A.1   Introducing a new ice sheet grid

This section describes what is needed when introducing a new ice sheet grid into the system. Much of the information in A.1.4   Generating the necessary inter-component mapping files is also relevant when introducing a new land grid (in which case new lnd-glc mapping files are needed) or a new ocean grid (in which case new ocn-glc mapping files are needed).

Note that local ice sheet grids must be rectangular; typically they are polar stereographic projections.

A.1.1   Prerequisites

Note

These instructions were written for the old yellowstone machine. The process on cheyenne will be very similar, but batch commands and some other details will differ.

The following instructions assume:

  1. You have a SCRIP grid file for your new grid
  2. You have the NetCDF operators (nco) in your path. On yellowstone, you can accomplish this with module load nco
  3. You are using the bash shell (otherwise some instructions will need to be modified slightly)
  4. You are working on yellowstone

A.1.2   Modifications needed for CISM

A.1.2.1   Add new entries in CISM’s xml file(s)

NEED TO FILL THIS IN WITH INSTRUCTIONS

A.1.2.2   For Greenland grids: Submerging land outside of Greenland

For Greenland grids, we typically need to submerge land outside of the island of Greenland. If we don’t do this (allowing for Ellesmere Island, Iceland, etc. to appear as land), we’ll run into at least two problems:

  1. CISM will try to dictate the land cover in those regions. Even if CLM’s surface dataset says there is glacier there, they will be overwritten with bare land if CISM doesn’t have any ice there.
  2. We can potentially get an ice sheet growing there if CLM dictates glacial inception.

A.1.2.2.1   New method, using the ‘mask’ field

Joe Kennedy’s new files have a ‘mask’ field on them that can be used for this purpose. Joe’s documentation of this field is:

  • 4 – floating ice
  • 3 – grounded ice
  • 2 – bare land
  • 1 – ocean
  • 0 – missing topg data
  • -1 – shallow paleo ocean
  • -2 – bare paleo land
  • -3 – shallow ocean or land outside the paleo domain

So, for the standard CISM input dataset, every point where mask < 0 should be sunk to -200 m a.s.l. For the paleo grid however, mask < -2 should be sunk and mask == -2 would become bare land (2) and mask == -1 would become ocean.

However, I also want to submerge the points with mask == 0, because I don’t trust the handling of missing values.

I used the following procedure:

  1. Submerge points with something like this (note: it’s important to have ‘mask’ in quotes; otherwise ncap2 thinks mask has a special meaning):

    ncap2 -s "where(('mask' == 0) || ('mask' < 0 && topg > -200)) topg = -200.;" greenland_4km_2017_02_23.epsg3413.nc greenland_4km_epsg3413_c170429.nc

  2. Confirm that there are now no missing values for topg, e.g., by loading it into python

  3. Remove the missing_value attribute from topg and add some metadata with something like this:

    ncatted -a missing_value,topg,d,, greenland_4km_epsg3413_c170429.nc
ncatted -h -a Note_170429,topg,c,c,"submerged all non-Greenland land to -200m with: ncap2 -s \"where(('mask' == 0) || ('mask' < 0 && topg > -200)) topg = -200.;\"; then removed now-unnecessary missing_value attribute" greenland_4km_epsg3413_c170429.nc
ncatted -h -a Note_170429,global,c,c,"Same as greenland_4km_2017_02_23.epsg3413.nc (provided by Joe Kennedy), except submerged all non-Greenland land to -200m with: ncap2 -s \"where(('mask' == 0) || ('mask' < 0 && topg > -200)) topg = -200.;\"; then removed now-unnecessary missing_value attribute of topg" greenland_4km_epsg3413_c170429.nc

A.1.2.2.2   Old method, using the ‘landcover’ field

Note: The method documented below is what I used before we had a ‘mask’ field on the input files.

Determining which grid cells are part of Greenland and which are outside Greenland is tricky. The easiest thing to do is to take an existing mask, which we have generated for some other grid, and regrid that to your new grid. Ideally, the existing mask will be at a resolution as close as possible to that of your new grid.

