The Precipitation Processing System (PPS) is recalling Level 3 GPROF data for GPM v08 because a minor bug was discovered in the Level 2 GPROF v08 program. GPM science data users should be aware that all GPM v08 L3 GPROF data will be reprocessed from v08A to v08B. L2 will not be reprocessed.
The bug in the L2 GPROF V08 program is related to the "precipitationYesNoFlag" variable value. This flag should have been set when the "ProbabilityOfPrecip" value was over 50%, but the limit of 0.5 was set when this probability was a fraction (0-1.0); it was not scaled up to 50 when it was converted to an integer (0-100). Consequently, the "precipitationYesNoFlag" is set more frequently than it should have been.
The GPROF-NN3D algorithm has a fairly tight distribution of probabilities around the precipitation estimate, so the impact of this error was not immediately apparent in the data. Since this problem can be bypassed with using "probabilityOfPrecip>50", PPS will not reprocess the L2 V08 GPROF products in the middle of the stream. Please use "probabilityOfPrecip>50" rather than the "precipitationYesNoFlag" variable for all L2 V08 GPROF products.
This wrong precipitationYesNoFlag value also caused the incorrect value of "npixPrecipitation" in the L3 V8 GPROF product. The "npixPrecipitation" is not recoverable from the file itself since it is already summed for the month.
PPS will reprocess all L3 V08A GPROF products into L3 V08B GPROF products. If the "npixPrecipitation" variable is used in your work, please wait for the L3 V08B GPROF products. We will inform you once the L3 GPROF reprocessing is completed.
Please direct any questions or concerns to the PPS Help Desk.