Help Request on MOZART - MOSAIC

[Indranil Nandi]

Respected members,

I am really struggling very hard for the last few months to run  WRF-Chem with MOZART-MOSAIC setup using a single as well as a nested domain approach. Each time I end up with no proper error message on rsl.error.0000 or rsl.out.0000 file. Sometimes in any random rsl file it shows  "FAtal Called: STOP: activate before 48000"  otherwise model just stops printing any message in rsl files but still running with a generation of fort.67 file containing lines  "ASTEM internal steps exceeded     200".

My domain ranges from -8.816597 to 49.22037 (N) 38.02636 to 116.7996 (E) (300*250*40 Grid dimension). In the case of a nested domain, the outer domain also has nearly the same area coverage. I have tried to follow the best practices to set up a MOZART-MOSAIC WRF-Chem (3.8.1) run, even tried with different grid point distances but each time ends up with the same error. 

To all the MOSAIC experts, please have a look at my attached namelist file and kindly please help me to complete my run successfully. I am unable to understand the cause bind this kind of run failure, although I am not very new with WRF-Chem but still anyhow unable to run the MOZART-MOSAIC setup. Please WRF-Chem support show me any way to resolve the error of  "ASTEM internal steps exceeded     200".

[Indranil Nandi]

Respected members,

Please find the attached rsl.error and rsl.out files which will make you people easy to understand the error I am facing to run WRF-Chem (3.8.1) using MOZART-MOSAIC parameterization.  

 rsl.error.0024

 rsl.error.0046

 rsl.out.0024

 rsl.out.0046

--

Indranil Nandi

Ph.D. Research Scholar
Centre for Atmospheric Sciences
IIT Delhi, Hauz Khas, New Delhi 110016, India

Contact No. 8902005948 

[Indranil Nandi]

Respected Members, 

If anyone has any idea how to get rid of the error "FATAL CALLED FROM FILE:  <stdin>  LINE:    2243 STOP: activate before 48000" please help. I am struggling with MOZART-MOSAIC setup "chem_opt = 202" for a very long time. I have tried every possible way to resolve and run the model but although model starts running but it just stuck after few hours of run with this error message on one of the rsl files. 

I am sharing my domain image and the namelist file. Please if anyone has any experience with  chem_opt = 202 and this kind of error please help. 

 namelist.input

 rsl.error.0000

 rsl.error.0133

 rsl.error.0148

 rsl.out.0000

 rsl.out.0133

 rsl.out.0148

[Stacy Walters]

Indranil,

Your problem is best tackled with online debugging.  This is a labor, time intensive task that is beyond the scope of ACOM's

WRF-Chem support.  We have pinpointed the source of the immediate error but without further effort can not say why you

are encountering the halt error.  Most likely some variables values are outside an acceptable range for convergence.  But

this is only a guess.

Is there anyone at your organization with advanced knowledge of WRF-Chem?

Stacy

[Luke Conibear]

Hi Indranil,

In a previous message, Gabriele noticed that your chemistry timestep (chemdt = 2 minutes) does not match your meteorological timestep (time_step = 30 seconds). This mismatch can cause issues. Maybe, you should try a test with these timesteps matching (i.e. reduce chemdt to 0.5).

Best, 

Luke

[Indranil Nandi]

Respected Stacy Sir,

Unfortunately, I don't have any support in my organization who can help me in this debug.  Yes, there is a possible chance that some variable values are outside an acceptable range for convergence as in my simulation I am trying to use  CMIP6, ECLIPSE statistically modified for 2019 emission files as emission input. But if this is the reason behind the model halt error then the model should halt in the same simulation position every time. But what I am facing that model starts running from 2020-01-01_00:00:00 then sometimes it runs up to 2020-01-02_08:30:00 then stops or up to 2020-01-03_01:30:00 then stops or something different. There is no fixed time for the model halt. But each time random rsl files contain the error message "FATAL CALLED FROM FILE:  <stdin>  LINE:    2243 STOP: activate before 48000".

What I am getting that all my prepared emission files have the minimum value assigned is 0 but the upper limit value might be for few files that have a very high positive fill value. I will work on it and update again if I get the same error.  

[Amit Sharma]

Hi Indranil,

In case you haven't been able to figure out the problem yet, can you also try a simulation with seas_opt=0 or 1. I remember seas_opt = 2 having been disabled in later versions due to error in that option. Also try to follow Luke's suggestion for matching chemistry timestep.

Interesting that your model doesn't stop at a fixed point. Can you also increase debug_level to say 100 to receive more information on what's going on in the model. Also give a random check to wrf outputs for some met variables and chemistry variables (T, PM, o3 etc) using ncview or something similar just to see if anything looks odd at any time.

Regards,

Amit

--
To unsubscribe from this group and stop receiving emails from it, send an email to wrf-chem+u...@ucar.edu.