Home

SeaWinds Product User Manual

1. ooicccinicicicnnccnccnnccnnnnnnnnnnnnenininonononenoninnnnnnninnnnnos PAPAT Ra pro ciar dio 2 3 Seaice detection cocccccccncccnnnnnnnennnnnnnnnnnnnonnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnrnnonennnneneneneness S Ale SULU AA angst ethene eae ne eeraa 3 1 Wind definition esas cebu an aig AABANG GALA Gs 3 1 1 Geophysical Model Function Xa ana AA 32 Wind Retrieval a GG GAGA Ah Laan habi Gd 3 2 1 Ambiguity Remoyal sabah n eese araea ad lanang adan 3 2 2 Q alty Control errian lor AEE E EA aE nian 3 3 Detailed SeaWinds algorithm descriptiON ooooooicccnnnnnnnnnccconccocnnnncnonananannnnnnnnncnnnnns S4 MONO NO cintia 4 PROCESSING AA redada 47 NWP COMOCALION iS A a 42 Maldad aa 4 3 Quality control and monitoring civic bo date ata AAs Sohware Configuratio Ni ieii A A bi Blanes 4 5 Dissemination and archivVe coccccccccncccnncnoncnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnineninnniss A BE raan aa E A aah Ana 5 1 Wind product characteristics 1 1 cnc 5 2 FIG TO MaS lan ic DAA QUIN e toba 61 a Relais dd 6 1 1 Usable 100 km WVC tas 6 2 Accuracy A inesi eee ea Aa 6 3 Acc racy 25 kr PIA UG Kina naban eie aea etA deni Maasahan NA Naan b4 Ambig ity SCISCUION sancion ep GAAN a aa ALAN 6 5 Quality indices and data use cuina A ida E GLOSSY E T E E 9 Literature ui E a a a a E 9 SeaWinds BUFR data descriptorS ooooooconnnnnnnnnnnocococcconcnoc nn conc 10 NetCDFEdata for
2. extreme sub footprint wind or wave variability and rain Concerning the latter point we compared our QC with the JPL rain flag and found Despite the fact that the JPL rain flag is set KNMI accepts a 2 category of high wind speed data However SSMI shows no rain in this category in about 85 of cases even if these winds are often found in the storm track region in the vicinity of intense low pressure systems We regard these SeaWinds data of great meteorological interest In the sweet swath region see figure 1 the QC is superior and sufficient but in the nadir part of the swath we use both the KNMI QC and the JPL rain flag 6 1 1 Usable 100 km WVC A 100 km WVC is used if for all four beams at least eight of sixteen backscatter values are present This choice was verified to provide satisfactory wind quality and quantity The front cover shows an example where close to the low and close to the cold front some WVC are rejected red dots The only further QC at 100 km resolution is performed during AR section 6 3 6 2 Accuracy 100 km product Figure 2 shows that SeaWinds wind variability at 25 km resolution is WVC number dependent This is due to the effect of instrument noise dominating the wind retrieval Obviously we would like the true wind field to dominate the wind variability and thus a variability which is independent of WVC At KNMI we chose to reduce the instrumental noise by spatial averaging and developed a 100
3. model speed units m s 1 short model dir NUMROWS NUMCELLS model dir long name model wind direction at 10m model_dir units degree short ice prob NUMROWS NUMCELLS ice prob long name ice probability ice prob units 1 short ice age NUMROWS NUMCELLS ice age long name ice age a parameter O amp SISAF SeaWinds Product User Manual version 1 6 August 2009 31 SAF OSI KNMI TEC MA 134 ice_age units dB int wvc_quality flag NUMROWS NUMCELLS wvc_quality flag long name wind vector cell quality wvc_ quality flag coordinates lat lon wvc_quality flag flag masks 64 128 256 512 1024 2048 4096 8192 16384 32768 65536 131072 262144 524288 1048576 2097152 4194304 wvc quality flag flag meanings distance to gmf too large data are redundant no meteorological background used rain detected rain flag not usable small wind less than or equal to 3 ms large wind greater than 30 m s wind inversion not successful some portion of wvc is over ice some portion of wvc is over land variational quality control fails knmi quality control fails product monitoring event flag product monitoring not used any beam noise content above threshold poor azimuth diversity not enough good sigma0 for wind retrieval short wind speed NUMROWS NUMCELLS wind speed long name wind speed at 10 m wind speed units m s 1 hort wind dir NUMROWS NUMCELLS wind
4. 06034 Cross Track Cell Number Numeric 021 21109 Seawinds Wind Vector Cell Quality Flag Flag Table 022 11081 Model Wind Direction At 10 M Degree True 023 11082 Model Wind Speed At 10 M m s 024 21101 Number of Vector Ambiguities Numeric 025 21102 Index of Selected Wind Vector Numeric 026 21103 Total Number of Sigma0 Measurements Numeric 027 21120 Seawinds Probability of Rain Numeric 028 21121 Seawinds NOF Rain Index Numeric 029 13055 Intensity Of Precipitation kg m 2 sec 030 21122 Attenuation Correction On Sigma 0 from Tb dB 031 11012 Wind Speed At 10 M m s 032 11052 Formal Uncertainty In Wind Speed m s 033 11011 Wind Direction At 10 M Degree True 034 11053 Formal Uncertainty In Wind Direction Degree True 035 21104 Likelihood Computed for Wind Solution Numeric 036 11012 Wind Speed At 10 M m s 037 11052 Formal Uncertainty In Wind Speed m s 038 11011 Wind Direction At 10 M Degree True 039 11053 Formal Uncertainty In Wind Direction Degree True 040 21104 Likelihood Computed for Wind Solution Numeric 041 11012 Wind Speed At 10 M m s 042 11052 Formal Uncertainty In Wind Speed m s 043 11011 Wind Direction At 10 M Degree True 044 11053 Formal Uncertainty In Wind Direction Degree True 045 21104 Likelihood Computed for Wind Solution Numeric 046 11012 Wind Speed At 10 M m s 047 11052 Formal Uncertainty In Wind Speed m s 048 11011 Wind Direction At 10 M Degree True 049 11053 Formal Unce
5. L 2 180 4 g lof oe CAR gt Lb E 3 5L A E 90 J amp Ls gt 0 0 Zi 0 5 10 15 20 0 90 180 270 36 Wind speed m s ECMWF Wind dir deg ECMWF N 313672 N 257964 mx 7 72 my 7 99 mx 180 49 my 182 49 m y x 0 27 s y x 1 39 m y x 1 99 s y x 15 72 cor_xy 0 92 cor_xy 0 99 Wind Histograms Wind Histograms _ 20t 3 _ 20 7 z z 4 106 1M 10t 3 ge 4 j CE f E 106 E iot o o o o gt 20 F ie 20 F 20 10 0 10 20 20 10 0 10 20 U comp m s ECMWF V comp m s ECMWF N 313672 N 313672 mx 0 41 my 0 17 mx 0 46 my 0 51 m y x 0 24 s y x 1 84 m y x 0 05 s y x 1 70 cor_xy 0 96 cor_xy 0 95 Figure 8 Contoured histograms of the 100 km OSI SAF wind product Both figures 8 and 9 give the wind speed direction and component statistics referenced against collocated ECMWF 10m wind analysis for the 100 km OSI SAF wind product and the 25 km NOAA product thinned to 100 km resolution respectively The data for these plots are from consecutive orbits from January 27 2002 until February 3 2002 The 100 km wind product was made with the QDP in default settings mode Variational Quality Control and KNMI quality control The thinning of the NOAA product was realized by taking winds from the nearest neighbour 25 km WVC closest to the centroid of the 100km super WVC The wind direction statistics are based on winds with wind speeds larger than 4 m s From the joint distributions and accompanying statistics in all plot
6. Surface Ss Back scattered component Slightly rough surface Rough surface 3 1 Wind definition A scatterometer measurement relates to the ocean surface roughness see figure 3 while the scatterometer product is represented by the wind at 10m height over a wind vector cell WVC It is important to realize that in the approach followed here the radar backscatter measurement oo is related to the wind at 10 meter height above the ocean surface simply O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 8 SAF OSI KNMI TEC MA 134 because such measurements are widely available for validation This means that any effect that relates to the mean wind vector at 10 meter height is incorporated in the backscatter to wind relationship As such air stability the appearance of surface slicks and the amplitude of gravity or longer ocean waves depend to some degree on the strength of the wind and may to the same degree be fitted by a geophysical model function GMF Ref 6 Chapter 1 Stoffelen Ref 6 Chapter IV discusses a unique method to determine the accuracy of scatterometer buoy and NWP model winds 3 1 1 Geophysical Model Function For the OSI SAF SeaWinds product the NSCAT 2 geophysical model function GMF for calculating equivalent neutral winds is used Wentz and Smith 1999 Portabella Ref 5 page 153 compares the QSCAT 1 Freilich et al 2002 and NSCAT 2 Ku band GMFs and found that the latt
