 Variable B level 1 22 0 25 6 Tk I 50 22 29 23 Variable B level 2 32 0 11 6 TZL 9 1 A269 43 Time level 2 Variable A 942 owes milk 11 5 23 9 8 3 Variable B level 1 rest of file not shown This file contains example data es the 2nd with 2 levels 2 7 6 Ist file header 2nd file header header for time 1 header for lst field data for A at t 1 header for 2nd field data for B at t 1 1st level header for 3rd field data for B at t 1 2nd level header for time 2 header for 1st field data for A at t 2 header for 2nd field 5 2 2 netCDF files If fileFormat nc the following information is read from a section that starts with the tag gt NC As noted earlier support for netCDF input is currently rather limited in JULES Table 5 Format of options used to specify the reading of netCDF format files Option name Type Notes nDim integer The number of dimensions used for a variable in the netCDF file This was set at the time the netCDF file was created dimName 1 nDi character The names of the netCDF dimensions for a variable m SDFname character The name u
19. character new or replace The status used when opening files This is the value given to the FORTRAN status argument of an OPEN statement e g open 1 status new Acceptable values are new file must not already exist If the code tries to create a file with the same name as an existing file the run will terminate replace If the file exists delete it and replace with a new version yrevoOut logical TRUE reverse the order of the rows in the output so that these are written in North to South order FALSE use the default South to North order with the southernmost row of data being the first in the file zrevOut logical Switch indicating if soil layer data are to be output in reverse order of levels compared with JULES s default TRUE reverse the order of the vertical levels in the output so that these are written in bottom to top order i e bottom layer first FALSE use the default top to bottom order i e top layer first This flag only applies to variables on the soil levels numMonth logical Switch controlling the date format used in file names TRUE months are represented by the numbers to 12 FALSE months are represented by 3 character strings jan feb mar useTemplate logical This relates to GrADS gridded files generated by outFormat bin Switch to activate the writing of template c
20. e g land points only soil layers If the desired diagnostic Page 92 of 100 does not fit any of the existing types the user may have to closely study the code to work out how to add a new type and or contact the JULES developers Finally code to load the values into the output space has to be added to subroutine loadout in module OUTPUT_MOD This code may have to calculate the diagnostic using other variables 7 2 2 Output of new variables Diagnostics that require variables that the user had added or that must be calculated in a section of the model code other than the output routines are more complex to add to JULES In such a case it may be easiest to declare a new variable in a FORTRAN module and to use this variable to hold the values of the diagnostic Space for the new variable will likely have to be allocated and the tidiest way to do this would be in the subroutine allocate arrays which is called at various points during initialisation The variable can then be accessed by the output procedures and the steps outlined in case above should be followed A more sophisticated scheme which only allocated space for a diagnostic if it was required and loaded the value from any subroutine avoiding the need to hold the variable in a module or pass it through the code has been implemented in some versions of JULES but is not available in the release versions because it was not compatible for use in the Unified Model If you ar
21. if 1 spec _albedo FALSE Must not be zero maskd real Used in exponent of equation weighting snow covered and snowfree albedo This is 0 2 in HCTN30 Eq 5 snowLoadLAI real Ratio of maximum canopy snow load to leaf area index kg m This is 4 4 in JULES1 Only used if can mode1 4 snowlIntercep tFact real Constant in relationship between mass of intercepted snow and snowfall rate This is 0 7 in JULES1 Only used if can _model 4 snowUnloadFa C real Constant in relationship between canopy snow unloading and canopy snow melt rate This is 0 4 in JULES1 Only used if can model 4 Page 55 of 100 6 12 INIT _TRIF Parameters for the TRIFFID model This section is used to read PFT parameters hat are only needed by the dynamic vegetation model TRIFFID Values are not read if TRIFFID is not selected TRIFFID also uses many other PFT specific variables that are also used in other parts of JULES and are read in Section 6 8 above gt INIT TRIF readFile fileName nnvginFile gt DATA dataVarl 1 dataVarl 2 dataVar2 1 dataVar2 2 a Be data values dataVarl nPft dataVar2 nPft Table 21 Description of options in the INIT_TRIF section Variable name Type and Notes permitted values readFile logical Switch controlling location of data TRUE read from an external file FALSE read from the run control file filename characte T
22. if several profiles are used they could be given names such as p1 p2 outPer integer The period for output seconds This must be a multiple of the timestep length except for the special cases lt 0 given below It must not exceed 30 days 2592000 seconds except for the special cases Special cases 0 generate output every timestep 1 monthly period 2 annual period calendar years outFilePer integer The period for output files seconds i e the time interval within which all output goes to the same file This must not exceed 30 days 2592000 seconds except for the special cases given below The file period must be consistent with the output period e g we can t have daily files for monthly output Output may be generated for only part of a run see outDateStart below and out FilePer controls how the data are stored during that part of the run when the output is active Special cases 0 output is every timestep and a new file is created every timestep 1 monthly files all output for a month goes to the same file 2 annual files calendar years 7 all output goes to one file but each cycle of spin up creates a separate file 8 all output goes to one file but all output during spin up goes to a separate file 9 all output for all times from this profile goes to one file outSamPer integer The sampling period seconds for time averages and accumulati
23. in files such as met _data humid_ data humid198207 dat while the surface pressure is held in the likes of met_data psfc_data psfcl198207 dat The ioPrecipType value of 2 shows that we read in two components of precipitation total solid and total liquid The liquid is considered to be convective precipitation when the temperature is above tForConv which here has a value of 298 2 K nfieldDriveFile 1 shows that each data file contains a single field which is consistent with the field number shown for each variable all 1 Page 68 of 100 6 16 INIT_IC Specification of the initial state The values of all prognostic variables must be set at the start of a run This initial state or initial condition can be read from a dump from an earlier run of the model or may be read from any other file Another option is to prescribe a simple or idealised initial state and this may be done via the run control file It is also possible to set some fields using values from a file e g a dump but to set others using idealised values from the run control file that is effectively to override the values in the external file gt INIT IC readFile fileFormat quoted fileName quoted Zrev gt ASCBIN nheaderFile nheaderField gt VARS varName 1 varFlag 1 constVal 1 varName 2 varFlag 2 constVal 2 Repeat for each variable gt ENDVARS gt NC nDim dimNam
24. lw nf sw nf liqp nf solp ndriveFileTime l1 indicates that all data are in one file readList FALSE indicates that the name of the file is read from the run control file not from a separate file noNewLineDrive TRUE shows that each variable is not on a new line in fact all variables are on one line The entries following gt VAR indicate where each variable lies in the input file Note that we can skip the unrequired time and obs1 fields in Figure 4 Time solar 1 3 3 2 898 5 3 142 3 long 187 8 0 0 185 8 0 0 186 4 data for later times rain 0 0 snow temp obs1 wind press humid 0 0 25 9 L0 83 0 3 610 102400 5 1 351E 03 0 0 259 45 24 1 3 140 102401 9 1 357E 03 0 0 259 85 56 9 2 890 102401 0 1 369E 032 Figure 4 Lines of an example file of meteorological driving data in ASCII format Page 66 of 100 Example 2 Driving data from binary files one variable per file The relevant entries in the run control file are shown below Only the lines in bold are relevant and irrelevant sections have been omitted gt INIT DRIVE 3600 0 driveDataPer 162 9 ndriveFileTime driveFilePer T readList file list txt fileName 19820701 03 00 00 driveFileDate 1 driveFileTime 1 bin driveFormat 2 F i0PrecipType 1_point_data Z2I75 0 tForSnow 298 2 0 3 tForConv conFrac 1 F io_rad_type ioWindSpeed F ioTotalRunoff 10 0 10 0 zl_uv
25. run when outDateStart 2 will output the initial state Thus the only way in which the initial state can be output is to have an output profile with outDateStart 2 All output at later times then has to be generated via another output profile This is a slight oversimplification see footnote 6 outTimeStart character Time of day in format hh mm ss at which output begins Not used 8 if outDateStart is one of the special cases outDateEnd integer Date on which output ends Not used if outDateStart is one of the special cases outTimeEnd character Time of day at which output ends Not used if outDateStart is 8 one of the special cases pointsFlag 1 integer Flag indicating how the points to be output are selected 0 1 2 0 all points in the model grid will be output 1 points in a rectangular subsection will be output 2 the points to be output will be listed individually pointsFlag 2 integer Flag indicating how the locations in the output grid of output points 0 to 4 will be calculated 0 the output grid will be the model grid 1 the output grid will be the rectangular subsection specified via pointsFlag 1 1 This option can only be used in conjunction with pointsFlag 1 1 2 the location of each output point will be listed individually This option can only be used in conjunction with pointsFlag 1 2 3 the output grid will be the smallest rectangle that contains
26. run WE and the columns SN hereafter referred to as an equal angle grid can be set up via a simple option All other grids including a vector of points require the latitude and longitude of all points to be input If regLatLon TRUE the input data must be presented in the default JULES order starting bottom left at regLat1 regLon1 and proceeding row wise If reghatLon FALSE the input data need not be in order of lat lon coordinates each point in the grid will use the lat lon read in for that point 6 4 4 Examples of grid description The latitude and longitude of the grid must be specified for all runs For many model runs the location of the grid is important since it controls important factors such as the angle of the sun Other more idealised runs might not need this information but in this case the location may still be required so that the model output can be correctly mapped If the location is not needed for either purpose the user should enter an arbitrary location e g 0 N 0 E Grid example 1 A single point run This covers the simplest case the input contains a single point We assume that nxIn 1 and nyIn 1 see Section 6 2 All values are obtained from the run control file no other file is involved Only the lines in bold are relevant and irrelevant sections have been omitted gt INIT GRID T F pointsList landOnly F F subArea subAreaLatLon Ly Zp Sip 4 l xcoord 1 2 ycoord 1 2 1 npo
27. satus carbon and qoiia parwinides ented ee INIT Ic Serre of the initial state E oid E E A 6 16 1 Examples of specification of initial Ea E A eo 6 17 INIT OUT Specification of output from the asad PEE E E E cease gt 6 17 1 INIT_OUT general values am maanantaina iini T 6 17 2 NEWPROF details of each ne profile RTE 6 17 3 Compression of the output grid E EE PE EOT ESEA EEIE ET EEA E E E 6 17 4 An example oe and mapp Bi E E O E E E 6 17 5 Notes on output aa EEE E E ead ee Page 4 of 100 DIe Fie omie PEA soa aS C ie ea E a iMate 85 ec Vondble name Tei Ie 25a a 619 Gee Ge aI I eI COI aienea EA r R N i 87 6 20 Example rin control Fles errre risien eenn S ATENEA EEEE a 90 P E E anni E E T EE SE E E T AAE e a E E E eee ee 91 72 How to implement new diagnostics for 0Uiputesressnn a T21 Omp ol existing variable ssnin a L Fe new vara D e neaei EEE ENEA EEA EEEE 92 8 Known limitations of and bugs in the code sisccccscicessvcccscenciscerectnanseiencrcaessienateibuerineeninnman JS o vales conte eee G 94 ile aa a a 1 fi Page 5 of 100 1 Introduction and what s new The Joint UK Land Environment Simulator JULES is a computer model that simulates many soil and vegetation processes It is based on the Met Office Surface Exchange System MOSES the land surface model used in the Unified Model of th
28. the 3 hour average over 06H to 09H centred at 07 30H should be treated as having timestamp 06H Sheng and Zwiers 1998 An improved scheme for time dependent boundary conditions in atmospheric general circulation models Climate Dynamics 14 609 613 Page 89 of 100 Time ocooz 0300Z 0600z 09002 1200z 15002 Driving data Mle e oe o o o o JULES run eeee0_ne egeoe0e eee e880 8 t 12 3 4 5 6 amp 9 10 11 12 13 14 15 16 intial Final orri dure Time 210Z ocooz 0300Z 0s00Z osoz 1200Z Driving data Me G i i iul JULES run e e eeeeee ee eee 8 E t 0L 2 34567889S 10 11 12 13 14 15 16 intial Final onal dune Figure 7 Schematic of JULES interpolation of driving variable from a 3 hour timestep to a 45 minute timestep Simulation start time is 0000Z on an arbitrary day and end time is 1200Z Blue circles indicate driving data required to complete a JULES simulation from t 0 to t 16 See text for discussion of requirements for driving variables that are forward or backward means Page 90 of 100 6 20 Example run control files Two example run control files come bundled with the JULES source code in the top level directory point loobos example jin for a single point simulation forced with weather station data This run requires a single input file meteorological data that is also included as part of the JULES distribution in the LOOBOS directory The results of running this code
29. the fraction cover is a prognostic variable and it must be read with the initial condition readFracIC TRUE readFile logical Switch controlling location of fractional cover data Only used if readFracIC FALSE TRUE read from an external file FALSE read from the run control file fileFormat character Format of data Only used if readFile TRUE See notes in Section 5 2 Page 38 of 100 filename character Name of file containing data Only used if readFile TRUE nheaderFile integer The number of headers at the start of the file gt 0 See Section 5 2 nheaderField integer The number of headers before each field gt 0 See Section 5 2 fieldNum integer The number of the first field to be used from the input gt file this represents the first surface type See discussion of fields in Section 5 1 nFracDim integer Data in SDF is held on array of nFracDim dimensions gt fracDim 1 nFra character array Names of dimensions in file cDim varName character The name of the variable containing data frac 1 nxIn 1 real array The fractional coverage of each surface type The nyIn gt 0 0 fractions should sum to 1 this is checked by the code These values are only read if readFile F and must be located after the tag gt DATA Note that all land points must be either soil points indicated by values gt 0 of the saturated soil moisture content or land ice points i
30. the second field of data is used for this variable Blank lines between fields in an ASCII input file can cause the code to read the wrong data and should be avoided If blank lines are present between fields they should be interpreted as header lines There are restrictions on what PP files JULES can read Each field must have no trailing extra data i e header 20 must be zero It is also assumed that the data are ordered as in the JULES GrADS model outlined above so for example we do NOT have all times of field 1 then all times of field 2 so that the required data can be found without using the information contained in the field headers The headers are used to check that the size of the field and the STASH code are as expected The STASH code for each variable is currently hardwired in the code At the time of writing the PP reading code has no known bugs but it has been used much less than other options so any more obscure bugs might not have been triggered 5 2 1 1 An example ASCII input file Table 4 shows part of an example ASCII file that could be read by JULES with nheaderFile 2 nheaderTime l nheaderField 1 The size of the input grid is assumed to be nxIn 3 nyIn 2 There are 2 variables A which has a single level and B which has 2 levels giving a total of 3 fields per time Annotation after any and shown in italics would NOT be present in the actual file The data are shown on 2 lines per field
31. the start of each time see gt 0 Section 5 2 nHeaderField integer The number of headers at the start of field see Section gt 0 5 2 noNewLineDrive logical Switch describing format of an ASCII data file TRUE variables are arranged across one or more lines and each variable does not necessarily start a new line This option should be used if all the driving data for each time are one line of the input file although it can also be used if the data are continued onto subsequent lines FALSE each variable starts on a new line Only used if there is only one point in the input grid and hence only one point in the model grid and driving data are in ASCII files name character The name of the variable This is used to identify the variable in the code and is set in the code Acceptable values are shown in Table 25 These must be entered exactly as listed in the table and are case sensitive fieldNumber integer The field number in the file that holds data for this gt variable See discussion of fields in Section 5 interpFlag character Flag indicating how variable is to be interpolated in time See Table 33 varNameFile character The substitution string used in the names of files that contain this variable Only used if variable name templating is used in file names gt NC If fi leFormat nc Page 64 of 100 ndim integer Number of dimensions in the netCDF file gt dimName
