CM2M_coarse_BLING_LPJ_05 is an experiment setup, where LPJmL is plugged into CM2M_coarse_BLING. land_lad resp. land_atlantes and LPJmL are run at 0.5x0.5 deg resolution. Ocean and atmosphere are unchanged from CM2M_coarse_BLING. Earlier attempts for LPJ coupling are documented in .../README-mom5_CM2_lad_05LPJ_05_Xcoarse # # begin of grid generation # mkdir -p exp/CM2M_coarse_BLING_LPJ_05/INPUT cd exp/CM2M_coarse_BLING_LPJ_05/INPUT ln -s ../../CM2M_coarse_BLING/INPUT/* . rm -f grid_spec.nc \ land_hgrid.nc \ land_mosaic.nc \ land_mosaicXocean_mosaic.nc \ atmos_mosaicXland_mosaic.nc \ atmos_mosaicXocean_mosaic.nc \ ocean_mask.nc \ land_mask.nc \ topog.nc #ln -s ../../../exp/CM2M_coarse_BLING/INPUT/atmos_hgrid.nc \ # ../../../exp/CM2M_coarse_BLING/INPUT/atmos_mosaic.nc \ # ../../../exp/CM2M_coarse_BLING/INPUT/ocean_hgrid.nc \ # ../../../exp/CM2M_coarse_BLING/INPUT/ocean_mosaic.nc \ # ../../../exp/CM2M_coarse_BLING/INPUT/ocean_vgrid.nc \ # . # Create a new land grid on 0.5x0.5 degree resolution. # The make_hgrid tool creates a ``supergrid'' on double resolution. ../../../src/tools/make_hgrid/make_hgrid --grid_type regular_lonlat_grid \ --nxbnd 2 --nybnd 2 --xbnd 0\,360 --ybnd -90\,90 \ --nlon 1440 --nlat 720 --grid_name land_hgrid # -> created land_hgrid.nc ../../../src/tools/make_solo_mosaic/make_solo_mosaic --num_tiles 1 --dir . \ --mosaic_name land_mosaic --tile_file land_hgrid.nc --periody 1 # -> created land_mosaic.nc # the make_coupler_mosaic tool requires a dimension named ntiles # but this is not contained in the GFDL-provided topog.nc files # sigh. the name topog.nc is hardcoded in the ocean topography module ncap2 -o topog-ntiles.nc -s 'defdim("ntiles",1)' ../../../exp/CM2M_coarse_BLING/INPUT/topog.nc mv topog-ntiles.nc topog.nc ../../../src/tools/make_coupler_mosaic/make_coupler_mosaic \ --ocean_mosaic ocean_mosaic.nc \ --land_mosaic land_mosaic.nc \ --atmos_mosaic atmos_mosaic.nc \ --ocean_topog topog.nc \ --check # created # ocean_mask.nc # land_mask.nc # atmos_mosaic_tile1Xland_mosaic_tile1.nc # atmos_mosaic_tile1Xocean_mosaic_tile1.nc # land_mosaic_tile1Xocean_mosaic_tile1.nc # mosaic.nc mv mosaic.nc grid_spec.nc # It seems that the land_mask.nc file is never used in the simulation run. # land mask and land area information seems to be obtained from the # atmos-to-land exchange grid information only. # However, the land area fraction information which land_lad receives in this way # is identical (modulo rounding errors) to the information in the land_mask.nc file. # The self-made tool netcdf2grid.c reads a netcdf file like the land_mask.nc that is generated by make_coupler_mosaic and writes out an LPJ coordinates file like grid.bin, which contains one entry with lat/lon values for each land cell. That file grid.bin can then be used as input to generate e.g. soil.bin (see above). About LPJ river routing: define axis/x=179.75w:179.75e:0.5 x05deg define axis/y=89.75s:89.75n:0.5 y05deg set data/ez/variables=ddm/grid=grid05deg/skip=6/columns=720 "/data/biosx/LPJ/input_new/DDM30.asc" ! this data set appears upside-down fill/vlimits=90:-90 ddm ..../src/land_atlantes/lpj_trunk/bin/drainage grid.bin DDM30.asc drainage.bin Data preparation for the coupled LPJ: Soil Data: HWSD Harmonized World Soil Database http://www.iiasa.ac.at/Research/LUC/External-World-soil-database/HTML/ download http://www.iiasa.ac.at/Research/LUC/External-World-soil-database/HWSD_Data/HWSD_RASTER.zip extract hwsd.bil from the zip file. Convert to NetCDF using ferret: define axis/x=179.99583333333334w:179.99583333333334e:0.00833333333333 x30sec define axis/y=89.99583333326137N:89.99583333326137S:-0.00833333333333 