Closed jaymes-kenyon closed 5 months ago
@jaymes-kenyon My UPP RTs indicate the impact only in 'CEIL' PRSLEV dataset of RRFS test as 827:761040843:CEIL:cloud ceiling:rpn_corr=0.975146:rpn_rms=1017.83 I would think that is expected change. Please validate my test at /home/Wen.Meng/stmp2/fv3r_2023062800 on Hera.
@WenMeng-NOAA: Thanks for the test! This is the expected change—looks good!
@WenMeng-NOAA: Thanks for the test! This is the expected change—looks good!
@jaymes-kenyon Thanks !
@FernandoAndrade-NOAA You may start the UPP RTs on Hera and Orion. Please refer to my testing at /home/Wen.Meng/stmp2 on Hera:
Wen.Meng@hfe07 stmp2$ ls -ltr */*diff
-rw-r--r-- 1 Wen.Meng stmp 75 May 2 12:50 rap_2020072316_pe_test/WRFPRS.GrbF16.diff
-rw-r--r-- 1 Wen.Meng stmp 541 May 2 12:52 fv3r_2023062800/PRSLEV10.tm00.diff
-rw-r--r-- 1 Wen.Meng stmp 1009 May 2 12:52 rtma_2023040400_pe_test/PRSLEV00.tm00.diff
-rw-r--r-- 1 Wen.Meng stmp 541 May 2 12:52 fv3r_2023062800_pe_test/PRSLEV10.tm00.diff
-rw-r--r-- 1 Wen.Meng stmp 1009 May 2 12:52 rtma_2023040400/PRSLEV00.tm00.diff
The UPP RTs have completed on WCOSS2. The baselines for fv3r and 3drtma will be recreated with this PR.
@FernandoAndrade-NOAA You may start the UPP RTs on Hera and Orion. Please refer to my testing at /home/Wen.Meng/stmp2 on Hera:
Wen.Meng@hfe07 stmp2$ ls -ltr */*diff -rw-r--r-- 1 Wen.Meng stmp 75 May 2 12:50 rap_2020072316_pe_test/WRFPRS.GrbF16.diff -rw-r--r-- 1 Wen.Meng stmp 541 May 2 12:52 fv3r_2023062800/PRSLEV10.tm00.diff -rw-r--r-- 1 Wen.Meng stmp 1009 May 2 12:52 rtma_2023040400_pe_test/PRSLEV00.tm00.diff -rw-r--r-- 1 Wen.Meng stmp 541 May 2 12:52 fv3r_2023062800_pe_test/PRSLEV10.tm00.diff -rw-r--r-- 1 Wen.Meng stmp 1009 May 2 12:52 rtma_2023040400/PRSLEV00.tm00.diff
Logs pushed up. RTs across Hera, Hercules, and Orion match with your results.
This PR is ready for merging.
[Addresses issue #944]
This updates the GSL "exp1" (experimental) ceiling diagnostic for possible use in RRFSv1. Recent RRFS retrospective tests with this code have produced improved verification statistics relative to the existing "exp1" code. In particular, this update accomplishes the following:
1) Reduces the threshold cloud fraction used to detect a ceiling, from 0.50 (current) to 0.41 (new).
2) Revises the code to detect shallow fog, and makes this code more readable / understandable. Any shallow fog that is detected will not yield a very low ceiling.
Testing of this code has been successful on Hera in a RRFS case (CONUS domain).