32. times of the data files The first line of this file will be skipped and so can be used for comments All other lines are to be of the form filename startDate startTime where fileName may contain variable name templating see Section 6 18 startDate is in the format yyyymmdd and time is in the format hh mm ss vegFileDate integer Date of first data in vegetation file in format yyyymmdd Only used if readList FALSE otherwise read from an external file vegFileTime character Time of first data in vegetation file in format hh mm ss Only used if readList FALSE otherwise read from an external file Page 48 of 100 vegEndTime logical Flag used with vegetation file templating TRUE means that time in filename refers to the final data in the file FALSE means the time in the filename refers to the first data in the file fileFormat character Format of vegetation data files See Section 5 2 The following are read only if readFile TRUE Only values for the appropriate file format are read gt ASCBIN If fileFormat asc Yorn ie pos nfieldFile integer Number of fields in each file nHeaderFile integer The number of headers at the start of each file see gt 0 Section 5 2 nHeaderTime integer The number of headers at the start of each time see gt 0 Section 5 2 nHeaderField integer The number of
33. 86400 vegDataPer vegUpdatePer 1 1 nvegFileTime vegFilePer T vegClim F readList lai_monthly dat fileName 20120115 00 00 00 wvegFileDate 1 vegFileTime 1 asc fileFormat gt ASCBIN 6 nfieldFile Page 50 of 100 1 0 nheaderFile nheaderField T noNewLineVeg lai t 2 i notused name flag field interp nameFile nvegVar 1 indicates that we only want to vary one vegetation characteristic vegFileDate 20120115 but since vegClim T the year is discarded effectively leaving 0115 15 January meaning that each time of data is valid on the 15 of the month nfieldFile 6 because we have data for each of 5 PFTs plus there is a timestamp variable that will not be used see Figure 3 The final line shows that we want to vary LAI as a function of time and PFT only The LAI data start with field 2 The T and vegUpdatePer 8 6400 indicate that the monthly data will be interpolated between the monthly values and updated once every 86400s once a day Page 51 of 100 6 10 INIT_NONVEG Parameters for non vegetation surface types gt INI T NONVE G readFile fileName nnvgInFile gt DATA data values dataVarl 1 dataVarl1 2 dataVar2 1 dataVar2 2 nFile nFile dataVarl nnvgI dataVar2 nnvgI Table 19 Description of options in the INIT_NONVEG section typeName
34. E 6 W 4 W To do this we use the following entries in the run control file Only the lines in bold are relevant and irrelevant sections have been omitted gt INIT GRID F T pointsList landOnly TT subArea subAreaLatLon 6 0 4 0 55 0 57 0 xcoord 1 2 ycoord 1 2 1 npoints F readFilePoints points dat fileNamePoints gt INIT LAND T readFileLand bin fileFormatLand grid gra fileNameLand pointsList F indicates that the model grid will be determined by the land fraction field and also latitude and longitude in this case Page 36 of 100 landOn1ly T indicates that only land points will be selected subArea T indicates a sub section of the input grid is requested subAreaLatLon T indicates that the sub section will be specified by a range of latitude and longitude shown by xcoord and ycoord to be 6 W to 4 E 55 N to 57 N npoints is irrelevant because the number of points will be determined as the number of land points the model finds within the given lat lon range readFileLand T indicates that the land fraction field is read from the binary file called grid gra which has no headers and contains land fraction as the first field With this information JULES determines that there are 4 gridpoints within the given lat lon range but only 3 are land As the 3 land points do not form a rectangle the model grid is a vector of 3 points As we are only modelling land po
35. Page 1 of 100 Joint UK Land Environment Simulator JULES Version 2 0 User Manual Page 2 of 100 Last revised DBC 03 Sept 2007 Page 3 of 100 i oat mid VOT Se ccc L ere o ULE aca nee eee eee 3 Building and running J ULES 3 1 The make utility EE E siete obs tenn E A S ES aan E EE E E 3 2 The netCDF interface e library ner ads obinia seas 3 3 ae baile lin 4 Overview ofthe JULES tah E AE EE EE eee 6 The JULES paler file 6 1 Introduction oe ae _OPTS 6 3 2 Examples of dates and times 6 3 3 Notes on spin up asc aaa hebeclada E 6 4 Grid description ASEA dst elena EON ENE 6 4 1 INIT_GRID setting up the JO AR 27 6 4 2 INIT LAND Land fraction m 6 4 3 INIT_LATLON Latitude and barik 6 4 4 Examples of grid description AEREE 6 5 INIT_FRAC Fractional coverage af land sut ice pes 65 1 fuse Reading frac from the run control file 6 6 INIT_DZSOIL Soil layer depths ai TT 6 7 INIT_SOIL Soil hydraulic and bel e Ss 5 2 6 8 INIT VEG PFT Time invariant parameters for plant functional woes i TA E 6 9 INIT_VEG_ VARY Time space varying parameters for plant functional wpe E Lo 6 9 1 Examples of INIT VEG VARY pees 6 10 INIT _NONVEG Parameters for notet surface types 6 11 INIT SNOW Snow parameters i 6 12 INIT_TRIF Parameters for the TRIFFID podel 6 13 INIT_AGRIC Fractional coverage by E PEE EEE 6 14 INIT_MISC Moe lan
36. TRUE may all be used to represent true In the sections beow the sizes of certain arrays are indicated using brackets e g myArray 1 20 requires values for the 20 elements numbered 1 to 20 Although a spatial field can be read from the run control file in practice this becomes unwieldy for large grids and most spatial fields are likely to be stored in separate files the names of which can be listed in the run control file In the following sections the first column lists the variables that are to be read from a line and subsequent columns give the type and a brief description of each variable The variable names given are generally those used to declare the corresponding FORTRAN variables except where the code uses temporary workspace and a more meaningful variable name is given in this documentation Page 18 of 100 6 2 INIT_OPTS General model options gt INIT OPTS pft nnvg tiles ftName 1 npft n n pf nvgname 1 nnvg nxiIn nyIn sm levels can _ model can rad mod ilayers l cosz l spec albedo L phenol 1 triffid 1 tif eq yrevin echo print step Table 7 Description of options in INIT_OPTS section Variable name Type and Notes permitted values np t integer The number of plant functional types to be modelled gt nnvg integer The number of non plant surface types to be modelled gt The total number of surface types to be modelled is called ntype
37. above Page 72 of 100 6 16 1 Examples of specification of initial state Example 1 A single point state from the run control file In this example we consider a run at a single point and read all data from the run control file The relevant entries in the run control file are shown below Only the lines in bold are relevant and irrelevant sections have been omitted gt INIT IC F readFile asc fileFormat quoted a0001 dump 19970105 fileName quoted F zrev gt ASCBIN 0 0 nheaderFile nheaderField gt VARS sthuf 1 0 9 varName varFlag constVal canopy 2 0 0 snow_tile 3 0 0 Note that none of these constVal tstar_ tile 4 275 0 values are used in this case t_soil 5 278 0 Instead values are listed after gt DATA cs 6 0 0 gs 7 0 0 gt ENDVARS Data fields to be read from this file should appear below here gt DATA 0 749 0 743 0 754 0 759 sthuf 9 0 0 canopy 9 0 0 snow_tile 9 276 78 tstar_tile 276 78 277 46 278 99 282 48 t_soil 12 100 cs 0 0 gs readFile FALSE indicates that all data will be read from the run control file no other file is involved In this case we use the gt ASCBIN section to describe the data The seven variables that are required to initialise this particular run are then listed The second entry in each line gives the position in the input data for each field Since all the data are to be read from the run control file which i
38. al water content and uses the soil temperature t_soil to partition between the phases Either sthuf or its components sthu and sthf are always required sthu land pts sm levels Unfrozen soil wetness for each soil layer This is the mass of unfrozen water expressed as a fraction of the water content at saturation See notes for sthf above Either sthuf or its components sthu and sthf are always required canopy land pts ntiles Amount of intercepted water that is held on each tile kg m Always required snow tile land pts ntiles Amount of snow on the vegetation canopy of each tile kg m Always required Note that if snow is allowed below the canopy can model 4 selected in INIT_OPTS see Section 6 2 the total amount of snow held on a vegetation tile is the sum of snow tile and snow grnqa see below For can model 4 Page 71 of 100 snow tile is the total snow on the tile tstar tile land_pts Temperature of each tile Always required ntiles K This is the surface or skin temperature t_ soil land pts Temperature of each soil Always required sm _ levels layer K cs land pts Soil carbon kg m Always required This is a prognostic variable only if TRIFFID is selected gs land_pts Stomatal conductance for Always required This is used to water vapour m s
39. alent to mm s assumed a density of 1000 kg m The input values of satcon are multiplied by multCon Conductivity is often expressed in terms of m s which would require Page 42 of 100 multCon 1000 Only used if readFile TRUE Table 15 List of soil parameters Names must be entered exactly as specified here must be in lower case Name Description albsoil Soil albedo A single averaged waveband is used b Exponent in soil hydraulic characteristics heap Dry heat capacity J m K hcon Dry thermal conductivity W m K satcon Hydraulic conductivity at saturation kg m s sathh If 1 vg_soil TRUE using van Genuchten model sathh 1 a where a is a parameter of the van Genuchten model If 1 vg _soil FALSE using Clapp and Hornberger model sathh is the absolute value of the soil matric suction at saturation m The suction at saturation is generally less than zero but JULES uses the absolute value sm crit Volumetric soil moisture content at the critical point m water per m soil The critical point is that at which soil moisture stress starts to restrict transpiration sm sat Volumetric soil moisture content at saturation m water per m soil Note that this field is used to distinguish between soil points and land ice points sm_sat gt 0 indicates a soil point sm wilt Volumetric soil moisture content at the wilting point m water per m soil T
40. ar patterns JULES can use a substitution template rather than requiring a potentially long list of file names Templating comes in two forms time templating and variable name templating which can be used separately or together Valid substitution strings are listed in Table 30 These are 3 character strings starting with Note that any file name that contains is assumed to use templating Table 30 Valid substitution strings for substitution templates Substitution string Description Time templating SLC 1 character representation of decade Met Office files Sy4 4 digit year Sy2 2 digit year SyYC 1 character representation of year Met Office files Sm2 2 digit month sml l or 2 digit month smc 3 character month abbreviation smm 1 character representation of month Met Office files d2 2 digit day of month Sal 1 or 2 digit day of month Sdm 1 character representation of day of month Met Office files Sh2 2 digit hour of day Shi 1 or 2 digit hour of day She 1 character representation of hour of day Met Office files Sn2 2 digit minute leading zero if needed Variable name templating SVV A character variable 6 18 1 Time templating Information about the time of each file is contained in the file name Valid substitution strings are listed in Table 30 and examples of the use of time templating are given in Table 32 7 JULES tem
41. are also provided in the same directory so the user can check that their installation of JULES produces results that are acceptably close to those of this standard run grid_gswp2 example jin for a gridded domain simulation forced with GSWP2 weather data This run requires a large amount of input data that is not distributed with JULES and merely serves as an example of a run control file for a gridded domain Page 91 of 100 7 Aspects of the code 7 1 Low level i o code In the course of adding to JULES a user may well want to read new variables into the model Most of the input output of spatial fields is handled by subroutines provided by the module READWRITE MOD Particularly important procedures that deal with input are summarised in Table 34 To use this code to read in a new variable the appropriate procedure should be identified based on the type of variable that is to be read in For example to read a field that is only defined on land points a call to readVar2dComp is appropriate All these procedures require arguments that define the mapping between the input grid and the model grid Note that the choice of procedure is governed solely by the type of variable and is not affected by the shape of the input grid The correct use of these procedures and the arguments required can be learned by studying the exiting code Table 34 Key procedures for reading input data Name Summary readVar2d Reads a variable that is de
42. arning is raised if an attempt is made to rewrite it 3 Driving data such as meteorological or vegetation data may not be correctly represented in output at the start of the first timestep of the run i e time 0 depending upon the frequency of data and any temporal interpolation The problem arises because the initial output is generated before the procedures that update the driving data are called Under some circumstances the driving data will already have been updated during the initialisation and so the output will be correct In other cases the initial output will have nonsense values such as zero for the driving data 4 The code that generates output contains many options and has to deal with a variety of possibilities in terms of output frequency run dates spin up and the likes Until the code has been thoroughly tested by the user community early versions of JULES are quite likely to contain bugs particularly in the output code If a user finds an error with the output the Page 85 of 100 bug should be reported but in the meanwhile JULES will hopefully run correctly if simpler output is requested Two simplifying options that may not always be practicable for the user are to request snapshot diagnostics rather than time averages in cases of extreme difficulty these snapshots should be every timestep and to send all output to a single file 6 18 File name templating If the names of input files follow particul
43. averages JULES can also be run at a point with inputs that are taken to represent conditions at that point this configuration might be used when field measurements of meteorological conditions are available Page 7 of 100 3 Building and running JULES Building a JULES executable requires two pieces of software e a Fortran 90 compiler e aversion of the make utility It may also be desirable but not essential to have available the following software e the Fortran 90 netCDF interface library 3 1 The make utility The Makefile supplied with the JULES code should be compliant with most versions of make but is only guaranteed to work with GNU Make also known as gmake which is available on most Linux and UNIX systems and also for Windows Once the Makefile is set up for the user s system JULES is built simply by entering make at the command prompt while in the directory containing the Makefile This will compile all of the JULES source code make a library libjules a and finally link the compiled source to create and executable file with a default name of jules exe To remove all the files created during the build process enter make clean at the command prompt The make utility uses architecture and compiler specific variables that must be set by the user to the appropriate values for their system These values may be set in the files Makefile common mk and Makefile comp The user should not have to ed
44. calculated by the model outGridNy integer As outGridNx but number of rows Flag indicating type of processing Acceptable values are S Instantaneous or snapshot value M Time mean value A Accumulation over time For time averaged variables the period over which each time average is calculated is given by outPer For time accumulation variables out Per gives the period for output of an updated accumulation 1 e how often the value if reported For both time averages and accumulations the sampling frequency is set via outSamPer NB A time accumulation is initialised at the start of a run actually at the start of each section of a run so that it is reinitialised after any spin up is completed see Section 6 3 3 and thereafter accumulates until the end of the run actually to the end of each section of a run This may mean that accuracy is lost particularly towards the end of long runs if small increments are added to an already large sum name character The name of an output variable one word This is the internal name as used in the model code A list of available variables is provided in Section 9 This list was correct at the Page 82 of 100 time of writing but the most reliable way to determine exactly which variables are available for a particular version of JULES is to look at the variables listed in the subroutine init out _varlist and which can be echoed to screen at the star
