Talk:Bugs and fixes

Problem with restarting runs

Havala Pye (havala@caltech.edu) wrote on 16 June 2008:

I've noticed a potential problem with some species when a run is restarted from a restart file. I noticed the problem in v7-4-11 and Daven has confirmed it in v7-4-13. The error occurs, for example, if you do a full-chemistry run for 10 days and specify a restart file (and tracer concentrations) to be written on the third to last day. If you restart the run on the third to last day using the restart file, and compare concentrations from that run to the continuous run, some species have significantly different concentrations. The most significant deviations appear to occur for RCHO. After a restart, RCHO can drop roughly 50% over S. America. Differences also occur for other species such as CO, but they are smaller. I didn't know if people were aware of this or if it has been addressed. Since not all species are saved to the restart file, there could be some minor differences between a continuous and restart run, but I was not expecting such large deviations. Let me know if you have any comments.