eric_wolf@colorado_edu
New Member
Hi, I am running the following configuration, CESM1.2.1, CAM4, SE dycore, grid ne16np4, aquaplanet with thermodynamic slab ocean zero oht custom compset.My problem is with the interpolated output. On my primary machine (Janus) the interpolated output from this configuration is perfectly fine and great. On a secondary (i.e. "user_defined") machine, the model seems to build and run just fine, but the interpolated output has the imprint of the cubed sphere as a remenant in the data in the near surface layers. Above the lowest 2or3 layers, the simulation appears fine, but the results are unusable due to the anonamlies. See the attached image. I have played around with numerous SE namelist variables, none of which seem to fully remove the imprinting.I was wondering if anyone has encounterd this issue before, and had any troubleshooting advice. The builds are identical, but on different machines. Could it be intel, netcdf or mpi version related? Something i am absent mindedly forgetting in my build?Thanks!Eric