45. can be read by the code for example a single line with 10 values or 2 lines with 5 values each Repeated numerical values can often be specified using the notation e g 100 values of 1 0 can be entered as 100 1 0 which can sometimes be useful in specifying a large but constant field Tags are used to indicate the start of each section and allow the code to skip directly to this point ignoring any intermediate lines Each tag is of the form gt SECTION NAME and must be included exactly as in the example run control files using capital letters and with no space before or after the initial gt These section tags are listed in Table 6 Table 6 Sections in a JULES control file N Described in manual Section name Description section INIT OPTS General model options 6 2 INIT TIME Start and end times for simulation timestep 6 3 lengths spin up INIT GRID Set up the model grid 6 4 INIT LAND INIT LATLON INIT FRAC Set gridbox tile fractional coverage options 6 5 INIT DZSOIL Set soil vertical level options 6 6 INIT SOTL Set model soil parameters 6 7 INIT VEG PFT Set uniform parameters for vegetation tiles 6 8 INIT VEG VARY Set parameters for vegetation tiles that vary in 6 9 Page 17 of 100 INIT VEG VARY2 either space o
46. ce conductance m s7 See HCTN30 p7 Soil conductance is modified by soil moisture according to HCTN30 Eq 35 infil_nvg real Infiltration enhancement factor The maximum infiltration rate defined by the soil parameters for the whole gridbox may be modified for each tile to account for tile dependent factors See HCTN30 p14 z0 nvg real Roughness length for momentum m See HCTN30 Table 4 ZOh_zOm real Ratio of the roughness length for heat to the roughness length for momentum This is generally assumed to be 0 1 See HCTN30 p6 Note that this is the ratio of the roughness length for heat to that for momentum It does not alter the roughness length for momentum which is given by z0 nvg above ch nvg real Heat capacity of this surface type J K m Used only if can model is3 or 4 See INIT OPTS Section 6 2 vf _nvg real Fractional coverage of non vegetation canopy Typically O lt vf_nv set to 0 0 but value of 1 0 used if tile should have a heat g lt l capacity in conjunction with can_model options 3 or 4 See INIT OPTS Section 6 2 Page 53 of 100 6 11 INIT_SNOW Snow parameters gt INIT SNOW rho_ snow snow _hcap snow_hcon r0 rmax snow_ggr 1 3 amax 1 2 dtland kland maskd snowLoadLAI snowInterceptFact snowUnloadFact Table 20 Description of options in the INIT_SNOW section HCTN30 refers to Hadley Centre technical note 30 available from http www metoffice gov uk resea
47. ction s ssessssesesesssseserieresesssresisterestsssnenenterenesnsnenenrerenett 43 able V7 LAst of PET parameters o5 2 85 E A beh scagse cide de obese egies E A 43 Table 18 Description of options in the INIT VEG _ VARY Section 0 cceceecseeeseeececseseeseseeececseseesesesesecseseeseseseseenees 46 Table 19 Description of options in the INIT NONVEG Section 0 cccceecccsseeseseeececesseseseeesecseseesesesessesssesseseseseenaes 51 Table 20 Description of options in the INIT SNOW section s ssssssesssiesisessssesesisesisssneseteresesssnenenterenenssnesenrenenentt 53 Table 21 Description of options in the INIT TRIF S ctiOn ccccccceseecccsssseeseeececsessesesesssecssssesesesessesssessesesseeesees 55 Table 22 Description of options in the INIT AGRIC section ccccceccccccsssseseseeececsessesesesececsessesesesessessseeseseseseenees 56 Table 23 Description of options in the INIT MISC SectiOn cccccccesseesccscsseeesesececsessesssesesscsesssessesesecsesssessesesaeees 58 Table 24 Description of option in the INIT DRIVE Section cccccceeceseeecscsseseseseeececseseeseseeeseesessesesesessesssesseseseseenees 60 Table 25 Names of meteorological driving variables ccccesceessesssessceescesecesecesecseceaeceaecaeecaeeeaeeeaeeseeeaeeeaeenneenseeereeas 64 Table 26 Description of options for INIT IC S CtiOn cece cccsseesesecscscssesesesecscsessssssesececsesssessesececessesesecsesaases 68 Table 27
48. d sea mask 1 0 more generally it can now be thought of as a map of points to be modelled e g land fraction can be set to 1 at all locations within a catchment and to zero or less at all other points such as land points outside the catchment Switch indicating if only land points are to be modelled If pointsList TRUE landOnly must be FALSE TRUE Only land points are modelled Sea points are excluded from the model grid FALSE All points are modelled land and sea subArea logical TRUE a subsection of the input grid will be used see xcoord and ycoord below FALSE the full input grid is considered subAreaLatLon logical If subArea TRUE this indicates how to interpret the coordinates xcoord and ycoord TRUE co ordinates are longitude and latitude FALSE co ordinates are x and y indices column and row numbers xcoord 1 2 real array x coordinates of the sub area to be considered Depending on subAreaLatLon these are longitudes in range 180 to 360 or column numbers See notes on grid definition in Section 6 4 If values are column numbers the code uses the nearest integer to the input value ycoord 1 2 real array As xcoord expect in latitudinal y direction npoints integer The number of points that are to be modelled Only used if pointsList TRUE readFilePoints logical Switch controlling source of list of point numbers Only used if poin
49. d with precipType 2 precipTS Snowfall rate kg m s 1 Used with precipType 2 pstar Air pressure Pa q Specific humidity kg kg Air temperature K u Zonal component of the wind m s Used with windSpeed FALSE v Meridional component of the wind m s7 Used with windSpeed FALSE wind Total wind speed m s7 Used with windSpeed TRUE 6 15 1 Examples of specifying driving data Example 1 single point driving data In this example we consider a case with one point in the input file and all driving data for each time held on a single line of an ASCII input file The input file is illustrated in Figure 4 The Page 65 of 100 relevant entries in the run control file are shown below Only the lines in bold are relevant and irrelevant sections have been omitted gt INIT DRIVE 3600 0 1 9 F datal dat 19970101 00 00 00 asc 2 T 275 0 375 0 0 2 1 T F 10 0 10 0 gt ASCBIN 10 0 0 0 T gt VARS pstar t q wind lw_down sw_down precipTR precipTS gt ENDVARS m OoP NWOADOAH WO driveDataPer ndriveFileTime readList fileName driveFileDate 1 driveFileTime 1 driveFormat driveFilePer ioPrecipType 1_point_data tForsSnow tForConv conFrac io_rad_type ioWindSpeed ioTotalRunoff zl_uv zl_tq nfieldDriveFile ndriveHeaderFile ndriveHeaderTime ndriveHeaderField noNewLineDrive nf psfc name field flag name nf t nf q nf u nf
50. ded in the options line of GrADS ctl files Acceptable values are little endian for little endian computers e g PCs big endian for big endian computers e g Suns dumpFreq integer Flag indicating how often the model state is to be O to 4 dumped written to a file Acceptable values are 0 no dumps are written 1 only the final state of the model at the end of the integration is dumped 2 dump initial and final model states 3 as 2 but also write a dump at the end of the spin up phase 4 as 3 but also write a dump at the end of each calendar year A model dump captures the state of the model at a given point in the integration If a final dump is saved the integration can later be extended by starting another run from this final dump For long integrations or large domains it is recommended that dumps are saved for every year so that in the event of any trouble such as a model crash the integration can be completed without having to start again from the initial state NB A run that is carried out in several steps each starting from Page 77 of 100 the model dump for the previous step will generally not evolve identically to a single run that proceeds without the intermediate dumps This is due in part to a loss of precision when the model state is written to the dump file dumpStatus Character new replace or The file status used when
51. dicating the file format Case sensitive Only used if readFile TRUE asc ASCII bin generic binary including GrADS ne netCDF pp PP format 5 2 1 ASCII or binary files If fileFormat asc bin or pp or pp some or all of the following information is read from a section that starts with the tag gt ASCBIN Exactly what information is needed Page 14 of 100 varies between cases for example it is assumed that there is a single time level in a file of soil properties so nheaderTime is not needed Table 3 Format of options used to specify the reading of ASCII binary and PP format files Variable name Type Notes nheaderFile integer The number of header records at the top of a file For an ASCII file a header record is a line in the file For a binary file a header record is an individual word or record e g a single real value Not used for a PP file nheaderTime integer The number of header records that precedes the data for each time level within a file Not used for a PP file nheaderField integer The number of header records that precedes each field x y plane of data Not used for a PP file fieldNumber integer This is used to locate a given field xy plane within all the fields available at each time level If there are nFieldFile fields of data at each time level and fieldNumber 2 for a particular variable
52. e fileFormat characte Format of data file Only used if readFile TRUE E filename characte Name of file containing data Only used E readFile TRUE nheaderFile integer The number of headers at the start of the file gt 0 nheaderField integer The number of headers before each field gt 0 fieldNum integer The field number of the first field to be used from the gt input file Page 57 of 100 nDim integer Data in SDF is held on array of nDim dimensions gt dimName 1 nDim character Names of dimensions in file array SDFName CHAR character The name of the variable containing data as it appear in the SDF frac _agr 1 nxIn 1_ real array The fraction that is agriculture nyIn Page 58 of 100 6 14 INIT_MISC Miscellaneous surface carbon and vegetation parameters gt INIT MISC hleaf hwood betal beta2 fwe_c3 fwe c4 q10 leaf kaps q10_ soil cs min co2 mmr frac min frac _ seed pow Table 23 Description of options in the INIT_MISC section HCTN24 and 30 refer to Hadley Centre technical notes 24 and 30 available from http www metoffice gov uk research hadleycentre pubs HCTN index html Variable name Type and Notes permitted values hleaf real Specific heat capacity of leaves J K per kg carbon HCTN30 p6 Typical value 5 7E4 hwood real Specific heat capacity of wood J K p
53. e 1 dimName 2 dimName ndim gt VARS varName 1 varFlag 1 constVal 1 SDFname 1 varName 2 varFlag 2 constVal 2 SDFname 2 Repeat for each variable gt ENDVARS Data fields to be read from this file should appear below here gt DATA Table 26 Description of options for INIT_IC section Variable name Type and Notes permitted values readFile logical Switch controlling location of initial state data TRUE read from an external file including a model dump FALSE read from the run control file fileFormat character Format of data Only used if readFile TRUE A model dump is indicated by fileFormat dump other valid formats are described in Section 5 2 Page 69 of 100 filename character Name of file used if readFile TRUE containing data Only Zrev logical Switch indicating if soil data are stored in reverse order of levels TRUE vertical order is reversed with data stored in bottom to top order i e bottom layer first FALSE standard vertical order with data stored in top to bottom order i e uppermost layer first gt ASCBIN or pp The following are used if fileFormat asc or if readFile FALSE Yorn chimio nheaderFile integer The number of headers at the start of the file See Section 5 2 nheaderField integer The number of headers at the sta
54. e UK Met Office This document describes how to run version 2 0 of JULES It primarily describes the format of the input and output files and does not include detailed descriptions of the science and representation of the processes in the model Further information can be found on the JULES website http www jchmr org jules What s new in this version The physical processes and their representation in version 2 0 have not changed from version 1 However version 2 0 is much more flexible in terms of input and output and allows JULES to be run on a grid of points New features include e Ability to run on a grid e Choice of ASCII or binary formats for input and output files also limited support of netCDF input More flexible surface types number and types can vary Optional time varying prescribed vegetation properties More choice of meteorological input variables Optional automatic spin up Enhanced diagnostics large choice of variables frequency of output sampling frequency etc Page 6 of 100 2 Overview of JULES This section provides a brief overview of JULES largely so as to provide background information and introduce terms used in the rest of the manual Further details on the science and process descriptions contained in JULES can be found at the JULES website http www jchmr org jules JULES views each gridbox as consisting of a number of surface types The fractional area of each surface type is eit
55. e field is to be initialised as spatially constant using the value given under constVal For example the temperature in each soil layer t_soil will be set to 278K at all locations in the model grid For soil wetness sthuf the field number is given as 7 meaning that soil wetness will be set using the data starting at field 7 in the named input file Since zrev TRUE these data are stored in the file in non standard order i e bottom to top so that field 7 is the deepest layer and assuming 4 soil layers field 10 will be used for the uppermost layer Page 74 of 100 6 17 INIT_OUT Specification of output from the model JULES separates output into one or more output profiles or streams Within each profile all variables selected for output are written to the same file with the same frequency although the time processing can differ between variables e g instantaneous values and time averages can appear in the same profile Each profile can be considered as a separate data stream By using more than one profile the user can for example Output one set of variables to one file and other variables to another file Write instantaneous values to one file and time averaged values to another Write low frequency output from the entire model grid to one file and high frequency output from a subset of points to another file Write low frequency output throughout the run to one file and high frequency output from a smal
56. e keen to get this code contact the JULES developers Page 93 of 100 8 Known limitations of and bugs in the code 1 Spin up over short periods The current code cannot cope with a spin up cycle that is short in comparison to the period of any input data For example a spin up cycle of 1 day cannot use 10 day vegetation data The code will likely run but the evolution of the vegetation data will probably not be what the user intended However it is unlikely that a user would want to try such a run 2 Lack of more generic i o code If a user wants to introduce new time varying data that cannot be made to fit into the existing code for vegetation or meteorological data for example the new data would need to have the same frequency as the other data type they may have a substantial job on their hands For many purposes a simple hack may suffice e g write code to read a particular data set for a particular run but this will lack generality and options such as automatic spin up will be hard to accommodate At present there is no good solution we don t have any flexible coupling code that can be told to fetch suitable values of an arbitrary field although JULES may move towards this in future Page 94 of 100 9 Variables available for output Variables that are available for output from JULES are listed in the tables of this section separated according to their type Types of variables are SINGLE a single value at al
57. eaf area index kg carbon m a ws Woody biomass as a multiple of live stem biomass b wl Allometric exponent relating the target woody biomass to the leaf area index This is 5 3 in HCTN24 Eq 8 eta_sl Live stemwood coefficient kg C m LAI G leaf 0 Minimum turnover rate for leaves 360days dgl dm Rate of change of leaf turnover rate with moisture availability dgl dt Rate of change of leaf turnover rate with temperature K This is 9 in HCTN24 Eq 10 glmin Minimum leaf conductance for HO m s dqcrit Critical humidity deficit kg H20 kg air See Eqn 17 of Cox et al 1999 fd Scale factor for dark respiration See HCTN 24 Eq 56 0 CI CA for DQ 0 See HCTN 24 Eq 32 fsmc of Moisture availability below which leaves are dropped neff Scale factor relating Vomax with leaf nitrogen concentration See HCTN 24 Eq 51 nlo Top leaf nitrogen concentration kg N kg C nr nl Ratio of root nitrogen concentration to leaf nitrogen concentration ns nl Ratio of stem nitrogen concentration to leaf nitrogen concentration r grow Growth respiration fraction sigl Specific density of leaf carbon kg C m2 leaf tleaf of Temperature below which leaves are dropped K tlow Lower temperature for photosynthesis deg C tupp Upper temperature for photosynthesis deg C Page 46 of 100 6 9 INIT_VEG_VARY Time space varying parameters for plant functional types This section describes prescribed characteristic
58. effectively there is no interpolation Thus the latitudes and longitudes of the model gridpoints specified in INIT GRID above must be consistent with those specified here for the full grid Page 83 of 100 If a package other than GrADS is being used to display the thinned data the user will have to either work out how to use the GrADS mapping between the vector and the full grid or create new mapping data 6 17 4 An example of output grids and mapping This example uses the grids shown in Figure 6 The model grid has nx 5 ny 4 as shown and is regular in latitude and longitude For simplicity we will assume that the input grid was identical to the model grid The user wishes to output the 3 shaded points to an output grid with nxOut 2 nyOut 2 maintaining their relative positions as given by latitude and longitude ae 16 7 T8 19 20 57 9 10 78 11 12 13 14 15 me 3 a A 3 oO a 6 7 8 9 10 096 4 E 2 E sal a8 1 2 3 4 5 1 2 7 8 9 10 ae 10 Longitude E iignaredae E model point output point Model grid Output grid Figure 6 An example of the grids used in output mapping The easiest way to achieve this is to use the following lines in the run control file irrelevant lines have been omitted 273 pointsFlag 1 2 Bey outCompress outLLorder F readFile 3 pointsOut 4 5 10 mapOut 1 pointsOut 1 pointsFlag 1 1