In order to regrid this mask from one CISM grid to another, follow this process. This assumes that you have some “old” CISM input file with the “landcover” field on it, and that you have a SCRIP grid file corresponding to that old input file as well as for your new CISM grid.

  1. Make a mapping file from the old grid to the new one: Modify the following script (see the FIXME note for what to change):

    #!/bin/bash
#
#
# Batch script to submit to create ESMF mapping file
#
# Set up for yellowstone
#
# yellowstone-specific batch commands:
#BSUB -P P93300601        # project number
#BSUB -n 8                # number of processors
#BSUB -R "span[ptile=16]"
#BSUB -W 1:00             # wall-clock limit
#BSUB -q caldera          # queue
#BSUB -o regrid.%J.out    # ouput filename
#BSUB -e regrid.%J.err    # error filename
#BSUB -J create_ESMF_map  # job name
#BSUB -N                  # send email upon job completion

#----------------------------------------------------------------------

#----------------------------------------------------------------------
# Set user-defined parameters here
#----------------------------------------------------------------------

# FIXME: Replace the following lines with paths to SCRIP grid files and names of your grids
filesrc="/glade/p/cesmdata/cseg/inputdata/glc/cism/griddata/SCRIPgrid_gland_4km_c161223.nc"
filedst="/glade/p/cesmdata/cseg/inputdata/glc/cism/griddata/SCRIPgrid_greenland_4km_epsg3413_c161223.nc"
namesrc='gland4kmOld'
namedst='gland4kmNew'

typesrc='regional'
typedst='regional'
maptype='aave'

#----------------------------------------------------------------------
# Done setting user-defined parameters
#----------------------------------------------------------------------

#----------------------------------------------------------------------
# Stuff done in a machine-specific way
#----------------------------------------------------------------------

# Determine number of processors we're running on
host_array=($LSB_HOSTS)
REGRID_PROC=${#host_array[@]}

#----------------------------------------------------------------------
# Begin general script
#----------------------------------------------------------------------

cmdargs="--filesrc $filesrc --filedst $filedst --namesrc $namesrc --namedst $namedst --typesrc $typesrc --typedst $typedst --maptype $maptype --batch"
env REGRID_PROC=$REGRID_PROC ./create_ESMF_map.sh $cmdargs

    Put this script in cime/tools/mapping/gen_mapping_files/gen_ESMF_mapping_file/, named regrid_cism_old_to_new.sh, then submit it with:

    bsub < regrid_cism_old_to_new.sh

  2. Extract the landcover field from your old CISM input file

    The landcover field is stored with a degenerate time dimension, but we need to remove that degenerate dimension. Run something like this, replacing the file path with the actual path to the CISM input file you’ll be using

    cd cime/tools/mapping/map_field
module load nco
ncks -v landcover /glade/p/cesmdata/cseg/inputdata/glc/cism/Greenland/glissade/init/greenland_4km_2015_06_03.mcb_trunk_c161025.nc landcover_old_with_time.nc
ncwa -a time landcover_old_with_time.nc landcover_old.nc

  3. Regrid the landcover field from your old CISM input file

    First, build the map_field tool (in cime/tools/mapping/map_field), by following the directions there.

    Then, from cime/tools/mapping/map_field, run something like the following, though replacing paths with the correct paths to your files. Note that, for this to work, you may need to source the env_mach_specific file that you sourced when building the map_field tool.

