NUMBER 62

DATE: 10/07/88

Sector Processing Center (SPC):

During August, NOA, ESA, and JMA collectively processed greater than 96% of A data into B1/B2 data. AES processed 73% of A data into B1/B2 data.

Normal operations were reported by NOA and JMA.

ESA reported that the transfer of imaging operations from METEOSAT-2 to METEOSAT-3 (renamed from P2) took place at 0800-0830 GMT on 11 August 1988. Data from this time through 1530-1600 GMT on 12 August are suspect because of calibration difficulties.

JMA also sent reports covering June and July operations; more than 99% of A data were captured in both June and July. These data sets were delivered to the GPC in B2 form.

CSU sent a report for July, but no report for August. CSU also reported that, although the new conversion table to change the AAA format 10-bit IR data to 8-bit data was implemented on 1 November 1987, the old conversion table was inadvertantly used on two occasions this past summer. Checking the consistency of these data may be more difficult as a result.

AES reported loss of data on 4-5 August caused by a failure in the front-end receiving system; a loss of data on 27 August was caused by an overflow in the B1 storage file.

Special Area Processing Center (SAPC):

No reports were received.

Satellite Calibration Center (SCC):

No report was received.

Global Processing Center (GPC):

The GPC continued to receive B2 data and correlative data in a routine manner.

Re-inspection of all B3 data to detect smaller, short-term calibration anomalies continues. Inspections and corrections for all B3 data for Jul 83 - Sep 83, for Jan 84, and Apr 85 and for NOAA and GOES B3 for Oct/Nov 86 and Jun/Jul 87 have been completed.

The new set of IR calibration tables, received from JMA, that were supposed to eliminate the spurious diurnal cycle in IR calibration, have been tested by examining their effect on tropical sea surface brightness temperatures (statistically, the warmest 10% of the image data), using the new calibration-checking procedures. While the corrections for most GMTs seem beneficial, the "midnight" image correction does not seem correct; the correction seems to be about twice as large as needed. This is being investigated further. Since corrections to most data for 1983 and 1984 are very small, these data will be corrected by implementing the new tables from JMA and adjusting the midnight image separately. Since the new calibration tables change the images used by the SCC for normalization, the GPC must also correct the BC coefficients to offset these changes.

The problem of diurnal variations of GOES IR calibrations gets more interesting. We can now confirm the suspicion that the problems are artificial, since we find that the "regular" daily variations are interrupted on weekends and holidays. This pattern of behavior does not occur in all months, however; but we find that the amplitude of the variations is similar for different, widely separated episodes (e.g., for Jan 84 and Apr 85).

NOAA-9 B3 re-calibration has been completed through February 1988 and the last of these data are being delivered to the ICA; however, the B3 data for October 1986 are still under investigation for the calibration anomaly observed by the SCC. The statistics that we examine do not show any sign of such a change and no report from NOAA has been received as yet.

A code error in the NOAA processing has been discovered in re-examining the calibrations. This error occurs in the changes of IR channel count values to allow use of a single calibration table for the whole image. The magnitude of the error for the 10.5 and 11.5 m channels is less than 2 counts (equivalent to about 0.5 K), but is several counts for the noisier 3.7 m channel. The error may introduce a small (about 1 K) systematic day-night difference in the data. The small magnitude of the error would not cause significant problems for most analyses; however, we plan to completely replace the NOAA B3 data over the next six months to eliminate this problem.

MET-2 B3 has been produced through May 86 and is being prepared for delivery to the ICA; production continues.

Production of TOVS data for 87 has been completed; the TOVS data for 86 still awaits replacement data sets. Production has also been extended through Aug 88 and will continue as data are received; deliveries will be on an annual basis.

C1 data for January 84 and April 85 have been completed; they will be delivered to the ICA next week after inspection. The January 84 C1 data were completed after re-processing the GOES-6 B3 data with a corrected IR calibration. Both the GOES-6 and GMS-3 B3 data for April 85 had to be corrected for spurious diurnal cycles of IR calibration before producing the C1 data. In addition, software changes in the cloud processing were required to handle the changed position of GOES-6, which alters all mapping routines and the hierarchy of selecting data for merging into C1. These procedures are being revised so that these changes occur automatically, given the nominal sub-satellite longitudes of the geostationary satellites for a given month.

The special FIRE C data sets from NOAA-9 for Oct/Nov 86 and Jun/Jul 87 have been re-done with corrected calibrations; the GOES-6 data sets for Oct/Nov 86 are in production. Design of a format for delivery of these data should be completed within two weeks.

ISCCP Central Archive (ICA):

The ICA continued to receive B1 data from NOA, ESA, JMA, CSU, and AES and B3 and C data from the GPC in a routine manner.