Issue with Long-Term CLM5 Simulation Using Custom Input Data

lfu · Feb 21, 2025

Hi,

I ran a single-point, long-term time-series simulation in CLM5 using my own input data.

I made these changes in env_run.xml.

./xmlchange DATAM_MODE=CLM1PT
./xmlchange PTS_LAT=10.8785
./xmlchange PTS_LON=274.6594
./xmlchange STOP_OPTION=nyears
./xmlchange STOP_N=86
./xmlchange DATM_YR_ALIGN=2015
./xmlchange DATM_YR_START=2100
./xmlchange DATM_YR_END=2014
./xmlchange CLM_USRDAT_NAME=monthly370
./xmlchange CLM_USRDAT_DIR=/glade/derecho/scratch/lfu/ssp_atm_input/IPSL-CM6A-LR/CR_single_sites

Additionally, I added the following input paths in user_nl_datm_streams:

CLM_USRDAT.monthly585:year_first=2015
CLM_USRDAT.monthly585:year_last=2100
CLM_USRDAT.monthly585:year_align=2015
CLM_USRDAT.monthly585:dtlimit=1.5
CLM_USRDAT.monthly585:offset=0
CLM_USRDAT.monthly585:tintalgo=nearest
CLM_USRDAT.monthly585:mapalgo=nn
CLM_USRDAT.monthly585:taxmode=cycle

Issue Description

The model ran successfully. However, when I loaded the output NetCDF file, I noticed that the time coordinate in the dataset spans from 2000 to 2086 (as shown in the attached figure).

Additionally, the simulated temperature suddenly decreases sharply after 15 years from the start of the simulation (The input does not happen this situation).

I am unsure what might be causing this issue. Could you please help me identify the problem?

Thank you!

slevis · Mar 4, 2025

The first piece that seems wrong is this, but maybe I have misunderstood, and you have a reason for these settings:
./xmlchange DATM_YR_START=2100
./xmlchange DATM_YR_END=2014

To change the time coord. in your output, you may need to try this:
./xmlchange RUN_STARTDATE=2015-01-01
or whatever date you want the model start with instead of 2000-01-01.

You can type
./xmlquery RUN_STARTDATE
to see what's there before and after the change.

oleson · Mar 4, 2025

It seems like there was a duplicate post for this. I think this was resolved here: