Using data for each component based on output of previous run

[eisenman@ucsd_edu]

We're using CESM1.1.2 on Yellowstone and trying to do a run with all model components specified from the output of a previous simulation (we'll follow this test run with runs in which model components are turned on individually). So we're using the "A_PRESENT_DAY" CompSet and replacing the observationally-based default values in datm,dlnd,dice,docn with output from a previously run case. For this initial test setup, we're using the preindustrial control run "b40.1850.track1.1deg.006".

We've run into a couple basic questions about setting this up:

(1) For datm, we found instructions for setting it to use coupler history forcing from a previous simulation rather than COREv2 at http://www.cesm.ucar.edu/models/cesm1.1/clm/models/lnd/clm/doc/UsersGuide/x9068.html . This requires coupler history files from the previous simulation. But it looks like most simulation output on hsi at /CCSM/csm (including /CCSM/csm/b40.1850.track1.1deg.006) doesn't include the history files from the coupler. Is there another place this is stored for most runs, or if not, is there a way to use forcing from the atm history output of a previous run?

(2) For dice, env_run.xml has an entry called "DICE_MODE" that says it can be switched from "ssmi" to "prescribed". But is setting this to prescribed correct? CESM1.0 documentation says running CICE in prescribed mode is similar to running dice, but I'm running dice, not CICE. Either way, I was unable to find where I can specify a previous run casename and path.

I've had a look at the Data Model User's Guide (http://www.cesm.ucar.edu/models/cesm1.1/data8/doc/book1.html) but wasn't able to figure out the answer to these questions.

Any advice on how to address these two issues, or where we might be able to find documentation on what to do here, would be most appreciated.

Thanks,
Ian

 

[eisenman@ucsd_edu]

I found the answer to one of these questions (thanks Nan Rosenbloom!) and wanted to post it here in case others have similar issues (note that any misconceptions here are surely my own and not Nan's).(1) datm wants very high frequency atmospheric forcing data, which is not typically saved in most runs - most runs don't save coupler history files. So you need to simulate another ~30 years from the end of the control run with the model specified to output 3hr coupler history fields. For some runs, including b40.1850, this has already been done. The ~30 year run that includes all the high-frequency output is called a MOAR run. For b40.1850, it's at /CCSM/csm/b40.1850.track1.1deg.006a/.
(2) Still working on this one ...