How to fix a certain type of PFT in CLM5.0

[JIAA]

Hello everyone! Thank you for seeing my questions.
My case is IHistClm50SpCru (HIST_DATM%CRUv7_CLM50%SP_SICE_SOCN_MOSART_CISM2%NOEVOLVE_SWAV).
I want to research how the land cover type influnce the soil temperature, so I want to fix one type of PFT and compare it with default test. But the PFT correlates with surface albedo, plant heights and so on, is it feasible for me to change a certain PFT only in fsurdat or flanduse_timeseries?
Thank all of you !

 

[slevis]

In the model's tools directory there is a fsurdat_modifier tool that allows you to modify fsurdat files (not flanduse_timeseries). If you are working with an older version of the model without access to this tool, then you should try to modify the fsurdat (and flanduse) files yourself. Yes, it's feasible.

 

[JIAA]

In the model's tools directory there is a fsurdat_modifier tool that allows you to modify fsurdat files (not flanduse_timeseries). If you are working with an older version of the model without access to this tool, then you should try to modify the fsurdat (and flanduse) files yourself. Yes, it's feasible.

Thanks for your reply! Do you mean the tool in clm compont? (my_cesm_sandbox/components/clm/tools/mksurfdata_map ?) If I change the fsurdata in user_nl_clm, do I still need to modify finidat, or flanduse_timeseries? Thank you sir!

 

[oleson]

I believe @slevis is referring to:

tools/modify_input_files/fsurdat_modifier

You don't necessarily need to change finidat if you change fsurdat. The model will interpolate the finidat to the new surface dataset if you set use_init_interp=.true. in your user_nl_clm.

 

[JIAA]

I believe @slevis is referring to:

tools/modify_input_files/fsurdat_modifier

You don't necessarily need to change finidat if you change fsurdat. The model will interpolate the finidat to the new surface dataset if you set use_init_interp=.true. in your user_nl_clm.

Thanks for your reply,sir! My cesm version is 2.2.0, but I still cannot find the fsurdat_modifier, maybe the tool is not available in this version?
My case is the IHistClm50SpCru, and in this case the LULCC changes over time.
The fsurdat is surfdata_1.9x2.5_hist_16pfts_Irrig_CMIP6_simyr1850_c190304.nc,which is a climatic state data and only 12 months. I'm guessing it's the flanduse_timeseries at work to make the land cover change over time？
Can I directly change the PCT_NAT_PFT in flanduse_timeseries to do sensitivity tests to explore the role of the PFT？By the way, in the process of modifying, I will ensure the sum is 100%.
Thank you sir!

 

[JIAA]

Or can I directly change PCT_NAT_PFT in fsurdat? But there are some variables like albedo,MONTHLY_HEIGHT_BOT etc., which are closely related to the PFTs.

 

[slevis]

For the version that you're working with:
- You will have to modify the netcdf files directly, as you said, keeping things consistent, so as to not trigger errors in the code.
- Alternatively, you could become familiar with the mksurfdata_map tool that generates fsurdat and landuse files from "raw" datasets:

1.3.5. Creating Surface Datasets — ctsm release-clm5.0 documentation

In this case you would modify the raw datasets and have the mksurfdata_map tool generate the fsurdat and landuse files for you.

The first option seems simpler but also more prone to error until you make everything consistent.

 

[xiaoxiaokuishu]

Or can I directly change PCT_NAT_PFT in fsurdat? But there are some variables like albedo,MONTHLY_HEIGHT_BOT etc., which are closely related to the PFTs.

Hi,
I am also have the question, can we only need to change the PCT_NAT_PFT , e.g. change forest to grass, i am not sure if we need to revise the albedo, LAI or other variables? Have you already solve it?

 

[slevis]

You may not need to change anything else. You could start with a simple change to PCT_NAT_PFT as a test.

After that you should look at other fsurdat variables, e.g. the height and lai/sai variables. Depending on the regions where you change forest to grass, you may need to introduce height values and lai/sai values for the grass.

 

[xiaoxiaokuishu]

You may not need to change anything else. You could start with a simple change to PCT_NAT_PFT as a test.

After that you should look at other fsurdat variables, e.g. the height and lai/sai variables. Depending on the regions where you change forest to grass, you may need to introduce height values and lai/sai values for the grass.

Yes, I only change the percent in PCT_NAT_PFT, i did not change the LAI value for grass, as it already predesribed for each PFT already,? i think.
After re-run the case after the deforestion, the ouput of LAI of my changed region has significantly decreased (because the land cover has changed from forest to grass). So it is not cesseary to revise the LAI in surfdata?

 

[oleson]

Yes, in SP mode, in most cases, LAI should already be prescribed for each PFT. There may be exceptions. For example, prescribing a pft type outside the region in which it normally exists. E.g., prescribing a boreal forest type in a tropical region. In that case, LAI may not be available. As @slevis suggested, you should look at fsurdat.

 

[xiaoxiaokuishu]

Yes, in SP mode, in most cases, LAI should already be prescribed for each PFT. There may be exceptions. For example, prescribing a pft type outside the region in which it normally exists. E.g., prescribing a boreal forest type in a tropical region. In that case, LAI may not be available. As @slevis suggested, you should look at fsurdat.

Hi, Oleson
Let me try to understand your reply. If my research is Amazon deforestion, the common land cover in this region is tropical forest, bare land, grass (only 3, i just take an example), if i deforested the forest to crop, the LAI is not avaliable, because the crop is not the common land cover in this region, if i deforest the forest to grass, it works well as grass is the common land cover in Amazon?

 

[oleson]

That is another example, yes. But you can verify by looking at fsurdat.