7. a more general way forward to improve AR which is followed up during IOP by tuning 2D VAR For SeaWinds this tuned version of 2D VAR is used For the 25 km wind product a somewhat different approach called MSS Multiple Solution Scheme is used In this approach the full circle of possible wind directions is divided in 144 equal sectors of 2 5 and for each wind direction an optimal wind speed with its corresponding probability is calculated in the wind retrieval step Subsequently the 144 solutions of each WVC are used in the AR to select the optimal solution In this way the full probability density function of the wind solutions is used Portabella and Stoffelen 2004 3 2 2 Quality Control Since the scatterometer wind retrieval problem is over determined this opens up the possibility of quality control QC by checking the inversion residual J The inversion residual is in theory inversely proportional to the log probability that a node is affected solely by a uniform wind Generally this probability is low when rain affects the WVC or there is substantial wind variability within the cell As such Portabella and Stoffelen 2002 found that the inversion residual is well capable of removing cases with extreme wind variability at fronts or centres of lows or with other geophysical variables affecting the radar backscatter such as rain QC is performed on the 25 km NASA or NOAA grids and rejection percentages vary between 1 5 In th
8. dir long name wind direction at 10 m n wind dir units degree hort bs distance NUMROWS NUMCELLS bs_distance long name backscatter distance n bs_distance units 1 global attributes title QuikSCAT SeaWinds Level 2 100 0 km Ocean Surface Wind Vector Product title short name SeaWinds L2 100km Conventions CF 1 4 institution EUMETSAT OSI SAF KNMI source QuikSCAT SeaWinds software identification level 1 0 instrument calibration version 0 software identification wind 0 pixel size on horizontal 1 100 0 km service type N A processing type O contents ovw granule name QS100 D08113 80827 E0956 B4604546 nc processing level L2 orbit number 46045 start date 2008 04 22 start time 08 28 05 stop date 2008 04 22 stop time 09 56 57 equator crossing longitude 332 828 equator crossing date 2008 04 22 equator crossing time 07 55 23 rev orbit period 6059 1 orbit inclination 98 6 history N A references SeaWinds Product Manual http www osi saf org http www knmi nl scatterometer comment Orbit period and inclination are constant values All wind directions in oceanographic convention 0 deg flowing North creation_date 2009 08 26 creation_time 11 47 52 O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 32
9. is monitored and the monitoring value is set to zero the product is valid otherwise it is rejected by the monitoring The monitoring flag and value are the same for all WVCs in one BUFR output file If the KNMI JPL QC flag is set in a 25 km WVC then the backscatter information is not useable for various geophysical reasons like rain confused sea state etc and the corresponding 100 km or 25 km WVC bit is set JPL rain flag information is incorporated in this flag for the nadir swath Wind Vector Cells Moreover this flag is set when the wind speed in one or more of the wind solutions is 50 m s The presence of such wind solutions has proven to be a reliable indicator for sea ice in the Wind Vector Cell WVCs in which the KNMI JPL QC flag is set are not used in the calculation of the analysis field in the ambiguity removal step However after the ambiguity removal the wind solution closest to the analysis field is chosen if wind solutions are present in the WVC This means that such a WVC may contain a selected wind solution but it is suspect Land Ice presence flag is set if at least one of the 25 km WVCs in a 100 km super WVC is flagged as containing land ice in the NOAA NRT BUFR product This may be a flag in one of the sigma0 surface flags or in the WVC quality flag of the input If more than half of the 25 km WVCs in a 100 km super WVC has bits 6 8 or 9 set the 100km WVC is rejected no wind is calculated If the Variational QC flag
10. is set the wind vector in the WVC is rejected during ambiguity removal due to spatial inconsistency For recommendations on the use of the Quality Flag see section 6 5 O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 17 SAF OSI KNMI TEC MA 134 6 Data Quality NOAA provides a near real time wind product that is used as basis for further processing at KNMI Our SeaWinds products are different from the NASA and NOAA products since special emphasis is put on increasing the reliability of the wind vectors by Quality Control and rejection of rain contaminated Wind Vector Cells improving accuracy by noise reduction through spatial averaging 100 km product and meteorological filtering Guaranteeing correctness by monitoring of the operational SeaWinds data and KNMI processing and Reduce wind direction selection errors by meteorologically balanced wind direction Ambiguity Removal AR These wind quality issues are elaborated here 6 1 Reliability QC improves the quality of winds however at some rejection rate useful wind information may be lost As such QC is a compromise between rejection of useful data and accepting inferior data quality Ref 5 chapter 2 describes a method developed at KNMI for QC of SeaWinds data at 25 km resolution based on similar methods used for the ERS and NSCAT scatterometers We show that these methods also work in case of SeaWinds for the rejection of cases with
11. km SeaWinds product O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 18 SAF OSI KNMI TEC MA 134 GSC AT 20021002 11 137 HIRLAM 2002 10020913 IR 12 IR 12 at LAT LON 15 amp ee E b F l y a F i 5 pp j k h j PL k k MI O LET Figure 7 OSI SAF SeaWinds product in the Gulf of Mexico for a tropical cyclone A handful of points are rejected by our QC but the cyclone structure remains clear Table 1 OSI SAF and NOAA DIRTH SeaWinds product Speed difference with ECMWF first guess winds for a set of triple collocated wind points Speed S D Direction 13 58 direction and wind component U difference standard deviations SD are shown OSI SAF winds verify better against ECMWF in particular for wind speed The 100 km resolution wind product is of reduced resolution compared to the original NOAA 25 km product which is obtained by means of oo averaging The product is more consistent with NWP model winds than the original product which makes it more suitable for use in NWP models Evidence of this is given in table 1 and figures 8 and 9 As can be seen in table 1 the 100 km product clearly compares better to the independent ECMWF winds than the NOAA SeaWinds Product O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 19 SAF OSI KNMI TEC MA 134 Wind Histograms Wind Histograms 2 201 15 Z L J pa 270 gt 15F J E
12. 2 Latitude Coarse Accuracy Degree 076 06002 Longitude Coarse Accuracy Degree 077 21118 Attenuation Correction On Sigma 0 dB 078 02112 Radar Look Azimuth Angle Degree 079 02111 Radar Incidence Angle Degree 080 02104 Antenna Polarisation Code Table 081 21105 Normalized Radar Cross Section dB 082 21106 Kp Variance Coefficient Alpha Numeric 083 21107 Kp Variance Coefficient Beta Numeric 084 21114 Kp Variance Coefficient Gamma dB 085 21115 Seawinds Sigma 0 Quality Flag Flag Table 086 21116 Seawinds Sigma 0 Mode Flag Flag Table 087 08018 Seawinds Land Ice Surface Flag Flag Table 088 21117 Sigma 0 Variance Quality Control Numeric 089 21112 Number of Inner Beam Sigma0 aft of sat Numeric 090 05002 Latitude Coarse Accuracy Degree 091 06002 Longitude Coarse Accuracy Degree 092 21118 Attenuation Correction On Sigma 0 dB 093 02112 Radar Look Azimuth Angle Degree 094 02111 Radar Incidence Angle Degree 095 02104 Antenna Polarisation Code Table 096 21105 Normalized Radar Cross Section dB 097 21106 Kp Variance Coefficient Alpha Numeric 098 21107 Kp Variance Coefficient Beta Numeric 099 21114 Kp Variance Coefficient Gamma dB 100 21115 Seawinds Sigma 0 Quality Flag Flag Table 101 21116 Seawinds Sigma 0 Mode Flag Flag Table 102 08018 Seawinds Land Ice Surface Flag Flag Table 103 21117 Sigma 0 Variance Quality Control Numeric 104 21113 Number of Outer Beam Sigma0 aft of sat
