Add GEFS programs#52
Conversation
Adds the four programs from the legacy GEFS system that I think will be needed in the combined global workflow. - ensadd Extends grib2 head information to include ensemble info - ensppf Calculates ensemble precip probability forecast - ensstat Computes mean and spread for atmosphere grib2 files - wave_stat Computes mean, spread, and prob for wave grib2 files Resolves NOAA-EMC#51
WalterKolczynski-NOAA
requested review from
BoCui-NOAA,
aerorahul and
bingfu-NOAA
| gfldo%ipdtmpl(1:15)=temp(1:15) | ||
| gfldo%ipdtmpl(16)=ienst | ||
| gfldo%ipdtmpl(17)=iensi | ||
| gfldo%ipdtmpl(18)=10 |
There was a problem hiding this comment.
line 130: gfldo%ipdtmpl(18)=10
I remember this line is to define the total ensemble number. Please correct it from 10 to 30.
| gfldo%ipdtmpl(1:15)=temp(1:15) | ||
| gfldo%ipdtmpl(16)=ienst | ||
| gfldo%ipdtmpl(17)=iensi | ||
| gfldo%ipdtmpl(18)=10 |
There was a problem hiding this comment.
Please modify line 152 from
gfldo%ipdtmpl(18)=10
to
gfldo%ipdtmpl(18)=30
There was a problem hiding this comment.
Is it that wrong in ops too then?
This should be turned into a parameter that is passed in instead of a hard-coded value. Right now this PR was just to get the GEFS code copied over to gfs-utils. Modifications can come later.
There was a problem hiding this comment.
I reviewed ush/gefs_prdgen.sh and checked why the ush/global_ensadd.sh is needed. It seems that global_ensadd.sh is used for GFS file only. The GFS forecast is thought to be one member of GEFS and has perturbation numbers ( 0 0 )
Add ensemble PDS header to GFS file
$ENSADD 0 0 pgb2afile.$ffhr epgbafile
Please correct me if I am wrong. For such a case, there is no need to modify the code.
|
Here are some clarifications for code ensadd.f The gfldo%ipdtmpl(18) come from Octect no 36 in https://www.nco.ncep.noaa.gov/pmb/docs/grib2/grib2_doc/grib2_temp4-1.shtml Use "wgrib2 -ens" to check ensemble files that don't display the ensemble size information. The ensemble size information will be saved and transferred to probability forecasts, mean or spread and ect , i.e how many members are used to do the calculation to get these products. For gfs forecast, choosing " gfldo%ipdtmpl(18)=30" will let the ensemble size information consistent with the GEFS member. However its setting won't affect the gefs mean and spread products' grib2 message for gfs files are not included to do mean/spread calculation. This is a small change in ensadd.f. you can fix it later. |
3 participants
Adds the four programs from the legacy GEFS system that I think will be needed in the combined global workflow. The source code is unchanged, though some directories and one source file has been renamed to remove prefixes.
Resolves #51
Refs NOAA-EMC/global-workflow#2296
Build tested and successful on: