Issue Reading Slope Input when Running LIS with Noah-MP 4.0.1

[mcthreems]

I am attempting to run a simulation of Noah-MP 4.0.1 in LIS using the 2nd groundwater option, but the simulation returns an error early in the run process and stops there. The specific error message in the log file (along with the surrounding lines) is:

[INFO] Reading mask from ./lis_INPUT_noahmp_4.0.1_nldas2.d01.nc
[INFO] Saving global land mask...
[INFO] Reading landcover map from ./lis_INPUT_noahmp_4.0.1_nldas2.d01.nc
[INFO] Reading surfacetype map from ./lis_INPUT_noahmp_4.0.1_nldas2.d01.nc
[INFO] Finished reading landmask and landcover data
[INFO] Reading elevation map from ./lis_INPUT_noahmp_4.0.1_nldas2.d01.nc
[INFO] Reading slope map from ./lis_INPUT_noahmp_4.0.1_nldas2.d01.nc
[ERR] Error in nf90_inq_dimid in read_slope Stopping.
[ERR] endrun is being called

The input data processed via LDT are roughly similar to previous simulations we have run that were successful, however we had not previously included the slope data input. Running LDT did not produce an error so we had assumed the issue is with LIS rather than LDT. Is this error indicative of some sort of mismatch between the input data we used for the slope and the other settings we have? I have attached both the ldt and lis configuration files for reference.
ldt.config_noahmp401_nldas2.txt
lis.config_S2S_gw2_test.txt

[dmocko]

Hi @mcthreems

Yes, there is a mismatch between the input data in LDT and your lis.config settings.

Your lis.config file contains this line:
Topographic correction method (met forcing): "lapse-rate and slope-aspect"

This indicates that you are attempting to use both the "lapse-rate" correction of these forcing variables:

• Near-surface air temperature
• Near-surface specific humidity
• Longwave radiation downward at the surface
• Surface pressure

.....as well as "slope-aspect" correction of this variable:

• Shortwave radiation downward at the surface

To use the "slope-aspect" correction, you need to change your ldt.config file to read in these fields.
Note that the slope here is NOT the same as the "slope type" config entries in your ldt.config file.

As an aside, Noah-MP LSM does not need/use "slope type". You only need those LDT config entries for Noah LSM.
For Noah-MP LSM, you should only have this single config entry in your ldt.config file:
Slope type data source: none

To add the slope and aspect fields into your lis_input file, add these lines to your ldt.config file and re-run LDT:

Slope data source:                      "SRTM_Native"
Slope number of bands:                  1
Slope map:                              ./input/LS_PARAMETERS/topo_parms/SRTM/SRTM30/raw_wgtopo30antarc
Slope fill option:                      none
Slope fill radius:                      5
Slope fill value:                       0

Aspect data source:                     "SRTM_Native"
Aspect number of bands:                 1
Aspect map:                             ./input/LS_PARAMETERS/topo_parms/SRTM/SRTM30/raw_wgtopo30antarc
Aspect fill option:                     none
Aspect fill radius:                     5
Aspect fill value:                      3.14159

[mcthreems]

The standard output does return an mpi error (copied below). I am not on Discover, but rather a server owned by our research group. If it is unlikely that this issue can be diagnosed I can look into the possibility of creating a Discover account and running there.

"--------------------------------------------------------------------------
MPI_ABORT was invoked on rank 14 in communicator MPI COMMUNICATOR 3 CREATE FROM 0
with errorcode 1.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on
exactly when Open MPI kills them.
--------------------------------------------------------------------------"

[dmocko]

Hi @mcthreems

Thanks for your patience with a reply. Several LIS team members were on leave in July/August, we have lost some staff,
and we had important annual meetings this month.

Did you resolve this issue? If not, would you mind attaching your ldt.config, lis.config, and LISF/lis/make/configure.lis files?

[mcthreems]

Thanks for following up on this. We hadn't made any progress on this particular front, other than finding that the blank log file only occurred when trying multiple processors. When trying just one processor there was actual output to the log, which is attached here along with the ldt and lis configuration files
ldt.config_slope_test.txt
lis.config_slope_test.txt
slope_test.0000.txt
configure.lis.txt

[dmocko]

Hi @mcthreems

Thanks for sharing these files. I was able to identify the issue. The end of your lislog (slope_test.0000) file shows:

 [ERR] No surface model tile fraction >= minp,  0.100000001    
   which is the 'Minimum cutoff percentage (surface type tiles):'
   value set in your lis.config file. Highest tile fraction value is:   9.02777761E-02
   Thus, surface tiles set to '0' for gridpoint:
 c,r,fgrd:           32          59   0.00000000       0.00000000       0.00000000       0.00000000       0.00000000       0.00000000       0.00000000       0.00000000       0.00000000       0.00000000       0.00000000       0.00000000       0.00000000       0.00000000       0.00000000       0.00000000       0.00000000       0.00000000       0.00000000       0.00000000       0.00000000    
 [ERR] endrun is being called

This error is occuring because you are using 2 surface model types in your ldt.config (and thus in your lis_input file):

Number of surface model types:          2
Surface model types:                    "LSM" "Openwater"

Your original lis.config file is only using 1 surface model type, which is what's causing this error message.
Please change your lis.config file lines to:

Number of surface model types:               2
Surface model types:                         "LSM"  "Openwater"

[mcthreems]

Thank you for identifying this. I adjusted the config files and it has resolved that issue but encountered a new one. This time it appears to be due to a missing input file. I have attached the log output for this most recent attempt here.
log_out.txt

[dmocko]

Because you are trying to do a "lapse-rate" correction using LIS, you need to have the NLDAS2 elevation difference file read in LDT.

NLDAS-2 is a special case for the "lapse-rate" correction - because the correction is already included the surface forcing.

Thus, LIS will use the elevation difference file to back out the included correction, and then perform a "lapse-rate" correction
using the elevation on your LIS running grid.

Please add these two lines to your ldt.config file, run LDT again to generate a new lis_input file, and then run LIS:

NLDAS2 elevation difference map: ../NARR_elev-diff.1gd4r
NARR terrain height map:         ../NARR_elevation.1gd4r

You can obtain these two files from: https://portal.nccs.nasa.gov/lisdata_pub/data/PARAMETERS/NLDAS_0.125/

Note that these two files are big_endian, so please check that you configured/compiled LDT using big_endian.
You can look at your LISF/ldt/make/configure.ldt file to see which "endian"-ness you used.

[mcthreems]

Thank you for your help with this, I will make these changes and try rerunning.