The WindSat Polarimetric Radiometer was developed by the Naval Research Laboratory (NRL) Remote Sensing Division and the Naval Center for Space Technology for the U.S. Navy and the National Polar-orbiting Operational Environmental Satellite System (NPOESS) Integrated Program Office (IPO). It was launched on January 6, 2003 aboard the Department of Defense Coriolis satellite. WindSat was meant to demonstrate the capabilities of a fully polarimetric radiometer to measure the ocean surface wind vector from space. Prior to launch, the only instrument capable of measuring ocean wind vectors were scatterometers (active microwave sensors). In addition to wind speed and direction, the instrument can also measure sea surface temperature, soil moisture, ice and snow characteristics, water vapor, cloud liquid water, and rain rate.
For a detailed description of the WindSat instrument, see P.Gaiser, "The WindSat Spaceborne Polarimetric Microwave Radiometer: Sensor Description and Early Orbit Performance". WindSat operates on the Coriolis satellite which is in a near-polar orbit. The local ascending node time is maintained at 6pm. This means the instrument passes over the equator at 6pm as it travels from south to north and at your location is measuring the surface near this time depending on your latitude.
The WindSat radiometer is well-calibrated and contains the lower frequency channels required for SST retrievals. The radiometer operates in 5 discrete channels: 6.8, 10.7, 18.7, 23.8 and 37.0 GHz. All are fully polarimeteric except the 6.8 and 23.8 GHz channels that have only dual polarization. Table 1 summarizes the channel details. The feedhorns of each frequency channel trace out different arcs along the scan, therefore the Earth Incidence Angles (EIA) are different for each frequency band. Unlike previous radiometers, the WindSat sensor takes observations during both the forward and aft looking scans. This makes the WindSat geometry of the earth view swath quite different and significantly more complicated to work with than the other passive microwave sensors. The RSS WindSat products are the only dataset available that uses both the fore and aft look directions. By using both directions, we obtain a wider swath and more complicated swath geometry as is visible in the browse images.
Table 1: Summary of WindSat channels, polarizations Earth incidence angles and channel resolutions. Polarizations are listed as V for Vertical, H for Horizontal, P for +45 deg, L for Left Circular, and R for Right Circular.
(3-dB footprint size)
[km x km]
|6.8||V, H||53.8||39 x 71|
|10.7||V, H, P, M, L, R||50.1||25 x 38|
|18.7||V, H, P, M, L, R||55.6||16 x 27|
|23.8||V, H||53.2||20 x 30|
|37.0||V, H, P, M, L, R||53.2||8 x 13|
After a post-launch checkout period, the WindSat data started on February 5, 2003. With some short periods of non-operation, the instrument continues to operate in the present time. Data gaps are generally due to missing data upstream from our processing facility, such as the instrument being turned off. Occasionally, there are delays in obtaining and/or processing recently recorded data. If the data are not present after several weeks, it is unlikely that the missing data will become available. We keep a record of missing data: List of WindSat gaps
We obtain WindSat antenna temperature data from NRL and NOAA. By using both these sources, we aim to provide scientists with as complete a set of data as possbile for their research. Using a consistent processing scheme and a robust radiative transfer model, we intercalibrate the WindSat data with the other microwave radiometers, obtain brightness temperatures, and then produce our ocean measurement products for distribution. The WindSat Ocean Products include Sea Surface Temperature (SST); a variety of Surface Wind Speeds and Directions; Atmospheric Water Vapor; Cloud Liquid Water; and Rain Rates.
Remote Sensing Systems performs a detailed processing of WindSat instrument data in two stages. The first stage produces a near-real-time (NRT) product (identified by "rt" within the file name) which we make available as soon as possible. This is generally within 3 hours of when the data are recorded, except for occasional delays in receiving or processing data. Although suitable for many timely uses, and the differences may appear subtle, the NRT products are not intended to be archive/research quality. They will be deleted within a week and should not be archived.
The "Final" data (identified by "v7.0.1" within the file name) are processed when we receive the NCEP FNL analysis. The NCEP wind directions are particularly useful for retrieving more accurate SSTs and wind speeds. The final "v7.0.1" products will continue to accumulate new swaths (half orbits) until the maps are full, generally within 7 days. Occasionally, data become available after more lengthy delays, and such data will be added to final "v7.0.1" products. Thus, final refers to the type of swath processing, and does not imply that a data file is finalized (as in, will never change). Newly available data swaths will be added to final "v7.0.1" products as we obtain data from our sources.