    ./map_field -m "/glade/p/work/sacks/cime/tools/mapping/gen_mapping_files/gen_ESMF_mapping_file/map_gland4kmOld_TO_gland4kmNew_aave.161223.nc" -if landcover_old.nc -iv landcover -of landcover_new.nc -ov landcover

  4. Round landcover to 0 or 1, and fix dimension names

    ncap2 -s 'landcover_int = int(round(landcover))' landcover_new.nc landcover_new2.nc
ncrename -d ni,x1 -d nj,y1 landcover_new2.nc
ncks -x -v landcover landcover_new2.nc landcover_new3.nc
ncrename -v landcover_int,landcover landcover_new3.nc

  5. Append landcover field onto input file

    Change the ‘today’ variable and file names to point to your actual file in the following:

    export today=161223
export path_to_input_file=/glade/p/cesmdata/cseg/inputdata/glc/cism/Greenland/glissade/init
export landcover_origfile=greenland_4km_2015_06_03.mcb_trunk_c161025.nc
export origfile=greenland_4km_2016_12_19.epsg3413.nc
export newfile=greenland_4km_epsg3413_c${today}.nc
cp $path_to_input_file/$origfile $path_to_input_file/$newfile
ncks -A -v landcover landcover_new3.nc $path_to_input_file/$newfile
ncatted -h -a no_data,landcover,c,i,0 -a has_data,landcover,c,i,1 -a Note_${today},landcover,c,c,"Regridded landcover from $landcover_origfile using area-conservative remapping then rounding to 0/1" $path_to_input_file/$newfile

  6. Submerge non-Greenland land with:

    export extra_info_on_origfile=" (provided by Joe Kennedy)"
ncap2 -s 'where(landcover == 0 && topg > -200) topg = -200.;' $path_to_input_file/$newfile tempfile.nc
mv tempfile.nc $path_to_input_file/$newfile
ncatted -h -a Note_${today},topg,c,c,"submerged all non-Greenland land to -200m with: ncap2 -s 'where(landcover == 0 && topg > -200) topg = -200.;'" $path_to_input_file/$newfile
ncatted -h -a Note_${today},global,c,c,"Same as ${origfile}${extra_info_on_origfile}, except (1) Includes landcover field, regridded from $landcover_origfile using area-conservative remapping then rounding to 0/1; (2) Submerged all non-Greenland land to -200m with: ncap2 -s 'where(landcover == 0 && topg > -200) topg = -200.;'" $path_to_input_file/$newfile

  7. Optional: Confirm the regridding of landcover.

    This step may not need to be done, but if you want to make sure landcover got regridded to the new grid properly, you can do it as follows. This uses python, with the NetCDF4 library. Note that dat_old points to the version of the dataset prior to modifying topg.

    dat_old = Dataset('greenland_4km_2016_12_19.epsg3413.nc')
dat_new = Dataset('greenland_4km_epsg3413_c161223.nc', 'a')
landcover = np.squeeze(dat_new.variables['landcover'][:])
topg_orig = np.squeeze(dat_old.variables['topg'][:])
category = dat_new.createVariable('category', 'i4', ('y1','x1'))
category_vals = np.zeros(landcover.shape)
land = np.logical_and(landcover==1, topg_orig>=0)
ocean = np.logical_and(landcover==1, topg_orig<0)
category_vals[ocean] = 1
category_vals[land] = 2
category[:] = category_vals
category.landcover_is_0 = 0
category.landcover_is_1_topg_lt_0 = 1
category.landcover_is_1_topg_ge_0 = 2
dat_new.close()

    Then, make sure:

    1. landcover = 0 points only occur off the coast of Greenland - not within or near Greenland

      First viewed this with a color scale that spanned 0 - 2 (so different colors for 0, 1 and 2), and viewing where the 0s are relative to the 1s and 2s. Ideally, there should be some 1 (ocean) between the 2 (land) and 0 (landcover = 0).

      Also viewed this by setting 0 to blue, 1-2 to white – making sure blue is only on periphery

    2. no topg > 0, landcover = 1 points outside of Greenland

      Viewed this by setting 2 to blue, 0-1 to white – making sure there is no blue on the periphery

A.1.3   Modifications needed for CLM

You need to ensure that the GLACIER_REGION field on CLM’s surface dataset is set up consistently with the new CISM grid. You should have a glacier region (or multiple glacier regions) encompassing the full CISM grid, whose glacier region behaviors are:

  • glacier_region_behavior = virtual: This is needed in order to provide downscaled forcings for all CISM grid cells.
  • glacier_region_melt_behavior = replaced_by_ice: This is needed in order to compute SMB throughout the CISM domain.

