Operational Calibration of the Imagers and Sounders
on the GOES-8 and -9 Satellites, Page 4

Michael Weinreb, Michael Jamieson, Nancy Fulton, Yen Chen, Joy Xie Johnson,
James Bremer, Carl Smith, and Jeanette Baucom

2. INFRARED CALIBRATION DATA

In orbit, the imagers and sounders periodically view space and their on-board warm blackbodies to acquire data for calibrating their infrared channels. Each instrument's blackbody is in front of its entire optical chain and fills its optical aperture. (This is in contrast to the Visible Infrared Spin-Scan Radiometer [VISSR] and VISSR Atmospheric Sounder [VAS] instruments, carried on earlier geostationary satellites. In those instruments, the calibration target was behind the fore-optics, necessitating modelling of the radiative contribution of the fore-optics in the calibration⁴.) The data acquired at the blackbody looks (and associated space looks) are used for inferring the instruments' calibration slopes (radiance increment per output count), which are the reciprocals of the responsivities. The data from the space looks allow us to infer the calibration intercepts (radiance at zero counts).

To preserve the precision of the measurements, the instruments are calibrated often, because temperatures on a three-axis stabilized satellite such as GOES vary diurnally by tens of degrees Kelvin. The diurnal variation of the intercepts, which measure the intensity of the radiation emitted by the instrument itself, is considerably greater than that of the slopes, which depend weakly on such quantities as the background flux on the detectors and the temperature of the electronics. Therefore, space looks are executed more frequently than blackbody looks. For the imagers, another reason to view space as frequently as possible is provided by the presence of 1/f noise⁵ in the channels that use photoconductive HgCdTe detectors. The 1/f noise manifests itself as a drift in the imager's output in the time between the space clamps. While viewing space, the imagers execute their DC signal restores ("space clamps"), which reset the zero-radiance output to a predetermined value (nominally 970 counts). The more frequent the clamps, then, the less severe the drifts. The sounders chop the signal from the scene at a frequency of approximately 50Hz against opaque "teeth" located between the filters on the filter wheels, which largely suppresses the effects of 1/f noise.

The intervals between calibration measurements are listed in Table 3. For the imager, the space looks for calibration all involve space clamps. Currently, the 2.2-sec space-look interval is used for imaging the full Earth, and the 36.6-sec interval for imaging smaller sectors. The 9.2-sec interval is not used for routine imaging. This selection of intervals represents a compromise between radiometric precision and scheduling requirements imposed by the NOAA/National Weather Service.

Table 3. Intervals between calibration measurements

For the sounder, a space look may interrupt a scan line in progress. A space look consists of the acquisition of 40 samples of data at a location at least 0.5 degrees away from the Earth. Typically, this requires a total of approximately 8 sec, of which 4 sec is devoted to taking the data and the rest to slewing and settling of the scan mirror. A blackbody sequence, which may interrupt a frame (a pre-defined rectangular target area on the Earth) in progress, consists of acquisition of 40 samples from a view of space (4 sec), data for a check of the linearity of the electronics (1.6 sec), and 40 samples during the view of the blackbody (4 sec). The entire sequence requires approximately 55 sec, of which approximately 45 sec are devoted to slewing and settling of the scan mirror. The view of the blackbody occurs approximately 23 sec after the view of space.

For the imager, space looks occur only during scan reversals, i.e., when the direction of the scan mirror's motion reverses between two scans in opposite directions. The location is at least 0.5 degrees from the edge of the Earth. The sequence of events at the reversal is acquisition of approximately 400 samples from a view of space, the DC restore (clamp), and another acquisition of 400 samples of data from a view of space immediately following the clamp. These two views of space are called the "pre-clamp" and "post-clamp" views, respectively. Acquisition of 400 samples requires approximately 73 msec, and an entire scan reversal, including the clamp and the two space views requires 200 msec. The purpose of acquiring two sets of data at each space look is to combat the effect of drifts during the period of data-taking on the Earth that occurs between any two space looks. The calibration intercepts are interpolated (see below) in time from the post-clamp of the first space look to the pre-clamp of the second.

Imager blackbody sequences occur every half hour between frames. A sequence consists of acquisition of 400 samples from the post-clamp phase of a space look (73 msec), 1000 samples during the view of the blackbody (183 msec), and 400 samples (73 msec) from the pre-clamp phase of the space look following the blackbody. The entire sequence requires approximately 44 sec, and almost all of that is dedicated to the slewing and settling of the scan mirror. The blackbody observations take place approximately 18 seconds after the first space look and 18 seconds before the second. To minimize the effect of drifts that will almost certainly affect the instrument outputs in the intervals between the views of the blackbody and space, we interpolate the outputs from the two space looks to the time of the blackbody look. This accounts for the linear component of the drifts but does nothing to correct for higher-order components⁵.

Next  |  Previous  |  Table of Contents

Contact Michael P. Weinreb at michael.weinreb@noaa.gov
Latest Revision: July 9, 1997