NUMBER 55

DATE: 03/03/88

Sector Processing Center (SPC):

During January, NOA, ESA, CSU and JMA collectively processed greater than 93% of A data into B1 and/or B2 data.

Normal operations were reported by NOA, ESA, JMA, and CSU.

AES reported that their shipment of January B2 data was returned by the post office in severely damaged condition. A new shipment is being prepared for the GPC.

CSU reports that the 0846, 1146, 1745, 2045 and 2345 GMT images are still being transmitted as partial images. They also indicate that NOAA's current estimate of the remaining lifetime of GOES-6 is until June - September 1988.

Special Area Processing Center (SAPC):

No reports were received.

Satellite Calibration Center (SCC):

The SCC reports completion of BC results for GOES-6 for December 84, January 85, December 85, January 86, February 85, March 85, and March 87. These results used the extra AC data received from UWS for January 85 and 86 and eliminate the backlog of BC for GOES-6. No AC data have been received from CSU since March 87 when the transmission format was changed.

The corrected AC data for GMS were received, covering the period March to November 87. New BC results for GMS-3 for April 87 have been completed.

BC data for METEOSAT-2 for July and October 87 have also been completed. Additional results were obtained for September and November 86 as well.

All of these new BC data will be shipped to the GPC shortly.

Global Processing Center (GPC):

The GPC continued to receive B1 and/or B2 data from JMA, ESA, CSU, AES and NOA and correlative data from NOAA/NESDIS in a routine manner.

An adjustment of the IR calibration of GOES-6 B3 data from 23 February through 2 March 1984 by 2 K has been done to offset the anomaly discovered in the QC inspection of these data. These revised B3 data will be sent to the ICA next week after repeating the QC inspection. Another anomaly in VIS data for GOES-6 was determined to be caused by an error in the compiler used for the software; a repeat run of the QC programs with a more recent compiler eliminated the anomaly.

This same problem also affected the QC results for METEOSAT from July 84 through January 85; these QC data have been reprocessed. Detailed inspection of the B3 data confirms that they are not affected.

All of the GMS B3 data from July 1983 through January 1985 have been shipped to the ICA, with the exception of the data for June 1984 which requires receipt of the navigation information to complete. QC inspection has revealed that the first 17 days of VIS B1 data for July 1984 contains the same error that the first AC data did, namely that the values are calculated including the channel on GMS-2 that had failed earlier. This error is corrected on the 18th day, causing a sudden increase in the magnitude of the VIS measurements. The B3 data for July 84 will be reprocessed to correct this problem.

All of the NOAA-8 B3 data (from October 83 through June 84) have been produced and are being prepared for shipment to the ICA. A preliminary examination of the calibration monitor results for NOAA-8, where the global surface reflectances retrieved from both NOAA-7 and NOAA-8 are compared for each month, shows no drift in the NOAA-8 VIS data. A final check of the NOAA-8 to NOAA-7 VIS normalization will be performed before shipment.

With the shipment of the NOAA-8 B3 data, all B3 for the first 19 months of ISCCP will be in the archive. QC inspection for all of these data is complete, with the exception of manual inspection of the last four months of the GMS B3 data and of all of the NOAA-8 B3 data. NOAA-9 B3 production will be re-started next week to provide a longer time record for determining whether the calibration of the VIS channel has drifted. Once the calibration for NOAA-9 has been determined, delivery of B3 data beyond January 1985 can begin.

The final tape version of the ice and snow data for July 1983 through December 1984 has been delivered, along with documentation, to the ICA.

Production of the ISCCP version of TOVS data has been completed for July 83 through June 84; however, two months during this period require replacement tapes before processing. A tape format design has been completed and documentation is in preparation. First delivery of these data to the ICA should begin this month.

Final modification of the cloud detection algorithm has accomplished several improvements in the results. One objective of these revisions was to control the effects of spurious hot IR values, since the use of the maximum value over a time period can provide more sensitive tests for cloud contamination of the clear IR radiances. These tests were added, along with a "spike" filter that reduces some residual effects of cloudiness, leading to a much improved clear IR radiance map. Tests of these results for summer and winter cases show that the best approach is to rely on long-term statistics over the ocean and short-term statistics over land.

We also devised a way to test for a bias in the clear sky IR and VIS values by varying the magnitude of the thresholds used and measuring the changes in the shape of the clear radiance distribution. This procedure is being performed on the January and July data to cover the seasonal range of conditions.

One change that has allowed for the reduction of the threshold magnitudes was to re-classify locations into four instead of two classes. Previously, we classified locations as land or water; the algorithm test values are different for these two surface types. Tests showed, however, that smaller thresholds were possible over most land and water locations if certain, more variable regions were treated separately. The more variable water regions are sea ice covered locations, near-coastal waters, and smaller lakes and seas (including the Mediterranean Sea). The more variable land regions are land ice (Greenland, Antarctica), deserts and high or rough topography.

All of these revisions have improved the detection of marine stratus with IR tests only (important to the nighttime analysis), removed the tendancy to "over-detect" cloudiness over the oceans (which produces fake cirrus), and has allowed for a small reduction of the radiance thresholds (increased sensitivity) without producing spurious clouds. Full production of this part of the cloud analysis started this week.

The radiation code output is being examined further to insure that results in more extreme geometries are accurate. In addition, some limited validation is being repeated to check the improvements in the surface temperature retrievals and the changes in the treatment of small optical thickness clouds. This part of the cloud analysis will be placed into production next week.

Changes in the diagnostic step that produces the C1 data are completed and are undergoing inspection. One significant revision is to include the first derivative of all cloud parameters, in addition to cloud amount. These additional values allow the user to determine, not only the change in cloud amount produced by doubling the detection threshold magnitudes, but also the changes in all the cloud paramters that would result. The new diagnostics also separate the VIS-dependent and IR-dependent values completely, making diurnal analyses easier.

Delivery of the first two months of C1 data (July 1983 and January 1984) is now expected to occur by the end of March or early April.

ISCCP Central Archive (ICA):

The ICA continued to receive B1 data from JMA, CSU, ESA, AES and NOA in a routine manner.

The health of GMS-3, METEOSAT-2, GOES-6, GOES-7, NOAA-9 and NOAA-10 remained good. ESA reported that the launch of METEOSAT P2 is delayed until after 22 May 1988.

Obtaining NOAA-10 B1/B2 data over the Indian sector starting in December 1986.

Arranging for the delivery of INSAT B1 data.

Delivery of all B3 data for 1985 and 1986.

Obtaining digital sea ice data for 1985 through 1987.

Beginning delivery of the final version of C1 data.

Delivery of special C data sets for the polar workshop, FIRE, WGNE, and ERBE.