CZCS MLAC Generation
NASA Ocean Biology Processing Group
CZCS data in the original CRTT (Calibrated Radiance and Temperature Tape) file format contains up to a maximum of 970 lines (approximately 2 minutes) of CZCS data, which was extracted from the periods of the orbit where the CZCS was turned on. The files created this way from the mission frequently overlap 2 or more times in coverage and quality. Thus, merging the data into orbit-sized files has the advantage of having fewer files to deal with, the elimination of duplicate data and the reduction of poorer data in the overalp regions.
The method for merging the files tries to follow 2 rules: keep contiguous data ranges from the contributing CRTT files, and merge the data so that the fewest dropped lines or bad data quality. The first rule is followed because individual CRTT files can be made with slightly different navigation and the mixing of lines from different CRTT files may introduce more navigation jumps than if contiguous data is used. The second rule ensures that the final merged file will contain the fewest number of missing or low quality lines.
The algorithm starts with a list of CRTT files for one orbit. It then orders the files in temporal order and for the first file, it constructs a line-by-line list of the time tag for each line, the existance of data for the line and any quality problem for the line. Any line gaps (seen as time gaps of more than 124 msec) are considered to be missing lines. Each subsequent file is then processed in the same way and the results are placed in comparison storage areas for the same time tag. If there is an overlap that is entirely contained in the previous data, the total number of missing or bad quality lines are compared for both stre4tches and the best stretch of lines is used. If the overlap stretches beyond the previous data, a break is determined where the fewest missing and bad quality lines remain and the lines from the new file are used beyond this break. This process goes on for all the files. In the end, a list exists of the lines to take from each file to make the merged file. The input files are opened and the determined lines are written into the output merged file.
The CRTT data contains several quality values but only a handfull were found to indicate data problems. If any of the following conditions existed, the line was assigned a bad quality:
- The parameter presence global attribute (name parm_presence) indicated that any of the first 5 bands (443, 520, 550, 670, or 750 nm) are not present in the file.
- The calibration summary (name cal_sum) values 3 and 4 (0 origin) are set to 1, indicating bad values. These seemed to indicate that the navigation was not valid for the line.
- The calibration scan quality (name cal_scan) is set to 1 (missing data) for the first 5 bands. This indicated that data for the band was not included on the line.