I am using the CMAQv4.7.1 version by adding the dust module.
In order to secure the optimal calculation time according to the change from the last 3-day forecast system to 7 days
There was work on changing the cpu (intel-> cray) and expanding the number of cores (240 -> 1140).
In the existing 3-day system, the blow-up phenomenon has never occurred, but the model often blows up after the change.
When looking at the core file with the gdb command, it was found that the thermodynamic equilibrium part did not satisfy the stability condition and was blow-up.
At the time of the blow-up, it is judged that the stability condition in the grid was not satisfied due to the influence of strong low-pressure vorticity (existence of large Vmax) in the domain…
Blow-up is sometimes solved by reducing the time step as a basic solution, but in some cases this is still a problem.
I think this problem would have been solved by the version upgrade. Since the model version cannot be upgraded immediately, is there a solution to block such blow-up problem in CMAQv4.7.1 version?
If you have any modifications to the source code, please let us know.