59. eorological driving data gt INIT DRIVE driveDataPer ndriveFileTime readList fileName AriveFilePer driveFileDate 1 driveFileTime 1 driveFormat i1oPrecipType 1 tForsSnow tForConv conFr _point data ac io rad_type ioWindSpeed zl uv z1 tq gt ASCBIN nfieldDriveFil e ndriveHeaderFile ndriveHeaderTime ndriveHeaderField noNewLineDriv gt VARS name 1 fieldNumber 1 interp 1 nameFile 1 name 2 fieldNumber 2 interp 2 nameFile 2 Repeat for each variable gt ENDVARS gt NC ndim dimName 1 ndim gt VARS name 1 SDFname 1 nameFile 1 interp 1 name 2 SDFname 2 nameFile 2 interp 2 Repeat for each variable gt ENDVARS Table 24 Description of option in the INIT_DRIVE section Variable name Type and Notes permitted values driveDataPer integer The period of the driving data s This must be a 1 86400 see multiple of the model timestep and must be at most notes 86400s one day 86400 must be a multiple of driveDataPer so that data are read at the same times each day ndriveFileTime integer The number of data files available for each variable gt each file holding data for different times If all variables are held together this is the number of data files If Page 61 of 100 variables are held in separate files this is the number of f
60. er kg carbon HCTN30 p6 Typical value 1 1le4 betal real Coupling coefficient for co limitation in photosynthesis model Cox et al 1999 Eq 61 Typical value 0 83 beta2 real Coupling coefficient for co limitation in photosynthesis model Cox et al 1999 Eq 62 Typical value 0 93 fwe c3 real Constant in expression for limitation of photosynthesis by transport of products for C3 plants This is 0 5 in Eq 60 of Cox et al 1999 fwe _ c4 real Constant in expression for limitation of photosynthesis by transport of products for C4 plants This is 2 0x10 in Eq 60 of Cox et al 1999 q10 leaf real Q10 factor for plant respiration Cox et al 1999 Eq 66 Typical value 2 0 kaps real Specific soil respiration rate at 25 degC and optimum soil Page 59 of 100 moisture s HCTN24 Eq 16 Typical vale 5e 9 q10 soil real Q10 factor for soil respiration HCTN24 Eq 17 Typical value 2 0 cs_min real Minimum allowed soil carbon kg m Typical value 1 0e 6 co2_ mmr real Concentration of atmospheric CO2 expressed as a mass mixing ratio frac min real Minimum fraction that a PFT is allowed to cover if TRIFFID is used Typical value 1 0e 6 frac_seed real Seed fraction for TRIFFID Typical value 0 01 pow real Power in sigmodial function used to get competition coefficients This is 20 0 in HCTN24 Eq 3 Page 60 of 100 6 15 INIT_DRIVE Met
61. er of fields in any time level must be the same in all files Table 32 Examples of the use of file name templating 8 Users of GrADS should note that for these shorter substitution string periods hours and minutes JULES can use files that cannot be described by a GrADS template control file GrADS at v1 9v4 insists that each file contains data that covers at most one period whereas JULES allows data for more than one period For example if the substitution template includes h2 GrADS insists that each file contains data for at most one hour whereas JULES allows each file to have 1 2 3 4 etc hours of data Page 87 of 100 Substitution template Description Valid Example file names Comments of files template data met_data y4 mc dat Monthly Yes data met_data_1990jan dat files data met_data_1990feb dat y4 met_data_ Y y4 mc dat Monthly Yes 1990 met_data_1990jan dat A substitution string can files appear more than once Here data for each year are stored in a separate directory vv_ y4 Yearly Yes Rain_1990 dat Variable name and time files with Wind_1990 dat templating used together each The strings that are to be variable in substituted for Yovv will be a separate provided by the user via file the run control file Data_ od2 dat Hourly No Each file can contain at data each most 1 day of data For file substitution strings that containing refer to years months or da
62. es North of the southernmost row of gridpoints in the input grid NOT Page 31 of 100 necessarily the model grid The gridpoint is in the centre of the gridbox regLonl real The longitude decimal degrees East of the westernmost 180 to 360 column of gridpoints in the input grid NOT necessarily the model grid regDlat real The size of a gridbox in the NS direction decimal gt 0 0 degrees of latitude Note regLatl and regLonl are only used if reghLatLon TRUE regDlat and regDlon may be used even if regLatLon FALSE if there is a need to establish the area of each gridbox which is needed by some parameterisations and to label output regDlon real The size of a gridbox in the EW direction decimal gt 0 0 degrees of longitude readFileLatLon logical Switch controlling source of latitude and longitude data Only used if pointsList FALSE and regLatLon FALSE TRUE read from an external file FALSE read from the run control file at the section marked gt DATA_ LATLON fileFormatLatl character Format of file containing latitude and longitude data on fileNameLatLon character Name of file containing latitude and longitude data nheaderFile integer The number of headers at the start of the lat lon file gt 0 See Section 5 2 nheaderField integer The number of headers before each field in the lat lon gt 0 file See Section 5 2 fieldLat
63. fined at all possible points both land and sea The result is a variable on the model grid this is considered to be a 2 dimensional variable on x y even if the model grid is effectively a vector with ny 1 For example air temperature is defined at all possible points both land and sea readVar2dComp Reads a variable that is only defined on a subset of points for example land points The result is a vector For example a land variable can be read from a 2 D x y map that may contain both land and sea points and the result is a vector on land points The Comp in the name is meant to suggest compression to a vector readVar3dComp As readVar2dComp but the variable is also a function of the vertical level e g a soil variable on several levels This 3d version works by looping over the vertical levels calling the 2d version for each level 7 2 How to implement new diagnostics for output The steps needed to add a new diagnostic vary according to what variables are needed in order to calculate the diagnostic These are covered in the next sections 7 2 1 Output of existing variables The data are already held in an existing FORTRAN variable in a module This is the easiest case since the data are easily accessed The name that is used to select the diagnostic should be added to subroutine init out varlist following the existing examples Care should be taken to specify the correct type of diagnostic
64. g the variables as options when make is invoked e g make CDF _ LIB PATH HOME mynetcdf lib CDF _ MOD PATH HOME mynetcdf mod If the user does not have access to a pre compiled netCDF library then JULES may be compiled by specifying CDFDUMMY true when make is invoked rather than setting the CDF _LIB PATH and CDF_MOD PATH variables This option compiles a set of dummy netCDF interface functions which merely allows the rest of the JULES code to compile correctly and provides no functionality When this option is used JULES will neither read from nor write to netCDF format files The user must ensure that netCDF input output options are not selected at any point in any JULES control file described in Section 6 used with an executable produced using this option 3 3 Example build lines To build JULES using the normal Sun compiler options and link with a netCDF library Page 9 of 100 make COMPILER sun BUILD run CDF LIB PATH SHOME mynetcdf lib CDF _ MOD PATH SHOME mynetcdf mod To build JULES using the fast Intel compiler options and do not link with a netCDF library make COMPILER intel BUILD fast CDFDUMMY true These command lines can become quite long and tedious to keep typing so it s a good idea to set the list of frequently used ones as environment variables export JULESBUILD COMPILER sun BUILD run CDF _ LIB PATH S HOME mynetcdf lib CDF MOD PATH SHOME mynetcdf mod
65. hat data are to be read from several files using one or more of these file formats For example soil data might be in an ASCII file while meteorological driving data are in netCDF files A self describing file SDF is one in a format that contains metadata describing the contents of the file For JULES only a netCDF file is presently considered to be a SDF Minimal use is made of any metadata contained within a file including SDFs and PP files For example a SDF might contain data that describes the grid or the times of data but these are not used by JULES Instead this information is provided via the run control file and all input data must be provided on the same grid For all non SDF files the data model is based on that used by GrADS Each variable is viewed as being 4 dimensional in x y z t on a regular grid Although we will talk of x and y in terms of West East and South North compass directions in general the grid can be any rectilinear grid with West East being replaced by left to right X varies in the direction from West to East y varies from South to North this default can be changed and z varies from bottom to top All variables in any one file must have the same grid size in x and y i e all variables are on a grid of nx ny points and have a value at all times although that value could indicate a missing datum The data are stored as a series of xy slices with x varying fastest then y then z and t var
66. he wilting point is that at which soil moisture stress completely prevents transpiration satcon and sathh may be adjusted through application of the multipliers multCon and multH described in this section This only applies if data come from an external file readFile TRUE Page 43 of 100 6 8 INIT_VEG_PFT Time invariant parameters for plant functional types This section reads the values of parameters for each of the plant functional types PFTs These parameters are a function of PFT only Parameters that also vary with time and or location are dealt with in control file section INIT VEG VARY see guide Section 6 9 Parameters that are only required if the dynamic vegetation TRIFFID or phenology sections are requested are read separately in control file section INIT _TRIF see guide Section 6 12 For many applications the best approach may be to read the PFT parameters from the standard parameter files provided with the JULES code readFile TRUE filename PARAM standard_pft_param dat since this removes the risk that values can be changed by an accidental edit to the run control file The description of INIT VEG PFT options is given in Table 16 and the list of required variables is given in Table 17 gt INIT VEG PFT readFile fileName npftinFile gt DATA varl 1 varl 2 varl npft var2 1 var2 2 var2 npft Table 16 Description of options in the INIT_VEG_PFT section Variable
67. he leaf area index LAI of each PFT The value read here is only used if TRIFFID is not active 1_trif FALSE If TRIFFID is active LAI is a prognostic variable and its initial value is read as described in Section 6 16 below Catch Minimum canopy capacity kg m This is the minimum amount of water that can be held on the canopy See HCTN30 p7 dcatch dlai Rate of change of canopy capacity with LAI kg m Canopy capacity is calculated as catch0 dcatch dlai lai See HCTN30 p7 dz0v_ dh Rate of change of vegetation roughness length for momentum with height Roughness length is calculated as dz0v dh canht ft See HCTN30 p5 zOh_ zOm Ratio of the roughness length for heat to the roughness length for momentum This is generally assumed to be 0 1 See HCTN30 p6 Note that this is the ratio of the roughness length for heat to that for momentum It does not alter the roughness length for momentum which is calculated using canht ft and dz0v dh see above ea A Infiltration enhancement factor The maximum infiltration rate defined by the soil parameters for the whole gridbox may be modified for each PFT to account for tile dependent factors such as marcro pores related to vegetation roots See HCTN30 p14 for full details rootd ft Root depth m An exponential distribution with depth is assumed with e folding depth rootd_ft Note that this means that generally some of the roots exist at depths greater
68. he name of the file to be read Only used if r readFile TRUE npftInFile integer The number of PFTs for which parameters are available in 2npft the input file gt DATA If readFile FALSE the dataVar parameters should be listed in the order given below crop integer Flag indicating whether the PFT is a crop 0 or Only crop PFTs are allowed to grow in the agricultural area 0 not a crop 1 acrop g_area real Disturbance rate 360days g grow real Rate of leaf growth 360days g_ root real Turnover rate for root biomass 360days g wood real Turnover rate for woody biomass 360days lai_max real Maximum LAI lai min real Minimum LAI Page 56 of 100 6 13 INIT_AGRIC Fractional coverage by agriculture If the TRIFFID vegetation model is used the fractional area of agricultural land in each gridbox is read in this section Otherwise this section is not used gt INIT AGRIC readFile fileFormat fileName gt ASCBIN nheaderFile nheaderField fieldNum gt NC ndim dimName 1 ndim varName Data fields to be read from this file should appear below here gt DATA frac_agr 1 nxIn 1 nyIn Table 22 Description of options in the INIT_AGRIC section Variable name Type and Notes permitted values readFile logical Switch controlling location of soil layer data TRUE read from an external file FALSE read from the run control fil
69. he run control file readFile TRUE and fileFormat dump these should now appear in the file in the order indicated by the value of flag for each variable see above For example if tstar is given a value of flag 1 and cs has flag 2 data for tstar and cs should then be listed with each variable starting on a separate line Page 70 of 100 Some of these variables may not be required for a particular run depending on the model configuration The size of each variable is defined in terms of four variables Land pts the number of gridboxes that contain any land sm_ levels the number of soil layers ntiles the number of tiles at each gridbox ntype the number of surface types see Section 6 2 for description of these last three variables Table 27 JULES variables that define the model state prognostic variables Name Shape Description Notes sthuf land pts sm levels Soil wetness for each soil layer This is the mass of soil water liquid and frozen expressed as a fraction of the water content at saturation Either sthuf or its components sthu and sthf are always required sthf land pts sm _ levels Frozen soil wetness for each soil layer This is the mass of frozen water expressed as a fraction of the water content at saturation Note that the partitioning of water between liquid and solid fractions may be altered during initialisation The procedure conserves the tot
70. headers at the start of each field see gt 0 Section 5 2 noNewLineVeg logical Switch describing format of ASCII file TRUE means that variables are arranged across one or more lines and each variable does not necessarily start a new line This option should be used if all the data for each time are one line of the input file although it can also be used if the data continue onto subsequent lines TRUE is only allowed if the fields are not functions of position i e vegFlag t see above FALSE means that each variable starts on a new line varName character The name of the variable This is used to identify the oanht variable in the code and is set in the code These lai must be entered exactly as listed and are case rootd sensitive Acceptable values lai for leaf area index canht for canopy height rootd for root depth Flag character Flag indicating how the characteristic varies t tx Acceptable values ix t function of PFT and time only tx function of PFT time and location x function of PFT and location only At present all nvegVar variables must have the same value for this flag rootd can only use flag t i e root depth cannot vary with location fieldNumber integer The field number of the first level of data in the input file that is to be used for a variable interpFlag character Flag indicating how variable is to be interpolated in See Table 33 t
71. her prescribed by the user or modelled by the TRIFFID sub model Each surface type is represented by a tile and a separate energy balance is calculated for each tile The gridbox average energy balance is found by weighting the values from each tile In its standard form JULES recognises nine surface types broadleaf trees needleleaf trees C3 temperate grass C4 tropical grass shrubs urban inland water bare soil and ice These 9 types are modelled as 9 tiles A land gridbox is either any mixture of the first 8 surface types or is land ice Note that with version 2 0 one is not limited to these 9 standard surface types Soil processes are modelled in several layers but all tiles lie over and interact with the same soil column Each gridbox requires meteorological driving variables such as air temperature and variables that describe the soil properties at that location It is also possible to prescribe certain characteristics of the vegetation such as Leaf Area Index to vary between gridboxes JULES can be run for any number of gridboxes from one upwards The number of gridboxes is limited by the availability of computing power and suitable input data When run on a grid JULES models the average state of the land surface within the area of the gridbox and most quantities are taken to be homogeneous within the gridbox with options to include subgrid scale variability of a few such as rainfall In that case the input data are also area
72. iles for any one variable If time templating is used see Section 6 18 ndriveFileTime should be 1 driveFilePer integer The period s of the files containing the data This is only used if time templating is used see Section 6 18 This must be at least as large as the period of the data driveDataPer and must be a multiple of the model timestep Special cases 1 monthly files 2 annual files readList logical Switch controlling how the names of the files containing the driving data and the times covered by each are read TRUE names are read from another file FALSE names are read from the run control file This option is only allowed if ndriveFileTime 1 filename 1 character If ndriveFileTime 1 this is the name of the single data file or the template name If ndriveFileTime gt 1 this is the name of a file that lists the names and times of the data files The first line of this file will be skipped and so can be used for comments All other lines are to be of the form filename startDate startTime where fileName may contain variable name templating see Section 6 18 startDate is in format yyyymmdd time isin format hh mm ss Starting time and date for first driving data file Only used if readList FALSE otherwise these values are read from an external file driveFileDate integer Date of first data in the driving data file in format yyyymmdd dr