Several steps are involved in the WindSat data processing:
- US Naval Research Lab provides RSS with Level-1A antenna temperatures also known as Temperature Data Records (TDR). We process the data into Level-2A calibrated top of the atmosphere brightness temperatures (TBs). The most important issues of this calibration procedure are discussed in the GMI ATBD.
- The top of the atmosphere TB are then optimally interpolated onto a fixed 0.125 degree Earth grid, from which the various ocean measurements are retrieved. The products available on our ftp site are then created by simple bin resampling to a 0.25 deg grid. Though each product is placed on this 0.25 deg grid, the effective resolution of the data varies. The following table gives an overview of the WindSat ocean measurements in the order they appear in the binary files, and lists each product's effective spatial resolution.
Table 2: WindSat ocean measurements in the order provided in RSS binary data files. Effective spatial resolution for each measurement varies from the 0.25 deg grid the data are mapped to. Channel polarizations are listed as V for Vertical, H for Horizontal, P for +45 deg, L for Left Circular, and R for Right Circular.
|Ocean Measurement||Acronym||Effective Spatial
Resolution [km x km]
|Sea surface temperature||SST||39 x 71||6.8 VH, 10.7 VH,18.7 VH, 23.8 VH, 37.0 VH|
|Ten-meter wind speed using low frequency channels||WSPD_LF||25 x 38||10.7 VH, 18.7 VH, 23.8 VH, 37.0 VH|
|Ten-meter wind speed using medium frequency channels||WSPD_MF||16 x 27||18.7 VH, 23.8 VH, 37.0 VH|
|Columnar atmospheric water vapor||VAPOR||16 x 27||18.7 VH, 23.8 VH, 37.0 VH|
|Columnar cloud liquid water content||CLOUD||16 x 27||18.7 VH, 23.8 VH, 37.0 VH|
|Rain rate||RAIN||8 x 13||18.7 VH, 23.8 VH, 37.0 VH|
|All-weather 10-meter wind speed||WSPD_AW||
25 x 38 no rain
39 x 71 in rain
|6.8 VH, 10.7 VH, 18.7 VH, 23.8 VH, 37.0 VH|
|Ten-meter wind direction (relative to north)||WDIR||25 x 38||10.7 VHPMLR, 18.7 VHPMLR, 23.8 VH, 37.0 VHPMLR|
If any of the required channels for a geophysical parameter are not available, then the retrieval is not completed. For example, this could result in VAPOR retrievals for a given location, but not WSPD_LF or SST in the same location. SST retrievals are often missing at the center of the swaths because there are no 6.8 GHz observations. Wind direction is not available at the edge of the swaths because there are no polarimetric observations. This "missing data" appearance is different than for other RSS radiometer products.
RSS WindSat Data Products
The current version of Remote Sensing Systems WindSat data products is v7.0.1 WindSat data were reprocessed in August 2013 to remove a slight processing error that affected wind speed and directions after Jan 2008. All other data within the product files are identical to the previous version-7 release. We provide a suite of data in one file for each day, 3-day period, week and month. The data suite contains the measurements listed in Table 2 above.
WindSat data products include the standard set of measurements (sea surface temperature, wind speed, water vapor, cloud liquid water, and rain rate) and differ from other RSS radiometer products in that we can include additional measurements:
- WDIR: Because of its polarimetric channels, WindSat can measure surface wind direction. WDIR is only provided if the wind speed exceeds 3 m/s and the rain rate is below 15 mm/hr. The wind directions in our product are in oceanographic convention.
- WSPD_AW: We have made significant progress in improving the retrieval of wind speeds in the presence of rain. See Wind-Vector Retrievals Under Rain With Passive Satellite Microwave Radiometers, for more details on the algorithm used. If the 6.8 GHz channels are available together with all the other vertically and horizontally polarized channels, we can retrieve WindSat surface wind speed in both rain-free and rainy atmospheres. We combine various algorithms, including a global wind speed through rain algorithm that works under all rain conditions and an H-wind algorithm that has been specifically trained for tropical cyclones. The final all-weather wind speed, WSPD_AW, provided in our binary files is a smooth blend between the standard wind speed obtained in non-raining conditions WSPD_LF, the global wind speed through rain and the H-wind derived wind speed.
