[ncl-talk] Error Generating Weight Files for ESMF regridding
Mary Haley
haley at ucar.edu
Tue Mar 27 22:44:15 MDT 2018
Hi Tim,
I believe the issue is simply that the ESMF_RegridWeightGen application,
which is called by ESMF_regrid to create the weights file, is asking for
more memory than it has access to, and it exits rather abruptly. It creates
the source and destination NetCDF files just fine (I plotted them to make
sure).
I'm running the script through qsub to see if I can increase the amount of
memory.
Meanwhile, I was able to get the data regridded by subsetting the original
7770 x 7563 grid with every 2nd point (I also tried nstep=5 and 10). This
also seems to point to a memory issue.
nstep = 2
src_lat = smfile->latitude(::nstep)
src_lon = smfile->longitude(::nstep)
. . .
var := sfile->$var_name$(::nstep,::nstep)
You should also do this, as your input data contains missing values:
Opt at SrcMask2D = where(.not.ismissing(var),1,0)
Since your output destination grid doesn't contain missing values, you
*don't* need this step:
Opt at DstMask2D = where(.not.ismissing(dst_lat).and.
.not.ismissing(dst_lon),1,0)
There are some other minor changes, but I'll wait and see if I can get the
full grid to be regridded and then I will post back here.
--Mary
On Mon, Mar 26, 2018 at 8:18 AM, Glotfelty, Timothy William <
twglotfe at email.unc.edu> wrote:
> Hi Mary,
>
>
>
> Thank you for getting back to me. I tried the netcdf4 option but that
> did not fix the issue. The destination_grid_file.nc is 3.5 MB but the
> source_grid_file.nc is very large at 4.6 GB. I’m using NCL version
> 6.4.0. My original input files are approximately 1.5 GB in total. Is that
> too large to upload to the frp site?
>
>
>
> Thanks,
>
> Tim
>
>
>
> *From:* Mary Haley [mailto:haley at ucar.edu]
> *Sent:* Friday, March 23, 2018 5:18 PM
> *To:* Glotfelty, Timothy William <twglotfe at email.unc.edu>
> *Cc:* ncl-talk at ucar.edu
> *Subject:* Re: [ncl-talk] Error Generating Weight Files for ESMF
> regridding
>
>
>
> Hi Tim,
>
>
>
> Hmmm, I'm not sure about this, but maybe the weights file is too large?
> You might need to set:
>
>
>
> Opt at WgtNetCDFType = "netcdf4"
>
>
>
> which will force a NetCDF4 file to be written, which allows for large
> variables. I'm a little dubious this will fix these, because it should have
> complained when it tried to write that file.
>
>
>
> How big are source_grid_file.nc and destination_grid_file.nc?
>
>
>
> Also, what version of NCL are you using?
>
>
>
> If you continue to have problems even when setting the WgtNetCDFType
> option, then could you provide the input files, if they are not excessively
> large? You can use our ftp:
>
>
>
> http://www.ncl.ucar.edu/ftp_files.shtml
>
>
>
> Thanks,
>
>
>
> --Mary
>
>
>
>
>
>
>
> On Fri, Mar 23, 2018 at 12:13 PM, Glotfelty, Timothy William <
> twglotfe at email.unc.edu> wrote:
>
> Hello,
>
>
>
> I’m trying to run the ESMF regridding function to regrid land use and
> land cover information from a rectilinear Albers equal area projection to a
> WRF lambert conformal projection. I’ve used a version of this script many
> times to regrid between different projections and it has worked perfectly
> but this time I get an error indicating that the weights cannot be
> generated. Any ideas as to why this might be occurring? Could it have
> something to do with the Albers projection?
