Soil C concentration

[Chr]

All,

I calculated Soil C concentration (%SOC) from CLM5 SoilC_vr and soil bulk density outputs as:

%SOC = SoilC_vr (g/m3) / soil_bulk_density (g/m3) * 100

The calculated values make sense, however, I noticed that bulk density and vertically resolved soil C are not connected to each other in the CLM5 code. In fact, soil bulk density appears to be soil bulk density of wet soil:

soil_bulkdensity(c,j) = bd(c,j) + h2osoi_liq(c,j)/col%dz(c,j)
bd = (1._r8-soilstate_inst%watsat_col(c,lev))*2.7e3_r8
soilstate_inst%watsat_col(c,lev) = 0.489_r8 - 0.00126_r8*sand

bd = bulk density of dry soil material [kg/m^3]
h2osoi_liq = liquid soil water (kg/m2)
watsat = volumetric soil water at saturation

Therefore my question, if this is a reasonable approach to acquire %SOC? And if not, if there is another way to calculate it.

Thanks a lot.

 

[wwieder]

Hi Christian,

I'm assuming you want to compare the soil C stocks in CLM to an observations that uses %C for some reason? I can think of a few options moving forward, neither of which may be very helpful...
1) estimate bulk density from soil properties. That's what we're doing in the bd calculation you cited, basically a function of porosity (1-watsat), which is just a function of soil texture (here sand).
2) pick a different pedotransfer function for bulk density and use the surface dataset (this is pretty similar to #1), but adds a challenge that many of these pedotransfer functions use %C too...
3) use a pedotransfer function to calculate the C stocks from the %C values for (basically make observations fit the model, instead of the model fitting the obs).

Does any of that help?

 

[Chr]

Will, your assumption is correct! I want to compare my simulated carbon density soilc_vr (g/m3) with observation data. I have both, bulk density (BD) and %SOC from field measurements, and carbon density as %SOC * BD.

My initial thought was that soilc_vr is connected to soil_bulkdensity, and was surprised that it is not.

My idea was to calculate CLM5 simulated %SOC as soilc_vr/soil_bulkdensity, and then see whether the simulation error originates from BD or %SOC. But this seems off the table.

I might pursue your first suggestion, maybe I can figure out a way to set the output variable soil_bulkdensity to bd insead of soil_bulkdensity.

 

[demibing]

May I ask how to use clm to estimate soil carbon stock? which ccomposet should I choose? Is there any references or documents？
Look forward to your reply
Thanks a lot !

 

[oleson]

Any compset with BGC or BGC-CROP on will give you soil carbon stocks. We recommend using the latest version of the model (CLM5.0).
The CLM User's Guide is here:

1. CLM5.0 User’s Guide — ctsm release-clm5.0 documentation

The CLM Technical Note is here:

2. CLM Technical Note — ctsm release-clm5.0 documentation

There is other CLM information here, including tutorials:

Land Component | Community Earth System Model

www.cesm.ucar.edu