13. Na KNMI SeaWinds Product User Manual Ocean and Sea Ice SAF Version 1 6 August 2009 SAF OSI KNMI TEC MA 134 DOCUMENT SIGNATURE TABLE Prepared by O amp SI SAF Project Aug 2009 Team Approved by O amp SI SAF Project Aug 2009 Manager DOCUMENTATION CHANGE RECORD Issue Revision Date Change Description Version 0 9 Draft version Version 0 99 QC flag correction Chapter 3 from UM incorporated Version 1 0 Version taking in account ORR 1 B RIDO37 www osi saf org mentioned in 81 1 overview Version 1 1 Adaptation to SDP included file name convention 2 Adapted web location of Ref 6 3 Use of ECMWF NWP data Version 1 4 Introduction of 25 km product Po 6 Version 1 Version 1 Version 1 Version 1 Version prepared for 25 km SeaWinds DRI NetCDF product format included KNMI de Bilt the Netherlands Reference SAF OSI KNMI TEC MA 134 O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 1 SAF OSI KNMI TEC MA 134 CONTENTS T IDO NA Aid A O AG BD ANAN NAG 1 2 Purpose and use of 100 km and 25 km products occcccccnncnnncnncinnnononennnennceninininnss 1 3 Access to data and helpdesk 2000000000000maaaaaaaaaaaaaaaannannaasnnasnasssnsssnsssnsssnsns 1 4 DISCIAIME ix A A a BANG 15 SRETEPENCES ALAN NAKASAAD KAG AL ede e a a e AAS 2 SeaWinds Instruments sssrinin cece cece cece eee a a a a a Eia aap era iaa oripa 2 1 SeaWinds I and SeaWinds Il
14. Numeric 105 05002 Latitude Coarse Accuracy Degree 106 06002 Longitude Coarse Accuracy Degree O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 29 SAF OSI KNMI TEC MA 134 Number Descriptor Parameter Unit 107 21118 Attenuation Correction On Sigma 0 dB 108 02112 Radar Look Azimuth Angle Degree 109 02111 Radar Incidence Angle Degree 110 02104 Antenna Polarisation Code Table 111 21105 Normalized Radar Cross Section dB 112 21106 Kp Variance Coefficient Alpha Numeric 113 21107 Kp Variance Coefficient Beta Numeric 114 21114 Kp Variance Coefficient Gamma dB 115 21115 Seawinds Sigma 0 Quality Flag Flag Table 116 21116 Seawinds Sigma 0 Mode Flag Flag Table 117 08018 Seawinds Land Ice Surface Flag Flag Table 118 21117 Sigma 0 Variance Quality Control Numeric O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 30 SAF OSI KNMI TEC MA 134 10 NetCDF data format The SeaWinds wind products are also available in the NetCDF format with the following characteristics The data format meets the NetCDF Climate and Forecast Metadata Convention version 1 4 http cf pcmdi lInI gov The data contain contrary to the BUFR data only Level 2 wind and sea ice information no sigma0 nor soil moisture information The aim was to create a compact and easy to handle product for oceanographic and climatological users The data contain only the selected wind solutions no ambig
15. ace in a circular pattern Due to the conical scanning a WVC is generally viewed when looking forward fore and a second time when looking aft As such up to four measurement classes called beam here emerge HH fore HH aft VV fore and VV aft in each wind vector cell WVC The 1800 km wide swath covers 90 of the ocean surface in 24 hours and represents a substantial O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 5 SAF OSI KNMI TEC MA 134 improvement compared to the side looking scatterometers where the largest coverage given by NSCAT is only half of SeaWinds coverage i e 90 of the ocean surface within 48 hours On the other hand the wind retrieval from SeaWinds data is not trivial In contrast with the side looking scatterometers the number of measurements and the beam azimuth angles vary with the sub satellite cross track location see figure 1 A detailed discussion is provided in Ref 5 pages 22 23 The wind retrieval skill will therefore depend on the position in the swath as illustrated in figure 2 QSCAT R1 N Trop KA 7 NG eh YAY Kai ma en pA Wen or 220 230 235 240 245 10m a Figure 2 SeaWinds product processed on a 25 km grid The wind retrieval on the grid results in four solutions of which one is selected by the ambiguity removal procedure Note the varying noise properties in the parts of the swath indicated in Figure 1 The blue patches are flagged as rain contamin
16. ame convention for SeaWinds 100 km BUFR data is QS100 Dyyddd Shhmm Ehhmm Bxxxxxxx QS100 is a fixed prefix denoting QuikSCAT and 100 km resolution Dyyddd_ yyddd contain year two digits and day of year 001 366 at start of data acquisition in this file GMT Shhmm_ hhmm contains hour and minute of start of data acquisition in this file GMT Ehhmm_ hhmm contains hour and minute of end of data acquisition in this file GMT Bxxxxxxx contains information about satellite orbit number Examples of file names are QS100 D05186 S2352 E0107 B3148182 or QS100_D05187_S1123_E1251_B3148889 The file name convention is applicable to the files which are available on the KNMI FTP server The files which are disseminated through EUMETCast have a fixed prefix S OSI KNMI before this file name for example S OSI_ KNMI QS100_D05186_S2352_E0107_B3148182 The file name convention for the 25 km data is identical but the QS100 part is replaced by QS025 denoting 25 km resolution In each node or wind vector cell WVC 118 data descriptors are defined In addition some extra information alterations have been put in place n the BUFR header the value for generating centre is set to 99 representing KNMI The value of byte 18 in BUFR section 1 identifies the generating application The products contain up to four ambiguous wind solutions with an index to the selected wind solution These are the up to four normal non MSS soluti
17. ated O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 6 SAF OSI KNMI TEC MA 134 In the outer swath region in figure 1 the two looks result in an ambiguous set of generally four wind solutions with an equal probability of about 25 Measurement noise here results in systematic wind direction errors which is why the outer swath is often well visible is figure 2 In the nadir swath region Ill in figure 1 insufficient azimuth views are available for wind retrieval and the measurement noise causes a rather noisy wind field As we enter the sweet swath this noise becomes smaller but does generally not disappear altogether At KNMI a spatially averaged product was developed at 100 km which strongly reduces the measurement noise as shown in section 6 Rather uniform and high quality winds are then obtained in regions Il and Ill Due to a more difficult QC in region I this region is not processed in the OSI SAF product at the time of writing this manual 2 2 Rain problem The NASA scatterometers work at a Ku band radar wavelength The atmosphere is not transparent at these wavelengths and in particular rain is detrimental for wind computation In fact moderate and heavy rain cause bogus wind retrievals of 15 20 m s wind speed which need to be eliminated by a quality control step Wind rain discrimination is easiest to manage in the sweet swath performs acceptable in nadir but is problematic in the outer swath 2 3 Sea i
