RRFS_dev1: save background in fcst/INPUT and run model executable from local run directory#287
Conversation
…le to run directory before run the model.
hu5970
requested review from
CurtisAlexander-NOAA,
JeffBeck-NOAA,
TrevorAlcott-NOAA,
christinaholtNOAA,
chunhuazhou,
guoqing-noaa and
terraladwig
as code owners
| # | ||
| #----------------------------------------------------------------------- | ||
| #SFC_CYC=2 | ||
| if_update_ice="TRUE" |
There was a problem hiding this comment.
Do we have an parameter in config.sh to turn on/off if_update_ic?
There was a problem hiding this comment.
Yes, We have. No true/false but use cycle time to turn it on. If you set 00-23, it will be on with those cycles. The off is 99.
There was a problem hiding this comment.
Got it. But here line 473 hardcodes "if_update_ice" to true. Is this temporarily?
There was a problem hiding this comment.
I did not decide if this will be an option or remove it. Right now, cycle surface and update_ice go together.
| cp_vrfy ${bkpath}/sfc_data.tile7.halo0.nc sfc_data.tile7.halo0.nc | ||
| ln_vrfy -s ${bkpath}/gfs_bndy.tile7.000.nc bk_gfs_bndy.tile7.000.nc | ||
| ln_vrfy -s ${bkpath}/gfs_data.tile7.halo0.nc bk_gfs_data.tile7.halo0.nc | ||
| ln_vrfy -s ${bkpath}/sfc_data.tile7.halo0.nc bk_sfc_data.tile7.halo0.nc |
There was a problem hiding this comment.
Does "ln_vrfy" works for the purpose to "save background in fcst/INPUT"?
There was a problem hiding this comment.
Yes, we only need to know where the background come from.
1) diag_table changes: turn on “ebu_smoke” and “ext550” 2) namelists changes to fix streak problem: fv_sg_adj = 7200; n_sponge = 65;sg_cutoff = -1 3) remove bufrsnd in 18-h cycles 4) add prepbufr as dependence in GSI prod cycle Co-authored-by: ming hu <ming.hu@clogin04.cactus.wcoss2.ncep.noaa.gov>
2 participants
Add function to save background in fcst/INPUT, so we can easily repeat the GSI or find analysis increment.
Copy model executable to run directory before run the model, so the model executable will not be locked for hours.