Issue with PM2.5 Output in WRF-Chem Model Run

[Akhtar Memon]

I've successfully run WRF-Chem  using the provided namelist, but i;m encountering an issue where PM2.5 results are not being generated properly. while the simulation seems to run without errors the output for PM2.5 concentrations appears incomplete or missing.

Has anyone faced similar issue ? could this be related to namelist or model configurations.

i would appreciate any advice, Thanks in advance !!

[Gabriele Pfister]

Hi

It seems you are only looking at the initial timestep here? Note that PM2.5 is a diagnostic variable and is zero at the initial timestep. Your subsequent simulation time steps should have values. 

Gabi

--

================================
Gabriele Pfister

Atmospheric Chemistry Observations & Modeling (ACOM) Lab

NSF National Center for Atmospheric Research (NCAR)

Email: pfi...@ucar.edu

Work Phone: +1 303 497 2915

Web: https://staff.ucar.edu/users/pfister

[Akhtar Memon]

My subsequent timesteps also do not have any values. Kindly see the attached file.

--

Regards

Akhtar Ali Memon

Cell # : +923322788727

[Akhtar Memon]

[Gabriele Pfister]

Hello Akhtar

I have not experiences any issues like this but have you looked at individual aerosol species and are they looking fine? And have you checked that you read in the emissions correctly and do you have aerosols assigned in your wrfinput and wrfbdy?  I also noticed in your nameslist that you have incorrect settings for emiss_opt and emiss_inpt opt. See below the correct combination. 

 chem_opt                            = 112,    
 emiss_inpt_opt                      = 111 
 emiss_opt                           = 8,  

Gabi

[Akhtar Memon]

I have observed that the real.exe is not writing chemistry parameters into wrfinput and wrfbdy while putting chem_opt=112, or any other. What would be the possible reason? What should I do?

[Gabriele Pfister]

Hello

real.exe will not write initial and lateral boundary conditions in your wrfinput and wrfbdy files. The WPS and created met_em files only operate on meteorology. You will have to use a tool like mozbc to set them from global model output. See our Tools page for more information on mozbc. 

Gabi

[Akhtar Memon]

1. (Just info) I have also successfully run anthro_emiss and bio_emiss (created wrfchemi and wrfbiochemi files)

2. mozbc is getting errors. pls see attached files and suggest.

3. (I think) Before running mozbc, I should  run real.exe with chem_opt=112 (I have done this). As a result, real.exe may write initial and boundary conditions in my wrfinput and wrfbdy files (but no changes occur in wrfinput and wrfbdy files). Please comment regarding this para.

[Gabriele Pfister]

Hello Akhtar

mozbc will add the chemical IC and BC to your files, so no need to run real.exe again afterwards. In fact you should not because then you would just again recreate a wrfinput and wrfbdy without chemical concentrations in it.
The error you receive is because you need to tell mozbc to add the chemical fields if they are not in your  wrfinput and wrfbdy.
def_missing_var = .true.

This all is explained in the README.
Gabi

[Akhtar Memon]

Please see attached files and suggest.

[Gabriele Pfister]

Hi

the error message indicates that it cannot open the global model input file: 

 failed to open /home/adrpc/Build_WRF/WRFV3/mozbc/h0001.nc
 NetCDF: Unknown file format                                                    
 fail to process netCDF file...

Maybe something went wrong in your download, so check that this file is a valid netcdf file. 

Gabi

[Akhtar Memon]

I have checked and downloaded a number of times but the error persists. Attached files are for your reference and suggestions, please.

[Gabriele Pfister]

Hi

can you issue: 

ncdump -v date,datesec /home/adrpc/Build_WRF/WRFV3/mozbc/waccm-20250414032202439119.nc > waccm.out

What is the output ? 

Gabi

[Akhtar Memon]

Thanks for your support and guidance. Your unwavering support has been a guiding lift - not only lifting me in moments of need, but also inspiring and strengthening the entire community with your kindness, grace, and quiet strength.

1. You were right. The issue was with downloaded files as I was downloading waccm subset from "https://www.acom.ucar.edu/waccm/download.shtml". And now that issue has been resolved as I have downloaded the global files from https://rda.ucar.edu/datasets/d312009/dataaccess/#2023-eachfire_modisviirs.

2. Now, the error appears while running mozbc with CBMZ-MOSAIC_4bins.inp. Supporting files are attached. Kindly suggest.

[Akhtar Memon]

The links in the above email are wrong. Please ignore them.

Sorry about that

[Akhtar Memon]

I want to study the source apportionment of PM2.5 in 3 districts of Pakistan (like the attached paper). I am currently using EDGAR for Anthro emissions, MEGAN for bio and FINN for fire. I want  to use MOZCART (or you may suggest the best for Pakistan) for IC/BC (for the purpose I am using mozbc). You are requested to guide me in details in this regard.

[Gabriele Pfister]

Hi Akhtar

I cannot tell what is wrong but possibly the length of the assignment for num_a0? is too long. If you comment these species out, does it run then?

Gabi

[Gabriele Pfister]

Hi Akhtar

this is much more a scientific question than technical and we cannot really help you with making this decision.  The choice of the chemical gas phase and aerosol scheme depends on your scientific question, the characteristics of the environment you intend to simulate and your available computing resources. So I suggest looking through the literature at past studies and determine what is best for you. 

E.g., if secondary aerosols are important for your study, then MOZCART (MOZART gas phase with GOCART aerosols) might not be the best choice since it does not include SOA or nitrate aerosols, and you might want to use MADE/SORGAM or MOSAIC. However, these are computationally more expensive schemes. 

Gabi

[Akhtar Memon]

I want to investigate PM2.5 for some districts of Pakistan. For the purpose, I have (i) run WPS and generated met files, (ii) run real by putting chem_opt=o, (iii) run EDGAR for Anthro emissions, (iv) run MEGAN for bio, (v) FINN for fire, (vi) MOZCART for ic/bc. While running wrf.exe, I have encountered the error attached in the rsl file. Supporting documents are attached for your guidance, please.

Furthermore, I have back traced the error, my system is providing 10% of the memory to the simulation (pic attached).

[Gabriele Pfister]

HI Akhtar

The error message is about a memory corruption but I cannot say what exactly is the issue. Try running with debug_level=120, which should give you more information on any errors. Does your simulation run fine without chemistry? This is the first test to do. Then I  suggest looking carefully at your emission inputs and making sure that you have no negative values or other unreasonable numbers. 

Further I noticed a few things in your namelist though that might need to be changed.  

time_step=216      change to 200 so that it fits within one hour, else your model time step and model inputs/outputs are on different intervals

phot_opt=3     change to 4 as discussed on our website

Comment out n2o5_hetchem, this option is not for chem_opt=112.

You are also running for a fairly small domain (40x50) and you will need to be aware of potential effects from your boundary conditions. 

Gabi