y30sec define grid/x=x30sec/y=y30sec g30sec file/var=hwsd/grid=g30sec/format=stream/type=i2 "hwsd.bil" set memory/size=15000 save/file="hwsd.cdf" hwsd download http://www.iiasa.ac.at/Research/LUC/External-World-soil-database/HWSD_Data/HWSD.mdb load it into MS Access (at PIK this can be done on the wincenter_X servers) and export the table HWSD_DATA to HWSD_DATA.csv Sibyll has R-scripts to regrid and remap the HWSD data base to LPJ soil input file at 0.5deg resolution in /iplex/01/elis/sibyll/hwsd/ Sibyll's LPJ soil map resulting from that procedure is in /data/biosx/./LPJ/cru06/soil_new_skiprock.bin During initialisation of the coupling wrapper for LPJ, the runoff mapping from LPJ into the ocean is computed. In LPJ, the grid cells which drain into the ocean have a negative value as index for the next drainage cell, indicating that the runoff is simply removed from LPJ. For such cells, a neighbouring ocean is searched and appointed as runoff destination cell in a separate runoff map. After each LPJ update time step, runoff from those cells can then be put into the apropriate ocean cells for passing to the FMS coupler. Runoff from Antarctica can be obtained from land_lad. ---- LPJ compile time option STORECLIMATE must be turned off for coupling with FMS. It would cause LPJ to try and read 30 years of spinup climate during init, which is not available in the coupled version. LPJ run time option PERMAFROST must be turned off. It would cause LPJ to try and read 30 years of soil temperatures during init. --- Synchronize dates between LPJ and FMS: The GFDL-provided example setup CM2M_coarse_BLING sets the initial date to 01-Jan-0002 . LPJ prefers as initial date the year 1901 (again 1st of Jan), see lpjml_fms.conf . FMS takes the initial date for the run from input.nml/coupler_nml . If INPUT/coupler.res exists, it contains the start date for the restarted run. Further, diag_table has in its first line the time stamp for the ``days since XXX'' time axis specification. Quick and dirty method: enter 1901 1 1 0 0 0 in input.nml and diag_table, and remove INPUT/coupler.nml Or enter the date 1901 1 1 0 0 0 also in the 3rd line of INPUT/coupler.nml . ---- Build a climatology for LPJ spin-up: Use CM2M_coarse_BLING. Write out lprec\,fprec\,lwnet\,swdown\,t_surf\,t_surf_max\,t_surf_min\,wind on land_lad grid. NOTE: this requires to add apropriate diag output to land_lad/land_model.F90 . It also requires to add allocation and de-allocation of %wind in land_lad/land_types.F90 . It is not used otherwise in land_lad, and thus not passed through the FMS coupler, if not allocated. Run it for 50 years. Regrid to 0.5x0.5 degree resolution for LPJmL ../../../src/tools/fregrid/fregrid \ --input_mosaic /iplex/01/climber3/petri/mom5.0.2/work/CM2M_coarse_BLING/INPUT/land_mosaic.nc \ --input_file /iplex/01/climber3/petri/mom5.0.2/work/CM2M_coarse_BLING/history/01020101.land_daily.nc \ --output_mosaic /iplex/01/climber3/petri/mom5.0.2/work/mom5_CM2_lad_05LPJ_05_Xcoarse/INPUT/land_mosaic.nc \ --output_dir /iplex/01/climber3/petri/mom5.0.2/work/mom5_CM2_lad_05LPJ_05_Xcoarse/INPUT \ --output_file CM2M_land_daily_regridded_05x05.nc \ --scalar_field lprec\,fprec\,lwnet\,swdown\,t_surf\,t_surf_max\,t_surf_min\,wind Now add lprec and fprec to prec: cdo merge -chname\,fprec,prec \ -add \ -selvar\,fprec CM2M_land_daily_regridded_05x05.nc \ -selvar\,lprec CM2M_land_daily_regridded_05x05.nc \ CM2M_land_daily_regridded_05x05.nc \ CM2M_land_daily_regridded_05x05+prec.nc ncatted -a long_name\,prec\,m\,c\,'total precipitation rate (input)' CM2M_land_daily_regridded_05x05+prec.nc ---- Now