*** ERROR ABORT in subroutine HRSOLVER on PE 013

CMAQ
1

Hi, I am running CCTM with CMAQv5.3.1 for PRD at a resolution of a 98x74 with 128 processors. It already ran this script several times before. But after turning on the CTM_LTNG with my own NLDNstrike files, I am getting an error. Then I checked my debug file, the reason might be the NO concentration.

*** ERROR ABORT in subroutine HRSOLVER on PE 013
ERROR: Stopping because of EBI convergence failures
PM3EXIT: DTBUF 3:00:00 May 29, 2015
Date and time 3:00:00 May 29, 2015 (2015149:030000)

Concentrations less than, but reset to  0.000E+00 in            HADV
     Date:Time     Col  Row Layer Species             Value before reset
  2015149:030000    3   10    1    2 (NO              )  -2.029E+25
  2015149:030000    2   11    1    2 (NO              )  -2.134E+25
  2015149:030000    5   11    1    2 (NO              )  -2.128E+24
  2015149:030000    2   12    1    2 (NO              )  -2.974E+24
  2015149:030000    5   12    1    2 (NO              )  -2.672E+23

It might be because of the NLDNSTRK file, but I have no idea what wrong with this file.
It would be a great help if anyone can help me.

Thanks,
Axueaha
CTM_LOG_013.v531_bisheng_M_9_2015_0529_20150529.txt (230 KB)
FLOOR_022.v531_bisheng_M_9_2015_0529_20150529.txt (18.0 KB)

2

@Axueaha
Since you have a doubt of your NLDNstrike files, you may want to first check if the files don’t contain any unreasonable values or NAs. It is not clear why the NO concentrations are that large negative numbers before reset.

3

Thanks for your reply, Dr Kang. I’ve checked my NLDNstrk file. There is no null value, and the value range is 0-8.3. Do you think this value is within a reasonable range？

4

Something is happening at 3:00:00 that is causing gas-phase species concentrations to blow up. Maybe there is a lightning strike then?

 After       COUPLE :  Gas  2.527E+03  | Aer  5.192E+13  | Non  1.058E+00
 After         HADV :  Gas  2.220E+24  | Aer  5.153E+13  | Non  1.048E+00

Please verify that if you setenv CTM_LTNG_NO N that the model runs ok. There still might be something wrong with your boundary conditions.

5

Thanks for your reply, @cgnolte. I fixed this error a few minutes ago. The original reason is that the data type of variables in my NLDN file is incorrect. The template provided is float32, and the variable type I ran before is also float32, which works well. My new file is float64. I just modified it and worked. I didn’t notice this detail. The cctm can run when I set CTM_LTNG N, so maybe there’s nothing wrong with my boundary conditions. The sudden increase in concentration at 3:00:00 may be due to other reasons, cause the frequency of lightning at this moment is not much. Thanks a lot.

6

I fixed this error, Dr Kang. The original reason is that the data type of variables in my NLDN file is incorrect. The template provided is float32, and the variable type I ran before is also float32, which works well. My new file is float64. I just modified it and worked. Thanks a lot.

7

I’m not sure how strong the lightning activities are in your file, but just to make sure that it is lightning flash density in flashes/km^2 in your input file. The values should mostly be less than 1, but occasionally it can be values larger than 1 when lightning activity is very strong at a fixed location. Just be mindful that the value is flashes/km^2.hr. Good luck!