Questions about irrigation from groundwater

[liliyao]

Hi, I was trying to figure out how to use groundwater for irrigation in CLM5BGC. This question is related to this post: Questions about increasing the irrigation amount?. But I still only have irrigation water from surface water and not any from groundwater even though I have added the following to user_nl_clm:
irrigate = .true.
limit_irrigation_if_rof_enabled= .true.
use_groundwater_irrigation= .true.

I ran my test case on a point scale with ctsm5.1.dev115 for 2 simulation years, and the compset is “I2000_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_MOSART_SGLC_SWAV”. I have attached the surface dataset and domain file.

I was thinking maybe it is because the river volume already supplies enough water to meet the soil water deficit. So I tried to change the “irrig_river_volume_threshold”. Based on my reading of this code: src/biogeophys/IrrigationMod.F90, a threshold of 0.1 means allow 90% of the river volume to be used. So, if I use a threshold of 1.0, I thought it means no river water will be used for irrigation. But my simulation results show that this threshold does not have any impact on the irrigation in my case. I am also curious to know whether I can use only groundwater for irrigation even though the river water is available.

Basically, my questions are:

• How to set the model correctly to allow irrigation water from groundwater?
• Can I use “irrig_river_volume_threshold” to control the irrigation water from river volume?
• Can I control the ratio of the irrigation water sources between surface water and groundwater?

Thank you very much!

 

[oleson]

I think the problem is that those settings are not active/used when the river routing model is off, as it is in your single-point simulation.
See the following code in biogeophys/IrrigationMod.F90:

limit_irrigation = (this%params%limit_irrigation_if_rof_enabled .and. rof_prognostic)
if (limit_irrigation) then
call this%CalcDeficitVolrLimited( &
bounds = bounds, &
check_for_irrig_col_filter = check_for_irrig_col_filter, &
deficit = deficit(bounds%begc:bounds%endc), &
volr = volr(bounds%begg:bounds%endg), &
deficit_volr_limited = deficit_volr_limited(bounds%begc:bounds%endc))
else
deficit_volr_limited(bounds%begc:bounds%endc) = deficit(bounds%begc:bounds%endc)
end if

In particular, although limit_irrigation_if_rof_enabled is true, rof_prognostic is false.

 

[liliyao]

Thank you, Dr. Oleson. Since the river routing model was not used in the single-point simulation, I tried to run the simulation at CONUS scale (still with ctsm5.1.dev115). But I was not able to get past an error “too much NH4 uptake predicted by FUN”. I found this error has been discussed here: ERROR: too much NH4 uptake predicted by FUN In this post, you mentioned that (1) the atmospheric forcing might be the problem, (2) try with different initial conditions. I think the atmospheric forcing is not the problem in my case, because when I switched to CESM2.1.3, the same case can run successfully. As for the initial condition (i.e., finidat), the default one used by the model is finidat_1, and I tried another two:
finidat_1: '/lnd/clm2/initdata_map/clmi.I2000Clm50BgcCrop.2011-01-01.1.9x2.5_gx1v7_gl4_simyr2000_c190312.nc'.
finidat_2: '/lnd/clm2/initdata_map/clmi.I2000Clm50BgcCrop.2000-01-01.1.9x2.5_gx1v7_gl4_simyr2000_c200428.nc'
finidat_3: '/lnd/clm2/initdata_map/clmi.I2000Clm50BgcCrop.2011-01-01.1.9x2.5_gx1v7_gl4_simyr2000_c180715.nc'.
The default paramfile used by the model is '/lnd/clm2/paramdata/clm50_params.c211112.nc'. I have also tried with a different one, i.e., lnd/clm2/paramdata/ctsm51_params.c211112.nc.
But I did not get past the error. Do you have any suggestions to fix this error? I have attached my log files. Thank you very much.

 

[oleson]

There is also a canopy energy balance error in the log file. The magnitude of that error (e.g., energy balance in canopy 134 , err= -3.141599526680122E+028) suggests that either there is probably a missing or fill value in the forcing file being used by the datm. I would check the mask in the mesh file to make sure it is consistent with valid forcing data.

 

[liliyao]

Thanks for your reply. It is highly possible that the mesh caused the problem because the mesh I was using is actually not generated from my forcing file. I will start learning how to make ESMF mesh and to see if I could fix the error.

 

[liliyao]

Hi Dr. Oleson, I created a CONUS scale test case in release-cesm2.2.0 to test how to activate irrigation using groundwater. The compset is HIST_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_MOSART_SGLC_SWAV_SIAC_SESP. I also have added the following to user_nl_clm

limit_irrigation_if_rof_enabled= .true.
use_groundwater_irrigation= .true.