18. ce detection Due to the availability of VV and HH polarisation measurements discrimination of water and ice surfaces is generally well possible and performed by NOAA Ref 7 On top of this an additional filter based on NWP Sea Surface Temperature is applied to prevent erratic winds over sea ice surfaces in case the NOAA ice screening fails O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 ji SAF OSI KNMI TEC MA 134 3 Algorithms Scatterometry was developed heuristically It was found experimentally that the sensitivity to wind speed and direction describe well the changes in backscatter over the ocean at moderate incidence angles due to changes in surface roughness as depicted in figure 3 Valenzuela 1978 In return backscatter measurements can be used to determine the wind speed and wind direction in a WVC In the NWP SAF development phase the ERS scatterometer SCAT processing has been successfully extended to SeaWinds A schematic illustration of the processing is given in figure 4 After defining the wind output and motivating the Geophysical Model Function that is used the algorithms developed at KNMI are described Reflected Figure 3 Schematic representation of microwave ING Waye scattering and reflection at a Na smooth a rough b and very Ng rough c ocean surface As the roughness increases more microwave power is returned towards the direction of the microwave source Incident wave Smooth
19. cumentation on http www osi saf org Ref 3 NWP SAF website http www metoffice gov uk research interproj nwpsaf index html Ref 4 NASA SeaWinds Documentation http podaac jpl nasa gov quikscat gscat doc htmIl Ref 5 Thesis Wind Field Retrieval from Satellite radar systems by Marcos Portabella available on http www knmi nl scatterometer publications Ref 6 Thesis Scatterometry by Ad Stoffelen available on http igitur archive library uu nl dissertations 01840669 inhoud htm Ref 7 Leidner M Hoffman R and Augenbaum J SeaWinds scatterometer real time BUFR geophysical data product version 2 3 0 NOAA NESDIS June 2000 available on ftp metroweb nesdis noaa gov seawinds bufr_v2 3 0 ps qz O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 4 SAF OSI KNMI TEC MA 134 2 SeaWinds Instruments I Figure 1 Earth surface coverage of 1 the scans of the horizontal HH and vertical polarisation VV pencil I I I II p lt p o beams of SeaWinds As the satellite propagates towards the top of the page the swath in grey is illuminated and three areas are na discriminated ne Outer swath only viewed once by the VV beam in the forward direction W LN and once in the aft direction 2 views 1 Y A Il Sweet inner swath Viewed both il by the VV and HH beam both in fore and aft direction 4 views lll Nadir inner swath As Il but
20. d Histograms T T T z T T TE Z a z 270 Z ik gt 15t q E 2 180 o 10f YY E o A 2 5 o 5 E 90 Ss 5 gv bl 0 06 Na AN 0 5 10 15 20 0 90 180 270 360 Wind speed m s ECMWF Wind dir deg ECMWF N 4540535 N 3855708 mx 7 50 my 7 26 mx 177 46 my 177 88 m y x 0 24 s y x 1 49 m y x 0 42 s y x 14 12 cor_xy 0 91 cor_xy 0 99 Wind Histograms Wind Histograms Es 20F E 205 z 10f 3x 10Ff 7 l E o Ig QO jp i E 10 3 Bot j o o o o z 20 F 7 20 E 3 20 10 0 10 20 20 10 0 10 20 U comp m s ECMWF V comp m s ECMWF N 4540535 N 4540535 mx 0 23 my 0 13 mx 0 74 my 0 70 m y x 0 10 s y x 1 60 m y x 0 04 s y x 1 58 cor_xy 0 97 cor_xy 0 96 Figure 11 Contoured histograms of the 25 km OSI SAF wind product The top left plot corresponds to wind speed bins of 0 4 m s and the top right plot to wind directions bins of 2 57 The latter are computed for ECMWF winds larger than 4 m s The bottom plots show the U and V wind component statistics N is the number of data mx and my are the mean values along the x and y axis respectively m y x and s y x are the bias and the standard deviation with respect to the diagonal respectively and cor xy is the correlation value between the x and y axis distributions The contour lines are in logarithmic scale each step is a factor of 2 and the lowest level outer most contour line is at N 64000 data points From these results it is clear that t
21. e nadir WVCs both the JPL rain flag and the KNMI residual check are used for QC 3 3 Detailed SeaWinds algorithm description SeaWinds processing is described in some detail in Ref 4 The OSI SAF extensions to the NOAA and NASA SeaWinds processing are described in Stoffelen 2000 and in Portabella Ref 5 3 4 Monitoring Automatic ways of monitoring backscatter data quality and wind products is of the utmost importance for using the data in particular for routine use in NWP for example The way of monitoring at KNMI is reported in de Vries Stoffelen and Beysens 2005 In short a multi step check is used for each product half orbit and if the number of QC rejections is above a threshold or the mean normalised residual J is above a threshold for those data where a wind solution can be calculated or O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 11 SAF OSI KNMI TEC MA 134 the wind speed bias against the NWP reference is above a threshold then the monitoring flag is raised and the output is suspicious The false alarm rate of the monitoring flag is about 0 001 O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 12 SAF OSI KNMI TEC MA 134 4 Processing scheme KNMI has an operational processing chain running in near real time with SeaWinds data including visualisation on the web http www knmi nl scatterometer gscat prod 100 km product and http www knmi nl scattero
22. eoffs in the design of a spaceborn scanning pencil beam scatterometer application to SeaWinds IEEE Trans Geosci Rem Sens vol 35 no 1 pp 115 126 1997 Stoffelen Ad A Generic Approach for Assimilating Scatterometer Observations ECMWF seminar 2000 available on http www knmi nl scatterometer publications Stoffelen Ad John de Vries and Aart Voorrips Towards the Real Time Use of QuikScat Winds BCRS project report KNMI de Bilt 2001 available on http www knmi nl scatterometer publications Stoffelen Ad Siebren de Haan Yves Quilfen and Harald Schyberg ERS Scatterometer Ambiguity Removal Comparison OSI SAF report 2000 available on http www knmi nl scatterometer publications O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 26 SAF OSI KNMI TEC MA 134 17 18 Valenzuela G R Theories for the interaction of electromagnetic and ocean waves a review Bound Layer Meteor 13 612 685 1978 Wentz F J D K Smith A model function for the ocean normalized radar cross section at 14 GHz derived from NSCAT observations J Geophys Res 104 C5 11499 11514 1999 19 Wentz F J S Peteherych and L A Thomas A model function for ocean radar cross 20 21 22 23 sections at 14 6 GHz J Geophys Res 89 3689 3704 1984 Portabella M and A Stoffelen A probabilistic approach for SeaWinds data assimilation Quart J Royal Meteor Soc 130 127 152 2004 Verh
23. er is of better quality since the former leads to more wind solutions during wind retrieval i e is more ambiguous At low wind speeds the wind direction and speed may vary considerably within the WVC Locally below a speed of roughly 2 ms calm areas are present where little or no backscatter occurs perhaps further extended in the presence of natural slicks that increase the water surface tension Donelan and Pierson 1987 However given the variability of the wind within a footprint area of 25 or 50 km it is even in the case of zero mean vector wind very unlikely that there are no patches with roughness in the footprint As the mean vector wind increases the probability of a calm patch will quickly decrease and the mean microwave backscatter will increase Also natural slicks quickly disappear as the wind speed increases and as such the occurrence of these is correlated to the amplitude of the mean vector wind over the footprint as modelled by the GMF Low scatterometer wind speeds are thus providing useful information Figure 4 Overview of SeaWinds algorithms QC and Spatial Averaging Wind computation by GMF Inversion l Fieldwise Ambiguity Removal Monitoring Module At high wind speeds wave breaking will further intensify causing air bubbles foam and spray at the ocean surface and a more and more complicated ocean topography Although theoretically not obvious it is emp
