PFT Input

[s_a_monks@leeds_ac_uk]

I am running the stand alone CLM model using the I_1948-2004 compset. I have a run using this compset looking at changes in biogenic emissions between 1948 and 2004. I am now wanting to do a comparable run but include the impact of historical changes in PFTs on biogenic emissions between 1948 and 2004.  In this compset the following namelist options are set as below:fpftcon = '/glade/p/cesmdata/cseg/inputdata/lnd/clm2/pftdata/pft-physiology.c110425.nc'fsurdat = '/glade/p/cesmdata/cseg/inputdata/lnd/clm2/surfdata/surfdata_1.9x2.5_simyr2000_c091005.nc' From what I can make out the fpftcon file is simply constants and the fsurfdat contains the fields I am wanting to change (probably PCT_PFT)?I found a file which seems to have historical PFT which varies with time:/glade/p/cesmdata/cseg/inputdata/lnd/clm2/surfdata/surfdata.pftdyn_1.9x2.5_hist_simyr1850-2005_c110114.ncso I tried setting this :fsurfdat = '/glade/p/cesmdata/cseg/inputdata/lnd/clm2/surfdata/surfdata.pftdyn_1.9x2.5_hist_simyr1850-2005_c110114.nc' However, it was missing some data such as CANYON_HWR so the run didn’t work.I thought I could create a netcdf file to include the missing data from surfdata_1.9x2.5_simyr2000_c091005.nc combined with the pft data from surfdata.pftdyn_1.9x2.5_hist_simyr1850-2005_c110114.nc, however, I didn’t know whether this was possible as maybe it would cause problems as some of the data is time-varying and some isn’t?So then I found the namelist setting ‘fpftdyn’ so tried setting:fpftdyn = '/glade/p/cesmdata/cseg/inputdata/lnd/clm2/surfdata/surfdata.pftdyn_1.9x2.5_hist_simyr1850-2005_c110114.nc' However, this didn’t work either and I get the following message in my log file:/glade/p/cesmdata/cseg/inputdata/lnd/clm2/surfdata/surfdata.pftdyn_1.9x2.5_hist_simyr1850-2005_c110114.ncpftdyn_init                     mismatch between input pctspec =   97.0000000000000       and that obtained from surface dataset   100.000000000000       at g=            4 Does anyone know of an easy way to do what I want?Many thanks for your time,Sarah

 

[erik]

Starting wtih the 1948-2004 compset is NOT the best place to start, as there are compsets that give you transient PFT out of the box. Use the "-list" option to ./create_newcase" to see all of the compsets that are available. The CLM UG also talks about appropriate compsets... http://www.cesm.ucar.edu/models/cesm1.2/clm/models/lnd/clm/doc/UsersGuide/c720.html#compset_choice
You can also use the CLM_NML_USE_CASE option to change your existing case to a use-case for transient PFT's. http://www.cesm.ucar.edu/models/cesm1.2/clm/models/lnd/clm/doc/UsersGuide/x1230.html#clm_script In the end you'll find that you need to set BOTH fsurdat AND fpftdyn, where fsurdat is the starting year of the transient period.