Isseu when running with a specific mesh

[sarriegi]

Hello, I am facing an issue when I run CoastalApp using a specific mesh. This mesh is generally the same as the HSOFS mesh, but it has very high resolution in the west coast of Florida. The run starts, but right after the first timesteps, the water level becomes really high in the open boundary: If I run exactly the same case but with a different mesh, then there is not this problem. Do you know what could be going wrong? Note that when I was working in a different cluster to the one I am working now, I was able to resolve this problem by adding the nodal attribute 'elemental_slope_limiter' in the fort.15 file. However, I don't have access to that cluster anymore. In the cluster where I am working now, when I add the 'elemental_slope_limiter' in the fort.15 file, I get this error: Do you have any idea of how I could resolve this issue?

@pvelissariou1 @saeed-moghimi-noaa

Angelos

[saeed-moghimi-noaa]

Hi @sarriegi

This seems to be a ADCIRC issue. Did you ran ADCIRC alone? Did you plot it results? Did you also changes the mesh on the wave side?

Thanks, -Saeed

@yunfangsun @pvelissariou1

[pvelissariou1]

@sarriegi I believe you need to specify the elemental_slope_limiter in fort.13 instead of fort.15. See: https://adcirc.org/home/documentation/users-manual-v53/input-file-descriptions/nodal-attributes-file-fort-13/ Example:

elemental_slope_limiter
none
1
.040000

[sarriegi]

Thank you very much for your responses @pvelissariou1 and @saeed-moghimi-noaa! When I run the same case with ADCIRC standalone at CoastalApp/ADCIRC, then I don't face any issues (even without the elemental_slope_limiter). When I run the CoastalApp as ADCIRC standalone however (by removing the line "WAV->OCN" in the nems.configure file), then again the water level gets very high values from the start. Also yes, the mesh is the same for the wave model. @pvelissariou1 I forgot to mention that I specified the elemental_slope_limiter in both the fort.13 and fort.15

Angelos

[pvelissariou1]

@sarriegi (@saeed-moghimi-noaa , @yunfangsun ) I don't think you should specify the "elemental_slope_limiter" in fort.15 as this attribute is directly read from fort.13 In nems.configure besides removing the WAV->OCN lines, you need to remove the WAV model lines as well. Running ADCIRC standalone and within NEMS (ATM->ADC) standalone you should get identical results. If I remember correctly, the only time I encountered an issue like yours was when my atm. data file was corrupted (I was getting NaNs at some nodes)

[sarriegi]

@pvelissariou1 @saeed-moghimi-noaa @yunfangsun Another clue could be that if I run the same case with cold start for ADCIRC, then the water level doesn't get high values and the run continues. I don't know what is wrong with the fort.68.nc file however, it works fine for ADCIRC standalone or for ADCIRC-SWAN.