This Readme file contains information pertaining to the Seasat scatterometer global 50km sigma-0 data produced by F.J. Wentz (PO.DAAC product #031). This file is meant to serve as an update to the original documentation, Description for Wentz Seasat Scatterometer Forward and Aft Sigma-naught binned 50km x 50km . Excerpts from Documentation for Program Order: Collocating SASS Sensor Data in 50 km Bins (Wentz, 1982) have also been included for clarity. To obtain the original versions of these documents, please contact the PO.DAAC User Services Office.

Sections in this file include:

• Data Background
• Description of Data Sets
• Data Quality Flags
• Program to Read the Data
• References
• Other Pertinent Information
  • Important Note
  • Contacting the PO.DAAC User Services Office

Originally, JPL had processed all the Seasat scatterometer wind vectors by pairing radar backscatter coefficients (sigma-0) measurements from forward and aft beam cells. These data are available on the SASS Sensor Data Tapes (referred to as PO.DAAC product 020). Later, F.J. Wentz of Remote Sensing Systems contracted with Atmospheric Environmental Service, Canada (AES) to reprocess the wind vectors onto a rectangular grid oriented along the satellite subtrack.

The first step in the reprocessing was to bin the sigma-0 values from the SASS Sensor Data Tapes into 50-km resolution cells so that both the forward and aft radar cross section measurements would be collocated. The 50-km resolution cells were organized into strips which were 2200 km long in the crosstrack direction and 50-km wide in the subtrack direction. Each strip consisted of fourty four 50-km cells. The strips were perpendicular to the satellite subtrack and were bisected by subtrack. Figure 1 shows the geometry of the strips. One orbit of scatterometer data consisted of 820 such strips if there were no data gaps. Each record corresponded to a single strip of data, and each file corresponded to a single orbit (Wentz, 1982).

Figure 1 Geometry for a Strip of SASS Sensor Data

These data were originally available on 16 sigma-0 tapes covering the entire Seasat mission. Each tape was a 9-track, 6250 bpi, with a RECORDSIZE of 1696 bytes and a BLOCKSIZE of 30528 bytes (where 1 byte is equal to 8 bits). Each tape contained about 6 days of data with a short EBCDIC header file at the beginning, followed by one data file for each rev on the tape, as shown below:

Table 1 Contents of the Original 16 9-Track Seasat Tapes

Tape	Days	Revs	Number of Files	Number of Dat Rec	Missing Revs
1	188 - 193	142 - 228	87	63282
2	194 - 199	229 - 314	86	62356
3	200 - 205	316 - 400	84	59972	(315), 320
4	206 - 211	401 - 486	86	66803
5	212 - 217	487 - 572	86	67039
6	218 - 223	573 - 657	84	64982	629
7	224 - 229	658 - 743	79	61560	660,671,702-706
8	230 - 235	744 - 829	81	62837	745-748,770
9	236 - 241	830 - 915	86	67416
10	242 - 247	916 - 1001	83	66882	998-1000
11	248 - 253	1002 - 1087	86	69022
12	254 - 259	1088 - 1172	85	68722
13	260 - 265	1173 - 1258	86	69266
14	266 - 271	1259 - 1344	86	68464
15	272 - 277	1345 - 1430	86	69551
16	278 - 283	1431 - 1501	71	53791

The number of files indicated for each tape did not include the header file. If an entire rev was missing, there was no file on the tape for it.

Changes to Original Data Format

• These data are no longer available on 16 9-track tapes. They are currently provided on either 8-mm or 4-mm tape in UNIX Tar format. These data are also available via anonymous FTP to podaac.jpl.nasa.gov.
  
  
• As described above, each of the original 16 tapes contained the same EBCDIC header file. This header has been converted from EBCDIC to ASCII without any other changes. Therefore, please be aware that the new header still refers to the old 9-track tape conventions.
  
  
• The data files have been renamed according to rev number. The new naming convention is as follows.

  s0revNNNN_50km.dat
  where:
  s0	denotes	sigma-naught
  NNNN	denotes	the four digit rev number
  50km	denotes	50 km cell
  dat	denotes	RAW Binary data format

  Please note that although the filenames have changed, the data contained in each file are the same as those originally produced by F.J. Wentz.

Each data record corresponds to one 50-km strip which contains forty-four 50-km boxes oriented perpendicular to the nadir track, and has the following format:

Table 2 Format of Each Data Record

First Byte for Type of Element	Number of Bytes Per Element	Number of Elements	Description of Quantity	Units and/or Offset	Note
1	4	1	Time of Nadir Point	seconds since beginning of 1978	A
5	4	1	Time of Last Ascending Node	seconds since beginning of 1978
9	4	1	Longitude of Last Ascending Node	.01 degrees East
13	4	1	Strip Number		B
17	4	1	Geodetic Latitude of Nadir Point	.01 degrees, offset by 9000	A
21	4	1	Longitude of Nadir Point	.01 degrees East	A
25	4	72	Times of Sigma-0 Measurements	seconds since beginning of 1978	C
313	2	44	Number of Sigma-0's in each 50-km bin in strip
401	2	72	Geod. Latitude of Sigma-0's	.01 degrees, offset by 9000	C
545	2	72	Longitudes of Sigma-0's	.01 degrees East	C,F
689	2	72	Mode + Antenna Cell + Polar + Antenna Number		C,D
833	2	72	Incidence Angles	.01 degrees	C
977	2	72	Azimuth Angle of Reference Antenna	.01 degrees	C,F
1121	2	72	Instrument and Atmospheric Corrected Sigma-0	.01 dB, offset by 30000	C,F
1265	2	72	Standard Deviation of Sigma-0	.01 dB, offset by 30000	C,F
1409	2	72	Attenuation	.01 dB, offset by 10000	C
1553	2	72	Quality Flags		C,E,F

Note:

1. The nadir point occurs in the center of the strip on the line between bins 22 and 23.
   
   
2. There are exactly 820 strips contained in each rev. The edge of the first strip in a rev intersects the equator at the nadir track (i.e., the first nadir point lies 1/1640th of a rev above the ascending node along the nadir track). The first strip in rev 1 is number 1, and the subsequent strip numbers increment by 1 throughout the mission. Therefore:

   strip number = strip number within rev + (rev - 1) * 820

   rev= 1 + (strip number -1)/820
   strip number within rev = strip number - (rev - 1) * 820
   
   

3. 72 spaces are provided to contain all the times (or some other variable) associated with the sigma-0 cells which fall into the 44 bins. For example, if the numbers of sigma-0 cells contained in the first 5 bins were 0, 2, 0, 1, and 3, then spaces 1 and 2 would contain the cell times for bin 2, space 3 would contain the cell time for bin 4 , and spaces 4, 5, and 6 would contain the cell times for bin 5. The entire 72 spaces are normally not filled up, nor are unused spaces zeroed out if they were filled in a previous strip. It should be noted that when AES processed the sigma-0's to wind vectors, they used only bins 6 through 39 for both 50-km and 100-km winds. Bins 1-5 and 40-44 virtually never contain enough valid sigma-0's for processing.
   
   
4. Mode, antenna cell, polarization and antenna number are packed decimally into one number:

   number = mode*1000 + ant. cell*10 + pol.*4 + ant. number

   mode=number/1000
   ant. cell = (number - mode*1000)/10
   i = number - mode*1000 - ant. cell * 10
   polarization = (i -1)/4, where 0 = H and 1 = V
   antenna number = i - pol. *4
   
   

5. The quality flags are collections of one-bit or multiple-bit flags. They are not the same as the JPL SASS basic sensor record flags.
   
   
6. Highest order bit may be filled.

Table 3 describes the bits in the cell data quality flags. By interogating these bits, one can determine the quality of a given SASS measurement.

To determine if a measurement should be included in subsequent wind vector processing the following bits are checked: B1, B2, B4, B5, B6, B7, B10, B11, B13, B16. If any of these bits are turned on, then the measurement is excluded. Furthermore, if B14 is off (indicating no new gain correction is made) and B9 is on, the cell is also excluded.

In going from the SASS Sensor Data Tapes to this data product, any data frames for which the following engineering data quality flags are turned on are excluded from these data: EDQF1(B8), EDQF2(B2), EDQF2(B6), EDQF3(B1), EDQF3(B4), EDQF3(B6).

Bits B5, B6, B8, and B10 are the same as on the original NOAA/EDIS tapes. If any of bits B4, B11, B12, or B13 are on, then B14 will be off, indicating no gain correction is made. Bit B4 indicates that there are not enough good cell temperature measurements to accurately estimate the mean noise power. Bit B9 indicates that a bad system noise temperature may have been used in the calcualtions on the SASS Sensor Data Tapes. Bit B13 indicates a channel overflow for the system noise temperature occurred in the NOAA/EDIS tapes. When this overflow happens, the original value for the mean noise power cannot be recovered from the data on the SASS Sensor Data tapes, and hence no gain correction can be made. Bit 16 rarely turns on, but when it does the cell is definitely suspect and should be excluded from any further processing.

Table 3 Description of Data Quality Flag

Bit Number*	When bit is on, it indicates:
1	Land in cell
2	Mixed land and water in cell or unknown surface type in cell
3	Data quality summary flag is on for frame
4	Less than 5 cells in the frame have a system noise temperature between 1175 K and 1375 K for h-pol (1200 K to 1400 K for v-pol) and a signal to noise ratio PR/NP (calculated before the new gain correction) less than or equal to 10.
5	Voltage VSPN for cell is less than or equal to 0.1 volts
6	Voltage VSPN for cell is greater than or equal to 7.0 volts
7	Received power and NRCS for cell are less than zero
8	Summary latest available data (LAD) from previous calibration sequence is used.
9	At least one cell in the frame has a system noise temperature (using old tg) outside the 900 K to 1600 K range
10	Antenna angle PHIONEk for cell is out of range
11	System noise temperature for cell is outside 1000 K to 1500 K range
12	Signal to noise ratio PR/NP (calculated before the new gain correction) for the cell is greater than 10
13	At least one off-nadir cell (cell 1 through 12) has a system noise temperature (using old tg) greater than 6500 K
14	New gain correction is made for cell
15	The equivalent temperature of PSN for cell is four standard deviations below the mean system noise temperature for frame
16	SIGZDB for cell equals flag value of -199.9 dB, or incidence angles for frame do not increase monotonically

*Bits are numbered from right to left, with bit 1 being the least significant bit.

A Fortran program entitled read_seasat_031.f is provided to read the Seasat scatterometer global 50-km sigma-0 data on a UNIX system. Unlike other Seasat products, these data were not produced on a DEC/VAX platform. Therefore it is not necessary to swap bytes to read this data on a UNIX system.

PO.DAAC Documentation, "Description for Wentz Seasat Scatterometer forward & aft sigma-naught binned 50km x 50km", pp. 3.

Wentz, F.J., "Documentation for Program Order: Collocating SASS Sensor Data in 50 KM Bins", RSS Technical Report 113082, 1982, pp. 23.

While the filenames, read software and the manner in which the data is provided has changed, the data are the same as those originally produced by F.J. Wentz and described in the Description for Wentz Seasat Scatterometer forward & aft sigma-naught binned 50km x 50km.

For further information about this data set or to place another data order, please contact: