Issue with PM at the restart points

Any issues with running WRF Chemistry

Issue with PM at the restart points

Postby georgeg » Thu Oct 05, 2017 3:43 am

Hello everyone.

I am running the WRF-Chem model with the RADM2-MADE/SORGAM mechanism (chem_opt = 2). Since I want to do annual simulations, I have to use restart.

The problem is that at the restart points, there is no continuity in PM2.5 (and PM10) concentrations. For some reason, when the new simulation starts, these concentrations go back to an initial condition (which I don't know where comes from), instead of getting the condition of the last time step of the previous simulation.

More specifically, the PM2.5 components (so4ai, so4aj, no3ai, no3aj, ect) are the ones which are not passed to the new simulation, although they look fine in the restart file.

I also have to say that I tried with 3.6.1, 3.7.1, 3.8.1, and 3.9.1 and still facing the same issue.

Did any of you came across this issue?

Thanks,
George.
George K. Georgiou
PhD Candidate
The Cyprus Institute
Nicosia, Cyprus
georgeg
 
Posts: 8
Joined: Tue Apr 28, 2015 3:32 am

Re: Issue with PM at the restart points

Postby Christian813 » Thu Oct 05, 2017 12:52 pm

Hi,

I have heard that using chem_opt = 300 (GOCART SIMPLE) you have to change the chemdt to 60 because it needs a time to recover. Maybe something like could be happening to you. I'm not pretty sure but you could paly with chemdt.
Christian813
 
Posts: 11
Joined: Thu Nov 17, 2016 6:54 pm

Re: Issue with PM at the restart points

Postby kwthomas » Thu Oct 05, 2017 6:18 pm

Is your restart from a boundary condition time? If not, try from there, assuming you do restart times that correspond the boundary times.

WRF 3.9 is documented to have this bug. Maybe you've found another form of this bug in many old versions.
Kevin W. Thomas
Center for Analysis and Prediction of Storms
University of Oklahoma
kwthomas
 
Posts: 196
Joined: Thu Aug 07, 2008 6:53 pm

Re: Issue with PM at the restart points

Postby georgeg » Fri Oct 06, 2017 5:00 am

Hello again and thanks for the replies. I think I found where the problem is...

It seems that the aerosol scheme I am using (MADE/SORGAM) initializes aerosol concentrations at the beginning each simulation (there is a subroutine called aerosols_sorgam_init in the module_aerosols_sorgam.F file). In this subroutine there is an IF statement which tells the model NOT TO INITIALIZE when chem_in_opt = 1. This is the case where cycled chemistry is used. There is no IF statement though for the case of restart. So I added it myself, re-compiled, and it seems that the problem is fixed.

I still need to perform some additional tests though just to make sure that everything works OK before sharing the solution.
George K. Georgiou
PhD Candidate
The Cyprus Institute
Nicosia, Cyprus
georgeg
 
Posts: 8
Joined: Tue Apr 28, 2015 3:32 am

Re: Issue with PM at the restart points

Postby georgeg » Thu Oct 12, 2017 8:18 am

Hello again. I performed several tests and I can now confirm my previous assumption - The MADE/SORGAM aerosol mechanism is set to initialize aerosol concentrations at the beginning of each simulation except when the option for cycling the chemistry is used. in order to include an exception for restart as well I modified the module_aerosols_sorgam.F file line 7525 (version 3.9.1) as follows:

Code: Select all
if(chem_in_opt == 1 .OR. config_flags%restart ) return


instead of

Code: Select all
if(chem_in_opt == 1 ) return
George K. Georgiou
PhD Candidate
The Cyprus Institute
Nicosia, Cyprus
georgeg
 
Posts: 8
Joined: Tue Apr 28, 2015 3:32 am

Re: Issue with PM at the restart points

Postby kwthomas » Fri Oct 20, 2017 4:08 pm

Hi George...

Be sure to report what you've found to wrfhelp@ucar.edu. They will see that it is fixed in the code for future releases.
Kevin W. Thomas
Center for Analysis and Prediction of Storms
University of Oklahoma
kwthomas
 
Posts: 196
Joined: Thu Aug 07, 2008 6:53 pm


Return to Runtime Problems

Who is online

Users browsing this forum: No registered users and 1 guest