In addition to the all-weather wind speeds, we produce two standard rain-free radiometer wind products: WSPD_LF (low-frequency) and WSPD_MF (medium frequency). The first, WSPD_LF is created using the frequency channels at 10.7 GHz and above (see Table 1) and is most similar to the AMSR-E wind. The second, WSPD_MF, uses frequency channels at 18.7 GHz and above and is most similar to the SSM/I wind. Each wind product has distinct advantages. The WSPD_LF is less affected by the atmosphere and rain, but is affected by 10.7 GHz RFI and sun glitter effects. The WSPD_MF has a higher effective spatial resolution, is less affected by ice and land contamination, and only slightly affected by sun glitter effects, and RFI. The WSPD_MF are a little noisier than the WSPD_LF.
Gridded Binary Data Files
We produce WindSat Daily and Time-Averaged (3-day, weekly and monthly) binary data files. The files contain no headers or metadata and are a simple set of single-byte arrays as described below. Since the data are stored as single bytes, no byte-swapping when transfering between machines is necessary.
Gridded data files are organized according to observation date. All dates and times are Coordinated Universal Time (UTC), also known as Greenwich Mean Time (GMT), Zulu Time (Z), Universal Time (UT), and World Time. Data products include daily and time averaged measurements as follows:
orbital data mapped to 0.25 degree grid
divided into 2 sets of maps based on ascending and descending passes
when adjacent passes overlap at high latitudes and daily "seam", earlier data is chosen. (See note below)
|3-Day||average of 3 days ending on and including file date|
|Weekly||average of 7 days ending on and including the Saturday file date|
|Monthly||average of all data within the calendar month|
Note that all the ocean measurements are gridded onto a 0.25 deg map (approximately 25 km grid). However, the effective spatial resolution for each geophysical parameter is determined by the frequency channels used in the retrieval as listed in Table 2 above.
The details of the ocean measurements in the gridded binary files are listed below.
Table 3: RSS WindSat ocean measurement product details. Measurements are provided in the order contained in the data files.
|Product Description||Scale||Offset||Valid Data Range||Reason for no data|
Minutes since midnight GMT
Fractional hour of day GMT
0 to 1440
0.0 to 24.0
|SST||Sea surface temperature||Temperature of top layer (skin) of water ~1 mm thick||0.15||-3.0||-3 to 34.5 deg||high winds (<20 m/s), sun glint, rain, RFI, near sea ice or land (~75 km)|
|WSPD_LF||10-m wind speed||Wind speed using 10.7 GHZ channels and above||0.2||0.||0. to 50.0 m/s||sun glint, rain, RFI, near sea ice or land (~50 km)|
|WSPD_MF||10-m wind speed||Wind speed using 18.7 GHZ channels and above||0.2||0.||0. to 50.0 m/s||sun glint, rain, RFI, near sea ice or land (~50 km)|
|VAPOR||Columnar atmospheric water vapor||Total gaseous water contained in a vertical column of the atmosphere||0.3||0.||
0. to 75.0 mm
|heavy rain or near land (~25 km)|
|CLOUD||Columnar cloud liquid water||Total cloud liquid water contained in a vertical column of the atmosphere||0.01||-0.05||-0.05 to 2.45||near land (~25 km)|
|RAIN||Rain rate||Rate of liquid water precipitation||0.1||0.||0. to 25.0 mm/hr||near land (~25 km)|
|WSPD_AW||All-weather 10-m wind speed||Wind speed derived using all channels and three separate algorithms to obtain winds in all weather conditions||0.2||0.||0. to 50.0 m/s||sun glint, RFI, near sea ice or land (~75 km in rain, ~50 km no rain)|
|WDIR||10-m wind direction||Oceanographic-convention wind direction, relative to north||1.5||0.||0. to 360.0 deg||low winds (< 3 m/s), RFI, rain > 15 mm/hr, sun glint, near sea ice or land (~50 km)|
The daily files consist of WindSat ocean measurements mapped to a regular grid complete with data gaps between orbits. Two maps exist for each parameter, one of descending orbit segments (local morning passes) and the other of ascending orbit segments (local evening passes). Daily data files contain time maps consisting of the UTC observation time for each set of passes (descending and ascending). Time-Averaged data files do not contain any time information.
