About LULCC

[MikeYuan]

Hello! If I want to explore the global GPP responses to Land Use and Land Cover Change (LULCC) for a period such as 1980-2010, should I choose the “IHIST-” compset? Then, to set up the comparative experiment of LULCC vs. no LULCC, how should I run the simulations for 'no LULCC'? Should I simply set 'flanduse_timeseries' to a blank string in 'user_nl_clm'：flanduse_timeseries = ' '? Or should I use the 'I2000-' compset, which doesn't have the 'flanduse_timeseries' file?

Looking forward to your reply~

Thank you very much
Mike

 

[oleson]

Yes, IHIST is a good choice. Yes, you can set flanduse_timeseries to blank.

 

[MikeYuan]

Thank you very much, Mr. Oleson, for your response.
By the way, I wonder: if I use IHIST- to simulate LULCC for a period between 1980 and 2010, can I specify the run to start in 1980 and simulate 30 years, or is it necessary to first simulate from 1850 to 1980 and then continue for 30 years? Starting from 1850 may consume a significant amount of computational power for us

Looking forward to your reply~

Thank you very much
Mike

Yes, IHIST is a good choice. Yes, you can set flanduse_timeseries to blank.

 

[oleson]

You could do that. You would need some initial conditions that are appropriate for 1980. There should be some initial conditions in inputdata/lnd/clm2/initdata_map that could work. For example, there is a 1deg initial condition for 1979 available:

inputdata/lnd/clm2/initdata_map/clmi.BHIST.2000-01-01.0.9x1.25_gx1v7_simyr1979_c200806.nc

This is from a CESM2 historical run. So, you could start the model at 1979 with this initial condition.

 

[MikeYuan]

Thanks for your help, Mr.Oleson. I set the "clmi.BHIST.2000-01-01.0.9x1.25_gx1v7_simyr1979_c200806.nc" as the finidat(compset is IHistClm50BgcCrop), but when I run the model, there occurred the following errors:
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
.....................
ENDRUN:
ERROR: BandDiagonal ERROR: dgbsv returned error code
ERROR: BandDiagonal ERROR: dgbsv returned error code
index: 39149
n,kl,ku,m 27 2 2 7
dgbsv info: 39149 2

ab matrix
1 0.0000000 0.0000000 NaN 0.0000000 NaN
2 0.0000000 0.0000000****************** 0.0000000 0.0000000
3 NaN****************** NaN -0.5213628 0.0000000
4 0.0000000 -1.2358723 1.9846611 -0.2957218 0.0000000
..................

Stack trace terminated abnormally.
*** stack smashing detected ***: /data3/home/tfyuan/clm_practice_1/cesm2.1.3/output/IHistClmBgcCrop50BgcCrop_testn2/bld/cesm.exe terminated
forrtl: severe (174): SIGSEGV, segmentation fault occurred
forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image PC Routine Line Source
cesm.exe 0000000001525391 Unknown Unknown Unknown
cesm.exe 00000000015234CB Unknown Unknown Unknown

Stack trace terminated abnormally.
Image PC Routine Line Source
cesm.exe 0000000001440B1D Unknown Unknown Unknown
cesm.exe 0000000001108642 shr_abort_mod_mp_ 114 shr_abort_mod.F90
cesm.exe 000000000050052F abortutils_mp_end 50 abortutils.F90
cesm.exe 0000000000BB055D banddiagonalmod_m 213 BandDiagonalMod.F90
cesm.exe 0000000000A0610D soiltemperaturemo 408 SoilTemperatureMod.F90
cesm.exe 00000000005068E9 clm_driver_mp_clm 720 clm_driver.F90
cesm.exe 00000000004F561E lnd_comp_mct_mp_l 456 lnd_comp_mct.F90
cesm.exe 0000000000430A54 component_mod_mp_ 728 component_mod.F90
cesm.exe 0000000000416209 cime_comp_mod_mp_ 2720 cime_comp_mod.F90
cesm.exe 00000000004306F7 MAIN__ 125 cime_driver.F90
cesm.exe 0000000000413EFE Unknown Unknown Unknown
libc-2.17.so 0000003DA5E1F84D __libc_start_main Unknown Unknown
cesm.exe 0000000000413E29 Unknown Unknown Unknown
...................
ENDRUN:
ERROR: BandDiagonal ERROR: dgbsv returned error code
Image PC Routine Line Source
cesm.exe 0000000001440B1D Unknown Unknown Unknown
cesm.exe 0000000001108642 shr_abort_mod_mp_ 114 shr_abort_mod.F90
cesm.exe 000000000050052F abortutils_mp_end 50 abortutils.F90
cesm.exe 0000000000BB055D banddiagonalmod_m 213 BandDiagonalMod.F90
cesm.exe 0000000000A0610D soiltemperaturemo 408 SoilTemperatureMod.F90
cesm.exe 00000000005068E9 clm_driver_mp_clm 720 clm_driver.F90
cesm.exe 00000000004F561E lnd_comp_mct_mp_l 456 lnd_comp_mct.F90
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
The related files are attached below, could you help me check it?
Thank you very much
Mike