24. essing o at output resolution Wind Retrieval Ambiguity B Removal 5 Monitoring i Figure 6 SeaWinds Data Processor de Vries 2003 ambiguous winds p QDP data product repository unique winds O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 14 SAF OSI KNMI TEC MA 134 5 Data description 5 1 Wind product characteristics Physical definition Horizontal equivalent neutral wind vector at 10 m height Units and range Wind speed is measured in m s The wind speed range is from 0 50 m s but wind speeds over 25 m s are generally not reliable cf Donelly et al 1999 In the BUFR products the wind direction in meteorological WMO convention relative to North 0 degrees corresponds to a wind flowing to the South with a clockwise increment In the NetCDF products the wind direction is in oceanographic convention 0 degrees corresponds to a wind flowing to the North with a clockwise increment Input satellite data SeaWinds BUFR data from NOAA are described in their user manual see Ref 7 Product computed from FTPed BUFR messages issued by NOAA to the UK Met Office and by the UK Met Office to KNMI containing geo located measurement quadruplets on a satellite swath grid of 25 km size Portabella Ref 5 section 2 4 shows that the backscatter data in the BUFR and the off line product are of similar quality despite processing differences The dela
25. he spread in the distributions is small The wind speed bias is quite small and it is clear from the bottom plots that the standard deviations in the components are well below 2 m s The quality of the 25 km winds is further elaborated in Verhoef Vogelzang and Stoffelen 2007 By passing more wind and probability information from the inversion step to the ambiguity removal step the 25 km product spatial noise filtering is effectively performed by O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 23 SAF OSI KNMI TEC MA 134 the meteorologically balanced 2D VAR This appears crucial for the wind quality that is obtained in the end product and it is shown in the above report that good quality winds showing small scale details are obtained 6 4 Ambiguity selection For SeaWinds a version of 2D VAR is used with minimal regional performance differences Vogelzang 2007 This version is based on the work by de Vries Stoffelen and Beysens 2005 which in turn was triggered by the findings of Stoffelen et al 2002 A variational QC step is performed to reject a few WVCs which are in meteorological unbalance with their neighbours This rejection rate often called gross error rate peaks at nadir and falls off quickly by a factor of two in the sweet swath 6 5 Quality indices and data use We summarise the product quality indices here 1 Rain sea state ice and land presence resp bits 6 9 and 8 of WVC_QUALITY_FLAG sec
26. irically found that o gt keeps increasing for increasing wind speed up to 25 m s and even higher and that a useful wind direction dependency remains Donelly et al 1999 O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 9 SAF OSI KNMI TEC MA 134 3 2 Wind Retrieval The GMF has two unknowns namely wind speed and wind direction so if more than two backscatter measurements are available then these two unknowns may be estimated using a maximum likelihood estimator MLE as the objective function for determining wind vector solutions Pierson 1989 1990 The MLE is defined by TD Om U Li YY Var om 10log cone distance FS f i E 20 u m s Figure 5 The wind retrieval objective function J as a function of wind vector components u and v for four collocated fore and aft beam measurement of radar backscatter The cross indicates the position of the most likely ambiguous solution The other solution is opposite where are the backscatter measurements on u y are the model backscatter values corresponding to the measurements and Var o Kyi 0 are the measurement variances Note that in the NOAA product Var 0 Ki On is used i e the noise variance estimated by the modelled rather than the observed backscatter We Ref 5 found the latter to perform better The local minima of J correspond to wind vector solutions The three or more independent measurements should well
27. isually display the confidence levels of the products and their evolution with time some of these graphs are available on the WWW site for product quality monitoring by the users O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 13 SAF OSI KNMI TEC MA 134 4 4 Software Configuration To keep record of system configuration and releases and the history of the source files in the processing system the development team at KNMI uses the Concurrent Version System CVS 4 5 Dissemination and archive SeaWinds BUFR and NetCDF products are put available on a password protected FTP site This password is provided to new users by E mail request to scat knmi nl A BUFR reader is available at www knmi nl scatterometer bufr reader BUFR Data are available also through the EUMETCast system Please consult http www eumetsat int under Access to Data for more information on EUMETCast dissemination and how to receive these and other EUMETSAT meteorological satellite products or contact ops eumetsat int The data will be archived in the EUMETSAT Unified Meteorological Archive and Retrieval Facility UMARF see http archive eumetsat org umarf For data not yet present in UMARF KNMI also keeps an off line archive of the products You can send a request to scat knmi nl ECMWF NWP data NOAA NRT products Quality Control Info Warnings Errors Statistics AJ logging Pre Proc
28. km winds may contain spatial correlation structures in the errors that have not been investigated in full detail and may prove detrimental when the product is assimilated in an NWP model Hence we do not recommend to use SeaWinds at 25 km for data assimilation 1 3 Access to data and helpdesk These wind products are disseminated through the EUMETCast system for access please contact EUMETSAT http www eumetsat int The products are also available on a password protected FTP server Contact the KNMI OSI SAF team scat knmi nl for access For a swift response management procedure user requests on these data products should be issued at the Ocean and Sea Ice SAF website http www osi saf org 1 4 Disclaimer All intellectual property rights of the OSI SAF products belong to EUMETSAT The use of these products is granted to every interested user free of charge If you wish to use these products EUMETSAT s copyright credit must be shown by displaying the words copyright year EUMETSAT on each of the products used The OSI SAF is much interested in receiving your feedback would appreciate your acknowledgment in using and publishing about the data and like to receive a copy of any publication about the application of the data Your feedback helps us in maintaining the resources for the OSI SAF wind services 1 5 References Ref 1 KNMI OSI SAF web site http www knmi nl scatterometer osisaf Ref 2 O8SI SAF wind product do
29. len 2000 NASA and NOAA put available SeaWinds products the former produces science products whereas NOAA provides a near real time wind product that is used as basis for further processing at KNMI Our SeaWinds products are different from the NASA and NOAA products since special emphasis is put on increasing the reliability of the wind vectors by Quality Control and rejection of rain contaminated Wind Vector Cells WVC improving accuracy by noise reduction through spatial averaging 100 km product and meteorological filtering Guaranteeing correctness by monitoring of the operational SeaWinds data and KNMI processing and Reduce wind direction selection errors by meteorologically balanced wind direction Ambiguity Removal AR These wind quality issues are further elaborated in this manual and discussed in section 6 Over the whole globe 80 of SeaWinds products are available within 3 hours after the last satellite data acquisition Availability includes the processing time at KNMI which is less than 10 minutes and a not very substantial part of the total delay The OSI SAF products are delivered on request through the FTP server to all users and through EUMETCast See also http www knmi nl scatterometer osisaf for real time graphical examples of the products and up to date information and documentation The easiest access to the SeaWinds products is through the OSI SAF as described in this manual Alternatively in
