DYCOMS-II Satellite: QuikSCAT Wind Vector Data 1.0 General Description The QuikSCAT Wind Vector Data is one of several satellite data sets collected by the University Corporation for Atmospheric Research/Joint Office for Science Support (UCAR/JOSS) as part of the Dynamics and Chemistry of Marine Stratocumulus Phase II: Entrainment Studies (DYCOMS-II) project. QuikSCAT is a microwave scatterometer launched in June 1999. The QSCAT data are stored in orbit files and include wind speed and direction at a height of 10 meters retrieved from measurements of the QSCAT backscattered power.. The data cover the period from 7 - 28 July 2001 and are in binary format. The data and readme information were acquired from Remote Sensing Systems (www.ssmi.com) and are from Version 3.0 of their processing algorithms. 2.0 Data Contact Deborah Smith smith@remss.com (707) 545-2904 ext.11 10am - 6pm (PST) 3.0 Data and Format Information 3.1 Introduction The microwave scatterometer QuikSCAT (or QSCAT) was launched in June 1999. It's primary mission is to measure winds near the sea-surface. It is also useful for some land and sea ice applications. QSCAT is the third in a series of NASA scatterometers that operate at Ku-band (i.e., a frequency near 14 GHz). The first Ku-Band scatterometer was flown on SeaSat in 1978. Eighteen years later, NSCAT was launched on Japan's ADEOS-1 spacecraft in August 1996. It operated for 10 months until the ADEOS-1 solar panel experienced a catastrophic failure. The Europeans also fly satellite scatterometers, which operate at C-band (approx. 5 GHz),on their ERS spacecrafts. QSCAT is essentially a radar that transmits radar pulses down to the Earth's surface and then measures the power that is scattered back to the instrument. This "backscattered" power is a measure of surface roughness. For water surfaces, the surface roughness is highly correlated with the near-surface wind speed and direction. Hence, wind speed and direction at a height of 10 meters are retrieved from measurements of the QSCAT backscattered power. 3.2 Geophysical Model Function: Ku-2001 Improvements have been made since the mid-1990s to the algorithms used to compute wind vectors. The geophysical model function (GMF) called Ku-2000 (used for version 2 data) was developed as an improvement over the NSCAT-2 model function. The model function improvements can be characterized in two ways. First, at low winds (below 7 m/s) the model is a more accurate representation of the directional modulation of the backscatter power. Second, at high wind speeds ( > 20 m/s) the relationship of backscatter versus wind speed has been modified to better agree with observations in tropical storms and hurricanes. The Ku-2000 model was capable of retrieving winds from 0 to 70 m/s. This is a major improvement over earlier models that consistently underestimated the strength of winds in storms and hurricanes. We would like to acknowledge a number of scientists whose research contributed to the development of Ku-2000: Simon Yueh, Linwood Jones, and James Carswell (high wind adjustment) and Michael Freilich and Naoto Ebuchi (low wind adjustment). Following further validation and study of high wind case studies, we have adjusted the Ku-2000 model function to better retrieve high winds (winds greater than 30 m/s). The new geophysical model function is called Ku-2001. No major changes were made at lower wind speeds. Changes were made to the sigma-0 measurements so that no correction is needed to use these winds ( that is, you can remove any 1.048 wind speed adjustment in your programs as was required for version 2 data). 3.3 Differences Between Version 2 and Version 3 This information is included for informational purposes. All data available for DYCOMS-II were from Version 3. In July 2001, all QSCAT wind vectors have been reprocessed using an improved geophysical model function (Ku2001). The main differences between the v02 and v03 data are the following: 1) about 0.4 dB was added to the sigma-0 so that the wind speed correction necessary for the version 2 data is no longer needed. 2) higher wind speeds have been changed to better reflect the winds observed in case studies used for calibration. You should see no change in winds below 30 m/s. The new v_03 QuikScat winds agree extremely well with buoys (mean dif=0.01 m/s, rms dif=0.75 m/s) and TMI winds (mean dif=0.05 m/s , rms dif=0.62 m/s ). Long gone are the days of +/- 2 m/s accuracy performance; +/- 0.5 m/s is more like it (assuming no rain is present and the wind speeds are below 20 m/s). 