Hi,
I’m running CMAQv5.3.3 with cb6r3m mechanism. The run of CCTM seemed to be fine, but the CTM_LOG output indicated that there was something wrong with time-steps for EBI solver. The log output gave warnings as follows:
WARNING: At solution attempt # 1
WARNING: EBI Euler convergence failure
Reducing EBI time step because of MAXPRED failure in
Cell (203, 129, 5) for the below species: : Init.Conc, Pred.Conc.
WARNING: EBI Euler convergence failure
Reducing EBI time step because of convergence failure in
Cell (203, 129, 5) Solution Attempt # 1
Below Species Causing Error: Init.Conc, Pred.Conc.
This warning occurred in many cells of my simulation domain and caused CCTM operation unnecessarily slow. So I’m wondering that is there any options in run_cctm scripts to modify to avoid causing EBI solver warnings, or is it wrong with my ICON outputs?
I’ve attached my run_cctm scripts and an example of CTM_LOG (which is pretty large, sry) that contained warnings of EBI solvers. I’d appreciate any help from you!
We have also encountered similar warnings in running CMAQ with the cb6r3m mechanism. The convergence failure occurs with regular time steps and the EBI solver then reduces the time step and repeats the procedure which slows down the model. We have also used the Rosenbrock solver and encountered similar problem though it provides fewer warnings. It occurs due to some of the very fast rate constants in halogen chemistry. There are no options in the current model to avoid these warnings. We plan to look into the issue in the future. Thanks for letting us know.
Thank you very much for the kind reply. It’s a pity that there are no options in the current model to avoid these warnings. Anyway my model would work but just slow, and I’m looking foward to future updates!
Besides, I’d like to ask that in #> Synchronization Time Step and Tolerance Options section of run_cctm scripts, is there any options that could make the model provides fewer warnings (with EBI solver?)