collect the data files for the ocean, atmosphere, sea-ice and land_lad components. cd INPUT ln -s ../../CM2M_coarse_BLING/INPUT/* . Careful: this has now created symlinks to the initial conditions for year 0002 that are distributed by GFDL with the CM2M_coarse_BLING model setup. Those conditions specify, among others, cold start for land_lad, with 288K soil temperature (260K for snow-or-glacier covered cells). But above, we have used output from years 0052 to 00101 as climatology for the LPJmL spinup. Thus use the RESTART files for year 0102 as initial conditions for the coupled model run. cd ../Data-For-LPJ/CM2M_coarse_BLING-history-for-LPJspinup tar xvjf tar xvjf 01020101.RESTART.tar.bz2 cd 01020101.RESTART for i in *.res* ; do rm ../../../INPUT/$i ; cp -p $i ../../../INPUT/$i ; done # -> this did not work, due to differences in coastlines / land-sea mask! # #Still not ready, the RESTART files for land_lad must be regridded from #the CM2M_coarse_BLING grid to the LPJ grid. # ## mg_drag_res.nc belongs to the atmosphere! #../../../../../src/tools/fregrid/fregrid \ # --input_mosaic /iplex/01/climber3/petri/mom5.0.2/work/CM2M_coarse_BLING/INPUT/land_mosaic.nc \ # --input_file mg_drag.res.nc \ # --output_mosaic /iplex/01/climber3/petri/mom5.0.2/work/mom5_CM2_lad_05LPJ_05_Xcoarse/INPUT/land_mosaic.nc \ # --output_dir ../../../INPUT \ # --output_file mg_drag.res.nc \ # --scalar_field ghprime # #../../../../../src/tools/fregrid/fregrid \ # --input_mosaic /iplex/01/climber3/petri/mom5.0.2/work/CM2M_coarse_BLING/INPUT/land_mosaic.nc \ # --input_file soil.res.nc \ # --output_mosaic /iplex/01/climber3/petri/mom5.0.2/work/mom5_CM2_lad_05LPJ_05_Xcoarse/INPUT/land_mosaic.nc \ # --output_dir ../../../INPUT \ # --output_file soil.res.nc \ # --scalar_field frac\,soil_temp_1\,soil_temp_2\,soil_temp_3\,soil_temp_4\,soil_temp_5\,soil_temp_6\,soil_temp_7\,soil_temp_8\,soil_temp_9\,soil_temp_10\,soil_temp_11\,soil_temp_12\,soil_temp_13\,soil_temp_14\,soil_temp_15\,soil_temp_16\,soil_temp_17\,soil_temp_18\,water\,snow\,groundwater\,fusion_1\,fusion_2\,fusion_3\,fusion_4\,fusion_5\,fusion_6\,fusion_7\,fusion_8\,fusion_9\,fusion_10\,fusion_11\,fusion_12\,fusion_13\,fusion_14\,fusion_15\,fusion_16\,fusion_17\,fusion_18 # # #../../../../../src/tools/fregrid/fregrid \ # --input_mosaic /iplex/01/climber3/petri/mom5.0.2/work/CM2M_coarse_BLING/INPUT/land_mosaic.nc \ # --input_file vegetation.res.nc \ # --output_mosaic /iplex/01/climber3/petri/mom5.0.2/work/mom5_CM2_lad_05LPJ_05_Xcoarse/INPUT/land_mosaic.nc \ # --output_dir ../../../INPUT \ # --output_file vegetation.res.nc \ # --scalar_field q_ca Do we also need to regrid the non-restart data files? -> No we dont, those are on arbitrary / odd resolutions anyway, and are regridded to the actual model resolution at run time. Sigh. When starting with these regridded restart files for land_lad , the coupler fails in the very first time step in the infamous water vapour saturation pressure calculation, showing that it got NaNs in the temperature input. The current, preliminary, suspicion is that in the above regridding there are some cells at the land/sea border that are treated incorrectly. Without these restart files, the model starts without problems. ---- we still have problems with runoff map. Many cells in DDM30 are marked as ocean wich are taken as land by LPJ, and are inside the land mask of land_lad. For those cells, we cannot easily find a near-by land_lad ocean cell (== MOM ocean cell). ---- Sigh again. Regridding the soil and vegetation restart file for land_lad from coarse resolution to 720x360 yields incompatibilities around the