Hello,
I am having an error when running the Othar section. Running the Monthly_othar_12US1_2016ff_16j.csh script, it directs me to the log file;
/projects/b1045/SMOKE.EMF/2016_beta/2016ff_16j/intermed/othar/logs/smkreport_othar_jan_2016ff_16j_invgrid.log
Which shows:
File "SCCDESC" opened for input on unit: 93
/projects/b1045/SMOKE.EMF/2016_beta/ge_dat/smkreport/sccdesc_2014platform_07dec2018_nf_v2.txt
Value for REPLABEL: 'othar_jan'
Value for REPLABEL: 'othar_jan'
Reading source data from inventory file...
Generating unique lists from inventory data...
Reading gridding matrix...
ERROR: Gridding matrix dimension is inconsistent with records count!
Delete gridding matrix and recreate it.
*** ERROR ABORT in subroutine RDGMAT
I’m using conus_canda_mexico_4k_beta spatial surrogates. I don’t know where the error message “delete gridding matrix and recreate it” comes from, and which gridding matrix it is referring to. I’ve tried to run this script for Dec 2015 and Jan 2016 and neither work.
Any help would be greatly appreciated, thank you!
Anastasia
We are not sure of the cause of this error. If you can provide the Grdmat log and the full Smkreport ‘invgrid’ log that would help track it down.
In the meantime, we suggest that you skip Smkreport altogether, by adding this to the run_settings.txt file:
(Replace “12US2” with the name of the grid that you are running)
othar, 12US2, smkreport, PART1, 0, 0, N
1 Like
Here is log:
smkreport_othar_jan_2016ff_16j_invgrid.log.txt (18.4 KB)
And the gridmat file says it’s too large for upload (5000 KB).
But the end of the gridmat file says:
File "SRGPRO_PATH" opened for input on unit: 93
/projects/b1045/SMOKE.EMF/2016_beta/ge_dat/gridding/CONUS_CANADA_MEXICO_4K_beta/./CA_1263_NOFILL.txt
WARNING: Re-normalizing county total surrogate fractions of Co/St/Ct: 000000124102 :: Surrogate: 01263 greater than 1 :: 1.00000298
WARNING: Re-normalizing county total surrogate fractions of Co/St/Ct: 000000135101 :: Surrogate: 01263 greater than 1 :: 1.00000048
Looping surrogate 1265 through sources to generate gridding matrix
File "SRGPRO_PATH" opened for input on unit: 93
/projects/b1045/SMOKE.EMF/2016_beta/ge_dat/gridding/CONUS_CANADA_MEXICO_4K_beta/./CA_1265_NOFILL.txt
WARNING: Re-normalizing county total surrogate fractions of Co/St/Ct: 000000124102 :: Surrogate: 01265 greater than 1 :: 1.00000024
Looping surrogate 9450 through sources to generate gridding matrix
File "SRGPRO_PATH" opened for input on unit: 93
/projects/b1045/SMOKE.EMF/2016_beta/ge_dat/gridding/CONUS_CANADA_MEXICO_4K_beta/./CA_9450_NOFILL.txt
Writing out GRIDDING MATRIX file...
Value for IOAPI_LOG_WRITE not defined;returning default: TRUE
Record written to AGMAT
GRIDDING-MATRIX statistics:
Total number of coefficients : 16508374
Max number of sources per cell : 571
Min number sources per cell > 0: 1
Mean number of sources per cell : 19.69
--->> Normal Completion of program GRDMAT
I have amended the Smkreport as well and resubmitted the job. This is the new smkmerge file:
smkmerge_othar_jan_2016ff_16j_20160104_CONUS4K_d02_cmaq_cb6.log.txt (25.2 KB)
Ok noticed that probably it’s the error between what the gdmat is putting out for the rdgmat to pull in to create the agmat file.
ncdump -h /projects/b1045/SMOKE.EMF/2016_beta/2016ff_16j/intermed/othar/agmat_othar_jan_CONUS4K_d02_2016ff_16j.ncf
netcdf agmat_othar_jan_CONUS4K_d02_2016ff_16j {
dimensions:
TSTEP = 1 ;
DATE-TIME = 2 ;
LAY = 1 ;
VAR = 1 ;
ROW-NO = 838530 ;
SP-COEF-NO = 16474970 ;
SP-COEF-NO is different than what’s being written out from the GRDMAT file:
Writing out GRIDDING MATRIX file...
Value for IOAPI_LOG_WRITE not defined;returning default: TRUE
Record written to AGMAT
GRIDDING-MATRIX statistics:
Total number of coefficients : 16508374
Max number of sources per cell : 571
Min number sources per cell > 0: 1
Mean number of sources per cell : 19.69
--->> Normal Completion of program GRDMAT
16508374 in gridmat log file, vs. 16474970 in agmat final file
I’ll mention that EPA has never run a full CONUS at 4km. The surrogate files are on a purposefully on a larger grid that has not been used for emissions or AQ modeling. Instead we might have a grid that covers a few states or is centered on a particular state. The same surrogate files can be used for any aligned 4km grid that is a subset of the surrogate grid You would likely need a very big computer to run a 4km CONUS run for the whole country. I don’t think that has ever been tried. You likely should be able to do a CONUS run for 12km if your computer is big enough – we do that routinely.
Hello,
Thank you for your help! I believe I have found a work around –
Because of all of the surrogate related warnings in the log file, this error was resolved by adding SMK_USE_FALLBACK "Y" into the runscript. Thank you @jschnell for suggesting this.