73. ime nameFile character The substitution string used in the names of files that Page 49 of 100 contain this variable Only used if variable name templating is used see Section 6 18 gt NC If fileFormat nc nvegDim integer Number of dimensions in SDF variable 21 vegDim 1 nvegDim character array Names of dimensions varName character See above under gt ASCBIN flag character See above under gt ASCBIN interpFlag character See above under gt ASCBIN SDFname character The name of the variable as it appears in a SDF nameFile character See above under gt ASCBIN 6 9 1 Examples of INIT_VEG_VARY Example 1 Time varying Leaf Area Index Leaf Area Index is to vary with time but not with position on the grid Climatological monthly data are to be used with values updated at the start of each day Note that the values are always assumed to be a function of PFT The ASCII input file is illustrated in Figure 3 and contains one month of data for all PFTs on a single line Month p1 p2 p3 p4 p5 1 0 5 4 0 1 0 2 0 1 0 2 0 7 4 0 wl 2 0 ao 3 0 9 4 2 5 2 0 2 0 4 2 0 4 5 2 0 2 0 2 5 rest of file not shown Figure 3 Schematic of an ASCII file with monthly LAI data The relevant entries in the run control file are shown below Only the lines in bold are relevant and irrelevant sections have been omitted gt INIT VEG VARY 1 nvegVar 1
74. in latitudinal y direction outCompress logical Switch indicating if output data are to be compressed so that only model points are output TRUE Only output model points Also output the mapping between the model points and the output grid e g how to scatter the output points across a larger grid The mapping is output in a form suitable for use with GrADS pdef FALSE If the output grid is larger than the number of points to be output the grid is filled with missing data or padding values See Section 6 17 3 for further discussion of output compression If the output grid is the same size as the number of points to be output so no compression is possible outCompress TRUE may still cause output to differ in format from out Compress FALSE the points may be written in a different order so outCompress should always be set to FALSE unless needed otherwise Note that if out Compress TRUE then yrevOut is ignored for the profile it becomes irrelevant outLLorder logical Switch indicating the coordinate system to be used to determine the locations of the output points in the output grid Only used if pointsFlag 2 lor3 TRUE use the latitude and longitude of each point to determine its location in the output grid FALSE use the row and column number in the INPUT grid to determine where each point goes in output grid This option is particularly useful if the input grid is rectilinear but is
75. integer The field number in the file that holds latitude data gt See discussion of fields in Section 5 1 fieldLon integer The field number in the file that holds longitude data gt nLatLonDim integer Lat lon data in a SDF are held in arrays of gt nLatLonDim dimensions latLonDim 1 nl character array Names of dimension variables in SDF atLonDim varNameLat character The name of the variable containing the latitude data varNameLon character The name of the variable containing the longitude data pointList l np oints integer array gt A list of the points that are to be modelled These are point numbers in the input grid NB If the input data run from North to South i e not the JULES S to N order the point numbers should still be Page 32 of 100 calculated following the JULES S to N convention Thus point number 1 is in the SW corner of the grid which will not be the first point in the input data if yrevIn T unless nyIn 1 flandg 1 nxiIn real array The fraction of each gridbox that is land l1 nyIn Latitude 1 nxI real array The latitude of each gridpoint n l nyIn Longitude 1 nx real array The longitude of each gridpoint All values should be in In 1 nyIn 180 to 360 the range of either 180 to 180 or 0 to 360 The special case of an equal angle grid all gridboxes have same extent in terms of latitude and longitude in which the rows
76. ints F readFilePoints points txt fileNamePoints gt INIT LAND F readFileLand ban fileFormatLand ce ileNameLand grid gra Page 33 of 100 gt INIT LATLON T regLatLon 40 0 50 0 regLatl regLonl 1 0 1 0 regDlat regDlon F readFileLatLon bin fileFormatLatLon latlon gra fileNameLatLon gt DATA_ POINTS 1 pointList gt DATA LAND 1 0 flandg gt DATA LATLON 0 0 latitude 5 0 longitude pointsList T indicates that the grid will be described by a list of points npoints 1 shows that this run is for a single point readFilePoints F indicates that the point numbers are read from the gt DATA POINTS section where point number 1 is indicated the only possibility for an input grid of one point readFileLand F indicates that the land fraction field is read from the gt DATA_ LAND section where the value 1 0 shows that the single gridbox is 100 land regLatLons T indicates that the grid is regular and will be described by its origin regLat1 regLon1 and gridbox size regDlat regDlon There is then no need for any further information about coordinates in particular the data at gt DATA_LATLON are not read Grid example 2 Selecting points in a given range of latitude and longitude The grids used in this example are shown in Figure 1 The input grid has nxIn 5 nyIn 4 and we wish to model the area 55 57 N 355 357
77. ints the land grid is identical to the model grid 57 9776 17 18 1920 a Cllti i Cd Z x nape 3 4 8 9 2 57 556 1 2 3 4 Z 6 7 8 D E 2 2 5 J O 56 TO 2E 3 Za 55 5553 354 355 356 357 358 Longitude E 59 i bonan E zee input point input point model point model point Input grid Model grid Land grid Figure 2 Example of grid selection based on longitude and latitude taking land points only Page 37 of 100 6 5 INIT_FRAC Fractional coverage of land surface types In this section we specify the fraction of the land area in each gridbox that is covered by each of the surface types Under certain circumstances described below this information may be acquired later via another section gt INIT FRAC readFracIC readFile fileFormat filename gt ASCBIN nheaderFil nheaderField fieldNum gt NC nfracDim fracDim l nfracDim varName gt DATA frac 1l nxIn 1 nyIn Table 12 Description of options in the INIT_FRAC section Variable name Type and Notes permitted values readFracIc logical Switch indicating location of fractional cover data TRUE fractional cover is provided as part of the initial condition see INIT FRAC in Section 6 5 and will not be read here FALSE fractional cover will be read here For runs with dynamic vegetation TRIFFID l _trif TRUE
78. it the file named Makefile There are two convenience options COMPILER and BUILD which should be passed to make from the command line to tell that program where the appropriate values should be taken from The COMPILER option allows the user to define a list of compiler specific variables including compiler flags without having to edit the Makefile The BUILD option allows the user to build with sets of predefined flags for different situations e g debugging The Type and permitted values for each of these options are described in Table 1 and additional information is given in the comments at the head of Makefile The compiler specific variables are specified in individual files named Makefile comp fora handful of common compilers e g Makefile comp sun The list of tested compilers includes two Intel and G95 that can be downloaded for no cost via the URLs in Table 1 To use a compiler that is not listed the user should replace the strings in the blank compiler file Makefile comp misc with values appropriate to their compiler and invoke make with the option COMPILER misc The netCDF interface library can be downloaded for no cost from http Awww unidata ucar edu software netcdf The GNU Make utility can be downloaded for no cost from http www gnu org software make Page 8 of 100 Table 1 Options that can be passed to make when building JULES Variable Type and Notes permitted values sun Default optio
79. iveFileTime character Time of day of first data in the driving data file in format hh mm ss driveFormat character Format of data files See Section 5 2 ioPrecipType integer Flag indicating which precipitation variables are input 1 2 3 or 4 and how they are treated Note that all precipitation in JULES is considered to be either rainfall or snowfall 1 A single precipitation field is input This represents the total precipitation rainfall and snowfall The total is partitioned between snowfall and rainfall and large scale and convective using tForSnow and tForConv see below 2 Two precipitation fields are input namely rainfall and snowfall The rainfall is partitioned between large scale and convective using t ForConv see below Page 62 of 100 3 Three precipitation fields are input namely large scale rainfall large scale snowfall and convective rainfall This cannot be used with l point data TRUE 4 Four precipitation fields are input namely large scale rainfall large scale snowfall convective rainfall and convective snowfall This cannot be used with l point data TRUE Note that this is the only option that considers convective snowfall which is set to zero in all other cases The concept of convective and large scale or dynamical components of precipitation comes from atmospheric models in which the precipitation from small scale convective and large scale motions is often calculated
80. l Soil thermal conductivity W m KS for each soil layer satCon Saturated hydraulic conductivity kg m s for each soil layer sathh Saturated soil water pressure m for each soil layer smcl Moisture content of each soil layer kg m soilWet Total moisture content of each soil layer as fraction of saturation sthf Frozen moisture content of each soil layer as a fraction of saturation sthu Unfrozen moisture content of each soil layer as a fraction of saturation tSoil Sub surface temperature of each layer K vsmcCrit Volumetric moisture content at critical point for each soil layer vsmcSat Volumetric moisture content at saturation for each soil layer vsmcWilt Volumetric moisture content at wilting point for each soil layer wFlux Downwards moisture flux at bottom of each soil layer kg m s1 Page 100 of 100 10 List of Tables Table 1 Options that can be passed to make when building JULES ec ccecceeecceseceseceeeeseeeneeeeeeeeeeeeceeeneeneeseesseenaes 8 Table 2 Format of frequently used control file Options c ceccceseeesceseceseeesecsnecseecseeeaeeeseeeeceseeesecesecsecaecnseceeeeeeseeess 13 Table 3 Format of options used to specify the reading of ASCII binary and PP format files ee eeeeeeeeeeeeeeeees 14 Table 4 Part of an example ASCII file that could be read by JULES o eceececccecceecceseceseeseceeeeaeceseceeecaeecneeeaeeseeeeseeereees 15 Table 5 Format of option
81. l gridpoints land and sea Table 35 LAND a single value at land gridpoints Table 36 PFT a value for each of npft PFTs at each land gridpoint Table 37 TILE a value for each of ntiles tiles at each land gridpoint Table 38 TYPE a value for each of nt ype surface types at each land gridpoint Table 39 SOIL a value for each of sm_levels soil layers at each land gridpoint Table 40 These tables were correct at the time of writing but the most reliable way to determine exactly which variables are available for a particular version of JULES is to look at the variables listed in the subroutine init out _varlist and which can be echoed to screen at the start of a JULES run by setting echo TRUE see Section 6 2 Page 95 of 100 Table 35 A list of output variables that have a single value at each gridpoint Name Description conRain Gridbox convective rainfall kg m s conSnow Gridbox convective snowfall kg m s CoOsSzZ Cosine of the zenith angle diffFrac Gridbox fraction of radiation that is diffuse ecan Gridbox mean evaporation from canopy surface store kg m s ei Gridbox sublimation from lying snow or sea ice kg m s esoil Gridbox surface evapotranspiration from soil moisture store kg ms fqw Gridbox moisture flux from surface kg m s ftl Gridbox surface sensible heat flux
82. le called grid gra which has no headers and contains land fraction as the first field regLatLon T indicates that the input grid is a regular grid with origin the gridpoint in the southwest corner shown by regLat1 regLon1 to be 55 5 N 355 5 E and gridbox size 1 x1 With this information JULES determines that there are 4 gridpoints within the given lat lon range and that the model grid will be a square of side 2 gridboxes The land fraction field shows that these are all land points meaning that the land vector also has 4 points Note that these points could also have been selected by providing a list of the point numbers indicated by pointsList TRUE npoints 4 and then entering the point numbers 3 4 8 9 after gt DATA_ POINTS Page 35 of 100 57 746 17 18 o0 gt 58 77 sia ts Z a cart 3 4 8 9 o 57 o 8 9 10 056 ilad a la 38 a 55 353 354 355 356 357 358 Longitude E Pts 356 357 enau e E input point input point model point model point Input grid Model grid Land grid Figure 1 Example of grid selection based on longitude and latitude Grid Example 3 Selecting only land points in a given range of latitude and longitude This example is similar to Example 2 but this time we only wish to model land points within a given area The grids used in this example are shown in Figure 2 and we wish to model land points in 55 57 N 354 356
83. le to that at the start of the next cycle Generally this does not cause a problem Depending upon the details of the input data and any temporal interpolation the driving data may vary rapidly at the end of a cycle of spin up causing an extreme response from the model In most cases the model will adjust possibly with large heat fluxes over a few hours but the user should be aware that unusual behaviour near the end start of a spin up cycle may be the result of this adjustment Consider the case of a spin up over 1 Jan 2005 to 31 Dec 2005 At or near the end of 31 Dec 2005 during the spin up the driving data will start to adjust to the values for 1 Jan 2005 which could be very different from conditions on 31 Dec 2005 The length of time over which the driving data adjust depends on the frequency of the data and the choice of temporal interpolation For example with 3 hourly data that is interpolated onto a one hour timestep the adjustment will take place over 3 hours However hourly data and an hourly timestep will force an instantaneous adjustment at the start of 1 Jan 2005 Although nspin specifies the maximum number of spin up cycles some of which might not be used if the model is considered to have spun up earlier it is possible to specify the exact number of cycles that will be performed This can be done by demanding an impossible level of convergence by setting spinTol lt 0 remember that spinTol is compared with the absolute change ove
84. ler part of the run e g a Special Observation period to another file This flexibility comes at the expense of having to set several values in the run control file However default values allow the user to select certain configurations relatively easily The first values in this section of the run control file concern general details of the output such as the file format that apply to all output profiles This is followed by a separate section for each output profile describing the variables the grid and time sampling for that profile 6 17 1 INIT_OUT general values This section starts with the tag gt INIT OUT gt INIT OUT un id utFormat utDir utStatus revOut zrevOut umMonth useTemplate nout unde fOut zsmc zst outEndian SK OOO 8F T 2 dumpF req dumpStatus Page 75 of 100 Table 28 Description of options in the INIT_OUT section Variable name Type and permitted values Notes runID character 10 A name or identifier for the run This is used to name output files and any model dumps outFormat character The format for output files Acceptable values are asc ASCII files bin flat binary files nc netCDF files outDir character 150 The directory used for output files This can be an 66 99 absolute or relative path Enter to write output to the directory from which JULES is run outStatus
85. m h land sf resist dustresb vgrav sf flux sea sf flux land stdevl_sea stdevl land sf orog_ gb sfl int sea phi m h sea sfl int tand OuEpUE bottom of timestep jules_ final sf impl hydrol VEG2 Vegiie loop Page 11 of 100 phi m h land im_sf pt sf evap sf melt screen tq qsat sice htf sfsnow surf_hyd frunoff sieve pdm calc_baseflow soil_ hyd hyd_con _vg darcy _ vg hyd_con _vg gauss cale fsat s611 Ate heat_con gauss ice_ hte soilme soilt ch4 wetl tilepts phenol triffid vegcarb growth lotka compete soilcarb decay tilepts sparm pft_sparm tilepts phenol sparm pft_sparm Page 12 of 100 5 File formats and the JULES grid 5 1 Overview of file formats JULES aims to support input and output in three formats ASCII netCDF and a generic binary format simply called binary below At present the ASCII and binary options are implemented but only a very limited implementation of input via netCDF exists and there is no netCDF output The netCDF options will be improved for future releases Input can also be read from many PP files a format used by the UK MetOffice The binary format is compatible with the GrADS package amongst others A run control file might indicate t