30. ls Delivery time A wind product is available for distribution 10 minutes after the input product reception Expected accuracy O amp SISAF SeaWinds Product User Manual version 1 6 August 2009 15 SAF OSI KNMI TEC MA 134 The expected accuracy is defined as the expected bias and standard deviation of the primary calculations The accuracy is validated against in situ wind measurements from buoys platforms or ship and against NWP data Even better the errors of all NWP model winds in situ data and scatterometer winds are computed in a triple collocation exercise Ref 6 Djepa 2002 The performance is pretty constant over the globe and depends mainly on the sub footprint wind variability The performance of the products issued by the OSI SAF is characterised by a wind vector RMS error smaller than 3 m s 5 2 File formats Wind products are in BUFR or NetCDF format A complete description of BUFR can be found in WMO publication No 306 Manual on Codes The graphical display of the wind products is available and explained on the web http www knmi nl scatterometer gscat prod and http www knmi nl scatterometer gscat 25 prod The 100 km and 25 km OSI SAF wind products are stored in exactly the same BUFR format as described in the SeaWinds BUFR manual from NOAA Ref 7 a list of descriptors fields contained in each WVC is provided in section 9 Data are organised in files containing approximately 75 90 minutes of data The file n
31. mativa a Dina ALAN andes Lala nn O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 SAF OSI KNMI TEC MA 134 1 Introduction 1 1 Overview The Ocean and Sea Ice Satellite Application Facility OSI SAF is producing a range of air sea interface products namely wind sea ice characteristics Sea Surface Temperatures SST and radiative fluxes Surface Solar Irradiance SSI and Downward Long wave Irradiance DLI This document is the SeaWinds product user manual dedicated to the OSI SAF wind product users It describes the global SeaWinds products available at KNMI Quality monitoring information on this product and more general information on the whole OSI SAF project is available on the OSI SAF web site at the following address http www osi saf org The user is strongly encouraged to register on this web site in order to receive the service messages and the latest information about the OSI SAF products KNMI has a long experience in scatterometer processing and is developing generic software for this purpose Processing systems for the ERS scatterometers were adopted for NSCAT SeaWinds and ASCAT The scatterometer is an instrument that provides information on the wind field near the ocean surface and scatterometry is the knowledge of extracting this information from the instrument s output Space based scatterometry has become of great benefit to meteorology and climate in the past years e g Isaksen and Stoffe
32. meter gscat 25 prod 25 km product This processing is performed using the NWP SAF software The processing includes monitoring and archiving functionalities A global overview of the modules of our SeaWinds scatterometer processor is given below Figure 6 gives an overview of the entire processing system of SeaWinds 4 1 NWP collocation KNMI receives NWP model data from ECMWF twice a day through the RMDCN NWP model sea surface temperature data is used to provide information about possible ice presence in the WVCs WVCs with a sea surface temperature below 271 56 K 1 6 C are assumed to be covered with ice and no wind information is calculated This ice check which is usually inactive is done in addition to the evaluation of the JPL ice flags which are present in the NOAA input product If the JPL ice screening fails the ice is still detected by the SST check NWP forecast wind data are necessary in the ambiguity removal step of the processing ECMWF wind forecasts are available twice a day 00 and 12 GMT analysis time with forecast time steps of 3h 6h 36h The 10 m model wind data are averaged with respect to time and location and put into the model wind part of each WVC where they replace the NCEP 1000 hPa forecast winds that are present in the NOAA input product This has a positive input on the output product quality and eases the product monitoring and validation see Verhoef and Stoffelen 2006 Note that the ECMWF winds
33. ncy Network Common Data Form National US Oceanic and Atmospheric Administration NASA Scatterometer Numerical Weather Prediction Ocean and Sea Ice SAF Quality Control US dedicated scatterometer mission to bridge ADEOS I and ADEOS II Regional Meteorological Data Communication Network Satellite Application Facility US scatterometer on board QuikSCAT and ADEOS II platforms Scatterometer Ocean Stress Surface Solar Irradiance Sea Surface Temperature West to east wind component South to north wind component Vertical polarisation send and receive mode Wind Vector Cell SeaWinds Product User Manual version 1 6 August 2009 25 SAF OSI KNMI TEC MA 134 10 11 12 13 14 15 16 8 Literature de Vries J Stoffelen A and Beysens J Ambiguity Removal and Product Monitoring for SeaWinds NWP SAF report NWPSAF_KN_TR_001 2005 available on http www knmi nl scatterometer publications Donelan M A and W J Pierson Radar scattering and equilibrium ranges in wind generated waves with application to scatterometry J Geophys Res 92 4971 5029 1987 Donelly William J James R Carswell and Robert E McIntosh Revised ocean backscatter at C and Ku band under high wind conditions J Geophys Res 104 11 485 11 497 1999 Figa J and A Stoffelen 2000 On the Assimilation of Ku Band Scatterometer Winds for Weather Analysis and Forecasting IEEE Transactions on Geoscience and Remote Sensing s
34. oef A and A Stoffelen Ambiguity Removal using different background wind fields OSI SAF Technical Report SAF OSI KNMI TEC TN 091 August 2006 available on http www knmi nl scatterometer publications Vogelzang J Two dimensional variational ambiguity removal NWP SAF report NWPSAF KN TR 004 2007 available on http Awww metoffice gov uk research interproj nwpsaf scatterometer index html Verhoef A J Vogelzang and A Stoffelen Validation of SeaWinds 25 km winds OSI SAF Technical Report SAF OSI KNMI TEC RP 143 January 2008 O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 27 SAF OSI KNMI TEC MA 134 9 SeaWinds BUFR data descriptors Number Descriptor Parameter Unit 001 01007 Satellite Identifier Code Table 002 01012 Direction of Flight Degree True 003 02048 Satellite Instrument Identifier Code Table 004 21119 Wind Scatterometer GMF Code Table 005 25060 Software Identification Numeric 006 02026 Cross Track Resolution m 007 02027 Along Track Resolution m 008 05040 Orbit Number Numeric 009 04001 Year Year 010 04002 Month Month 011 04003 Day Day 012 04004 Hour Hour 013 04005 Minute Minute 014 04006 Second Second 015 05002 Latitude Coarse Accuracy Degree 016 06002 Longitude Coarse Accuracy Degree 017 08025 Time Difference Qualifier Code Table 018 04001 Time to Edge Second 019 05034 Along Track Row Number Numeric 020
35. ons in the selected wind solution the wind speed and wind direction have been replaced by the wind speed and wind direction of the selected MSS solution The Formal Uncertainty in Wind Direction does not contain the uncertainty but the normalised inversion residual referred to as Rn in Ref 5 The Wind Vector Cell Quality Flag table 021109 is redefined and now has the following definitions Description BUFR bit Fortran bit Not enough good sigma 0 available for wind retrieval 1 15 O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 16 SAF OSI KNMI TEC MA 134 Description BUFR bit Fortran bit Not used 2 3 14 13 Monitoring flag 4 12 Monitoring value 5 11 KNMI JPL Quality Control including rain 6 10 Variational QC 7 9 Land presence 8 8 Ice presence 9 7 Not used 10 6 Reported wind speed is greater than 30 m s 11 5 Reported wind speed is less than or equal to 3 m s 12 4 Not used 13 3 Rain flag algorithm detects rain 14 2 Data from at least one of the four possible beam view 15 1 combinations are not available Not used 16 0 Missing value All 17 set All 17 set In Fortran if the Wind Vector Cell Quality Flag is stored in an integer I then use BTEST I NDW NB 1 to test BUFR bit NB where NDW 17 is the width in bits of the data element in BUFR If the monitoring flag is set to zero the product is monitored If the product
