Bill Sacks November 11, 2013 This file documents the procedures that were used to create the following 11 initial conditions files, which are the defaults starting in clm4_5_43: clmi.I1850CRUCLM45BGC.0241-01-01.0.9x1.25_g1v6_simyr1850_c130926.nc clmi.I1850CRUCLM45BGC.0241-01-01.1.9x2.5_g1v6_simyr1850_c130927.nc clmi.I1850CRUCLM45BGC.0241-01-01.360x720cru_hcru_simyr1850_c131011.nc clmi.I1850CRUCLM45BGC.0241-01-01.ne30np4_g1v6_simyr1850_c131011.nc clmi.I1850CRUCLM45BGCDV.0241-01-01.0.9x1.25_g1v6_simyr1850_c131011.nc clmi.I1850CRUCLM45SP.0521-01-01.0.9x1.25_g1v6_simyr1850_c130926.nc clmi.I2000CLM45CRUBGC.2000-01-01.0.9x1.25_gx1v6_simyr2000_c130926.nc clmi.ICRUCLM45BGCCROPmp24.0241-01-01.10x15_USGS_simyr2000_c131028.nc clmi.ICRUCLM45BGCCROPmp24.0241-01-01.1.9x2.5_g1v6_simyr2000_c131028.nc clmi.ICRUCLM45BGCCROPmp24Irrig.0241-01-01.10x15_USGS_simyr2000_c131028.nc clmi.ICRUCLM45SP.2000-01-01.0.9x1.25_g1v6_simyr2000_c130926.nc --- REASON FOR THE ABOVE FILES --- These new initial files were needed starting in clm4_5_43 simply because this tag expands the 1-d memory structures to include a lot more points (for the sake of dynamic landunits). --- FILES DESIGNED TO GIVE NEARLY BIT-FOR-BIT RESULTS WITH PREVIOUS VERSIONS --- Where possible, I used a hacked version of interpinic that fixes a number of bugs in the tool, and is able to produce interpinic'ed files that give nearly identical results to the old files. HOWEVER, PLEASE NOTE THAT THIS VERSION OF INTERPINIC IS NOT GENERAL, AND SHOULD ONLY BE USED IN THE LIMITED CASE OF INTERPOLATING A FILE TO ANOTHER WITH THE SAME RESOLUTION, AND PROBABLY THE SAME OPTIONS (e.g., crop vs non-crop). The following files were created from the version of interpinic in https://svn-ccsm-models.cgd.ucar.edu/clm2/branch_tags/landunit_percent_CLM_tags/landunit_percent_CLM_14_clm4_5_07/. For a given file, I interpolated from the old version of that file (i.e., the version used prior to clm4_5_43) to an output template corresponding to the desired resolution. clmi.I1850CRUCLM45BGC.0241-01-01.0.9x1.25_g1v6_simyr1850_c130926.nc clmi.I1850CRUCLM45SP.0521-01-01.0.9x1.25_g1v6_simyr1850_c130926.nc clmi.I2000CLM45CRUBGC.2000-01-01.0.9x1.25_gx1v6_simyr2000_c130926.nc clmi.ICRUCLM45SP.2000-01-01.0.9x1.25_g1v6_simyr2000_c130926.nc clmi.I1850CRUCLM45BGC.0241-01-01.1.9x2.5_g1v6_simyr1850_c130927.nc The following crop initial conditions files required some additional steps, in order to fix a crop-related bug in interpinic that requires some additional metadata to be on both the input & output files for interpinic: clmi.ICRUCLM45BGCCROPmp24.0241-01-01.10x15_USGS_simyr2000_c131028.nc clmi.ICRUCLM45BGCCROPmp24.0241-01-01.1.9x2.5_g1v6_simyr2000_c131028.nc clmi.ICRUCLM45BGCCROPmp24Irrig.0241-01-01.10x15_USGS_simyr2000_c131028.nc The necessary changes, for both CLM & interpinic, are here: https://svn-ccsm-models.cgd.ucar.edu/clm2/branch_tags/interpinic_hacks_for_bfb_tags/interpinic_hacks_for_bfb_n01_clm4_5_39 I did two short runs for each of the above 3 files: (1) A run from the above tag (which gave a trunk-like version of the above files, but with extra metadata for column type), and (2) A run from my branch that added many landunits, but with a SourceMod to add the necessary crop metadata, as follows: Index: src/clm4_5/main/initGridCellsMod.F90 =================================================================== --- src/clm4_5/main/initGridCellsMod.F90 (revision 54277) +++ src/clm4_5/main/initGridCellsMod.F90 (working copy) @@ -769,9 +769,9 @@ ci = ci + 1 pi = pi + 1 - col%itype(ci) = 1 pft%itype(pi) = m pft%mxy(pi) = m + 1 + col%itype(ci) = (istcrop*100) + pft%itype(pi) col%gridcell (ci) = gi col%landunit (ci) = li I then copied the necessary metadata from (1) onto the existing, out-of-the-box initial conditions files with: ncks -A -v cols1d_ityp ... and then ran interpinic from the above tag (edu/clm2/branch_tags/interpinic_hacks_for_bfb_tags/interpinic_hacks_for_bfb_n01_clm4_5_39) Finally, after running interpinic, I restored the old colsd1d_ityp values onto the output file. I did this by running three runs like the above, from my new branch (the one used to produce the output/template files for interpinic), but now WITHOUT the above SourceMods. --- FILES GENERATED USING STANDARD INTERPINIC --- For a few files, it did not work to use the above, hacked version of interpinic (for reasons I don't understand). For these, I used the standard, trunk version of interpinic (with all of its bugs): clmi.I1850CRUCLM45BGC.0241-01-01.360x720cru_hcru_simyr1850_c131011.nc clmi.I1850CRUCLM45BGC.0241-01-01.ne30np4_g1v6_simyr1850_c131011.nc clmi.I1850CRUCLM45BGCDV.0241-01-01.0.9x1.25_g1v6_simyr1850_c131011.nc The input file for these was: clmi.I1850CRUCLM45BGC.0241-01-01.0.9x1.25_g1v6_simyr1850_c130926.nc and the interpinic version was: dynlu_allocate_memory_n09_clm4_5_37 (from the branch dynlu_allocate_memory). Compared with the earlier trunk version of interpinic, this has the following bug fix; this was brought to the trunk in clm4_5_43: Index: src/interpinic.F90 =================================================================== --- src/interpinic.F90 (revision 51769) +++ src/interpinic.F90 (working copy) @@ -850,8 +850,8 @@ nmin = 0 do n = 1, numcols + calcmin = .false. if (wti(n) > 0.0_r8) then - calcmin = .false. if (typei_urb(n) == nonurbcol) then if (typei(n) == typeo(no)) calcmin = .true. else