86. make SJULESBUILD It is then possible to override options specified in that variable by adding revised ones at the end make SJULESBUILD BUILD debug 3 4 Running JULES A JULES executable is run by redirecting standard input to a file that contains all the information needed to describe a run e g jules exe lt runl jin The format of this input file is described in Section 6 with some example runs described in Section 6 20 The file extension jin is meant to suggest JULES Input File but there is no need to use this or any other extension 4 Overview of the JULES code Page 10 of 100 The general structure of the JULES source code including the order in which routines are called is illustrated below For the sake of clarity the full details are not shown here In particular the initialisation and output steps subroutines init and output can call several routines jules init init calls various initialisation routines top of timestep loop drive_ update veg_ update control tile albedo albpft albsnow sf expl tilepts physiol root_frac smc_ext raero sf stom qsat canopy Leaft c3 leaf c4 soil_ evap leaf lit cancap microbe heat_con s exch qsat sf orog sf_ resist sf rib sea sf rib land 8f Or6g fcdch_sea phi m h sea fcdch_land phi
87. means that the chosen points will be listed individually pointsFlag 2 3 means the output grid is to be the smallest rectangle that includes all output points Page 84 of 100 outCompress F means the output grid will be padded as necessary In this case it means that output point 3 in Figure 6 will be filled with the missing data value outLLorder T means the location of each point in the output grid is calculated using the latitude and longitude of the point pointsOut 3 indicates that 3 points are to be output mapOut 1 pointsOut 1 indicates that the points to be output are numbers 4 5 and 10 in the input grid which is identical to the model grid in this case The model uses the latitude and longitude of each point to establish that the chosen points should occupy locations 1 2 and 4 in the output grid and that location 3 should be filled with the missing data flag undefOut The same effect could be achieved by using pointsFlag 1 2 pointsFlag 2 2 mapOut 2 1 2 4 outGridNxy 2 2 i e the user can completely specify the mapping and grid shape Calculating mapOut 2 is trivial in this example but would involve the user in more and unnecessary work if many more points were to be output 6 17 5 Notes on output 1 A warning is raised if any output is not generated because the output period is never completed This can occur when a run starts or end partway through an output period or if a spin up cycle e
88. n use options for Sun Studio compiler series previously known as Workshop and Forte intel Use options for Intel Fortran compiler for Linux Windows and MacOS http www3 intel com cd software products asmo na eng compilers 284132 htm Version 9 0 was used for testing and it was found that two lines in the source code had to be oe changed find these and the suggested replacements by searching the code for Intel g95 Use options for G95 compiler http www g95 org nag Use options for NAGWare compiler pgf Use options for Portland Group compiler misc Use options for an unlisted compiler run Default option for normal compilation of JULES BUILD debug Switch on compiler debug flags fast Switch on compiler optimisation flags for faster execution false Use a precompiled netCDF library re true Use the dummy netCDF library provided with JULES 3 2 The netCDF interface library To build JULES the user must also pass make some information about the netCDF interface library If the user has access to a pre compiled netCDF interface library then they should pass make the options CDF_LIB PATH and CDF MOD PATH The values for these options are the directories in which the pre compiled netCDF library libnc a and Fortran 90 module files those with mod extension are located respectively This can be done also by editing the Makefile itself but the recommended method is by specifyin
89. name gt ASCBIN nheaderFil gt VARS name 1 f f1eldNumber 1 nheaderField Repeated for each variable gt ENDVARS gt NC nsoilDim soilDim 1 nsoilDim name 1 SDFname 1 gt VARS Repeated for each variable gt VARS gt INIT SOIL2 multH multCon gt DATA data values Table 14 Description of options in the INIT_SOIL and INIT_SOIL2 sections Variable name Type and Notes permitted values l vg_soil logical Switch for van Genuchten soil hydraulic model TRUE use van Genuchten model FALSE use Clapp and Hornberger model constz logical Switch indicating if soil characteristics are to be uniform with depth at each gridbox TRUE soil characteristics do not vary with depth FALSE soil characteristics vary with depth zrev logical Switch indicating if input data are stored in reverse order of levels compared with JULES s expectation Page 41 of 100 TRUE vertical order is reversed with data stored in bottom to top order i e bottom layer first FALSE standard vertical order with data stored in top to bottom order i e uppermost layer first readFile logical Switch controlling location of soil layer data TRUE read from an external file FALSE read from the run control file fileFormat character Format of data file Only used if readFile TRUE filename character Name of file containing data
90. name Type and Notes permitted values readFile logical Switch controlling location of data TRUE read from an external file FALSE read from the run control file filename character The name of the external file containing the data Only used if readFile TRUE npftInFile integer The number of PFTs for which parameters are given in gt npft the input file Table 17 List of PFT parameters Each parameter has a separate value for each PFT npftInFile values are read for each parameter All values are of type REAL unless stated otherwise Parameters for the TRIFFID or phenology modules are described in Section 6 12 Page 44 of 100 HCTN24 and 30 refer to Hadley Centre technical notes 24 and 30 available from http www metoffice gov uk research hadleycentre pubs HCTN index html Variable name Description typeName Character Name of each PFT This list must include the PFTs used in this run see pftName in section INIT OPTS Section 6 2 These names are for the user s convenience and do not have any special significance within JULES amp 3 Integer Flag indicating whether PFT is C3 type 0 not C3 i e C4 1 C3 canht_ft The height of each PFT m also known as the canopy height The value read here is only used if TRIFFID is not active 1_trif FALSE If TRIFFID is active canht ft is a prognostic variable and its initial value is read as described in Section 6 16 below LAT T
91. ndicated by the fractional coverage of the ice surface type if used being one The fractional cover of the ice surface type in each gridbox must be either zero or one there cannot be partial coverage of ice within a gridbox 6 5 1 Example Reading frac from the run control file We assume nxIn nyIn npoints l and ntype 9 Only the lines relevant to this case are shown gt INIT_FRAC F readFracIcC F readFile gt DATA 0 55 0 15 0 20 0 00 0 05 0 00 0 05 0 00 frac readFracIC F indicates that frac is read here rather than as part of the initial condition readFile F indicates that data will be read from the run control file not from another file The 9 values of frac are positioned after the gt DATA tag 6 6 INIT_DZSOIL Soil layer depths Page 39 of 100 gt INIT DZSOIL gt DATA dzsoil 1 sm_levels Table 13 Description of options in the INIT_DZSOTL section Variable name Type and Notes permitted values dzSoil 1 sm_levels real array The soil layer depths m starting with the uppermost layer Note that the soil layer depths and hence the total soil depth are constant across the domain In its standard setup JULES uses layer depths of 0 1 0 25 0 65 and 2 0m giving a total depth of 3 0m Page 40 of 100 6 7 INIT_SOIL Soil hydraulic and thermal characteristics gt INIT SOIL l vg_soil constZ zrev readFile fileFormat file
92. nds partway through an output period For example if monthly average diagnostics are requested but the run ends on the 10 day of a month the final monthly average is incomplete In such cases a value is still written to the output file but the details of this value vary between cases In short a monthly or annual average is calculated if a large fraction of the month or year has been simulated but averages over shorter periods are not calculated and a missing data value is output For details see the code 2 GrADS output A control file ct1 file that describes GrADS output includes a specification of the number of times of output that are contained in the associated data files the TDEF line When a data file is first opened a control file is written with an estimate of the expected number of times that will be written Sometimes this initial estimate will prove wrong for example if the model is spinning up the number of spin up cycles may not be known in advance and the ct1 file is later rewritten when the data file is complete Under most circumstances this procedure is carried out without any problem However if the user opens the ct1 file in GrADS while the integration is still underway it may not correctly specify the number of times In that case the ct 1 file will be correct if reopened at a later time However if the user has moved the ct1 file while the integration is underway it cannot be rewritten and a w
93. not regular in latitude and longitude e g it could be a rotated grid The output can then be placed on the same rectilinear grid readFile logical Switch controlling location of output mapping Only used if pointsFlag 1 2 i e a mapping is to be input TRUE read from an external file FALSE read from the run control file Page 81 of 100 character 1 S M orA filename character The name of the file that contains output mapping Only used if readFile TRUE pointsOut integer The number of points to be output This is only used if 1 to size pointsFlag 1 2 of grid mapOut 1 poin integer A list of the points that are to be output The list gives the locations tsOut 1 array point numbers in the INPUT grid which need not be the same as the model grid Only used if pointsFlag 1 2 mapOut 1 poin integer A list giving the destination location in output grid for each output tsOut 2 array point The list gives the point number in the output grid Only used if pointsFlag 2 2 outGridNx integer Number of columns in the output grid This is the full uncompressed output grid If compression is applied the actual output may be smaller but can be scattered across a grid with this number of columns Only used if pointsFlag 2 2 in which case the user specifies all aspects of the output grid and mappings Otherwise the size of the output grid is
94. on and time However if 1 cosz FALSE the user might prefer to use Local Time particularly if this is used for input or validation data as the timestamp on model output will then match that on the other data 6 3 2 Examples of dates and times 1 A run without spin up 19970101 19970101 00 00 00 19990101 01 00 00 dateMainRun 19970102 0 dateSpin timeRun nspin This specifies a run from midnight on 1 January 1997 until 01 00 GMT on 1 January 1999 nspin 0 means there is no spin up 2 A run with spin up over a period that immediately precedes the main run 19970101 19960101 00 00 00 19990101 01 00 00 dateMainRun 19970101 5 dateSpin timeRun nspin This specifies a spin up period from midnight on 1 January 1996 to midnight on 1 January 1997 the time of day is taken from the first line This spin up will be repeated up to 5 times before the main run from midnight on 1 January 1997 until 01 00 GMT on 1 January 1999 3 A run with spin up over a period that overlaps the main run 19970101 19970101 00 00 00 19990101 01 00 00 dateMainRun 19980101 5 dateSpin timeRun nspin This specifies a spin up period from midnight on 1 January 1997 to midnight on 1 January 1998 the time of day is taken from the first line This spin up will be repeated up to 5 times before the main run f
95. ons This must be a factor of the output period outPer Special case 0 means sample every timestep In some cases sampling every timestep adds a considerable computational burden and acceptable output can be achieved by sampling less frequently For example with a large domain many output diagnostics and a timestep of 30 minutes a monthly average would be calculated from several hundred values if every timestep was used For variables that evolve relatively slowly an acceptable monthly average might be obtained by sampling only every 12 hours outDateStart integer Date in format yyyymmdd Output from this profile is first Many variables that are input in terms of seconds such as out Per and out FilePer are converted within the code to a number of model timesteps Page 79 of 100 generated at the date and time indicated by outDateStart and outTimeStart These must be within the main run except for the special cases noted below Note that output is only generated at the end of a timestep except for the special cases noted below Special cases for outDateStart 0 output all times through the run including any spin up 1 output at all times after spin up is complete 2 output only at the start of the first timestep of the run used to output the initial state only Note that at present the only time at which output can be generated at the start of a timestep is at the start of the
96. p www metoffice gov uk research hadleycentre pubs HCTN i ndex html Mercado L M et al 2007 Tellus 59B 553 565 ilayers integer gt l Number of layers for canopy radiation model Only used if can_rad_mod 2 or 3 These layers are used for the calculations of radiation interception and photosynthesis I GOSZ logical Switch for calculation of solar zenith angle For land points this switch is only relevant if 1 _ spec_albedo TRUE otherwise it is better set to FALSE to prevent unnecessary calculations TRUE calculate zenith angle FALSE assume constant zenith angle of zero meaning sun is directly overhead l spec _ albedo logical Switch for albedo model TRUE use spectral albedo This includes a prognostic snow albedo FALSE use a single averaged waveband albedo l phenol logical Switch for vegetation phenology model TRUE use phenology model FALSE do not use phenology model T trittid logical Switch for dynamic vegetation model TRIFFID TRUE use TRIFFID FALSE do not use TRIFFID i trif eg logical Switch for equilibrium vegetation model i e an equilibrium solution of TRIFFID This is only used if l triffid TRUE TRUE use equilibrium TRIFFID FALSE do not use equilibrium TRIFFID Page 21 of 100 yrevin logical Switch indicating if the order of the rows in the input data is not the JULES standard TRUE Input data a
97. pin up variable this is the maximum change over a repetition of spin up that is allowed if the model is to be considered as spun up If the absolute value of the change or the magnitude of the percentage change if spinTolPercent TRUE is less than or equal to spinTol the variable is considered to have spun up For example spinTol 0 1 means that the variable in question must change by less than 0 1 over a cycle of spin up if it is to be considered spun up Spin up is assessed using the difference between instantaneous values at the end of consecutive cycles of spin up For example if the spin up period is from Page 24 of 100 15 Jan 2005 to 15 Jan 2006 every time the model gets to 15 Jan 2006 the spin up variables are compared with their value at the end of the previous cycle phenol period integer Period for calls to phenology model days Only gt relevant if 1 phenol TRUE tritfid period integer Period for calls to TRIFFID model days Only gt relevant if one of L TRIFFID is TRUE 1 360 logical Switch indicating use of 360 day years TRUE each year consists of 360 days This is sometimes used for idealised experiments FALSE each year consists of 365 or 366 days 6 3 1 Note on time convention If a run requires that the solar zenith angle be calculated 1_cosz TRUE then times must be in Greenwich Mean Time UTC so that the code can calculate the zenith angle at each locati
98. plating is similar to that used by GrADS with a few important differences JULES only allows a subset of the GrADS substitution strings not including the ch string used with chsub but is more flexible in how it deals with time templating Page 86 of 100 The substitution template must be compatible with the period frequency of the data files If a substitution template includes a substitution string that refers to a period of a day or longer each file must contain data for no more than one period For example if Sm2 appears in the template each file must contain data from at most one calendar month For periods less than one day i e hours and minutes data for more than one period can be held in the same file but the file period must be a factor of one day The start time of each file must also follow slightly complicated rules that are laid out in Table 31 The rules ensure that the first data in a file represent the first time that the time templating expects to find in that file Essentially they require that each file holds all possible data for the time period there cannot be any missing times Some of these rules are demonstrated in the example section below If these rules are not followed the code will detect an error and stop In Table 31 dataPerUnits and filePerUnits are the time units that are used to describe the period of the data and the files respectively chosen from 1 year month days hours and minutes If
99. r a cycle and setting spinFail FALSE so that the integration continues when spin up is judged to have failed after nspin cycles Although it is expected that a spin up phase will be followed by the main run in the same integration it is possible to do the spin up and main run in separate integrations This can be done by demanding an impossible level of convergence by setting spinTol lt 0 setting spinFail TRUE so that the integration stops when spin up is judged to have failed and setting Page 26 of 100 dumpF req see Section 6 17 1 to any value that writes a final dump The final state of the model after nspin cycles of spin up will be written to the final dump and a subsequent simulation started from this dump A limitation of the current code is that it cannot cope with a spin up cycle that is short in comparison to the period of any input data For example a spin up cycle of 1 day cannot use 10 day vegetation data The code will likely run but the evolution of the vegetation data will probably not be what the user intended However it is unlikely that a user would want to try such a run Occasionally the model fails to diagnose a spun up state when in fact the integration has reached a quasi steady state that is not detected by the procedure of assessing spin up through comparison of instantaneous values at the end of consecutive cycles of spin up An example of this is period 2 behaviour where the model state repeats i