3.4 Using Microwave Radiometers to Improve QSCAT Products QSCAT data processing uses contemporaneous microwave radiometer measurements for rain flagging and sea ice detection. Remote Sensing Systems processes both microwave scatterometer and radiometer data in an operational , near-real-time (NRT) environment. Thus the various data sets can be combined to obtain improvements in the individual products. For the case of QSCAT, we use 4 satellite microwave radiometers (F13 SSMI, F14 SSMI, F15 SSMI, and TMI) to determine if rain is present at the location of the QSCAT observation. In addition, the three SSMI's are used to detect sea ice. Using the SSMI daily observations of sea ice, the QSCAT observations can be properly flagged so that reliable wind vectors can be obtained immediately next to the marginal ice zone. 3.5 Further Information and Reading Routines For further information on these data please see the Remote Sensing Systems sites at: http://www.ssmi.com/qscatinfo.html ftp://ftp.ssmi.com/QSCAT/ Included here are readme files and fortran and IDL reading routines. The files are in a simple binary format. Remote Sensing Systems has made available two subroutines to assist in reading the data files: IDL: get_qscat_orbit_v03.pro Fortran: decode_qscat_orbit_v03.f They have also provided a sample fortran program that uses the above subroutine (rd_qscat_orbit_v03.f). 3.6 Decoded File and Data Quality Information The data files, once decoded contain: ATIME is the 21 character time string as given in the L2A files PHI_TRACK is the direction of the subtract velocity vector relative to North XLAT is the geodetic latitude XLON is the east longitude ICLASS indicates the expected quality of the vector retrieval ICLASS=0 denotes no retrieval was done (either no observations or only one flavor of observation) ICLASS=1 denotes 2 flavors of observations in wind vector cell ICLASS=2 denotes 3 flavors of observations in wind vector cell ICLASS=3 denotes 4 flavors of observations in wind vector cell We suggest using just cases for which ICLASS.GE.2 NUM_AMBIG is the number of ambiguities (0 to 4) ISELECT is the selected ambiguity (0 to 4) IRAIN_SCAT is the rain flag derived from the scatterometer measurements IRAIN_SCAT=1 indicates rain WINAL is the wind speed for the various ambiguities PHIAL is the wind direction for the various ambiguities (oceanographic convention) SOSAL is the normalized rms after-the-fit residual of the observation minus model sigma-nought Large SOSAL values indicate the observations did not fit the geophysical model function. We suggest discarding observations for which SOSAL.GT.1.9. WINDS is the smoothed version of WINAL(ISELECT) PHIWS is the smoothed version of PHIAL(ISELECT) WINGCM is the general circulation model wind speed used for nudging (either NCEP or ECMWF) DIRGCM is the general circulation model wind direction used for nudging (either NCEP or ECMWF) RAD_RAIN is the TMI or SSMI columnar rain rate (rain rate times rain column height, km mm/hour) RAD_RAIN=-999 no TMI or SSMI rain available RAD_RAIN=0 no rain RAD_RAIN=0.1 possible rain RAD_RAIN=0.2 through 25.4 definite rain and the give value is the columnar rain rate "no rain" means no rain was detected within +/- 50 km and +/- time given in MIN_DIFF. "possible rain" means some rain was detected within +/- 50 km and +/- time given in MIN_DIFF. "definite rain" means rain was detected within +/- 25 km and +/- time given in MIN_DIFF. We suggest discarding observations for which RAD_RAIN.GT.0.15 MIN_DIFF is the time difference in minutes between the scatterometer and the collocated radiometer. A value of 255 means that no radiometer observation was collocated within 254 minutes of the scatterometer. NUDGE_NCEP NUDGE_ECMWF Each data file contains two INTEGER(1) variables at the end of the file. Both variables supply information on which GCM was used for the nudging field. The first variable shows if NCEP data were used (0=used, 1=not used), and the second shows if ECMWF data were used (0=used, 1=not used). Since GCM data are tri-linearly interpolated to the QuikScat data, it is possible that one of each maps was used. If both ECMWF and NCEP data were missing for the required orbit, the orbit is not processed since no nudging field would be available. It is important to not use the data if the quality is suspect. We recommend the following conditions be satisfied for the **best** quality data: ICLASS(icel,iscan) gt 1 SOSAL(1,icel,iscan).le.1.9 (yes, you are checking the SOS of the lowest ambiguity) IRAIN_SCAT(icel,iscan).ne.1 MIN_DIFF(icel,iscan).le.30 .and. RAD_RAIN(icel,iscan).le.0.15 4.0 Quality Control Procedures UCAR/JOSS conducted no quality checks on these data. 5.0 File Naming Convention The QSCAT data are stored in orbit files and the file names have the form: qscat_RRRRR_yyyymmddhhmm_winvec.dat where qscat is the satellite used RRRRR is the five digit orbit number. yyyy is the four digit year mm is the two digit month dd is the two digit day hh is the two digit hour (UTC) mm is the two digit minute (the time is the time of equator crossing. winvec is the data type The date and time information for each rev# is provided in the QSCAT_INFO.txt file. The columns in this file represent: Rev Number, Number of good WVC rows in file, EQCROSSDATE, EQCROSSTIME, EQCROSSLONG, ORBITPERIOD 6.0 References None.