I am experiencing something very weird with RegCM 4.1. I am trying and running RegCM 4.1 for a 6-month integration over the period Oct. 2004 to Apr. 2005.

The horizontal resolution is 43.452km, over a domain covering Southern Africa and the South West Indian Ocean. With regard to the horizontal resolution, I have set up the time step dt in regcm.in to 120s.

The run performs correctly for the first month of simulation (October 2004) but crashes around mid-November 2004 with a CFL violation. I have two problems:

- the output files for November are empty, although I have asked for output saving every 6 hours

- more surprising, in the regcmMPI.out file that was created during the simulation, time step dt and dtau in October are 120s while dt and dtau in November are 240s

Where am I getting wrong?

Many thanks by advance for your response.

Best regards,