The value of glacier_region_ice_runoff_behavior can be whatever makes the most sense scientifically.

A.1.4   Generating the necessary inter-component mapping files

A.1.4.1   Generating lnd <-> glc mapping files for a new CISM grid

  1. Build the check_maps tool

    This isn’t entirely necessary, but allows the maps you generate to be checked by this tool. To build this, follow the instructions in cime/tools/mapping/check_maps/README.

  2. Modify the following script that will create the necessary mapping files. Make sure to fill in the correct values for -fglc and -nglc where it says ‘FIXME’:

    #!/bin/bash
#
#
# Batch script to submit to create suite of ESMF mapping files
#
# Set up for yellowstone
#
# yellowstone-specific batch commands:
#BSUB -P P93300601        # project number
#BSUB -n 8               # number of processors
#BSUB -R "span[ptile=16]"
#BSUB -W 24:00            # wall-clock limit
#BSUB -q caldera          # queue
#BSUB -o regrid.%J.out    # ouput filename
#BSUB -e regrid.%J.err    # error filename
#BSUB -J gen_cesm_maps    # job name
#BSUB -N                  # send email upon job completion

#----------------------------------------------------------------------

#----------------------------------------------------------------------
# Set user-defined parameters here
#----------------------------------------------------------------------

# CISM grid
# FIXME: Fill this in with the path to your SCRIP grid file and the name of your grid
glc_grid=" -fglc /PATH/TO/SCRIPgrid.nc -nglc gland4km "

# CLM grids
clm_f09=" -flnd $CESMDATAROOT/inputdata/lnd/clm2/mappingdata/grids/0.9x1.25_c110307.nc -nlnd fv0.9x1.25 "
clm_f19=" -flnd $CESMDATAROOT/inputdata/lnd/clm2/mappingdata/grids/1.9x2.5_c110308.nc -nlnd fv1.9x2.5 "
clm_T31=" -flnd $CESMDATAROOT/mapping/grids/T31_040122.nc -nlnd T31 "
clm_hcru=" -flnd $CESMDATAROOT/inputdata/lnd/clm2/mappingdata/grids/SCRIPgrid_360x720_nomask_c120830.nc -nlnd 360x720 "
clm_4x5=" -flnd $CESMDATAROOT/inputdata/lnd/clm2/mappingdata/grids/SCRIPgrid_4x5_nomask_c110308.nc -nlnd fv4x5 "
clm_10x15=" -flnd $CESMDATAROOT/inputdata/lnd/clm2/mappingdata/grids/SCRIPgrid_10x15_nomask_c110308.nc -nlnd fv10x15 "

# This grid is identical to $CESMDATAROOT/inputdata/lnd/clm2/mappingdata/grids/SCRIPgrid_ne120np4_nomask_c101123.nc
clm_ne120=" -flnd /glade/p/cesmdata/cseg/mapping/grids/ne120np4_pentagons_100310.nc -nlnd ne120np4 "

# This grid is identical to $CESMDATAROOT/inputdata/lnd/clm2/mappingdata/grids/SCRIPgrid_ne30np4_nomask_c101123.nc
clm_ne30=" -flnd /glade/p/cesmdata/cseg/mapping/grids/ne30np4_091226_pentagons.nc -nlnd ne30np4 "

# This grid is identical to $CESMDATAROOT/inputdata/lnd/clm2/mappingdata/grids/SCRIPgrid_ne16np4_nomask_c110512.nc
clm_ne16=" -flnd /glade/p/cesmdata/cseg/mapping/grids/ne16np4_110512_pentagons.nc -nlnd ne16np4 "