100. r time INIT NONVEG Set parameters for non vegetation tiles 6 10 INIT SNOW Set snow related parameters 6 11 INIT TRIF Set parameters for TRIFFID dynamic vegetation 6 12 model INIT AGRIC Set fraction of each gridbox that is agriculture for 6 13 use with TRIFFID INIT MISC Set miscellaneous carbon cycle parameters 6 14 INIT DRIVE Set input driving data options 6 15 INIT IC Set initial conditions of all prognostic variables 6 16 INIT OUT Set options for model output 6 17 NEWPROF Set up an output profile 61 The user can annotate the run control file for example to add comments but these must not interfere with the reading of the rest of the file Depending upon the details of the run there are various locations in which it is safe to include annotation but the only really safe location is on the lines immediately preceding a tag described above Annotation can also often be placed on the same line as the data at the end of any data field i e so that the code reads the values required and will not see the annotation Values of character variables such as file names should be enclosed within quotation marks either single or double Character variables have a maximum length specified in the code which are sometimes given in this documentation e g character 8 indicates a variable of length 8 Logical values can be entered in any of the formats understood by FORTRAN e g T true or
101. rch hadleycentre pubs HCTN index html Variable name Type and Notes permitted values rho snow real Density of lying snow kg m snow _hcap real Thermal capacity of lying snow J K m Typical value 0 3e6 snow_hcon real Thermal conductivity of lying snow W m K See HCTN30 Eq 42 Typical value 0 265 ro real Grain size for fresh snow um See HCTN30 Eq 15 A typical value is 50 0 Only used if 1 spec albedo TRUE rmax real Maximum snow grain size um See HCTN30 p4 A typical value 2000 0 Only used if 1 spec albedo TRUE snow ggr 1 3 real array Snow grain area growth rates um s Only used if l1_spec_albedo TRUE See HCTN30 Eq 16 The 3 values are for melting snow cold fresh snow and cold aged snow respectively Typical values are 0 6 0 06 0 23e6 amax 1 2 real array Maximum albedo for fresh snow Only used if l spec _ albedo TRUE Values 1 and 2 are for VIS and NIR wavebands respectively Typical values 0 98 0 7 Page 54 of 100 dtland real Degrees Celsius below zero at which snow albedo equals cold deep snow albedo This is 2 0 in HCTN30 Eq4 Only used if 1 spec albedo FALSE kland real Used in snow ageing effect on albedo This is 0 3 in HCTN30 Eq4 note the last term of that equation should be divided by dtland i e kland as specified here includes a factor dtland in the denominator Only used
102. re arranged in North to South order i e first data are from northernmost row FALSE Input data are arranged in South to North order the JULES standard Note that this does not affect how JULES numbers points on its internal grids within JULES the numbering always runs from South to North This switch applies to all input files echo logical Switch controlling output of messages to standard output e g screen TRUE print messages to screen This will print the values of parameters and also print messages when files are opened or closed This is useful while checking that a run is correctly set up but can result in a large volume of data if the model grid is large FALSE suppress printing of most messages to screen print step integer gt The number of timesteps in between the printing of timestep information Every print step timesteps the model prints the current timestep number and date to standard output While this can be a useful way to follow the progress of a model integration frequent messages can generate a large amount of unnecessary output during long integrations Page 22 of 100 6 3 INIT_TIME Date and time information gt INIT TIME timestep dateMainRun 1 dateSpin 1 2 spinFail spinFlag 1 spinFlag 2 nspin phenol period 1 360 spinTolPercent spinTolPercent timeRun 1 triffid period dateMainRun 2 1
103. rom midnight on 1 January 1997 until 01 00 GMT on 1 January 1999 Page 25 of 100 4 Example of specifying requirements for spin up T terminate run if spin up fails T F T Fy 1 0 soil moisture spinVar spinTolPercent spinTol Tee T Og Tsoil The first value soinFail TRUE means that if the spin up has not converged after nspin cycles the run will end Convergence is measured using moisture content and temperature of each soil layer At every point and in every layer soil moisture must change by less than 1 kg m 2 or mm of water while soil temperature must change by less than 0 1 6 3 3 Notes on spin up Note that at present the analysis of whether the model has spun up or not is limited to aspects of the physical state of the system and does not explicitly consider carbon stores making it less useful for runs with interactive vegetation TRIFFID During the spin up phase of a run the JULES code provides the correct driving data for example meteorological data as the model time cycles round over the spin up period Consider the case of a spin up over 1 Jan 2005 to 31 Dec 2005 At or near the end of 31 Dec 2005 during the spin up the driving data will start to adjust to the values for 1 Jan 2005 The calculated driving data may vary slightly between the start or end of the first cycle and similar times in later cycles because of the need to match the data at the end of each cyc
104. rt of a field See Section 5 2 varName character See Table 3 The name of the variable varFlag integer gt 1 Flag indicating how the variable is initialised Acceptable values gt 0 The field number in the file that holds data for this variable See discussion of fields in Section 5 1 If a dump file is being read any integer gt 0 is accepted and then effectively ignored this indicates that the field is to be taken from the dump and the exact field number is not required 1 The field will be set to the value constVal see below at all points This option can be used to specify an idealised initial condition constVal real The value to be used at all points Only used if flag 1 gt NC The following are used if fileF ormat nen ndim integer Number of dimensions in the netCDF file gt dimName 1 ndim character List of names of dimensions in the netCDF file array varName character The name of the variable See Table 5 varFlag integer Flag indicating how the variable is initialised Acceptable gt 1 values gt 0 Default effectively is ignored 1 The field will be set to the value constVal see below at all points This option can be used to specify an idealised initial condition constVal real See under gt ASCBIN above SDFvarName character The name of the netCDF variable that is to be used If further initial data are to be read from t
105. s easily edited it is easiest to list these variables in the order in which the data will be presented i e field numbers should be 1 2 3 In this example all the field numbers are gt 0 indicating that the data will be read from the gt DATA section and that the constVal entries will be ignored Note that data for soil variables are presented in the order top to bottom i e surface layer first Example 2 Initial state specified as a mixture of spatial fields and constant values Page 73 of 100 In this example we consider a run at a single point and read all data from the run control file The relevant entries in the run control file are shown below Only the lines in bold are relevant and irrelevant sections have been omitted gt INIT IC T readFile bin fileFormat quoted a001 initial state gra fileName quoted T 7 7 zrev gt ASCBIN 0 0 nheaderFile nheaderField gt VARS sthuf 7 0 9 varName varFlag constVal canopy 1 0 0 snow_tile 1 0 0 tstar_tile 1 275 0 t_soil 1 278 0 cs 1 10 0 gs 1 0 0 gt ENDVARS readFile TRUE indicates that the binary file a001_initial_state gra will be used to set the initial state for some variables The seven variables that are required to initialise this particular run are then listed The second entry in each line gives the position in the input data for each field For most variables the value 1 indicates that th
106. s in the input grid is thus nxIn nytIn If the input data consists of a single point nxIn nyIn 1 A vector of points is specified by setting nyIn 1 Although the notation may suggest a regular rectangular grid the model can be run at any number of arbitrary locations the most likely way of doing so being to set nyIn 1 nxIn number of points sm levels integer Number of soil layers gt A value of 4 is often used can_ model integer Choice of canopy model for vegetation 1 2 3 or 4 1 No canopy 2 Radiative canopy with no heat capacity 3 Radiative canopy with heat capacity 4 As 3 but with a representation of snow beneath the canopy e NB can_model 1 does not mean that there is no vegetation canopy It means that the surface is represented as a single entity rather than having distinct surface and canopy levels for the purposes of radiative processes Page 20 of 100 can rad mod integer 1 2or3 Switch for treatment of canopy radiation 1 Beer s law for light penetration and big leaf for scaling leaf to canopy scale photosynthesis 2 Multi layer approach for both radiation interception 2 stream approach of Sellers et al 1992 and canopy photosynthesis 3 As 2 but photosynthesis calculated separately for sunlit and shaded leaves References Sellers P et al 1992 Remote Sens Environ 42 187 216 Jogireddy V R et al 2006 Hadley Centre technical note 63 Available from htt
107. s of the vegetation that vary with time and or location in addition to varying with PFT gt INIT VEG VARY nvegVar vegDataPer vegUpdatePer nvegFileTime vegFilePer vegClim readList fileName vegFileDate 1 vegFileTime 1 vegEndTime fileFormat gt ASCBIN nfieldFile nheaderFile nheaderField noNewLineVeg varName 1 flag 1 fieldNumber 1 interp 1 nameFile 1 Repeated for each of nvegVar variables gt NC nvegDim vegDim 1 nvegDim varName 1 flag 1 interp 1 SDFname 1 nameFile 1 Repeated for each of nvegVar variables Table 18 Description of options in the INIT_VEG_VARY section Variable name Type and Notes permitted values nvegVar integer The number of prescribed characteristics that vary 0 lt nvegVar lt with time and or location The three characteristics 3 that may vary are vegetation height leaf area index and root depth If nvegVar 0 nothing more is read from this section vegDataPer integer The period s of time varying data If there are no time varying fields enter 0 Special cases 1 indicates monthly data VegUpdatePer integer The period s between updates of time varying fields This must be less than or equal to the data period vegDataPer For example Page 47 of 100 vegDataPer 86400 vegUpdatePer 3600 indicates that the data are daily values and these should be updated by interpolation on an ho
108. s used to specify the reading of netCDF format files 0 0 ee eeeeeeeeeceseeeeeecseeeenecaeeeeeneees 15 Table 6 Sections ina JULES control file oreinaren eini an A onto tages dae ols bez ent E as ab eiteat 16 Table 7 Description of options in INIT OPTS section cccceeeeecceseeseeececsescsesesesecsssssseseececsesssseseessesseeseceessseaes 18 Table 8 Description of the INIT TIME Section 2 ccccccscsssessssveeeecccnssssestvsnseccatessussvesesesseveussonsnstuveseseesansstdvevestsvedesesvens 22 Table 9 Description of options in the INIT GRID S CtiON 0 ccc cceeeseeecscscnseseeececsesssesesececsesssessesesecsesesenecsesasaees 27 Table 10 Description of options in the INIT LAND SeCtiON cccccecesesecsccsseseseeececsesesesesesecsesssessesececsesesessesesaees 29 Table 11 Description of options in the INIT LATLON Section 0 cccceccccsseeeseeececsesseseseessecsessssesesececsssesseseseceenees 30 Table 12 Description of options in the INIT FRAC S CtiON cccccceeseescscscssseesesececeessssesesececsesssessesececsesssessesesaeases 37 Table 13 Description of options in the INIT DZSOTL Section ccccccccccssseeseescecsseeseeceseessseesesesessesssesseseseseenees 39 Table 14 Description of options in the INIT SOIL and INIT SOIL2 sections 0 0 eee eeeeseceeeeecneeeeceseeeeeseeaeeeeeseeaees 40 Table 15 List of soil parameters neirinne aa a a e a a a aa e aa Eaa ea pesega 42 Table 16 Description of options in the INIT VEG PET Se
109. se is from timeRun 1 on dateMainRun 1 to timeRun 2 on dateMainRun 2 Examples are given below nspin integer gt 0 The maximum number of times the spin up period is to be repeated 0 no spin up gt 0 at least 1 and at most nspin repetitions of spin up are used After each repetition the model tests whether the selected variables have changed by more than a specified amount over the last repetition see below If the change is less than this amount the model is considered to have spun up and the model moves on to the main run spinFail logical Switch controlling behaviour at the end of spin up period if the model has not passed the spin up test Only used if nspin gt 0 TRUE End the run if model has not spun up FALSE Continue the run spinFlag 1 2 spinTolPercent 1 2 logical array logical array Switch indicating whether the variable is included in the decision as to when spin up is achieved TRUE include this variable in analysis of spin up FALSE exclude this variable The 2 elements refer to 1 moisture content of each soil layer kg m 2 temperature of each soil layer K Switch indicating whether the tolerance for this variable is expressed as a percentage TRUE tolerance is a percentage FALSE tolerance is an absolute value The elements refer to different variables see spinFlag spinTol 1 2 real array Tolerance for spin up For each s
110. sed in the netCDF file for the chosen variable Page 16 of 100 6 The JULES control file 6 1 Introduction Each run of the JULES code is controlled by a text file that is called the run control file Broadly speaking the run control file holds three types of information e the general details of the run such as start and end dates e the values for parameters of the model such as albedo e the specification of the required output The JULES code is designed to be moderately flexible in that there are switches that allow the user to select between different configurations and it can accommodate input data in several different file formats This flexibility means that the run control file may be relatively long and the user has to check that all values are set correctly The documentation below aims to help the user in this task Example input files can be found as described in Section 6 20 The run control file has a particular format in that the lines must be in a particular order and must contain various headers The file is read by various routines arranged under the subroutine INIT using FORTRAN list directed input i e the format is given as in a READ statement of the form READ unit The JULES executable is run with standard input redirected to this control file e g jules exe lt control file jin The use of list directed input means that there may be more than one arrangement of input values that
111. separately If JULES is to be driven by the output from such a model the driving data might include these components l point data logical Flag indicating if driving data are point or area average values This affects the treatment of precipitation input and how snow affects the albedo TRUE driving data are point data Precipitation is not distributed in space see FALSE below and is all assumed to be large scale in origin The albedo formulation is suitable for a point FALSE driving data are area averages The precipitation inputs are assumed to be exponentially distributed in space as in UMDP25 and can include convective and large scale components The albedo formulation is suitable for a gridbox tForSnow real gt 0 If ioPrecipType is 1 or 2 tForSnow is the near surface air temperature K at or below which the precipitation is assumed to be snowfall At higher temperatures all the precipitation is assumed to be liquid tForConv real gt 0 IfioPrecipType islor2 tForConv is the near surface air temperature K at or above which the precipitation is assumed to be convective in origin At lower temperatures all the precipitation is assumed to be large scale in origin Also see conFrac tForConv is not used if 1 point data is TRUE since then there is no convective precipitation tForSnow must be less than tForConv implying that all solid precipitation is large scale in origin
112. ss TRUE the output files will contain data only for the land points and a mapping is defined so that the land points can be plotted in their correct positions on the Earth This leads to considerable saving on disc space The data in the output file is written as a vector of 15000 points in this case in the order that they are held in the model grid The mapping is written to a binary file that contains 3 fields on the full expanded grid 360x180 points in this example starting from the southwest corner proceeding across each row then onto next row i e the default JULES order The first field is integer and gives the location in the output vector of 15000 points that should be plotted at this location in the globe If there are no data for a point i e a sea point in this case the missing data value is inserted The second field is real and is 1 0 at points with data elsewhere 0 0 The third field is not used by JULES it deals with wind rotation and will consist of the missing data value GrADS pdef option can be used to display just such a thinned grid i e the full grid is populated with values from the thinned grid with missing data values inserted at all other points Note that outCompress as implemented in JULES is a subset of the full pdef available in GrADS namely where pdef is used with a supplementary file and each point in the thinned output grid maps onto a single point in the full grid