coasts. Doing a cold start for the land model yields to instabilities in the coupler. Thus: Use the CM2M_coarse_BLING initial conditions on the above-created grid, where land_lad runs on 720x360 cells. Run it for the usual 50 years, then for another 50 years to create a climatology for LPJ spinup. Then use the high-res RESTART files for soil and vegetation to start the coupled model. For this, we set namelist option do_lpj = .false. , so that the LPJ is compiled in, but not run. Sigh. This crashes after ca. 7 years. Next try: use the original CM2M code, i.e. with land_lad instead of land_atlantes, i.e. without Matteos snow age albedo (and without any trace of LPJ). On 06/23/16 17:23, Stefan Petri wrote: > Hi! > > Das waren jetzt 10 Jahre das fast-Original CM2M_coarse_BLING setup, aber > mit land-Modell auf dem Halbgrad-Gitter statt dem groben Gitter. > Immerhin sehen in den Restart-Files die Land-Masken OK, sie stimmen (mit > ncview betrachtet) mit denen in der Gitter-Definition ueberein. Also > besser, als die Restart-Files, die vom groben auf das Halbgrad-Gitter > regridded wurden. > Ich lasse das jetzt mal weiterlaufen, vielleicht bekomme ich diemsal > mehr Diagnostik warum / wo es abstuerzt (odr vielleicht tut es das > diesmla ja auch gar nicht ...) This crashes after ca. 22 years. Looking at grid definitions, grid generating process, and reading of land model restart files, Erik and I discover in soil.F90: if( file_exist('INPUT/soil.res.nc') .or. file_exist('INPUT/soil.res.tile1.nc') ) then call read_data ( fname, 'frac', soil%frac) soil%mask = (soil%frac >= 0) endif For ``normal'' runs, with unchanged resolution, we should expect that the restored frac is identical to the frac that was derived from the grid_spec in the land_model init. However, for regridded restart files, we see that the regridded frac is very different from the land mask in the grid definition :-( On 06/30/16 15:33, Stefan Petri wrote: > wie ich vor ein paar Tagen geschrieben habe, crasht das CM2M_coars_BLING, wenn > man das dadrin haengende land_lad auf der LPJ-Aufloesung laufen laesst statt > auf der Origirnalen groben Aufloesung. Der Coupler findet im exchange-Grid an > 4 Stellen Temperaturen von 390K. Hier ein Schnappschuss vom letzten > Zeitschritt, fuer den noch NetCDF-Output geschrieben wird, das ist ca. der > vor-vor-letzte vor dem Crash. > > Im Himalaya findet sich in der Oberflaechentemperatur des Land-Modells eine > Eiskalte Zelle, dementsprechend mit sehr steilem Temperatur-Gradienten zu den > Nachbarn. Am Antarktis-Rand und im Polarmeer treten aehnliche Phaenomene auf. > Bis kurz davor sehen sowohl global mean surface temperature als auch die > surface temperature in diesem fraglichen Bereich normal aus, sehr aehnlich > unserem Referenz-Lauf , dem original-CM2M_coarse_BLING. > > Aehnliche Phaenomene hatten wir mit Levke beim Aeolus-Debugging gesehen, da > hat Reduzierung der Zetschrittlaenge geholfen. Allerdings gibt es m.W. im > Landmodell eigentlich keine Prozesse, bei denen das Verhaeltnis zw. > Zellengrosse und Zeitschrittlaenge kritisch sein sollte (CFL-Kriterium o.ae.) > > Ich probiere das trotzdem mal aus. On 07/01/16 12:02, Stefan Petri wrote: > mit halbiertem Zeitschritt (sowohl atmosphaere+land als auch Ozean) taucht das > Problem nicht auf. On 07/01/16 13:46, Stefan Petri wrote: > ich habe jetzt den OZean-Zeitscrhitt bei 3h gelassen, und nur den Atmos > (+land)-Zeitschritt von 1.5h auf 1h runtergesetzt, da tritt der Crash auch > nicht auf. Wed Jul 6 09:37:35 CEST 2016 Let us try 50 years from 00020101 with dt_atmos=3600 On 07/07/16 14:51, Stefan Petri wrote: > Mit