Each daily binary data file available from our ftp site consists of eighteen 0.25 x 0.25 degree grid (1440,720) byte maps. Nine descending (morning) maps are provided in the following order: Time (TIME), Sea Surface Temperature (SST), 10-meter Surface Wind Speed at Low Frequency (WSPD_LF), 10-meter Surface Wind Speed at Medium Frequency (WSPD_MF), Atmospheric Water Vapor (VAPOR), Cloud Liquid Water (CLOUD), Rain Rate (RAIN), All-Weather 10-meter Surface Wind speed (WSPD_AW), and Surface Wind Direction (WDIR). Nine ascending (evening) maps follow in the same order.
Daily, 3-day and monthly files are located in the directory /windsat/bmaps_v07.0.1/yYYYY/mMM. The respective file name conventions are wsat_yyyymmddv7.gz, wsat_yyyymmddv7_d3d.gz, and wsat_yyyymmv7.gz. Weekly files are located in the directory /windsat/bmaps_v07.0.1/weeks/, with the file name convention wsat_yyyymmddv7.gz where:
|yyyy||year||2010, 2011 etc.|
|mm||month||09, 10, etc.|
|dd||day||01, 02, etc.|
The center of the first cell of the 1440 column and 720 row map is at 0.125 E longitude and -89.875 latitude. The center of the second cell is 0.375 E longitude, -89.875 latitude.
The data values fall between 0 and 255. Specific values have been reserved as shown below:
|0 to 250||=||valid geophysical data|
|251||=||missing because of bad data quality (sun glint; RFI); missing SST, WSPD_LF or WSPD_MF because of rain; missing data because near sea ice; missing VAPOR because of heavy rain; missing SST because of high wind speed (greater than 20 m/s); missing data due to land proximity.|
|252||=||sea ice, or likely sea ice contamination|
|253||=||observations exist, but are bad (not used in composite maps)|
The data values between 0 and 250 need to be scaled to obtain meaningful geophysical units. To scale the data, multiply by the scale factors (and add the offsets) listed in Table 3.
All binary data files have gzip compression to reduce size and decrease transfer time. They should be downloaded in binary mode.
Consecutive swaths of measurements from polar orbiting satellites partially overlap at high latitudes ( > 60 degrees) and at the beginning/end of the day. When producing daily maps in which swath data are placed on an Earth-grids we can either average observations from different swaths, choose earlier observations or overwrite early observations with later observations. For all RSS satellite products (SSM/I, SSMIS, AMSR-E, TMI, QuikSCAT) except WindSat, daily maps are created by overwriting earlier observations when the time difference between the observations is greater than 50 minutes. For WindSat we use the early observations and do not overwrite unless an observation is missing. This is done so that all geophysical parameters for a given 0.25-deg cell refer to the same observational time. WindSat is more complicated than other satellites because each frequency channel (and thus each geophysical parameter) has a distinct viewing geometry and therefore a slightly different time. If we were to overwrite early WindSat observations with later observations, it would cause large data gaps at high latitudes in the low resolution channels, especially at swath edges.
Browse Imagery on Web Site
Each daily, 3-day, weekly and monthly graphic image on our web site displays one ocean measurement: Sea Surface Temperature (SST), 10 meter Surface Wind Speed at low frequency resolution (WSPD_LF), 10 meter Surface Wind Speed at medium frequency resolution (WSPD_MF), Columnar Water Vapor (VAPOR), Cloud Liquid Water (CLOUD), Rain Rate (RAIN), 10 meter surface All-Weather Wind Speed (WSPD_AW), and Wind Directions (WDIR). The date of the data displayed is the UTC date/day when WindSat made the observation. The scale for each browse image is located next to the map for reference. Though the valid data range (minimum to maximum) is given in Table 3, the scale bars in the browse images are set to visually enhance the data and may vary.
In the browse images white areas represent regions of likely sea ice contamination as determined by the WindSat field of view. In the time-averaged maps, ice is indicated when ice is present over 50% of the time, or when the number of times a cell is identified as containing ice is greater than the number of times the cell contains data. Land regions are colored gray. Those areas where WindSat data are not available are black. The black color includes areas where the satellite did not pass over and no data were collected, areas where data were collected but were determined to be bad, coastal areas, areas affected by sun glint on the SST and wind speed maps, areas near sea ice on the SST and wind speed maps, areas of rain on the SST and wind speed maps, areas of heavy rain on the water vapor map, areas of high wind speed (>20 m/s) on the SST map, and areas of RFI (radio frequency interference).
Browse imagery on our web site display NRT ("rt") data while it is available, and final data thereafter.
The gaps within the daily and 3-day products also reduce the number of observations in weekly and monthly average products. When you browse through the images we have on our web site, the navigation may skip dates for which there are no data, or you may see a blank map stating that no data are available for that time. Binary data files for dates with completely missing data are not produced and they will be absent from our FTP server.
WindSat data products are available for the time period from just after launch, Feb 2003 to the present. As with all the other RSS satellite products, the data are provided in a gridded, binary format file. Graphic browse images of the ocean measurements can be viewed on our web page and the binary data files can be downloaded from our ftp server: ftp.ssmi.com/windsat.
Binary file read routines and verification files are available on our ftp server in the windsat/support directory.
Related Data Sets
The WindSat data are also available from NASA PO.DAAC in netCDF format. The same ocean measurements exist in these files. Metadata are included.
WindSat water vapor values are included in the RSS TPW 1-deg gridded product.
Sea Surface Temperatures from WindSat are extracted and included in the RSS OISST daily product and the L2P GHRSST files.
Gaiser, P. et al, 2004, The WindSat space borne polarimetric microwave radiometer: sensor description and early orbit performance, IEEE Trans. on Geosci. and Remote Sensing, vol. 42 (11), pp. 2347-2361.
Meissner, T., L. Ricciardulli and F. J. Wentz, 2010, The RSS WindSat version 7 all-weather wind vector product, paper presented at International Ocean Vector Winds Meeting, Barcelona, Spain.
Meissner, T., L. Ricciardulli and F. J. Wentz, 2010, Wind Measurements from Active and Passive Microwave Sensors: High Winds and Winds in Rain, paper presented at URSI-F Microwave Signatures Meeting 2010, Florence, Italy.
Meissner, T. and F. J. Wentz, 2006, Ocean retrievals for WindSat: Radiative transfer model, algorithm, validation, paper presented at 9th Specialist Meeting on Microwave Radiometry and Remote Sensing Applications, Puerto Rico, USA.
Meissner, T. and F. J. Wentz, 2009, Wind Vector Retrievals Under Rain With Passive Satellite Microwave Radiometers, IEEE Transactions on Geoscience and Remote Sensing, 47(9), 3065-3083.
Meissner, T., L. Ricciardulli, and F. J. Wentz, 2011, All-weather wind vector measurements from intercalibrated active and passive microwave satellite sensors, paper presented at 2011 IGARSS meeting, Vancouver, BC, Canada.
WindSat data are produced by Remote Sensing Systems and sponsored by the NASA Earth Science MEaSUREs DISCOVER Project and the NASA Earth Science Physical Oceanography Program. RSS WindSat data are available at www.remss.com. We are grateful to Peter Gaiser at the Naval Research Laboratory in Washington, DC and Paul Chang at NOAA/NESDIS/STAR for providing us with WindSat TDR data.
How to Cite These Data
Continued production of this data set requires support from NASA. We need you to be sure to cite these data when used in your publications so that we can demonstrate the value of this data set to the scientific community. Please include the following statement in the acknowledgement section of your paper:
"WindSat data are produced by Remote Sensing Systems and sponsored by the NASA Earth Science MEaSUREs DISCOVER Project and the NASA Earth Science Physical Oceanography Program. RSS WindSat data are available at www.remss.com. "
Sign up for our instrument alert email list
If you would like to be alerted when we have problems delivering these data, please add your name to our email list by submitting a request to firstname.lastname@example.org. Use "add me to the mail list" for the subject line and be sure to specify the instruments you wish to be informed about in the body of the message.