113. start of a file at the start of each time within the file and at the start of each field Note that this means that JULES reads and writes data in terms of maps all values of one field then all values of another field rather than using a point by point data model all fields for one point then all fields for another point A related concept used in JULES is that of the point number in input or output files This is used to select individual points from a larger grid The point number runs from 1 at the gridpoint in the SW corner of the grid across rows so the bottommost row contains gridpoints 1 to nx and then from South to North up the grid Examples and further discussion of JULES grids can be found in Section 6 4 5 2 Describing the format of a file Variables that describe how data are arranged in files are used in several sections later in this document These variables are summarised in Table 2 Often the information that JULES will read and use from the control file depends on the file format of any one data file The information required for ASCII binary or PP files is generally fairly similar while netCDF files are rather different Table 2 Format of frequently used control file options Variable name Type Notes readFile logical Switch that indicates source of data TRUE data are read from a named external file FALSE data are read from the run control file fileFormat character Flag in
114. t of a JULES run by setting echo TRUE in INIT OPTS see Section 6 2 A variable may appear more than once in an output profile as long as each time it appears with a different time flag e g instantaneous and time average values useName character The name to be used in the output one word This variable need not be specified If useName is not provided the code will substitute name instead This facility allows the user to choose to call output variables by names other than those used in the code for example to use names that are more memorable or shorter names to avoid typing Although the name should be a single word characters such as underscore _ may be used 6 17 3 Compression of the output grid As noted above outCompress TRUE can be used to compress the output data so that any missing points are not written and file size is reduced Although this facility was designed to work with the pdef option of GrADS it might be useful with other packages too with the proviso that the user may have to tell another package how to use the available information This facility will be described by considering an example in which we have global input data on a 1 grid and JULES is run at land points only We would like to visualise the output plotted on the full globe The input grid is of size 360x180 64800 points of which only about 25 are land points at which the model is run If we set out Compre
115. t1 files A template ctl file allows GrADS to access several data files via one ctl file TRUE all suitable ctl files will use the template option FALSE generate a separate ctl file for each data file Page 76 of 100 Note A template ctl file will not work be able to describe the data if there are any missing times at the start of a file this is a limitation of the current JULES code rather than GrADS For example if daily data are to be written to monthly files with a template ctl but the run starts midway through the month JULES will only write output data for the latter part of the month GrADS will look for data for all days in the month but not be able to find them so the user will not be able to plot the first month nout integer The number of output profiles Each profile generates a separate stream of data as explained above undefOut real The value written to output files to represent missing or undefined data ZSMC real If a depth averaged soil moisture diagnostic is requested the average is calculated from the surface to this depth m Zst real If a depth averaged soil temperature diagnostic is requested the average is calculated from the surface to this depth m outEndian character Only used for GrADS output files little endian outFormat bin this describes the byte or big endian ordering of the computers on which JULES is run It is inclu
116. ta for 10 days more than one days year month day of data can be stored in each file Data_ h2 dat Hourly Yes Data_00 dat For substitution strings data each Data_06 dat that refer to hours or file Data_12 dat minutes more than one containing Data_18 dat hour or minute of data can data for 6 be stored in each file hours Data_ mc dat Hourly data Yes Data_jan dat in monthly Data_feb dat files The time of the first data is 00H 01Jun1990 Data_ mc dat Hourly data No Data_jan dat Similar to the previous in monthly Data_feb dat case but with first data files The one hour later In this case time of the the first data in each file first data is must represent 00H on the 01H 1 of a month These data 01Jun1990 cannot be described by a time template and instead the name and time of each data file must be listed see appropriate section Data_ oy4 dat Monthly Yes Data_1990 dat In this case the time of the data in Data_1991 dat first data must be in yearly files January Here it is shown The time of to be a vale at the first approximately mid month data is given as 00H 15Jan1990 6 19 Notes on temporal interpolation Page 88 of 100 Time varying input data to JULES require the user to specify how the data should be interpolated onto the model timestep The permitted interpolation flags are shown in Table 33 These flags are case sensitive Table 33 Time interpolation flags
117. than rootd_ ft See HCTN30 Eq 32 snowCanPFT Flag indicating whether snow can be held under the canopy of each PFT Only used if can_model 4 see Section 6 2 The model of snow under the canopy is currently only suitable for coniferous trees Acceptable values are 0 snow cannot be held under the canopy 1 snow can be held under the canopy albsnc_max Snow covered albedo for large leaf area index used if l spec albedo FALSE See HCTN30 Eq 2 Only albsnc min Snow covered albedo for zero leaf area index Only used if 1 spec albedo FALSE See HCTN30 Eq 2 albsnf max Snow free albedo for large LAI Page 45 of 100 Only used if 1 spec albedo FALSE See HCTN30 Eq 1 kext Light extinction coefficient used with Beer s Law for light absorption through tile canopies See HCTN30 Eq 3 kpar PAR Extinction coefficient m leaf m ground orient Flag indicating leaf angle distribution 0 spherical 1 horizontal alpha Quantum efficiency mol CO per mol PAR photons alnir Leaf reflection coefficient for NIR HCTN30 Table 3 alpar Leaf reflection coefficient for VIS HCTN30 Table 3 omega Leaf scattering coefficient for PAR omnir Leaf scattering coefficient for NIR a wl Allometric coefficient relating the target woody biomass to the l
118. the file that holds data for the first gt level of this variable See discussion of fields in Section 5 1 Page 30 of 100 nLandDim integer Land mask data in a SDF is held in an array of gt nLandDim dimensions landDim 1 nlan character array Names of land mask dimension variables in SDF dDim varNameLand character array The name of the variable containing the land fraction 6 4 3 INIT_LATLON Latitude and longitude gt INIT_LATLON regLatLon regLatl regLonl regDlat regDlon readFileLatLon fileFormatLatLon fileNameLatLon gt ASCBIN nheaderFile nheaderField fieldLat fieldLon gt NC nLatLonDim latLonDim varNameLat varNameLon gt DATA_ POINTS pointList 1l npoints gt DATA_ LAND E gt landg l nxIn 1 nyIn DATA _LATLON latitude 1 nxIn 1 nyIn longitude 1 nxIn 1 nyIn Table 11 Description of options in the INIT_LATLON section Variable name Type and Notes permitted values regLatLon logical Switch indicating if the input grid is regular and will be described by origin and increment or if latitude and longitude fields are to be read TRUE the grid is regular and can be specified by its origin and gridbox size There is then no need to read lat lon values for each gridpoint FALSE read latitude and longitude values for each gridpoint regLatl real The latitude decimal degre
119. tsList TRUE TRUE read from an external ASCII file FALSE read from the run control file Points are specified at the sub section marked gt DATA POINTS see Section 6 4 3 fileNamePoints character Name of file containing list of points Only used if Page 29 of 100 pointsList TRUE 6 4 2 INIT_LAND Land fraction This section describes how the land fraction field is set Land fraction describes the fraction of each gridbox that is land gt INIT LAND readFileLand fileFormatLand fileNameLand gt ASCBIN nheaderFileLand nheaderFieldLand fieldLand gt NC nlandDim landDim 1 nlandDim varNameLand Table 10 Description of options in the INIT_LAND section Variable name Type and Notes permitted values readFileLand logical Switch controlling source of land fraction data Only used if pointsList FALSE TRUE read from an external file FALSE read from the run control file at the section marked gt DATA LAND see Section 6 4 3 fileFormatLand character Format of file containing land fraction data See Section 5 2 fileNameLand character Name of file containing land fraction data nheaderFile integer The number of headers at the start of the land fraction gt 0 file See Section 5 2 nheaderField integer The number of headers before each field in the land gt 0 fraction file See Section 5 2 fieldLand Integer The field number in
120. tself over a period of 2 cycles Such behaviour should be apparent in the model output during spin up and the user can opt to repeat the integration over a given number of spin up cycles and not to wait for a spun up state to be diagnosed Page 27 of 100 6 4 Grid description The process of setting up the model grid involves three parts of the run control file INIT GRID INIT LAND and INIT LATLON INIT GRID is used to select how the model grid will be specified e g all points within a given range of latitude and longitude INIT LAND is used to set a land sea mask INIT LATLON specifies the latitude and longitude of each point These three sections are then followed by the DATA POINTS DATA LAND and DATA LATLON sections which provide input data if that is to come from the run control file Each run of JULES involves two grids the input grid and the model grid The input grid is the grid on which all input data are held The model grid is the set of points on which the model is run The model grid is a subset of the input grid The JULES grid is a rectangle of size nx ny points including the special case of ny 1 when the grid is a vector of points The points to be modelled may be selected from a larger input grid by specifying one or more of 1 a list of point numbers 2 a range of latitude and longitude 3 that only land points are to be selected The grid may contain both land and sea points altho
121. ugh sea points can be omitted A vector of points can be used to select locations that are not adjacent in the real world for example one might only want to run the model at locations within a catchment for which observations are available In this case although the model could be run on a grid that included the whole catchment it is more efficient to run only at the selected points 6 4 1 INIT_GRID setting up the grid gt INIT GRID pointsList landOnly subArea subAreaLatLon xcoord 1 2 ycoord 1 2 npoints readFilePoints fileNamePoints Table 9 Description of options in the INIT_GRID section Variable name Type and Notes permitted values pointsList logical Switch indicating whether the model grid is to be specified as a list of points TRUE Points to be modelled are selected from the input grid via a list of points In this case the points to be modelled are selected via a list of point numbers see Page 28 of 100 landOnly logical below Land fraction is not read FALSE All points in the input grid are modelled subject to elimination by subArea or landOnly see below Land fraction is read Land points to be modelled are indicated by land fraction gt 0 sea points by land fraction lt 0 A sub area can be selected see below The value of npoints is set by the model and equals the number of land points Although this field was originally intended to be a lan
122. urly basis Special cases 0 update every timestep 1 update once a month nvegFileTime integer gt 1 The number of data files available for each variable each file holding data for different times If all variables are held together this is the number of data files If variables are held in separate files this is the number of files for any one variable vegFilePer integer The period s of the files containing the data This must be at least as large as the period of the data vegDataPer and must be a multiple of the model timestep Special cases 1 monthly files 2 annual files vegClim logical Switch indicating if time varying vegetation data are to be treated as climatological in the sense that the same data are to be used regardless of the year TRUE data are climatological The year given for each file is ignored FALSE data are not climatological readList logical Switch controlling how the names of the files containing the vegetation data and the times covered by each are read TRUE a list of names and times is read from another file This is required if nvegFileTime gt l FALSE a single name and file are read from the run control file This option is only allowed if nvegFileTime 1 see above filename If nvegFileTime 1 this is the name of the single data file or the template If nvegFileTime gt 1 this is the name of a file that lists the names and
123. writing a model dump Acceptable values are new if a file with the same name already exists the run will terminate replace if a file with the same name already exists it will be overwritten 6 17 2 NEWPROF details of each output profile Each of the nout output profiles requires a section that describes that profile such as the times when output is to be generated which points are to be output which variables are to be output and more The size of a regular latitude longitude gridbox input as regDlat regDlon in control file see Section 6 4 3 is also used as the size of a gridbox in the output gt NEWPROF utName utPer outFilePer tSamPer utDate 1 outTime 1 utDate 2 outTime 2 OO000 0 pointsFlag 1 2 outAreaLL outRangeX 1 2 outRangeY 1 2 outCompress outLLorder readFile fileName pointsOut mapOut 1l pointsOut 1 mapOut 1l pointsOut 2 gt GRID outGridNx outGridNy gt VARS flag name useName repeat for each output variable gt ENDVARS Page 78 of 100 Table 29 Description of options for each output profile Variable name Type and permitte d values Notes outName character The name of this output profile This is used in file names and should be specified even if there is only one profile The names might reflect the variables in the file e g soil or
124. x for land tiles kg m s EELT Tile surface sensible heat flux for land tiles W m ger Tile surface conductance to evaporation for land tiles m s leT Tile surface latent heat flux for land tiles W m qlp5mT Tile specific humidity at 1 5m over land tiles kg kg radnetT Tile surface net radiation W m rgraintT Tile snow grain size um snowCanT Tile lying snow on canopy kg m snowCanMeltT Tile melt of snow on canopy kg m s snowGroundMeltT snowGroundtT Tile melt of snow under canopy kg ms Tile snow below canopy kg m snowMassT Tile lying snow total kg m snowMeltT Tile total snow melt kg m s tlp5mT Tile temperature at 1 5m over land tiles K tstarT Tile surface temperature K ZOT Tile surface roughness m Table 39 A list of output variables that have a single value for each tile type at each land gridpoint Name Description frac Fractional cover of each surface type tileIndex Index gridbox number of land points with each surface type Page 99 of 100 Table 40 A list of output variables that have a single value for each soil level at each land gridpoint Name Description bSoil Clapp Hornberger exponent for each soil layer ext Extraction of water from each soil layer kg m s1 hCapSoil Soil heat capacity J K m for each soil layer hConSoi
125. x total soil moisture in column kg m snowCan Gridbox snow on canopy kg m snomltSubHtf Gridbox sub canopy snowmelt heat flux W m snowF rac Gridbox snow covered fraction of land points snowGround Gridbox snow below canopy kg m snowMelt Gridbox rate of snowmelt kg m s soilIndex Index gridbox number of soil points subSurfRoff Gridbox sub surface runoff kg m s surfRoff Gridbox surface runoff kg m s surfRoffInf Gridbox infiltration excess surface runoff kg m s1 swetLiqTot Gridbox unfrozen soil moisture as fraction of saturation swetTot Gridbox soil moisture as fraction of saturation Tfall Gridbox throughfall kg m s1 Trad Gridbox effective radiative temperature K assuming an emissivity of 1 wFluxSfec Gridbox downwards moisture flux at soil surface kg m s Page 97 of 100 Table 37 A list of output variables that have a single value for each PFT at each land gridpoint Name Description cVegP PFT total carbon content of the vegetation kg C m canhtP PFT canopy height m ciP PFT internal CO2 pressure Pa fsmcP PFT soil moisture availability factor gLeafP PFT leaf turnover rate 360days gLeafDayP PFT mean leaf turnover rate for input to PHENOL 360days gLeafDrOutP PFT mean leaf turnover rate for driving TRIFFID
126. ying slowest For example say we have a file with two variables A and B on a grid with nx 2 ny 2 A has nz 1 and B has nz 2 In the JULES GrADS model the data must be stored in the input file in the order A x 1 y 1 z 1 t 1 1st xy plane of A at t 1 A x 2 y 1 z 1 t 1 A x 1 y 2 z 1 t 1 A x 2 y 2 Z 1 t 1 B x 1 y 1 z 1 t 1 lst xy plane of B at t 1 B x 2 y 1 z 1 t 1 B x 1 y 2 z 1 t 1 B x 2 y 2 Z 1 t 1 B x 1 y 1 z 2 t 1 2 xy plane of B at t 1 B x 2 y 1 z 2 t 1 B x 1 y 2 2 2 t 1 B x 2 y 2 Z 2 t 1 A x 1 y 1 z 1 t 2 lst xy plane of A at t 2 In brief input via netCDF has been enabled for vector input from files with certain netCDF dimensions This allows GSWP2 data held at CEH to be used A user who is familiar with the netCDF library will be able to modify the existing code to read other netCDF files 4 The GrADS software can be downloaded for no cost from http grads iges org grads gadoc index html Page 13 of 100 A x 2 y 1 z 1 t 2 SweLCy nd For clarity this example has shown each datum on a separate line but in fact any number of data within a single field see below can be on the same line A data field is considered to be a single x y plane of data i e nx ny values Header records can be present at the

JULES v2.0 User Guide - NERC Open Research Archive

Contents

Download Pdf Manuals

Related Search

Related Contents