dt_atmos=3600 (statt 5400) hat es jetzt 50 Jahre ohne Crash gerechnet, vom > Atmosphaeren-Kaltstart an. > Vergleiche der Surface temperature zwischen dem Lauf mit feiner Aufloesung > (schwarz) und dem Original mit grober Land-Aufloesung sind angetackert. > Stellenweise laufen die Versionen deutlich auseinander, aber fuer Erik und > mich sah keine unplausibler aus als die andere... > > Ich habe das jetzt fuer weitere 30 Jahre gestartet, um damit eine Klimatologie > fuer ein LPJ-Spinup mit von vorherein passenden Land-Ozean-Grenzen zu erzeugen. Georg suggested to compare not individual timesteps but climatology of the runs with different timesteps, to exclude comparison of local and temporal weather phenomena. That climatology comparison looks qite good. Thus we now take the 30-year daily output and prepare it as input for LPJ spinup: Add lprec and fprec to prec: cd CM2M_coarse_BLING_lad05_dtatmos3600-history-for-LPJspinup cdo -v merge \ -setattribute\,prec@long_name='total precipitation rate (input)' \ -chname\,fprec\,prec \ -add \ -selvar\,fprec 00820101.land_daily.nc \ -selvar\,lprec 00820101.land_daily.nc \ 00820101.land_daily.nc \ 00820101.land_daily+prec.nc #ncatted -a long_name\,prec\,m\,c\,'total precipitation rate (input)' 00820101.land_daily+prec.nc To create the spinup configuration, we first copy the original config file from the trunk: cp -p ../../src/land_atlantes/lpj_trunk/lpjml.conf lpjml_spinup.conf edit it to use param_spinup.conf and input_spinup.conf cp -p ../../src/land_atlantes/lpj_trunk/input.conf input_spinup.conf cp -p ../../src/land_atlantes/lpj_trunk/param.conf param_spinup.conf cdo -s outputtab\,year\,value -fldmean -yearmean -selvar\, Sigh. There is no CO2 output in the created diagnostic files. The only usable thing that is shown in the diag_field_log is rrvco2. It turns out that for this run, rrvco2 is globally constant at 286 . (which is to be expected from the entry 1860.5 in file INPUT/co2_gblannualdata , which is read and used according to the namelist parameter &radiative_gases_nml verbose = 3 gas_printout_freq = 240 time_varying_co2 = .false., co2_variation_type = 'linear', co2_dataset_entry = 1860,1,1,0,0,0, co2_specification_type = 'time_series' co2_data_source = 'input' [..] / for y in `seq -5000 2100` ; do echo $y 286.0 ; done > co2_spinup.txt ------------ Wed Jan 16 15:22:36 CET 2019 Apparently, the land model gets evaporation and dedq from the the coupler as _input_ , and computes an updated surface humidity. To emulate that with LPJ, we need a climatology with tr_flux(:,:,:,isphum) and dfdtr(:,:,:,isphum). # for creating a climatology for LPJ spinup "land_mod", "lprec", "lprec", "land_daily", "all", .true., "none", 2 "land_mod", "fprec", "fprec", "land_daily", "all", .true., "none", 2 "land_mod", "lwnet", "lwnet", "land_daily", "all", .true., "none", 2 "land_mod", "swdown", "swdown", "land_daily", "all", .true., "none", 2 "land_mod", "t_surf", "t_surf", "land_daily", "all", .true., "none", 2 "land_mod", "t_surf", "t_surf_max", "land_daily", "all", max, "none", 2 "land_mod", "t_surf", "t_surf_min", "land_daily", "all", min, "none", 2 "land_mod", "wind", "wind", "land_daily", "all", .true., "none", 2 "land_mod", "qflux", "qflux", "land_daily", "all", .true., "none", 2 "land_mod", "dedq", "dedq", "land_daily", "all", .true., "none", 2 "land_mod", "q_ca", "q_ca", "land_daily", "all", .true., "none", 2 "land_mod", "evap_surf","evap_surf", "land_daily", "all", .true., "none", 2 Added apropriate diagnostic output to land_lad/land_model.F90 . Run CM2M with land_lad at 0.5deg from 00520101 for 30 years with daily output for