### Not bothering with this one: seems to not be used any more
### clm_f02=" -flnd /glade/p/cesmdata/cseg/mapping/grids/fv0.23x0.31_071004.nc -nlnd fv0.23x0.31 "

#----------------------------------------------------------------------
# Done setting user-defined parameters
#----------------------------------------------------------------------

#----------------------------------------------------------------------
# Stuff done in a machine-specific way
#----------------------------------------------------------------------

# Determine number of processors we're running on
host_array=($LSB_HOSTS)
REGRID_PROC=${#host_array[@]}

#----------------------------------------------------------------------
# Begin general script
#----------------------------------------------------------------------

for lnd_grid in "$clm_f09" "$clm_f19" "$clm_T31" "$clm_hcru" "$clm_4x5" "$clm_10x15" "$clm_ne120" "$clm_ne30" "$clm_ne16"; do
    cmdargs="$glc_grid $lnd_grid --batch"
    echo "=============================================================================="
    echo "About to execute gen_cesm_maps with: $cmdargs"
    env REGRID_PROC=$REGRID_PROC ./gen_cesm_maps.sh $cmdargs
done

  3. Name the script cism.regridbatch.sh, and put it in cime/tools/mapping/gen_mapping_files

  4. Run:

    bsub < cism.regridbatch.sh

    You can ignore errors in the .err file that look like this:

    ATTENTION: 0031-408  8 tasks allocated by Resource Manager, continuing...
ATTENTION: 0031-408  8 tasks allocated by Resource Manager, continuing...
Abort(0) on node 0 (rank 0 in comm -2080374782): application called MPI_Abort(comm=0x84000002, 0) - process 0
ERROR: 0031-300  Forcing all remote tasks to exit due to exit code 1 in task 0
forrtl: error (78): process killed (SIGTERM)
Image              PC                Routine            Line        Source
libpthread.so.0    0000003F7240F4B5  Unknown               Unknown  Unknown
libpoe.so          00002B1CF8267AE2  Unknown               Unknown  Unknown
libpthread.so.0    0000003F724079D1  Unknown               Unknown  Unknown
libc.so.6          0000003F718E88FD  Unknown               Unknown  Unknown

  5. Look through output in the .out file telling you about the results of running check_maps on all of your new mapping files.

    Ideally, you’ll see a lot of output that looks like this:

    1: map_gland4km_TO_fv0.9x1.25_aave.161222.nc
 All           21  tests passed!
-----
2: map_fv0.9x1.25_TO_gland4km_aave.161222.nc
 All           21  tests passed!
-----
3: map_fv0.9x1.25_TO_gland4km_blin.161222.nc
 All           14  tests passed!
-----

    However, you should expect to see errors when checking the very coarse-resolution fv10x15 grid, like this:

    1: map_gland4km_TO_fv10x15_aave.161222.nc
 ERROR: the test did not successfully map any values
 from the source grid to the destination grid
           0  of            0  tests failed. See above for details.
-----
2: map_fv10x15_TO_gland4km_aave.161222.nc
 FAILED: L1 error =   9.028874246726999E-003  in test            1
 FAILED: L1 error =   2.228991274441720E-002  in test            3
           2  of           21  tests failed. See above for details.
-----

    In addition, you may see additional errors like that for other CLM grids, particularly if you have a higher-resolution CISM grid: The tolerances in check_maps are set such that errors can be expected when checking mappings between regional grids and relatively coarse-resolution global grids.