>
>
>
> My script is as follows:
>
>
>
> load "$NCARG_ROOT/lib/ncarg/nclscripts/csm/gsn_code.ncl"
>
> load "$NCARG_ROOT/lib/ncarg/nclscripts/csm/gsn_csm.ncl"
>
> load "$NCARG_ROOT/lib/ncarg/nclscripts/csm/contributed.ncl"
>
> load "$NCARG_ROOT/lib/ncarg/nclscripts/esmf/ESMF_regridding.ncl"
>
>
>
> begin
>
> vars = (/"2006_2001_Loss","2006_2001_Gain","2006_2001_Shift"/)
>
> outfile = "MODIS.2006-2001.Africa.nc"
>
> fon = addfile(outfile,"c")
>
>
>
> do v=0,dimsizes(vars)-1
>
>
>
> ;---Data file containing source grid
>
> src_file = "Change.2001_2006.nc" ;;---Change (likely)
>
> src_map = "MODIS_2001.nc"
>
> sfile = addfile(src_file,"r")
>
> smfile = addfile(src_map,"r")
>
>
>
> ;---Get variable to regrid
>
> var_name = vars(v) ;;---Change (likely)
>
> var = sfile->$var_name$ ;;---Change (likely)
>
> src_lat = smfile->latitude ;;---Change (maybe)
>
> src_lon = smfile->longitude ;;---Change (maybe)
>
>
>
> ;---Data file containing destination grid
>
> dst_file = "map.Africa.nc" ;;---Change (likely)
>
> dfile = addfile(dst_file,"r")
>
> dst_lat = dfile->XLAT(0,:,:) ;;---Change (likely)
>
> dst_lon = dfile->XLONG(0,:,:) ;;---Change (likely)
>
>
>
> ;---Set up regridding options
>
> Opt = True
>
>
>
> ;---"bilinear" is the default. "patch" and "conserve" are other options.
>
> Opt at InterpMethod = "neareststod" ;;---Change (maybe)
>
>
>
> Opt at WgtFileName = "rec_to_curv.nc" ; optional
>
>
>
> Opt at SrcGridLat = src_lat ; source grid
>
> Opt at SrcGridLon = src_lon
>
> Opt at SrcRegional = True ;;--Change (maybe)
>
> Opt at SrcInputFileName = src_file ; optional, but good idea
>
> ; missing
> values.
>
>
>
> Opt at DstGridLat = dst_lat ; destination grid
>
> Opt at DstGridLon = dst_lon
>
> Opt at DstRegional = True ;;--Change (maybe)
>
> Opt at DstMask2D = where(.not.ismissing(dst_lat).and.\
>
> .not.ismissing(dst_lon),1,0) ; Necessary if
> lat/lon
>
> ; has missing
> values.
>
>
>
> Opt at ForceOverwrite = True
>
> Opt at PrintTimings = True
>
> Opt at Debug = True
>
>
>
> var_regrid = ESMF_regrid(var,Opt) ; Do the regridding
>
>
>
> fon->$var_name$ = var_regrid
>
>
>
> printVarSummary(var_regrid)
>
> delete(var_name)
>
> delete(var)
>
> delete(var_regrid)
>
> delete(dst_lat)
>
> delete(dst_lon)
>
> delete(src_lat)
>
> delete(src_lon)
>
> delete(Opt)
>
> end do
>
> end
>
>
>
> The script output with the error message is as follows
>
>
>
> (0) get_src_grid_info: source lat dims = (7770)
>
> (0) get_src_grid_info: source lon dims = (7563)
>
> (0) get_src_grid_info: source grid type is 'rectilinear'
>
> (0) curvilinear_to_SCRIP: calculating grid corners...
>
> (0) curvilinear_to_SCRIP: no lat values are at the poles, so
>
> (0) calculating grid corners using
>
> (0) calc_SCRIP_corners_noboundaries...