LPJ. lrwxrwxrwx 1 petri climber3 14 Jul 12 2016 INPUT -> INPUT-00520101 SLURM Job_id=11680232 Name=CM2M_Coarse_BLING_LPJ_05 Ended, Run time 14:01:22, COMPLETED, ExitCode 0 WorkDir=/p/projects/climber3/petri/POEM/work/CM2M_coarse_BLING_LPJ_05 -rw-r--r-- 1 petri climber3 65719901 Jan 17 05:48 00820101.RESTART.tar.bz2 -rw-r--r-- 1 petri climber3 24970928 Jan 17 04:35 00820101.atmos_month.nc -rw-r--r-- 1 petri climber3 136236085008 Jan 17 05:44 00820101.land_daily.nc -rw-r--r-- 1 petri climber3 6719555088 Jan 17 05:48 00820101.land_month.nc -rw-r--r-- 1 petri climber3 319124 Jan 17 05:48 00820101.out.tar.bz2 Also, in the usual LPJ configurations it is required to start the climatology in year 1901. One possibility is to set the start date to 1901 01 01 in INPUT/couper.res and at the top of diag_table. The other possibility is to change the time axis of an already written output file. Further: it was observed that with some versions of the NetCDF library, LPJ is very slow in reading from large files with record time axis. A workaround is to convert the time axis to a ``fixed'' axis. #cdo setreftime\,1901-01-01\,00:00:00 00820101.land_daily+prec.nc 19010101.land_daily+prec.nc cdo -v merge \ -shifttime\,$((1930-82+1))years \ -setattribute\,prec@long_name='total precipitation rate (input)' \ -chname\,fprec\,prec \ -add \ -selvar\,fprec 00820101.land_daily.nc \ -selvar\,lprec 00820101.land_daily.nc \ 00820101.land_daily.nc \ 19300101.land_daily+prec.nc ncks -4 19300101.land_daily+prec.nc 19300101.land_daily+prec-nc4.nc ncks --fix_rec_dmn time 19300101.land_daily+prec-nc4.nc 19300101.land_daily+prec-nc4-fix.nc -------------- Mon Jul 15 13:28:28 CEST 2019 Added surf.res to INPUT-00020101 to have surface orography in the atmosphere. Note that this is on the original CM2Mc land-sea-mask, with the coarse-resolution land model setup. > I have now run a variation of CM2M_coarse_BLING , with the orography > derived from the restart file fv_rst.res.nc from https://groups.google.com/forum/#!topic/mom-users/GhmoD6IYvxM > > Note that this is still an atmosphere cold start, but now with non-flat > orography. It ran for 50 years. After the 1st year, it looks already > like a quite stable state to me. I have started to prepare a climatology for LPJmL spinup with the newly introduced orography height for the atmosphere. As we found out earlier, for the preparation of the climatology I must run the land model at LPJ resolution already. The above-mentioned orography height, however, is read in the atmosphere code, from a file and in a piece of code which does not know anything about the coupler's land-sea-mask nor the land model's resolution. Thus, the orography height as seen by the atmosphere and the land-sea-mask as seen by the coupler and the land model differ slightly, especially at west coast of the Andes, and few places around antarctica. See attached plot. However, for now I tend to just ignore those differences. The real fix, of course, would be to read topo height in the land model on its grid resolution and land-sea-mask, and pass it to the atmosphere via the coupler's grid interpolation. The FMS infrastructure already has the code to read topography and interpolate to land resolution. The coupler would need to be augmented to pass on this information, possibly only once at init time. And, however, I think that topo-reading code currently does not obey any land-sea mask. Sigh. cd /p/projects/climber3/petri/POEM/work/CM2M_coarse_BLING_LPJ_05/ mv history history-CM2M_coarse_BLING_land05deg-dt3600+oro cd history-CM2M_coarse_BLING_land05deg-dt3600+oro # there is a typo in the metadata