8 August 2011 - Project: This project downscales output from a low-resolution (T31) transient simulation (22ky long TraCE) to create the initial conditions for an AGCM simulation of the same time period at high resolution (T85). Author: Nan Rosenbloom Contributions from: Christine Shields Sam Levis Master path: /glade/proj2/cgd/ccr/paleo/INCITE/T31-T85 Master source: /glade/proj2/cgd/ccr/paleo/INCITE/T31-T85/src-low2high Jerry Potter Source: /glade/proj2/cgd/ccr/glpotter --- Definitions: ------------------------------------------------------------------------------------------- CASEID This refers to the TraCE name for discrete simulation periods. e.g., b30.16_0kaDVTd, b30.14_5kaDVTd, b30.14_0kaDVTd, b30.13_5kaDVTd % This symbol indicates a command line argument. $mypath This environment variable is set in Step 1.1. \ This symbol indicates that a command line argument continues to the next line. You do not type this symbol when you are writing in a command line argument. --- Steps: ------------------------------------------------------------------------------------------- 1. CLIMATOLOGY machine: mirage0 Create a multi-year climatology from the low-resolution (T31) atmospheric history files. We use the cam3 surface temperature (TS) as a proxy for sea surface temperature (SST). 1.1 create a new copy of the T31-T85 directory in your personal space. (Rename your previous directory to T31-T85_OLD if you don't want to remove it). e.g., % cd /glade/proj2/cgd/ccr/glpotter % cp -pR /glade/proj2/cgd/ccr/paleo/INCITE/T31-T85 . % setenv mypath /glade/proj2/cgd/ccr/paleo/ % cd $mypath 1.2 Choose the discrete TraCE simulation that includes your time period of interest. The TraCE simulations were run as discrete time slices (e.g., b30.14_5kaDVTd, b30.15_0kaDVTd). Each time slice represents a 500-1000 year sealevel/ice configuration that together produce a step-wise integration of rising sea level and diminishing land ice during the deglaciation. Retrieve the history files from the TraCE simulation that is close to your period of interest, but still older than the time period that you are interested in. For example, if you want to simulate the calendar year 15.8ka, you would use the model simulation for the period starting at 16ka B.P. (b30.16_0kaDVTd), not the period starting at 15ka B.P. (b30.15_0kaDVTd). The model will use the land ice reconstructions for 16.0ka, but the modeller will set the orbital year and greenhouse gas (GHG) concentrations for 15.8ka. GHG orb_year ice sheet SST b30.16_0kaDVTd |------[15.8]-----------------|b30.15_0kaDVTd|-- .... ----->|Present dayr| TIME ----------------> Counts down [ 22ka ---> 16ka ---> 10ka ..... ---> 0 CE ---> 2005 CE] MODEL YEAR ----------> Counts up for every time slice [16ka: 1,2,3, .... 1000] [15ka: 1,2,3, .... 1000] To choose your time period, notice that the TraCE naming convention counts down from 22ka B.P. to present day (passing through 0 CE). e.g., b30.16_0kaDVTd b30.14_5kaDVTd b30.14_0kaDVTd b30.13_5kaDVTd b30.13_0kaDVTd .... However, each individual 1000 yr model time slice counts forward from year 1. e.g., Notice how the run years progress from year 1 -> 1000. for the time period from 16ka to 15ka: b30.16_0kaDVTd model year 1 == 16.0ka BP b30.16_0kaDVTd model year 200 == 15.8ka BP b30.16_0kaDVTd model year 800 == 15.2ka BP e.g., -15800 = -16000 + 200y 1.3 Decide what years you would like to include in your multi-year climatology, and then bracket the period around your chosen model year: e.g., use model years 190-209 (to create a 20 yr climatology around year 200) use model years 375-424 (to create a 50 yr climatology around year 300) 1.4 Download the TraCE atmosphere history files from the HPSS to a local directory: 1.4.1 cd to your scratch directory on mirage0.ucar.edu. You will need a lot of space for these files. % cd /ptmp/glpotter/test_run 1.4.2 download the files from the HPSS using hsi: % hsi -P 'get /CCSM/paleo/b30.16_0kaDVTd/atm/hist/b30.16_0kaDVTd.cam2.h0.019[0-9].tar' % hsi -P 'get /CCSM/paleo/b30.16_0kaDVTd/atm/hist/b30.16_0kaDVTd.cam2.h0.020[0-9].tar' 1.4.3 Untar the annual history files. % tar -xvf b30.16_0kaDVTd.cam2.h0.0191.tar % tar -xvf b30.16_0kaDVTd.cam2.h0.0192.tar .... 