Main menu

Navigation

Running

10 posts / 0 new
Last post
fyke@...
Running

Hi there,

 

I am attempting to get CLM set up on a Regional Arctic System Model grid for the first time and am struggling with getting the necessary tools to work on Yellowstone.  As a first step, when I run gen_ESMF_mapping (called from within gen_ESMF_mapping):

 

./gen_cesm_maps.sh --fileocn $WorkDir/wr50a_090614.nc \
--fileatm $WorkDir/wr50a_090614.nc \
--typeocn regional --typeatm regional \
--nameocn wr50a_ocn --nameatm wr50a_atm

 

I receive this error:

 

./gen_ESMF_mapping_file/create_ESMF_map.sh
Mon Mar 25 16:21:16 MDT 2013
type_src is regional
type_dst is regional
map_type is patc
fsrc is /glade/u/home/jfyke/work/CLM_on_RASM_grid/wr50a_090614.nc
fdst is /glade/u/home/jfyke/work/CLM_on_RASM_grid/wr50a_090614.nc
Running interactively

Creating map_wr50a_atm_TO_wr50a_ocn_patc.130325.nc
regional source
regional destination
cmd is /glade/p/work/svasquez/install/esmf-5.3.0-O/bin/ESMF_RegridWeightGen --ignore_unmapped -m patch -p all -w map_wr50a_atm_TO_wr50a_ocn_patc.130325.nc -s /glade/u/home/jfyke/work/CLM_on_RASM_grid/wr50a_090614.nc -d /glade/u/home/jfyke/work/CLM_on_RASM_grid/wr50a_090614.nc --src_regional --dst_regional

ERROR: Program was not launched by POE. Abort!!!
Error status returned from create_ESMF_map script

 

Furthermore, after gen_ESMF_mapping fails, this message is returned from check_maps.sh:

 

1) For use on bluefire only!

 

Which makes me wonder if this tool is ready for Yellowstone use?

I also tried using the regridbatch.yellowstone.sh script to call gen_cesm_maps.sh.  However, I was unsuccessful in getting it to run, error:

 

mpirun.lsf: LSF_PJL_TYPE is undefined. Exit

 

Has this script been successfully used on Yellowstone?  I do note that within this script, it states:

 

"# Set up for bluefire"

 

Thanks for your help, and for providing these tools and support.   Please call directly, if it seems easier to do so and if you have any time.

 

Jeremy Fyke

505-606-0025

Jeremy Fyke

tcraig
Jeremy, First, what version of CESM are you using? Second, you cannot run this script interactively on yellowstone. You have to login to geyser. To login to geyser from yellowstone, do something like bsub -Is -q geyser -W 20:00 -n 1 -P P93300606 tcsh then try running the script again. this script worked for me recently on yellowstone using the latest version of mapping. please post if you have other problems.
fyke@...
Hi Tony, Works, now, thanks. FYI I am using clm tools from cesm1_2_beta02, with mapping tools updated to https://svn-ccsm-models.cgd.ucar.edu/tools/mapping/trunk/. I think the update to the latest mapping tools helped, as did the point to run interactively on geyser.

Jeremy Fyke

erik
Hey Jeremy The similar CLM tool mkmapdata.sh had similar issues documented here... http://bugs.cgd.ucar.edu/show_bug.cgi?id=1630 that might help you to solve your issues.

Erik Kluzek ...............