definition in this program code ncatted -a units\,dedq\,m\,c\,'kg/m^2/s' 00820101.land_daily.nc cdo -v -f nc4 merge \ -shifttime\,$((1930-82+1))years \ -setattribute\,prec@long_name='total precipitation rate (input)' \ -chname\,fprec\,prec \ -add \ -selvar\,fprec 00820101.land_daily.nc \ -selvar\,lprec 00820101.land_daily.nc \ 00820101.land_daily.nc \ 19300101.land_daily+prec.nc ncks --fix_rec_dmn time 19300101.land_daily+prec-nc4.nc 19300101.land_daily+prec-nc4-fix.nc ----------------- Mon Sep 21 11:00:41 CEST 2020 Create climatology file and restart conditions with the namelist and cross-land changes from Elizabeth Maroon (i.e. merge configuration changes from Galbraith et al into the ``official'' CM2M_coarse_BLING configuration. See exp/CM2M_coarse_BLING/README, work/CM2M_coarse_BLING-EM2) ------------------ cd /p/projects/climber3/petri/POEM/work/CM2M_coarse_BLING_LPJ_05/ mv history history-CM2M_coarse_BLING_land05deg-EM2 cd history-CM2M_coarse_BLING_land05deg-EM2 # there is a typo in the metadata definition in this program code ncatted -a units\,dedq\,m\,c\,'kg/m^2/s' 00820101.land_daily.nc cdo -v -f nc4 -z zip_9 merge \ -shifttime\,$((1930-82+1))years \ -setattribute\,prec@long_name='total precipitation rate (input)' \ -chname\,fprec\,prec \ -add \ -selvar\,fprec 00820101.land_daily.nc \ -selvar\,lprec 00820101.land_daily.nc \ 00820101.land_daily.nc \ 19300101.land_daily+prec.nc ncks -4 -L 9 --fix_rec_dmn time 19300101.land_daily+prec-nc4.nc 19300101.land_daily+prec-nc4-fix.nc ------------------------- Tue Oct 20 09:07:50 CEST 2020 To match the code changes in src/ocean_shared/generic_tracers/generic_BLING.F90 committed in rev 3754 on 2020-10-14 16:51:32, remove the name suffix _b from the variable names in INPUT/ocean_bling.res.nc . This follows Eric Galbraith's suggestions in his mail from 22.09.20 19:31 (see CM2Mc/README) to remove all the _b suffixes from all tracer names. ncrename -v dic_b\,dic \ -v alk_b\,alk \ -v po4_b\,po4 \ -v o2_b\,o2 \ -v dop_b\,dop \ -v fed_b\,fed \ -v htotal_b\,htotal \ -v co3_ion_b\,co3_ion \ -v irr_mem_b\,irr_mem \ ocean_bling.res.nc --------------- Fri Oct 30 15:50:26 CET 2020 To produce consistent initial conditions and atmosphere climatology forcing, performed runs with the -EM namelist settings, and the cross-land changes in the field_table, and with the BLING tracer names corrected without _b suffix. -> INPUT-00520101-EM3-no_b -> ./history-CM2M_coarse_BLING_land05deg-EM3-no_b/ # there is a typo in the metadata definition in this program code ncatted -a units\,dedq\,m\,c\,'kg/m^2/s' 01020101.land_daily.nc cdo -v -f nc4 -z zip_9 merge \ -shifttime\,$((1930-82+1))years \ -setattribute\,prec@long_name='total precipitation rate (input)' \ -chname\,fprec\,prec \ -add \ -selvar\,fprec -seldate\,0052-01-01T00:00:00\,0081-12-31T23:59:59 01020101.land_daily.nc \ -selvar\,lprec -seldate\,0052-01-01T00:00:00\,0081-12-31T23:59:59 01020101.land_daily.nc \ -selvar\,t_surf\,t_surf_max\,t_surf_min\,lwnet\,swdown\,wind\,qflux\,dedq\,q_ca \ -seldate\,0052-01-01T00:00:00\,0081-12-31T23:59:59 01020101.land_daily.nc \ 19300101.land_daily+prec.nc ncks -4 -L 9 --fix_rec_dmn time 19300101.land_daily+prec.nc 19300101.land_daily+prec-nc4-fix.nc ------------- On 25.10.21 16:41, Stefan Petri wrote: > In > /p/projects/climber3/petri/POEM/exp/CM2M_coarse_BLING_LPJ_05/ > CM2M_coarse_BLING_lad05_dtatmos3600-history-for-LPJspinup/ > I had prepared a spinup climatology in summer 2016. Since then, several > fixes and updates were applied to the CM2M_coarse_BLING setup, and > new spinup climatologies created. Thus I am now going to free those 282GB > disk space.