>
> (0) calc_SCRIP_corners_noboundaries
>
> (0) min/max original lat: -35.23254905085637/37.99549894885637
>
> (0) min/max original lon: -18.96852865285641/52.3084051828564
>
> (0) calc_SCRIP_corners_noboundaries
>
> (0) min/max Extlat2d: -35.24197472314363/38.00492462114362
>
> (0) min/max Extlon2d: -18.9779543251436/52.3178308551436
>
> (0) calc_SCRIP_corners_noboundaries
>
> (0) min/max ExtGridCenter_lat: -35.237261887/38.00021178499999
>
> (0) min/max ExtGridCenter_lon: -18.973241489/52.313118019
>
> (0) =====> CPU Elapsed Time: rectilinear_to_SCRIP: 49.0142 seconds
> <=====
>
> (0) get_dst_grid_info: destination lat dims = (169,252)
>
> (0) get_dst_grid_info: destination lon dims = (169,252)
>
> (0) curvilinear_to_SCRIP: calculating grid corners...
>
> (0) curvilinear_to_SCRIP: no lat values are at the poles, so
>
> (0) calculating grid corners using
>
> (0) calc_SCRIP_corners_noboundaries...
>
> (0) calc_SCRIP_corners_noboundaries
>
> (0) min/max original lat: -35.2259/19.128
>
> (0) min/max original lon: -24.1202/64.4362
>
> (0) calc_SCRIP_corners_noboundaries
>
> (0) min/max Extlat2d: -35.5159/19.4431
>
> (0) min/max Extlon2d: -24.5117/64.8278
>
> (0) calc_SCRIP_corners_noboundaries
>
> (0) min/max ExtGridCenter_lat: -35.3709/19.2855
>
> (0) min/max ExtGridCenter_lon: -24.3161/64.6321
>
> (0) =====> CPU Elapsed Time: curvilinear_to_SCRIP: 0.0331116 seconds
> <=====
>
> (0) ESMF_regrid_gen_weights: number of processors used: 1
>
> (0) --------------------------------------------------
>
> (0) ESMF_regrid_gen_weights: the following command is about to be
> executed on the system:
>
> (0) 'ESMF_RegridWeightGen --source source_grid_file.nc --destination
> destination_grid_file.nc --weight rec_to_curv.nc --method neareststod
> --src_regional --dst_regional -i'
>
> (0) --------------------------------------------------
>
> (0) ESMF_regrid_gen_weights: output from 'ESMF_RegridWeightGen':
>
> (0) Starting weight generation with these inputs:
>
> (1) Source File: source_grid_file.nc
>
> (2) Destination File: destination_grid_file.nc
>
> (3) Weight File: rec_to_curv.nc
>
> (4) Source File is in SCRIP format
>
> (5) Source Grid is a regional grid
>
> (6) Source Grid is a logically rectangular grid
>
> (7) Destination File is in SCRIP format
>
> (8) Destination Grid is a regional grid
>
> (9) Destination Grid is a logically rectangular grid
>
> (10) Regrid Method: nearest source to destination
>
> (11) Pole option: NONE
>
> (12) Ignore unmapped destination points
>
> (13) Norm Type: dstarea
>
> (14)
>
> (0) --------------------------------------------------
>
> (0) ESMF_regrid_gen_weights: 'ESMF_RegridWeightGen' was not successful.
>
>
>
> I checked the weights log file and the only information it has is the
> following
>
>
>
> 20180323 140041.361 INFO PET0 Running with ESMF Version
> 6.3.0rp1
>
> 20180323 140041.376 INFO PET0 Running with ESMF Version
> 6.3.0rp1
>
> 20180323 140041.408 INFO PET0 Running with ESMF Version
> 6.3.0rp1
>
>
>
> Thanks,
>
> Tim Glotfelty
>
> Postdoctoral Research Associate
>
> Department of Environmental Science and Engineering
>
> University of North Carolina at Chapel Hill
>
>
> _______________________________________________
> ncl-talk mailing list
> ncl-talk at ucar.edu
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20180327/2c00e1ef/attachment.html>
More information about the ncl-talk
mailing list