CESM Land Model (CLM) Software Liason

CESM Software Engineering Group, NCAR

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

fyke@...
Hi Erik, Thanks for the tip, however in the meantime I have gotten things (almost completely) working. But you may be able to help me with something else, very related. I receive errors when i try to run mksurfdata.pl with glc_nec>0 (which I require). Has this routine been tested with glc_nec>0, lately? Could you let me know if the problem is on my end (eg, you can successfully generate datasets with glc_nec>0). Thanks again,

Jeremy Fyke

sacks
Hey Jeremy, I have run mksurfdata.pl with -glc_nec > 0 recently. Can you post the exact command you're using and the error message? And also the tag you're using. Bill

Bill Sacks

CESM Software Engineer

fyke@...
###Exact command### ./mksurfdata.pl -res usrspec -usr_gname wr50a -usr_gdate 130326 -glc_nec 10 ###Error message### Attempting to make percent elevation class and mean elevation for glaciers ..... domain_read read lon and lat dims domain_read initialized domain domain_read read LANDMASK Open glacier file: /glade/p/cesm/cseg/inputdata/lnd/clm2/rawdata/mksrf_glacier_3x3min_simyr2000_me rgeGreenland.c120921.nc (gridmap_map_read) reading mapping matrix data... (gridmap_map_read) * file name : ../mkmapdata/map_3x3min_GLOBE-Gardner-mergeGIS_to_wr50a_nomask_aave_da_c130326.nc * matrix dimensions rows x cols : 25920000 x 56375 * number of non-zero elements: 3894140 Level: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 no,pctglc,pctglac= 25222 99.9999893097221 100.000000000000 no,pctglc,pctglac= 26323 99.9999887998122 100.000000000000 no,pctglc_gic,pctglc_icesheet,pctglac,pctglac_uncorrected,lon,lat= 25222 0.000000000000000E+000 99.9999893097221 100.000000000000 100.000000000000 -42.1524807525730 72.0282282018085 no,pctglc_gic,pctglc_icesheet,pctglac,pctglac_uncorrected,lon,lat= 26323 0.000000000000000E+000 99.9999887998122 100.000000000000 100.000000000000 -36.2413098388904 72.0651849886316 abort: ERROR in mksurfdata_map: 34304 ###Tag I am using### cesm1_2_beta02 In case you want to investigate the actual input files, this is checked out to working directory: /glade/u/home/jfyke/work/CLM_on_RASM_grid/cesm1_2_beta02

Jeremy Fyke

sacks
Hey Jeremy, Looks like I was simply too strict in my error checking... there is a roundoff-level difference slightly bigger than was allowed (but is not worrisome). I recommend simply applying these diffs, which will let this roundoff error slide through. This file is in mksurfdata_map/src Index: mkglcmecMod.F90 =================================================================== --- mkglcmecMod.F90(revision 45444) +++ mkglcmecMod.F90(working copy) @@ -414,7 +414,7 @@ ! Check that sum over pctglcmec_o (from 1 to nglcec) is equal to pctglac_o(no) ! - ! Note: We use a threshold of 1e-5 rather than 1e-6 or smaller because there is single- + ! Note: We use a threshold of 2e-5 rather than 1e-6 or smaller because there is single- ! precision roundoff error in the storage of the flat (i.e., non-multi-level) ! PCT_GLACIER variable on the input dataset (which is used to compute pctglac_o) (since ! the input dataset uses single-precision floats) @@ -423,7 +423,7 @@ do m = 1,nglcec glc_sum = glc_sum + pctglcmec_o(no,m) end do - if (abs(glc_sum - pctglac_o(no)) > 1.e-5) then + if (abs(glc_sum - pctglac_o(no)) > 2.e-5) then write(6,*)'no,pctglc,pctglac= ',no,glc_sum,pctglac_o(no) errors = .true. end if @@ -431,12 +431,12 @@ ! Check that GIC + ICESHEET = total glacier, for variables summed over elevation classes ! - ! Note: We use a threshold of 1e-5 rather than 1e-6 or smaller because there is single- + ! Note: We use a threshold of 2e-5 rather than 1e-6 or smaller because there is single- ! precision roundoff error in the storage of the flat (i.e., non-multi-level) ! PCT_GLACIER variable on the input dataset (which is used to compute pctglac_o) (since ! the input dataset uses single-precision floats) do no = 1,ns_o - if (abs((pctglc_gic_o(no) + pctglc_icesheet_o(no)) - pctglac_o(no)) > 1.e-5) then + if (abs((pctglc_gic_o(no) + pctglc_icesheet_o(no)) - pctglac_o(no)) > 2.e-5) then write(6,*)'no,pctglc_gic,pctglc_icesheet,pctglac,pctglac_uncorrected,lon,lat=',no,pctglc_gic_o(no),& pctglc_icesheet_o(no),pctglac_o(no),pctglac_o_uncorrected(no),tgridmap%& xc_dst(no),tgridmap%yc_dst(no)