36. pecial issue on Emerging Scatterometer Applications 38 4 pp 1893 1902 Freilich M H B A Vanhoff and R S Dunbar Empirical determination of a Ku band wind model function from SeaWinds scanning scatterometer J Geophys Res 107 C 2002 Graham R D Anderson A Hollingsworth and H B ttger Evaluation of ERS 1 wind extraction and ambiguity removal algorithms meteorological and statistical evaluation ECMWF report ECMWF Reading England 1989 Isaksen L and A Stoffelen 2000 ERS Scatterometer Wind Data Impact on ECMWF s Tropical Cyclone Forecasts IEEE Transactions on Geoscience and Remote Sensing special issue on Emerging Scatterometer Applications 38 4 pp 1885 1892 JPL QuikSCAT science data product user s manual version 2 2 Jet Propulsion Laboratory D 12985 pp 89 December 2001 JPL NASA scatterometer science data product user s manual version 1 1 Jet Propulsion Laboratory D 18053 pp 68 April 1997 Mastenbroek Kees Wind Wave Interaction thesis at the Delft University of Technology Delft the Netherlands 12 December 1996 Pierson W J Probabilities and statistics for backscatter estimates obtained by a scatterometer J Geophys Res 94 9743 9759 1989 Correction J Geophys Res 95 809 1990 Portabella M A Stoffelen EUMETSAT fellowship report 2002 available on http www knmi nl scatterometer publications Spencer M W Wu C and Long D G Trad
37. rtainty In Wind Direction Degree True 050 21104 Likelihood Computed for Wind Solution Numeric 051 02104 Antenna Polarisation Code Table O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 28 SAF OSI KNMI TEC MA 134 Number Descriptor Parameter Unit 052 08022 Total Number w r t accumulation or average Numeric 053 12063 Brightness Temperature K 054 12065 Standard Deviation Brightness Temperature K 055 02104 Antenna Polarisation Code Table 056 08022 Total Number w r t accumulation or average Numeric 057 12063 Brightness Temperature K 058 12065 Standard Deviation Brightness Temperature K 059 21110 Number of Inner Beam Sigma0 fwd of sat Numeric 060 05002 Latitude Coarse Accuracy Degree 061 06002 Longitude Coarse Accuracy Degree 062 21118 Attenuation Correction On Sigma 0 dB 063 02112 Radar Look Azimuth Angle Degree 064 02111 Radar Incidence Angle Degree 065 02104 Antenna Polarisation Code Table 066 21105 Normalized Radar Cross Section dB 067 21106 Kp Variance Coefficient Alpha Numeric 068 21107 Kp Variance Coefficient Beta Numeric 069 21114 Kp Variance Coefficient Gamma dB 070 21115 Seawinds Sigma 0 Quality Flag Flag Table 071 21116 Seawinds Sigma 0 Mode Flag Flag Table 072 08018 Seawinds Land Ice Surface Flag Flag Table 073 21117 Sigma 0 Variance Quality Control Numeric 074 21111 Number of Outer Beam Sigma0 fwd of sat Numeric 075 0500
38. s of both figures it is evident that the random error of the difference distribution s y x is smaller for the 100 km product In addition the systematic error in the difference distribution m y x is also smaller for the 100 km product except for the wind direction The asymmetric peak in the U component distribution in both figures is due to the systematic sampling of the trade winds The jaggedness of the wind speed distributions is due to the fact that the retrieved scatterometer wind speeds stem from the GMF lookup table Both figures show that currently in the trade off between resolution and noise QuikSCAT winds at 100 km resolution are the preferred product for NWP data assimilation O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 20 SAF OSI KNMI TEC MA 134 z Wind Histograms l S S 3 20 A T 15f gt 9 E 10 DE J f 8 EZ E Sf J 9 ES a 2 0 0 5 10 15 20 Wind speed m s ECMWF N 313672 mx 7 72my 8 39 m y x 0 67 s y x 1 68 cor_xy 0 89 Wind Histograms 3 20 3 S 5 10 E x ai E lt 10F 3 g 5 gt a J E L L 20 10 0 10 20 U comp m s ECMWF N 313672 mx 041 my 0 16 m y x 0 25 s y x 2 05 cor_xy 0 96 Wind dir deg JPL thinned V comp m s JPL thinned Wind Histograms 270 180 90 Nga A 0 90 180 270 360 Wind dir deg ECMWF N 257964 mx 180 68 my 182 22 m y x 1 55 s
39. sample the azimuth variation of the GMF in order to resolve the wind direction albeit ambiguously O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 10 SAF OSI KNMI TEC MA 134 The problem of ambiguity illustrated in figure 5 does generally not disappear if more measurements become available However multiple azimuth views and polarisation measurements do increase the probability that one of the ambiguous solutions becomes more likely than the other ones Stoffelen de Vries and Voorrips 2001 describe that SeaWinds retrievals at 100 km resolution are much less ambiguous than retrievals at 25 km due to this see also section 6 3 2 1 Ambiguity Removal SeaWinds scatterometer winds have a multiple ambiguity and there are up to four wind solutions in each wind vector cell WVC on the earth s surface These ambiguities are removed by applying constraints on the spatial characteristics of the output wind field such as on rotation and divergence Several ambiguity removal AR schemes were evaluated for ERS data Stoffelen de Haan Quilfen and Schyberg 2000 In the OSI SAF Development Phase some schemes that were developed for the SCAT were compared In addition to the subjective comparison of AR schemes a method for the objective comparison of AR performance among the different schemes was used Stoffelen et al 2002 show that this way of comparison is effective to evaluate the shortcomings of AR schemes but also reveals
40. stored in the wind products are real winds rather than neutral winds 4 2 Validation Each step in the processing is validated separately and also the product as a whole by a quality control and monitoring scheme The product validation step is controlled by visual inspection and a statistical analysis is performed to control the validation steps The inversion step is controlled in the same way For ambiguity removal schemes an objective scheme exists that relies on initialisation with a one day lead NWP forecast and validation of the ambiguity selection against NWP analyses as in Stoffelen et al 2002 Moreover Stoffelen de Vries and Voorrips 2001 describe subjective comparison of the 2D VAR and PreScat schemes by routine operational meteorologists 4 3 Quality control and monitoring The quality of the delivered products is controlled through an ad hoc visual examination of the graphical products and the automatic production of control parameters The examination of the products is done at KNMI by experts Specific tools have been developed to help this analysis User queries obviously lead to the inspection of suspect products The ad hoc and user queried inspections are used for quality assurance An information file is made for each product The content of the file is identical whatever the product and results from a compilation of all the global information concerning this product From these files various graphs are produced to v
