ISCCP CATALOG OF DATA AND PRODUCTS Section 5

Table of Contents, Acronyms, Previous Section, Next Section

5. BT-DATA DESCRIPTION AND HOLDINGS

BT data consist of calibration look-up tables to convert count values to physical units for each image that ISCCP has processed into B3 format. Only images which have actually been processed by ISCCP have calibration tables included. All information on these tapes is written as ASCII text.

A B3 image may have radiances for up to a maximum of five channels; the number of calibration tables varies from satellite to satellite according to the number of channels actually available and is indicated in the file header record. For each channel present, there are six calibration tables reported: 1) Nominal Radiance, 2) Normalized Radiance, 3) Absolute Radiance, 4) Nominal Reflectance or Temperature, 5) Normalized Reflectance or Temperature, and 6) Absolute Reflectance or Temperature. The Absolute tables represent the best available calibration information used by ISCCP. (Note that ISCCP calibration are only available for the standard VIS = 0.6 micron and IR = 11 micron channels; for other spectral channels, the normalized and absolute calibrations are the same as the nominal calibration.)

In general, the only calibration information that is available for VIS data is the pre-launch instrument calibration with no post-launch monitoring of instrument sensitivity. In contrast, all IR measurements are actively re-calibrated by routine observations of an on-board calibration standard; however, the calibration procedures and the frequency of re-calibration vary among the different satellite systems. Moreover, there is usually no attempt made to relate the calibrations of the different radiometers in the same satellite series nor is there information relating the calibrations for different satellite systems. Thus, the ISCCP has implemented procedures to document radiometer characteristics, to monitor the performance of all radiometers throughout their operational life, to normalize all radiance measurements to a single calibration standard, and to obtain accurate absolute calibrations for the reference standard. Taken together, the results of these procedures provide a single, uniform calibration for all of the measured radiances.

The ISCCP Radiance Calibration Data set reports the results of the ISCCP calibration procedures in the same form as in the Stage B3 data set: tables for each satellite image listing the physical radiance values for each radiance count value. There are two purposes for collecting these tables into this separate data set. First, these tables can be used to calibrate other versions of the radiance images obtained from the same satellites, including the original resolution images. Although the ISCCP results are reported only for those images actually present in the ISCCP Stage B3 data set, the time intervals between images are short enough that linear interpolation between tables can provide useful estimates of calibration for any image in between.

Second, as calibration continues to be studied and refined, changes in calibration can be "published" by issuing a separate, revised calibration data set more easily than changing the entire Stage B3 data set. Thus, the entire ISCCP Radiance Calibration data set may be replaced to supply an improved calibration of these radiances; however, all versions will be archived. The first version (Version No. = 0) of the calibration data set reports the calibration originally placed in the Stage B3 data set, which uses NOAA-7 VIS and IR in July 1983 as its calibration reference standard. This calibration was used to produce the first version of the cloud climatology with one change: the VIS values were multiplied by 1.2 to account for later aircraft calibration results (Brest and Rossow 1992). Version 1 is the first revision of the whole ISCCP calibration, which includes, in addition to small corrections to some specific images, a refinement of the NOAA series VIS calibrations to change the calibration reference standard to NOAA-9 (particularly for NOAA-7 and NOAA-11) and improved normalization of the NOAA IR calibration. This version of the calibration will be used to produce the second version of the ISCCP cloud climatology.

As a final step in the calibration, the statistics of the cloud and surface properties retrieved from co-located visible and infrared radiances are compared after processing for each geostationary satellite with the reference polar orbiter. If these statistics still show a systematic difference of more than 1% in reflectivity or more than 1K in temperature, the calibration is revised and the processing repeated. For data processed after June 1991, this results in small changes in the calibration, so there will generally be a higher version number for the BT calibration data representing this last processing step.

From 1996 onwards, two additional spectral channels became available for almost all satellites: a "split-window" channel at about 12 microns wavelength and a "water vapor" channel at about 6.5 microns wavelength. These extra channels are being routinely normalized by the SCC. Also about this time, the normalization procedure was changed to be performed every month, instead of every third month. The first version of B3 and BT data to be delivered has both the original and normalized calibrations for these extra channels; however absolute calibrations have not yet been obtained (Rossow et al 1996a). These will be delivered in a later version of the BT data.