Is it possible to change grid resolution and eco-physiological parameter values for I2000clm50FatesCruRsGs?

[aatreyee_iirs]

Hello everyone,
I have recently started to work on CESM and so far I have been able to port CESM to my machine. My prime objective is to work on GPP using CLM compset I2000clm50FatesCruRsGs as I am interested in DGVM. However, I see that it is a defined compset but not a tested one. So I have a few questions-

1. I want to run regional case with 0.1x0.1 degrees resolution for this compset, would that be possible if I am able to prepare my own surface data and forcing data?
2. Is it possible to change eco-physiological parameter values e.g. canopy average specific leaf area, CN ratios etc if I have calibrated values for my region?

Thank you.

 

[slevis]

1) It should be possible.
2) Assuming that our model includes these parameters, then yes.

I want to make sure that by "DGVM" you mean CLM-FATES, because clm's "first generation dgvm" has not been tested for several years and is unlikely to work, unless you spend time fixing it.

 

[aatreyee_iirs]

Thank you for your reply.
Yes I mean the CLM-FATES. I have gone through the CLM5.0 technical note, I could find these parameters in different chapters e.g FATES or Carbon and Nitrogen allocation etc. I hope to be able to incorporate some of these calibrated values.

 

[aatreyee_iirs]

Hello everyone,
I have a rather fundamental question regarding my work. I am a beginner with CLM and I was wondering whether GPP can also be simulated using other compsets besides FATES, e.g. BGC or SP.
Currently I am trying to run I1850Clm50Sp as I came across under the tutorials and hoping to understand how it works. Would this also allow me to simulate GPP?

P.S. I am following the documentation regional grid simulations in order to understand the workflow for FATES.

Thank you.

 

[slevis]

I am sure that GPP will appear in the model's history output when you use a Bgc compset. Possibly not when you use an Sp compset. You can always try a 1-month simulation to confirm.

 

[aatreyee_iirs]

Thank you for your response, I will try this.

 

[aatreyee_iirs]

Hello everyone,
I am working on compset I2000Clm50FatesCruGs with resolution f09_g16. I could compile it successfully. But I am unsure as to how to spin-up the model and what kind of data would be required. I have followed some other threads in this forum but I could not understand. I would be grateful for any suggestions.
Thank you.

 

[slevis]

In the CLM5 Technical Note I found this section that seems a good starting point for your work:

2.28. Dynamic Global Vegetation — ctsm release-clm5.0 documentation

 

[aatreyee_iirs]

Thank you for your response, I am following this. I also went through Spinup of CLM5.0-BGC-Crop and tried to follow a similar path as the example for my compset. It involves two steps. For now I have done the following,

./create_newcase --case /home/ubuntuvm/CESM/cesm/scratch/FATES_spinup --res f09_g16 --compset I2000Clm50FatesCruGs --machine ubuntuvmfed --run-unsupported
./xmlchange CLM_ACCELERATED_SPINUP="on"
./case.setup
./case.build
./xmlchange RESUBMIT=3,STOP_N=50,STOP_OPTION=nyears,STOP_DATE=02010101
./case.submit

Currently I am running this, after which I have to run for 400+ years according to the example.
I would be grateful for suggestions on whether it is a correct procedure for FATES or it would require some changes.

Thank you.

 

[afoster]

Hi,

Generally with FATES we would just run the model for a period of time to allow the vegetation to come up to equilibrium. Depending on your complexity mode this could take a couple decades or a couple centuries of runtime. The accelerated spinup is unlikely to help much, since what you really need is to spin up the vegetation.

 

[aatreyee_iirs]

Thank you for your response.
So, instead of using ./xmlchange CLM_ACCELERATED_SPINUP="on", what should I go for? Should I just run it for 200 years and then 400+ years?

Thank you.

 

[aatreyee_iirs]

I want to simulate GPP from 2000 onwards, so if I want to spin up from 400 years ago, what data or conditions are used here?
I apologize I am still unclear about how the spin-up works.
I would be thankful for any suggestions.
Regards,

 

[slevis]

You would use CLM_ACCELERATED_SPINUP to spin up a Bgc case. For a Fates case @afoster says above that there is no special setting and that you run for as long as necessary for the vegetation to reach equilibrium.

If you're spinning up for 2000 onwards, a method often used (generally, not specifically with Fates) is to run a few decades or centuries for 1850 until you consider your output spun up. Then run a historical simulation from 1850 to 2000. Then use that as your initial state.

 

[aatreyee_iirs]

Thank you very much for your response. I will work on this.
Also, if I follow this, will the compset remain the same for the spinup?

Thank you

 

[oleson]

Generally, with the method @slevis suggested you would run an 1850 compset for the spinup and then an IHIST simulation from 1850 to 2000 and if you want to go beyond 2000, just extend the IHIST.

 

[aatreyee_iirs]

Thank you very much for your response.
Regards,

 

[aatreyee_iirs]

In case I want to run any of the following compsets

• 2000_DATM%CRUv7_CLM50%FATES_SICE_SOCN_MOSART_SGLC_SWAV
• 2000_DATM%GSWP3v1_CLM50%FATES_SICE_SOCN_MOSART_CISM2%NOEVOLVE_SWAV

Can I use the following for the 1850 compset and the IHIST compset?

• 1850_DATM%GSWP3v1_CLM50%BGC_SICE_SOCN_MOSART_CISM2%NOEVOLVE_SWAV
• HIST_DATM%GSWP3v1_CLM50%BGC_SICE_SOCN_MOSART_CISM2%NOEVOLVE_SWAV

 

[oleson]

You could try combining FATES with those 1850 and HIST compsets using the long name approach, e.g.,

1850_DATM%GSWP3v1_CLM50%FATES_SICE_SOCN_MOSART_SGLC_SWAV

although actually I'm not sure they will work.
The supported compsets that include FATES in the latest version of the model (release-cesm2.2.2) seem to be all I2000 compsets.

Alternatively, you can use one of the I2000 compsets and spinup using that and then use the same compset to simulate GPP. It will be for circa-year-2000 however, there won't be any transient aspects to the simulation.
@afoster , are you able to provide any guidance regarding the suitability of FATES for 1850 and HIST compsets?

 

[aatreyee_iirs]

You could try combining FATES with those 1850 and HIST compsets using the long name approach, e.g.,

1850_DATM%GSWP3v1_CLM50%FATES_SICE_SOCN_MOSART_SGLC_SWAV

although actually I'm not sure they will work.
The supported compsets that include FATES in the latest version of the model (release-cesm2.2.2) seem to be all I2000 compsets.

Alternatively, you can use one of the I2000 compsets and spinup using that and then use the same compset to simulate GPP. It will be for circa-year-2000 however, there won't be any transient aspects to the simulation.
@afoster , are you able to provide any guidance regarding the suitability of FATES for 1850 and HIST compsets?

So if I am running 2000_DATM%CRUv7_CLM50%FATES_SICE_SOCN_MOSART_SGLC_SWAV, I can use the same compset for spinup?
In that case, I wanted to confirm whether I'll have to start from a cold start and then use the file as the finidat file for the GPP simulation.

Thank you.

 

[oleson]

Yes.

When I do this:

./create_newcase --case clm50_cesm222_1deg_CRUv7_FATESspinup_2000 --compset 2000_DATM%CRUv7_CLM50%FATES_SICE_SOCN_MOSART_SGLC_SWAV --res f09_g17 --run-unsupported

finidat is blank, so yes you'll have to start from a cold start.