  6. Put mapping files in correct directories in the inputdata space

    The mapping files should go in $CESMDATAROOT/inputdata/cpl/gridmaps/RES where RES is the from resolution. e.g., map_fv0.9x1.25_TO_gland4km_aave.161223.nc goes in $CESMDATAROOT/inputdata/cpl/gridmaps/fv0.9x1.25, whereas map_gland4km_TO_fv0.9x1.25_aave.161223.nc goes in $CESMDATAROOT/inputdata/cpl/gridmaps/gland4km. You can accomplish this with the following code in bash:

    for fl in map_*.nc; do
    IFS='_' read -ra fname_split <<< "$fl"
    from_res=${fname_split[1]}
    mv -v $fl $CESMDATAROOT/inputdata/cpl/gridmaps/${from_res}/
done

A.1.4.2   Generating glc -> ocn mapping files

See also https://github.com/NCAR/cism_misc-runoff_mapping_inputs

  1. Build the runoff_map tool in cime/tools/mapping/gen_mapping_files/runoff_to_ocn by following the directions there

  2. Create a namelist file like the following, but changing the details to match your new grid:

    &input_nml
   gridtype     = 'scrip'
   file_roff    = '/glade/p/cesmdata/cseg/inputdata/glc/cism/griddata/SCRIPgrid_greenland_4km_epsg3413_c161223.nc'
   file_ocn     = '/glade/p/cesm/cseg/mapping/grids/gx3v7_120309.nc'
   file_ocn_coastal_mask = '/glade/p/cesm/cseg/mapping/grids/gx3v7_coast_180430.nc'
   file_nn      = 'map_gland4km_epsg3413_to_gx3v7_nn.nc '
   file_smooth  = 'map_gx3v7_coast_to_gx3v7_sm.nc '
   file_new     = 'map_gland4km_to_gx3v7_nnsm_e1000r500_171024.nc'
   title        = 'runoff map: gland4km -> gx3v7, nearest neighbor and smoothed '
   eFold        = 1000000.0
   rMax         =  500000.0
   restrict_smooth_src_to_nn_dest = .true.
   step1 = .true.
   step2 = .true.
   step3 = .true.
  /

    Name this file runoff_map.nml

  3. Run

    ./runoff_map < runoff_map.nml

    Note that it may be necessary to have the same environment that you used for building (e.g., via sourcing src/.env_mach_specific.sh before running this executable).

  4. Run

    ./run_merge_mapping_files.sh \
--map_in_oo map_gland4km_epsg3413_to_gx3v7_nn.nc \
--map_in_ms map_gland4km_to_gx3v7_nnsm_e1000r500_171024.nc \
--region_mask /glade/p/cesmdata/cseg/inputdata/ocn/pop/gx3v7/grid/region_mask_20090831.ieeei4 \
--map_out map_gland4km_to_gx3v7_nn_open_ocean_nnsm_e1000r500_marginal_sea_171024.nc

  5. Repeat the above process with the gx1v6 grid, changing the input namelist appropriately.

    • For rMax, use 300000.0
    • As of 2017-10-24, for file_ocn_coastal_mask, use gx1v6_coast_170503.nc
    • As of 2017-10-24, for --region_mask, use /glade/p/cesmdata/cseg/inputdata/ocn/pop/gx1v6/grid/region_mask_20090205.ieeei4

  6. Repeat the above process with the gx1v7 grid, changing the input namelist appropriately.

    • For rMax, use 300000.0
    • As of 2017-10-24, for file_ocn_coastal_mask, use gx1v7_coast_170322.nc
    • As of 2017-10-24, for --region_mask, use /glade/p/cesmdata/cseg/inputdata/ocn/pop/gx1v7/grid/region_mask_20151008.ieeei4

  7. Run the check_maps tool on each of the resulting final mapping files (the files with the date stamp at the end)

    • To build and run this, follow the instructions in cime/tools/mapping/check_maps/README
    • Examine the output from this tool to make sure there are no major mapping errors
      • For these runoff mapping files, messages about “L1 error” and “L2 error” can be ignored.
      • Until https://github.com/ESMCI/cime/issues/2014 is resolved, you need to add dimensions to the merged files with something like ncap2 -s 'defdim("ni_a",416);defdim("nj_a",704);defdim("ni_b",320);defdim("nj_b",384)' YOUR_MAP_NAME.nc (where you can find the correct dimensions on the non-merged (i.e., nnsm) mapping files).
    • If you’d like, you can also visually examine the mapped fields. Open the file named test_YOUR_MAP_NAME.nc; a useful field to view is dst02, which is the result of mapping a uniform field (with value 2) from the glc grid to the ocn grid.