41. the azimuth view direction is close to the satellite propagation direction or just opposite to it 1800km Scatterometers fly on polar orbiting satellites for SeaWinds these are the NASA NOAA QuikSCAT and Japanese ADEOS II platforms Data are read out once per orbit usually which lasts about 100 minutes This means that a delay of up to 100 minutes is already present at read out Dedicated transmission lines and fast processing are needed to limit further delays 2 1 SeaWinds l and SeaWinds Il Two SeaWinds scatterometers are developed and flown The SeaWinds on QuikSCAT mission from NASA NOAA is a quick recovery mission to fill the gap created by the loss of data from NSCAT when the ADEOS 1 satellite lost power in June 1997 It was launched in June 1999 and is still operational A similar version of the instrument SeaWinds 2 flew on the Japanese ADEOS 2 satellite launched in December 2002 which was regrettably lost in October 2003 For detailed information on the QuikSCAT instrument and data we refer to Spencer et al 1997 JPL 1997 2001 and Ref 7 A brief description is given below The SeaWinds instrument is a conically scanning pencil beam scatterometer as depicted in figure 1 It uses a rotating 1 meter dish antenna with two spot beams of about 25 km size on ground a horizontal polarisation beam HH and a vertical polarisation beam VV at incidence angles of 46 and 54 respectively that sweep the surf
42. the context of the Numerical Weather Prediction SAF see http Awww metoffice gov uk research interproj nwpsaf index html software is developed and maintained to produce scatterometer winds from raw backscatter data which may be implemented at the user site if special processing features are desired The NWP SAF software is used by the OSI SAF production system In the context of the climate CM SAF KNMI is developing the Scatterometer Ocean Stress SOS product from the OSI SAF wind product O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 3 SAF OSI KNMI TEC MA 134 This user manual outlines user information for the OSI SAF products based on the SeaWinds scatterometers Ref 2 Ref 4 Section 2 presents a brief description of the SeaWinds instruments section 3 the processing algorithms and section 4 gives an overview of the data processing configuration Section 5 provides detailed information on the file content and format while in section 6 the product quality is elaborated 1 2 Purpose and use of 100 km and 25 km products The OSI SAF SeaWinds scatterometer winds are available in two resolutions 100 km and 25 km The 100 km product is especially suitable for assimilation in Numerical Weather Prediction models It provides up to 4 wind solutions in each Wind Vector Cell The 25 km product is more suitable for nowcasting it contains more detail in the wind field that is not present in the 100 km product However the 25
43. tion 5 2 the land flag may be ignored since it indicates land contamination if a too high land fraction is present in the WVC no winds are calculated 2 Monitoring flag and bit resp bits 4 and 5 of WVC_QUALITY_FLAG sections 5 2 and 3 4 a selected wind may be given but it should be suspected if bits 4 and 5 are set 3 Variational QC bit 7 of WVC_QUALITY_FLAG section 5 2 if this bit is set the selected wind is spatially inconsistent with its neighbours and suspect O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 24 SAF OSI KNMI TEC MA 134 ADEOS AR ASCAT BUFR CM SAF DLI EPS ERS EUMETCast EUMETSAT GOES HDF HH IR JPL KNMI MSS NASA NetCDF NOAA NSCAT NWP OSI SAF QC QuikSCAT RMDCN SAF SeaWinds SOS SSI SST lt WVC O amp SI SAF 7 Glossary Japanese Advanced Earth Observation Satellite Ambiguity Removal Advanced Scatterometer Binary Universal Format Representation Climate SAF Downward Long wave Radiation EUMETSAT Polar Satellite European Remote Sensing Satellite EUMETSAT s Digital Video Broadcast Data Distribution System European Organisation for the Exploitation of Meteorological Satellites Geostationary Operational Environmental Satellite Hierarchical Data Format Horizontal polarisation send and receive mode InfraRed Jet Propulsion Laboratory NASA Royal Netherlands Meteorological Institute Multiple Solution Scheme National US Air and Space Age
44. uity information The wind directions are in oceanographic rather than meteorological convention see section 5 1 The format is usable for ASCAT ERS QuikSCAT and any other scatterometer data The data has file sizes comparable to those of the corresponding BUFR data e g one file of SeaWinds 100 km wind data is 300 kB in BUFR and 180 kB in NetCDF When compressed with gzip the size of one file in NetCDF reduces to 60 kB The NetCDF data in near real time are only available on the KNMI FTP server but EUMETCast dissemination can be considered on user request The file name convention for the gzipped NetCDF product is QS100 Dyyddd Shhmm Ehhmm Bxxxxxxx nc gz QS025 Dyyddd Shhmm Ehhmm Bxxxxxxx nc gz where the meaning of the fields is identical to those in the BUFR file names see section 5 2 Below are some meta data contained in the NetCDF data files dimensions NUMROWS 359 NUMCELLS 19 variables int time NUMROWS NUMCELLS time long_name time time units seconds since 1990 01 01 00 00 00 int lat NUMROWS NUMCELLS lat long name latitude lat units degrees north int lon NUMROWS NUMCELLS lon long name longitude lon units degrees east short wvc index NUMROWS NUMCELLS wvc_index long_ name cross track wind vector cell number wvc_index units 1 short model speed NUMROWS NUMCELLS model speed long name model wind speed at 10 m
45. y of the input SeaWinds data is up to 3 hours delay from observation time Geographical definition Satellite swath projection Swath is located below the satellite The QuikSCAT satellite proceeds in a near polar orbit at 98 degrees inclination at 800 km orbit height Equator ascending crossing time is 10 UTC Swath width is 76 19 25 km 100 km size WVC each side At KNMI we currently only use WVC for which four backscatter measurements are available WVCs 9 68 nominally resulting in 15 100 km size WVCs This substantially reduces the overlap of subsequent files Products are organised in batches of about half an orbit along track Time resolution Polar satellites have the capability to provide data twice daily The SeaWinds swath width of 1800 km provides full coverage twice a day for latitudes above 50 degrees KNMI processes a 1400 km swath where above 60 degrees full coverage at least twice daily is provided The dissemination frequency of the scatterometer data is about every 50 minutes in half orbit files Every useful input backscatter product has a corresponding output wind product Coverage Global see web visualisation Output product The input product in BUFR is processed into a BUFR output product with a unique wind solution chosen and its corresponding ambiguities quality information probability of solution quality flag e g monitoring bit The products are also available in NetCDF format see section 10 for more detai
46. y x 15 30 cor_xy 0 99 Wind Histograms 205 J 105 J 0 E Aq 10f J 20F J 20 10 0 10 20 V comp m s ECMWF N 313672 mx 0 46 my 0 55 m y x 0 09 s y x 1 87 cor xyz 0 94 Figure 9 As figure 8 but for the thinned NOAA product In our verification the NOAA DIRTH processing tends to deliver reasonably smooth wind direction see figure 10 but rather variable wind speeds On the other hand the 100 km product exhibits similar scales and variability in both wind speed and direction This is what one may anticipate from the way both processing algorithms work The noise in the 100 km product is smaller than in the NOAA 25 km product and the performance more uniform over the swath consistent with the initial objective O amp SI SAF SeaWinds Product User Manual version 1 6 August 2009 21 SAF OSI KNMI TEC MA 134 NOAA product bottom the trough appears weaker due to the wind direction spatial smoothing Blue winds are HIRLAM in the top panel while here the background image is Meteosat IR Black winds did not pass the JPL rain check in the bottom panel O amp SISAF SeaWinds Product User Manual version 1 6 August 2009 22 SAF OSI KNMI TEC MA 134 6 3 Accuracy 25 km product Figure 11 shows two dimensional histograms of the OSI SAF retrieved 25 km winds versus ECMWF forecasts The data are from consecutive orbits from 1 5 July 2007 Wind Histograms Win

SeaWinds Product User Manual

Contents

Download Pdf Manuals

Related Search

Related Contents