You could do that. You would need some initial conditions that are appropriate for 1980. There should be some initial conditions in inputdata/lnd/clm2/initdata_map that could work. For example, there is a 1deg initial condition for 1979 available:

inputdata/lnd/clm2/initdata_map/clmi.BHIST.2000-01-01.0.9x1.25_gx1v7_simyr1979_c200806.nc

This is from a CESM2 historical run. So, you could start the model at 1979 with this initial condition.

 

[slevis]

@MikeYuan have you run the IHist case with no changes to make sure that it works for you? If that doesn't work, I would try an even simpler I1850 or I2000 case. This way we can narrow down whether the problem is with the specific case that you tried or with any of these cases.

 

[MikeYuan]

@MikeYuan have you run the IHist case with no changes to make sure that it works for you? If that doesn't work, I would try an even simpler I1850 or I2000 case. This way we can narrow down whether the problem is with the specific case that you tried or with any of these cases.

Thanks for your reply, Mr. Slevis.
Yeah, I've tried the IHist case with no changes, and it worked normally. However, if I have to choose the I1850 or I2000, can I still simulate the LULCC from 1980 to 2010?It seems that the 'flanduse_timeseries' file may not be available when using the transient case.

Thank you very much
Mike

 

[slevis]

The IHist compset will automatically include the flanduse_timeseries and other transient simulation features, so it's easier to use it.

Since the simple IHist case with no changes worked, could you now list in detail all the changes from the simulation that did not work?

 

[MikeYuan]

The IHist compset will automatically include the flanduse_timeseries and other transient simulation features, so it's easier to use it.

Since the simple IHist case with no changes worked, could you now list in detail all the changes from the simulation that did not work?

Thanks for your reply, Mr. Slevis.
The only changes I made are in the 'use_nl_clm' file mentioned above; no other modifications were made. When I removed 'clmi.BHIST.2000-01-01.0.9x1.25_gx1v7_simyr1979_c200806.nc' as the finidat, the model ran successfully.

Thank you very much
Mike

 

[slevis]

Ok, that's a good starting point. I think this means that, if you need a finidat that differs from the one provided as default, then you will need to generate it yourself.

 

[MikeYuan]

Ok, that's a good starting point. I think this means that, if you need a finidat that differs from the one provided as default, then you will need to generate it yourself.

It means I have to generate the finidat by starting the case from 1850 to 1979? It may consume a significant amount of computational power for us

 

[slevis]

Since the 1979 file that @oleson recommended failed, maybe you wish to try another 1979 file in
/glade/p/cesmdata/cseg/inputdata/lnd/clm2/initdata_map

If none of them work, then you may be stuck doing your own spinup...

 

[MikeYuan]

You could do that. You would need some initial conditions that are appropriate for 1980. There should be some initial conditions in inputdata/lnd/clm2/initdata_map that could work. For example, there is a 1deg initial condition for 1979 available:

inputdata/lnd/clm2/initdata_map/clmi.BHIST.2000-01-01.0.9x1.25_gx1v7_simyr1979_c200806.nc

This is from a CESM2 historical run. So, you could start the model at 1979 with this initial condition.

Mr. Oleson,
By the way, I have a question about how to distinguish the effects of the land use and land cover changes in CLM5(offline model). I know the LUH2 file represents the anthropogenic land use, but how can we define the effect of land cover?

Thank you very much
Mike

 

[slevis]

I'm not certain that I understand the question, but if you mean, how does one isolate the effect of land cover change from other changes, then one possible way is to compare a simulation with land cover change to a simulation without land cover change.

 

[MikeYuan]

I'm not certain that I understand the question, but if you mean, how does one isolate the effect of land cover change from other changes, then one possible way is to compare a simulation with land cover change to a simulation without land cover change.

Thanks for your reply, Mr.slevis.
I mean how do we differentiate the impacts of anthropogenic and natural sources on land cover change? I know that LUH2 represents changes caused by human activities.

 

[slevis]

That's an interesting question. I think you're asking how might we simulate the effects of a boreal forest advancing northward, for example, right? I don't think that our landuse datasets can do that right now. You would have to modify them to answer your specific science question.