  8. Put mapping files in correct directories in the inputdata space

    The mapping files should go in $CESMDATAROOT/inputdata/cpl/gridmaps/RES where RES is your new CISM resolution.

    There are two final mapping files that need to be kept for each glc-ocn grid combination: The two files with the date stamp at the end.

A.1.4.3   Adding new grid and mapping files in config_grids.xml

In order for your new grid and mapping files to be recognized by the CESM scripts, you need to add entries in config_grids.xml (cime/cime_config/cesm/config_grids.xml).

  1. Add new grid definition

    You’ll need to add a section like this:

    <domain name="gland4">
  <nx>376</nx> <ny>701</ny>
  <desc>4-km Greenland grid, for use with the glissade dycore</desc>
</domain>

  2. Point to new mapping files: lnd <-> glc

    You’ll need to add a section like this for each land grid, in the section “lnd to glc and glc to lnd mapping”:

    <gridmap lnd_grid="0.9x1.25" glc_grid="gland4" >
  <map name="LND2GLC_FMAPNAME">cpl/gridmaps/fv0.9x1.25/map_fv0.9x1.25_TO_gland4km_aave.161223.nc</map>
  <map name="LND2GLC_SMAPNAME">cpl/gridmaps/fv0.9x1.25/map_fv0.9x1.25_TO_gland4km_blin.161223.nc</map>
  <map name="GLC2LND_FMAPNAME">cpl/gridmaps/gland4km/map_gland4km_TO_fv0.9x1.25_aave.161223.nc</map>
  <map name="GLC2LND_SMAPNAME">cpl/gridmaps/gland4km/map_gland4km_TO_fv0.9x1.25_aave.161223.nc</map>
</gridmap>

  3. Point to new mapping files: glc -> ocn

    In the section “GRIDS: glc to ocn mapping”, add a section like this:

    <gridmap ocn_grid="gx1v6" glc_grid="gland4" >
  <map name="GLC2OCN_LIQ_RMAPNAME">cpl/gridmaps/gland4km/map_gland4km_to_gx1v6_nn_open_ocean_nnsm_e1000r300_marginal_sea_171105.nc</map>
  <map name="GLC2OCN_ICE_RMAPNAME">cpl/gridmaps/gland4km/map_gland4km_to_gx1v6_nnsm_e1000r300_171105.nc</map>
</gridmap>
<gridmap ocn_grid="gx1v7" glc_grid="gland4" >
  <map name="GLC2OCN_LIQ_RMAPNAME">cpl/gridmaps/gland4km/map_gland4km_to_gx1v7_nn_open_ocean_nnsm_e1000r300_marginal_sea_171105.nc</map>
  <map name="GLC2OCN_ICE_RMAPNAME">cpl/gridmaps/gland4km/map_gland4km_to_gx1v7_nnsm_e1000r300_171105.nc</map>
</gridmap>
<!-- POP's estuary box model is currently not active for gx3v7, so
     we need nnsm maps for liquid as well as ice. -->
<gridmap ocn_grid="gx3v7" glc_grid="gland4" >
  <map name="GLC2OCN_LIQ_RMAPNAME">cpl/gridmaps/gland4km/map_gland4km_to_gx3v7_nnsm_e1000r500_171105.nc</map>
  <map name="GLC2OCN_ICE_RMAPNAME">cpl/gridmaps/gland4km/map_gland4km_to_gx3v7_nnsm_e1000r500_171105.nc</map>
</gridmap>
    • Important note for gx3v7 grid: The estuary box model is not active for the gx3v7 grid for now, so point to the nnsm file for both ice and liquid runoff. (However, to follow what’s done for the rof2ocn mapping files, you can still put the merged maps in the inputdata repository, so that they can be used if the estuary box model is ever activated for gx3v7.)