Bad ozone data

[andythewxman]

Good afternoon. I am currently trying to run a 25km-5km-1km nest over an area that has complex topography. When I run the model, the CAM radiation scheme errors out after the first couple timesteps, with the message: "RADOZN: Bad ozone data: non-monotonicity suspected". I am not experiencing this issue over "less-complex" terrain. I have tried adjusting several of the namelist options, but have not found anything that works yet. Unfortunately, I am not able to share the namelist file here, but I was wondering if you had any ideas on what I can try. I did see that some users have commented out this line in the code (see line 9716, linked below). Could this be a viable solution or would you not recommend doing so? I appreciate any information you can provide.

WRF/phys/module_ra_cam.F · 9f346e3601a68def021fcbb82fc257ed173a7773 · coawstmodel / COAWST · GitLab

Coupled-Ocean-Atmosphere-Wave Sediment Transport Modeling System src code

code.usgs.gov

 

[Ming Chen]

Hi,
Inside the code "phys/module_ra_cam.F", I can find the folloiwng piece of code:

      do kk=kkstart,levsiz-1
         do i=1,ncol
            if (pin(kk).lt.pmid(i,k) .and. pmid(i,k).le.pin(kk+1)) then
               kupper(i) = kk
               kount = kount + 1
            end if
         end do

We expect that the pressure monotonically decreases with increasing height. However, when the model goes wrong for some reasons, the monotonicity will be violated. This will lead to the error in your case.

 

[andythewxman]

I am still trying to track down this issue with no luck.

In the namelist, I see a couple variables: levsiz and paerlev. I see the recommended values for levsiz=59 and paerlev=29. I have also seen another site that recommends setting these equal to e_vert, which is 51 in my case. Which one is correct?

 

[Ming Chen]

Please stay with levsiz=59 and paerlev=29.

Again, I don't think your issue is derived from ozone data. It is more like that in your simulation, the hydrostatic condition is no longer maintained.

 

[andythewxman]

ok, thanks. Do you have a way that I can send you my namelist.wps and namelist.input without sharing them here?