1.5 Create a multiyear climatology of atmospheric surface temperature (TS). This is your SST proxy. This code reads the history files you have just downloaded from the HPSS (Step 1.3) and writes a local file with the SST climatology (still at T31 resolution. Step #3 interpolates to T85). 1.5.1 Edit the source code to point to the scratch directory that contains your history files (from Step 1.3). Set the model years for your period of interest. 1.5.2 Make a personal copy of the original code. % cp mk_SST_climo_from_histFiles.pl mk_SST_climo_from_histFiles.myrun.pl 1.5.3 edit your copy of the source code to customize the caseid, pathnames, and model years. % nedit mk_SST_climo_from_histFiles.myrun.pl $caseid $wkdir $fyr $lyr 1.5.4 make the source code executable % chmod u+x mk_SST_climo_from_histFiles.pl 1.5.5 run the script. % ./mk_SST_climo_from_histFiles.pl 1.5.6 example of input (ifile) and output (ofile): ifile: b30.16_0kaDVTd.cam2.h0..nc ofile: b30.16_0kaDVTd.cam2.climo..nc ---------------------------------------------------------------------------------------------- 2. Restart.tar machine: mirage0 Download the restart tar file from the time period of interest. You need the restart file in order to retrieve the clm.i and cam.i files. 2.1 cd to your scratch directory. % cd /ptmp/$LOGNAME/ % mkdir mytestcase % cd mytestcase 2.2 Find the restart file that centers around your time period. e.g., to look for the time period, 15800, which is year 200 of the 16ka simulation, type: % hsi -P 'cd /home/fenghe/CCSM/b30.16_0kaDVTd/restart.tars/; ls -l ' | less 2.3 Download the restart file closest to model year 200 (which is simulated year 15.8ka BP): % hsi -P 'get /home/fenghe/CCSM/b30.16_0kaDVTd/restart.tars/b30.16_0kaDVTd.ccsm.r.0201-01-01-00000.080607-222728.tar' 2.4 Untar the restart file. % tar -xvf b30.16_0kaDVTd.ccsm.r.0201-01-01-00000.080607-222728.tar 2.5 Copy the cami and clmi file to local directories. % cp b30.16_0kaDVTd.cam2.i.0201-01-01-00000.nc $mypath/T31-T85/inputdata/cami/ % cp b30.16_0kaDVTd.clm2.i.0201-01-01-00000.nc $mypath/T31-T85/interpinic-DGVM/in/ ---------------------------------------------------------------------------------------------- 3. SST machine: mirage0 Interpolate the SST climotology created in Step #1 from T31 to T85. This is your 'bndtvs' file in the AGCM/CAM namelist. Note: Resist the temptation to rewrite this code. The file requirements are extremely precise and any inconsistencies will cause the model to fail. 3.1 make a personal copy of the original code: % cp interpolate_sst_t31-t85.ncl interpolate_sst_t31-t85.myrun.ncl 3.2 edit the source code % nedit interpolate_sst_t31-t85.myrun.ncl caseid fyr lyr 3.3 run the script % ncl interpolate_sst_t31-t85.myrun.ncl 3.4 input and output files ifile1 = b30.16_0kaDVTd_ic_48x96_climo.nc ! from Step 1 ifile2 = sst_HadOIBl_bc_128x256_clim_pi_c101028.nc ! NCAR template at T85 ofile = sst_ic_128x256_b30.16_0kaDVTd..nc ! output file with TS (SST) at T85 Your output file (ofile) will be written to "../inputdata/sst_ic/" ---------------------------------------------------------------------------------------------- 4. TOPOGRAPHY machine: mirage0 Next you need to create high resolution (10min) topography that matches your time period of interest. This is your 'bnd_topo' file in the AGCM/CAM namelist. 4.1 make a personal copy of the original code: % cp mkTopo.10min.ncl mkTopo.10min.myrun.ncl 4.2 edit time period in source code % nedit mkTopo.10min.myrun.ncl timePeriod 4.3 run the script % ncl mkTopo.10min.myrun.ncl 4.4 input and output files and relative paths ifile1: ../inputdata/Peltier_ice5g_v1.2_10min/ice5g_v1.2_