Grid shift for version4.6

SMOKE
#1

Hi everyone,

I’m upgrading SMOKE for my runnings from v3.7 to v4.6. I encountered a grid shift error for the point source. Specifically, almost every grid with high emission has moved two grids up in v4.6 comparing with v3.7. I checked lat lon information of one point from the input inventory and found that the grid in v3.7 is relatively closer to this point.

In addition to the above question, there is about a 3% loss in terms of LFRAC total for v4.6 comparing with v3.7.

I’m worried about these differences could affect the accuracy of the newer version. Has anyone encountered similar problems?

Many thanks, Stella

HK_narrow inventroy
HK_narrow inventroy634×692 50.5 KB

The blue point I added is the major point source in this region based on lat lon information

#2

There was a map projection issue discovered in SMOKE 4.6 that was corrected in SMOKE 4.7. Can you move to SMOKE 4.7 instead of 4.6?

#3

Hi, thanks for you reply. I try to run v4.7 test case but meet an error “** ERROR ABORT in subroutine MRGVNAMS” for smkmerge.area.nctox.20050710.ys12-nc.log and " ERROR: Speciation joiner “~” is not found for speciation variable" series errors for snjreort.temporal log. Do you have any ideas about it? really appreciate your help!

#4

Have you tried to remove any output files from v4.6 and then run v4.7 fresh?

#5

yes, they are installed in different directories

#6

sorry, I missed an error message from Smkreport.area.temoporal

" *** ERROR ABORT in subroutine BLDREPIDX
Problem setting up input data to output columns."

I complied v4.7 for PGI since we don’t have intel installed. I got all executables and a bunch of .o files in my Linux2_x86_64pg.

#7

Can you use the precompiled executables here? ftp://newftp.epa.gov/air/emismod/2016/v1/smoke_2016v1_platform_core_02jan2020.zip

Or do they not work because you don’t have the same compiler available?

#8

@wheaj. The error message you are getting is from Smkreport which is a QA program that helps you to generate various types of reports. This error won’t prevent you from generating CMAQ ready gridded spectated hourly ncf emissions file. Have you checked Smkmerge program log file? Did you get the normal completion? If you did, check the output file to see whether this map projection correction fix in SMOKE v4.7 the issue you are having or not.

#9

Hi BH, I didn’t get temporal profile, error message in temporal log file is "NOTE: Output grid “US12NC_66X52” set; described as
No description available

 *** ERROR ABORT in subroutine MRGVNAMS
 ERROR: Variable descriptions do not contain proper separator in spec matrix."
#10

Hmmm… Can you send me (bbaek@unc.edu) all log files from SMOKE v4.7 run? Also, did you start SMOKE v4.7 test case run by download all downloadable tar.gz files from CMAS website and followed the Chapter 12?

#11

By checking smkmerge log file, I found it may be caused by no var descriptions for agmat_l file. Detail information is sent.

same happened for agmat_s, psmat_l and psmat_s, may related to grdmat.

#12

AGMAT output file from Grdmat should not have that joiner (~). Smkmerge does not expect that while processing AGMAT intermediate output file from Grdmat but only ASMAT output file from Spcmat. Both output files from Grdmat and Spcmat look good to me. However, there is something wrong in Smkmerge run. Meanwhile, you can use Smkmerge from SMOKE v4.6 since the issue you are describing is not caused by Smkmerge. You can simply replace Smkmerge with the one from SMOKE v4.6. Let me know how it goes.

#13

Thanks! everthing goes well now. I’m planning write down the whole process of updating from 3.7 to 4.7 to see if can help others.