An error in CLM5.0 mksurfdata process

[USER_M]

When I run ./mksurfdata.pl -r usrspec -hirespft -usr_gname 0.05x0.05 -usr_gdate 210328 to make surface map at 0.05 degree, there exist an error below
-------------------------------------
Attempting to make Topography statistics.....
domain_read_dims_1d read 1-d length from num_pixels
domain_read initialized domain
(gridmap_map_read) reading mapping matrix data...
(gridmap_map_read) * file name : ../mkmapdata/map_1km-
merge-10min_HYDRO1K-merge-nomask_to_0.05x0.05_nomask_aave_da
_c210328.nc
* matrix dimensions rows x cols : 125599447 x 25920000
* number of non-zero elements: 83811663
(gridmap_map_read) ERROR: frac_dst out of bounds
max = 1.0528475567504021 min = 0.0000000000000000
Program aborted. Backtrace:
#0 0x14fd535b8d01 in ???
#1 0x14fd535b978d in ???
#2 0x14fd5380cfac in ???
#3 0x56385c89e3b2 in ???
#4 0x56385c90e71a in ???
#5 0x56385c8f1249 in ???
#6 0x56385c8fa6bf in ???
#7 0x14fd5327e0b2 in ???
#8 0x56385c88d52d in ???
#9 0xffffffffffffffff in ???
Aborted (core dumped)
ERROR in mksurfdata_map: 34304
-----------------------------------------------------------
According to mkgridmapmod.F90 file, this error is caused by frac_b variable in the map_1km-merge-10min_HYDRO1K-merge-nomask.nc file.
Although I use ESMF 8.1.0 to build the UGRID file from SCRIPgrid_1km-merge-10min_HYDRO1K-merge-nomask_c130402.nc, the error still exist(this time frac_dst max≈1.05, if I use the UGRID file download from the CESM inputdata sever, frac_dst max≈19).
Could anyone help me solve this problem? I am deeply grateful for your help!

 

[erik]

Hello.

From our experience this problem usually points out a problem in your originating SCRIP grid file to define the grid you want to run on. The fraction should be less than or equal to 1.0, but sometimes it can be slightly larger than that and still be OK. So we let it to be a little larger. Here it's 1.05, so more than I'd expect it to be. And example of when this happens is if there are a gridcell (or grid cells plural) that have area covered by another grid-cell. Hence, it's good to go through you SCRIP grid file and make sure that's not the case. For a fine grid that's a bit of a pain to do, but for a regular grid like this you might be able to setup a script to loop through your gridcells and look for this type of problem.

 

[USER_M]

Hello.

From our experience this problem usually points out a problem in your originating SCRIP grid file to define the grid you want to run on. The fraction should be less than or equal to 1.0, but sometimes it can be slightly larger than that and still be OK. So we let it to be a little larger. Here it's 1.05, so more than I'd expect it to be. And example of when this happens is if there are a gridcell (or grid cells plural) that have area covered by another grid-cell. Hence, it's good to go through you SCRIP grid file and make sure that's not the case. For a fine grid that's a bit of a pain to do, but for a regular grid like this you might be able to setup a script to loop through your gridcells and look for this type of problem.

Is there any way to solve this problem at present？ Should I must use a lower resolution? I tried to use NCO to build the map file (map_1km-
merge-10min_HYDRO1K....nc) by bilinear method, the max frac_dst is 1, but the slope is more than 90, there still exist an error.