Bill Sacks

CESM Software Engineer

sacks
It looks like a related change is needed in the CLM source code, in main/surfrdMod.F90: Index: surfrdMod.F90 =================================================================== --- surfrdMod.F90(revision 45515) +++ surfrdMod.F90(working copy) @@ -761,11 +761,11 @@ ! (and there are arguments for putting it here anyway). do nl = begg,endg - ! We need to use a fairly high threshold for equality (1.0e-5) because pctgla + ! We need to use a fairly high threshold for equality (2.0e-5) because pctgla ! and pctglc_mec are computed using single precision inputs. Note that this ! threshold agrees with the threshold in the error checks in mkglcmecMod: ! mkglcmec in mksurfdata_map. - if (abs(pctgla(nl) - pctglc_mec_tot(nl)) > 1.0e-5) then + if (abs(pctgla(nl) - pctglc_mec_tot(nl)) > 2.0e-5) then write(iulog,*) ' ' write(iulog,*) 'surfrd error: pctgla not equal to sum of pctglc_mec for nl=', nl write(iulog,*) 'pctgla =', pctgla(nl) @@ -800,7 +800,7 @@ ! If pctglc_mec_tot is very close to 100%, round to 100% do nl = begg,endg - ! The inequality here ( <= 1.0e-5 ) is designed to be the complement of the + ! The inequality here ( <= 2.0e-5 ) is designed to be the complement of the ! above check that makes sure pctglc_mec_tot is close to pctgla: so if pctglc= ! 100 (exactly), then exactly one of these conditionals will be triggered. ! Update 9-28-12: Now that there is a rescaling of pctglc_mec to bring it more @@ -809,7 +809,7 @@ ! - or perhaps even get rid of this whole block of code. But I'm keeping this as ! is for now because that's how I tested it, and I don't think it will hurt ! anything to use this larger tolerance. - if (abs(pctglc_mec_tot(nl) - 100._r8) <= 1.0e-5) then + if (abs(pctglc_mec_tot(nl) - 100._r8) <= 2.0e-5) then pctglc_mec(nl,:) = pctglc_mec(nl,:) * 100._r8 / pctglc_mec_tot(nl) pctglc_mec_tot(nl) = 100._r8 endif

Bill Sacks

CESM Software Engineer

flocolleoni@...

Hi everybody,


I am actually trying to redo all my initial conditions for a glaciation to run with CESM 1.1.2.

So far, I created an run successfully the glaciations initial conditions files for B1850 case running CESM 1.0.5.


I was trying to update a little bit the procedure to change all the necessary land surface initial conditions

and I got a problem using the PCT_GLACIER file in the mksurfdata_map tool.

 

In the old version of CESM 1.0.5, the mksrf-glaciers.nc only contained PCT_GLACIER variable.
 From that and from the mksrf-topo.nc file, the mksurfdata tool produced the needed glacier elevation classes
 by its own. For CESM 1.1.2, i need to provide a glacier file which is already distributed over the vertical elevation classes.

How can I do that? How can I convert my old mksrf-glacier.nc file to the new type of mksrf-glacier.nc file needed by the mksrufdata_map tool?

 

Cheers

 

Florence

Florence Colleoni
CMCC (Bologna, Italy)

Log in or register to post comments

Who's new

  • praveenmaniyatt@...
  • arjunbabun11@...
  • lama@...
  • sisi393@...
  • 1658093099@...