This site is an archive of model forecast precipitation vs. 24-h (12Z-12Z) observed precipitation analysis. Near the top of each day's page is an index of available comparisons (models and forecast lengths) for that day. Forecast lengths vary, but the accumulation period is always 12Z-12Z. For example, on 1 Dec 2001, the 06-30h forecast of the 10km Eta Western Nested run (which has a starting time of 06Z) was verified against the 12Z-12Z precipitation analysis.
Models being verified:
Log of changes
Log of problems
| 2001 | ||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Dec | 01 | 02 | 03 | 04 | 05 | 06 | 07 | 08 | 09 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | |
20011201: Starting this page
20011206: Added verification for 48h forecasts
20020123: Switched to the CPC 0.125 deg daily precipitation analysis for verification
20020123-20020423: 32km Eta parallel runs (etav, etal, etay, etaz) were plotted incorrectly (wrong grid was used)
20020626: Extended verification from 48h to 84h
20020901: Started verifying the three CONUS 8km NMM (non-Hydrostatic Mesoscale Models)
Changed the name AVN to GFS (for Global Forecasting System)
Started plotting the verifying precipitation (the CPC 1/8 degree analysis of RFC gauge data, on the right-hand-side column) by first mapping it to the 12-km AWIPS grid 218, for easier viewing and comparison.
20050125: ETA is renamed NAM (North American Mesoscale)
20020423: up this date, the verification plots for the 32km Eta parallel runs
are wrong, because the plotting program was using some incorrect
grid specifications.
20080101-20080226: ECMWF plots were flipped in the N-S direction. The ECMWF
precipitation files have (1,1) in the "northwest" corner rather than
the "southwest" corner. For precipitation verification, we have
been changing the order of the ECMWF array to the more conventional
form [with (1,1) on the southwest corner] and that had worked well
when plots were generated by NCARGraphics. Model QPF plots began
to be generated by GEMPAK when precipitation verification became
operational. GEMPAK software anticipates the "reverse" order of the
original ECMWF data and is hard-wired to flips the data in the N-S
direction when encountering GRIB header saying that the data are
originated from ECMWF. This causes our plots to be reversed in the
N-S direction. The solution is to change the 'originating center'
info in our flipped ECMWF array into something not recognized by
GEMPAK as 'from ECMWF' so the array is not flipped again during
plotting.