"%m2/$MM" in fullchem simulation cannot recognize the current month
Closed this issue · 5 comments
Name and Institution (Required)
Name: liumy
Institution: CAMS
Confirm you have reviewed the following documentation
Description of your issue or question
Hi all, I want to run the fullchem simulation, starting from July 2019, and I encountered a strange problem.
The wildcard, "%m2" in ExtData.rc or "$MM" in HEMCO_Config.rc, which seems to be unable to find the correct simulation time.
As shown in the ERROR log, the month of the file name is "01".
As shown in the HEMCO_Config.log and ExtData.rc, the month of the file name is wildcard.
As showin in the GCHP log, the "Current Date" is "2019/07/01".
There are these two files in the “./HcoDir/C2H6_2010/v2019-06/” path
Could someone tell me where I am not configured correctly? I would be very grateful.
Please provide as much detail as possible. Always include the GCHP version number and any relevant configuration and log files.
logfile.000000.out.txt
job.2529349.err.txt
HISTORY.rc.txt
HEMCO_Diagn.rc.txt
HEMCO_Config.rc.txt
geoschem_config.yml.txt
GCHP.rc.txt
gchp.20190701_0000z.log.txt
ExtData.rc.txt
cap_restart.txt
allPEs.log.txt
Hi @helpyuan, I wonder if ExtData is looking for the previous month. Unable to find it, it may be searching all the way back to January. Try downloading the June 2010 file and run again. Also turn on ExtData prints to get more information in allPEs.log. See GCHP ReadTheDocs Debugging page for instructions on how to do that. We keep ExtData debug prints off by default because it slows down the model.
Thanks @lizziel. I downloaded the data for June 2010, but unfortunately, GCHP still reported the same error. I turn on ExtData prints to get more information in allPEs.log, as shown below:
allPEs.log.txt
I'm not sure if this is related to fixed year data like this. Because after I changed the C2H6_global_anth_biof.2010%m2.2x25.nc in Extdatac.rc to C2H6_global_anth_biof.201007.2x25.nc and the C2H6_global_anth_biof.2010$MM.2x25.nc in HEMOrc to C2H6_global_anth_biof.201007.2x25.nc, GCHP seemed to recognize this emission. However, similar errors have been reported in other emissions, and the relevant log files are as follows:
job.2541503.err.txt
gchp.20190701_0000z.log.txt
HEMCO_Config.rc.txt
ExtData.rc.txt
allPEs-Bry.log.txt
Hi @helpyuan, I dug a bit more into the MAPL by following the traceback in your error log:
nf90_open: returned error code (2) opening ./HcoDir/STRAT/v2015-01/Bry/GEOSCCM_Bry.200701.day.nc [No such file or directory]
pe=00003 FAIL at line=00280 NetCDF4_FileFormatter.F90 <status=2>
pe=00003 FAIL at line=00097 DataCollection.F90 <status=2>
pe=00003 FAIL at line=02799 ExtDataGridCompMod.F90 <status=2>
pe=00003 FAIL at line=02054 ExtDataGridCompMod.F90 <status=2>
It looks like during intitialization MAPL just checks if the January file is available for monthly climatology files, and uses that to make metadata for the import. This is probably because it assumes you have all 12 months of data for your simulation, which makes sense because it assumes you will be running a full year, or at least any month of the year. I bet if you download the full year of data it will work. Alternatively, you can turn off the climatology flag in ExtData.rc and simply use the frequency already set (F2016-%m2-01T00:00:00). This should work equally well since it bypasses some of the logic specific to climatology data.
Also note that the filename in HEMCO_Config.rc is not actually used. I saw your message that you updated that line as well. That is not necessary for GCHP. We keep it consistent with ExtData.rc simply for reference and comparison with GC-Classic HEMCO_Config.rc files.