The SeaWinds/Quikscat (Qscat) instrument is a radar scatterometer designed to measure the sigma-0 of the ocean's surface. In making measurements of sigma-0, Qscat also measures the system noise power. These noise measurements can be converted into measurements of the apparent brightness temperature of the Earth. Over the ocean, rain results in higher brightness temperatures than the ocean alone. Thus, brightness temperature measurements can be useful in locating rain.
The Qscat design is optimized for operation as a radar rather than as a radiometer. As a radiometer, it design is sub-optimum though good quality radiometric measurements are obtained. The following paragraphs consider some of the known limitations of the data.
Because the radiometric and radar measurements are made simultaneously, some contamination of the radiometric measurement by the radar signal can be expected, particularly for large attitude excursions which might shift some of the return echo signal outside of the signal measurement spectrum. However, very little signal `spill over' has been observed in the data studied thus far.
The absolute calibration of the measurements is somewhat uncertain since no provisions for absolute radiometric calibration of the Qscat receiver are incorporated into the design or operation since they are not needed for radar operation. The short-term (hours to days) relative calibration is considered to be accurate to a few tens of Kelvin. Long-term calibration will be established by comparison to other satellite radiometer systems (e.g., SSMI and TRMM-TMI). The radiometric precision (delta-T) for a single pulse observation is approximately 25 K. This is reduced by averaging multiple pulses with an improvement being 1/sqrt(number of pulses averaged).
Tb measurements extracted from the noise-measurements are averaged onto the wvc grid used for Qscat L2A/L2B products. The grid resolution is 25 km. The product is rev-based, with one file per rev.
This product is generated using information from both L1A and L2B files. In generating this product the individual noise measurements contained in the L1A product files are converted to noise-only measurements by subtracting out the signal power contained in the slice measurements. The resulting noise-only power estimate is converted to effective brightness temperatures (Tb's) using calibration coefficients determined by L. Jones.
In order to Earth-locate the Tb measurements, the lat/lon location of the corresponding pulse in the L1B file is extracted. The wind vector cell (wvc) in which the measurement center falls into is then determined by computing the along-track/cross-track grid index. Note that only Tb measurements corresponding to 'usable' (according the Qscat L1B SIS) sigma-0 values are retained; the others are discarded. Frame quality and other flags are also examined, with measurements discarded which do not pass very conservative flag checking.
The mean, standard deviation, and count of the Tb measurements (for a given beam) falling into each wvc are then determined. This results in two Tb measurements, one for each polarization or beam, for each wvc along with the counts and standard deviations. Note that Tb measurements for both forward-looking and aft-looking azimuth directions are averaged into the final product.
In generating this product, Tb measurements computed from each pulse are accumulated in a swath-based grid without regard to azimuth angle or measurement time, separated only by the beam (inner=h pol, outer=v pol). Thus Tb measurements for both forward-looking and aft-looking azimuth directions are averaged into the final product. Only data for a single rev is averaged.
The product is stored in HDF format. Tb measurements are stored in the same grid as the L2B wind product. The Tb, count, and standard deviation arrays are stored as scientific data sets. Global attributes are copied from selected L1B global attributes and are stored in the same manner. In addition, two new global attributes are included. These give the full file names of the L1A and L1B data files used to create the brightness temperature product.
The 6 scientific data sets in each file are 1624 by 76 arrays containing the mean Tb values, the counts, and standard deviations for each wvc. The SDS names are Tb_h, Tb_v, Tb_hcnt, Tb_vcnt, Tb_hstd, and Tb_vstd. The cnt's are stored as 4 byte integers while the other variables are stored as floats.
The global attributes (with typical values) are:
| Source_L1A_file* | QS_S1A00678.19992301242 |
| Source_L1B_file* | QS_S1B00678.19992301325 |
| LongName | QuikSCAT L2B Radiometer Measurements in 25km Swath Grid |
| ShortName | QSCAT_RadMode_L2 |
| producer_agency | NASA |
| producer_institution | Brigham Young University |
| PlatformType | spacecraft |
| InstrumentShortName | SeaWinds |
| PlatformLongName | NASA Quick Scatterometer |
| PlatformShortName | QuikSCAT |
| project_id | QuikSCAT |
| data_format_type | NCSA HDF |
| QAPercentOutofBoundsData | 0 |
| QAPercentMissingData | 0 |
| build_id | QS_revrad Version 3 |
| ProjectionDateTime | 19992450910 |
| RangeBeginningDate | 1999-217 |
| RangeEndingDate | 1999-218 |
| RangeBeginningTime | 23:17:50.690 |
| RangeEndingTime | 00:58:54.107 |
| rev_number | 678 |
The global attributes marked with `*' (Source_L1A_file and Source_L1A_file) are new. They are the file names for the L1A and L1B data files used in creating this product. All other attributes are identical to those documented in the Qscat L1B SIS.
Questions concerning data distribution should be be directed to the PO.DAAC. Issues related to data quality or processing should be directed to Profs. David Long or Linwood Jones. Please note that e-mail is the preferred means of communication.
Prof. W. Linwood Jones
Central Florida Remote Sensing Laboratory
Dept. of Electrical and Computer Engineering
P.O. Box 162450
Orlando, FL 32816-2450 USA
e-mail: ljones@iu.net
Telephone: (407) 823-6603 FAX: (407) 823-5835