I'm using data from MIROC model as initial and boundary conditions in REGCM-4.1. I've tested on three different computers, but REGCM always stops on the same day and time (1961021706). I'm using SST data from EH5OM (20C3M scenario). I compared input files generated by ICBC to the original data from MIROC and everything seems to be ok. Runs always result in CFL violation. I initially used 120 sec, then 90, 60 and 30 sec in dt variable but model always stopped on the same day and time 1961021706.