I ran the case for two simulation years. But QIRRIG_FROM_GW_CONFINED and QIRRIG_FROM_GW_UNCONFINED are still 0 in the output. Do you have any idea on this? Thank you!

 

[oleson]

Despite specifying MOSART in the compset, I don't think mosart will be active by default in a regional case because you would have to create mapping/mesh files and frivinp_rtm file for your region. You can check for the presence/absence of rof log files to see if mosart is active/inactive. Also, ./xmlquery MOSART_MODE should tell you that.

 

[liliyao]

Despite specifying MOSART in the compset, I don't think mosart will be active by default in a regional case because you would have to create mapping/mesh files and frivinp_rtm file for your region. You can check for the presence/absence of rof log files to see if mosart is active/inactive. Also, ./xmlquery MOSART_MODE should tell you that.

Thanks! ./xmlquery MOSART_MODE tells me MOSART_MODE: NULL. Could you please let me know what mapping/mesh files you are referring to? Does NCAR have a sample one? For the frivinp_rtm file, I guess I can use this one https://svn-ccsm-inputdata.cgd.ucar.../mosart/MOSART_routing_0.125nldas2_c190415.nc ?

 

[oleson]

In order to answer your questions, I'll need to reproduce the steps you've taken to this point.
From your previous posts I gather you are using release-cesm2.2.0 with an IHIST compset at 0.125nldas2 resolution? It may be that we provide the necessary files to run MOSART at that spatial resolution. But,

I tried setting that case up using:

./create_newcase --case cesm220_nldas2_IHISTClm50BgcCrop --compset HIST_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_MOSART_SGLC_SWAV_SIAC_SESP --res nldas2_rnldas2_mnldas2 --project P08010000 --run-unsupported

Then I ran ./case.setup.
./xmlquery MOSART_MODE gives me "ACTIVE" (not "NULL" as in your case).

However, running ./preview_namelists gave me this error:

ERROR: Command /glade/work/oleson/release-cesm2.2.0/components/clm/bld/build-namelist failed rc=255
out=
err=ERROR : CLM build-namelist::CLMBuildNamelist::add_default() : No default value found for flanduse_timeseries.
Are defaults provided for this resolution and land mask?

And so the namelists are not generated. So I'm wondering what the exact steps you've taken are.

 

[liliyao]

Thanks! Yes, I am using release-cesm2.2.0 with an IHIST compset at 0.125nldas2 resolution. But I use "CLM_USRDAT" resolution when I created the case. The defaults don't provide the flanduse_timeseries and land domain file (I am assuming the land mask you are referring to is the land domain file), but I provided them in user_nl_clm and env_run.xml, respectively, sorry for the confusion. I have attached the steps that I have taken.

 

[oleson]

Ok thanks. I think you are on the right track, although I'm not sure why mosart is getting turned off in your particular case. I tried a slightly different approach in that I used the nldas2_rnldas2_mnldas2 resolution instead of CLM_USRDAT (I also used ctsm5.1.dev115 but I assume things would work similarly in cesm220).

./create_newcase --case ctsm51d115_nldas2_IHISTClm50BgcCrop --compset HIST_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_MOSART_SGLC_SWAV --res nldas2_rnldas2_mnldas2 --project P08010000 --run-unsupported

That results in some runoff model specific variables being set appropriately, e.g.,

./xmlquery MOSART_MODE
MOSART_MODE: ACTIVE

env_build.xml: <entry id="ROF_GRID" value="0.125nldas2">
env_build.xml: <entry id="ROF_NX" value="464">
env_build.xml: <entry id="ROF_NY" value="224">
env_run.xml: <entry id="ROF_DOMAIN_MESH" value="$DIN_LOC_ROOT/share/meshes/0.125nldas2_ESMFmesh_cd5_241220.nc">

frivinp_rtm = "/glade/p/cesmdata/cseg/inputdata/rof/mosart/MOSART_routing_0.125nldas2_cdf5_c200727.nc"

So it looks here as if mosart is supported for this particular regional grid.
I then added in fsurdat and flanduse_timeseries files and:

limit_irrigation_if_rof_enabled=.true.
use_groundwater_irrigation=.true.

The case built and is running. I'm getting non-zero irrigation from groundwater.

 

[liliyao]

When I use 0.125nldas2 resolution instead of CLM_USRDAT in CESM2.2.0, I also got the non-zero irrigation from groundwater. Thanks!