On-orbit analyses show that the DN (digital number, or count) bin-width fill-factors frequently vary by factors of two and greater from one DN value to the adjacent DN value. These characteristics are most prominent for Bands 31 - 36, but may be present in all bands. These patterns are channel-dependent for Bands 31 – 36.
Impact to L1B Data Set
The non-uniformity is apparent in DN patterns of 2’s, 4’s, 8’s, 16’s and sometimes 32’s. Pre-launch DN quantitative uncertainty was assessed by the instrument builder as up to +2.7/-1.0.
Estimated Impact to L2 Science Products
Analyses of sea surface temperature have indicated that this uncertainty due to non-uniform bin-fill is likely to be the limiting source of error in the SST retrieval.
Cause of Effect
The effect is related to the presence of system noise "at the moment" an analog signal is converted to its digital equivalent. The presence of a noise equivalent to one bit or less will appear as different DN patterns depending on the noise that is present in the moment for which a bit is being determined.
Mitigation Approach/Added Testing to Characterize Effect
Pre-launch sensor thermal vacuum testing of the Terra MODIS showed that this characteristic was less of a problem for the B-side electronics than the A-side electronics. Therefore, the decision was made to switch from the A-side (primary) electronics to the B-side (backup) electronics. The Terra MODIS configuration was switched to the B-side on October 30, 2000 and analyses show a reduction in ADC noise as a result. The so-called "bin-filling" effects appears to have been reduced by a factor of 2 for the LWIR bands used in the Pathfinder SST product.
Estimated Schedule to Complete
Reassess in new side-A era, summer 2001.
History of Characteristic
|Time Span||Version Number at|
start of epoch
|1||Feb. 24, 2000 - Oct. 30, 2000|
2000055.1550 - 2000304.1400
|188.8.131.52||Ubiquitous, but most prominent in LWIR PC bands.||Not applicable.|
|2||Oct. 30, 2000 - June 15, 2001|
2000304.1600 - 2001.
|184.108.40.206||Corrected bug in emissive bands PC_XT pre-processing. PC LUTs updated based on 2000084 moon observation.||Not prominent on side-B.|
|3||July 2, 2001 - present|
2001183.2100 - present
|220.127.116.11||Ubiquitous, but most prominent in LWIR PC bands.||Not applicable.||MsWG histograms, July 25, 2001|