dsc254@cornell_edu
New Member
Following the instructions on creating a SCAM IOP in the CAM6 User's Guide, I've run CESM in the F2000climo compset with default settings for a year, outputting the appropriate dynamics variables to one of the history tapes. I then use the provided NCO routines to produce an IOP file, and remap one of the .i. files from the CESM run to the Eulerian grid to serve as the initial conditions file (passed in with the ncdata namelist variable). Typically, I create an IOP file from about a month of the CESM output, and use the .i. file corresponding to the start of the month. I then run SCAM with that IOP.On both the CESM and the SCAM runs, I also output radiative flux variables with and without dust, by defining appropriate aerosol modes and rad_diag variables in the namelist. I use these to compute the dust radiative forcing as computed by each run (for CESM I just look at the fluxes during the month of the IOP and at the location where SCAM is run).CESM and SCAM give very disparate dust RF values in this experiment. I've tried a number of variations on this theme, including adding the CESM monthly average aerosols to the ncdata initial conditions file, and setting the SCAM relaxation to be very strong so that the SCAM aerosol column differs very little from the CESM average, but to no avail.Can SCAM reproduce CESM radiation results over a short period of time, with prescribed dynamics, aerosols, and so on? I'd love to hear whatever thoughts anyone might have. Thanks,Dave