soil organic matter fraction (om_frac) change from squared to unsquared

[gretam]

Is there a reason why CLM4.5 used a squared formulation for the soil organic matter fraction (om_frac), and then CLM5 onwards uses an unsquared formulation? Is the squared formulation in CLM4.5 just a bug, or was there a particular reason for squaring om_frac in CLM4.5?

In the current CLM, there is a flag to use the old CLM4.5 formulation:
CTSM/src/biogeophys/SoilStateInitTimeConstMod.F90 at 7731d83c7fdb4bb31a3d0b5fc6eead3ea7ff4b29 · ESCOMP/CTSM

L56: logical, private :: organic_frac_squared ! If organic fraction should be squared (as in CLM4.5)

L480: om_frac = min(params_inst%om_frac_sf*organic3d(g,j+1)/organic_max, 1._r8)

L497: if (organic_frac_squared) then
om_frac = om_frac**2._r8
end if

The original paper, Lawrence and Slater 2008, Equation 1 uses an unsquared formulation, and CLM documentation also shows an unsquared formulation. But I can't find any papers/documentation/forums on how or why the CLM4.5 squared formulation originated.

Thanks for your help!

 

[oleson]

Interesting. The CLM4.5 technical note, which was published July, 2013, shows it as being unsquared, which is consistent with Lawrence and Slater (2008).
And the CLM4 technical note, published in April, 2010, also shows it as being unsquared, again consistent with Lawrence and Slater (2008).
The namelist flag with the CLM4.5 squared version came in later, Nov 2014, with clm4_5_1_r096.
The CLM4.5 simulations documented in the Lawrence et al. (2019) CLM5 paper did use the squared formulation (I did those simulations and was able to look at the related code), so it's possible we decided that was better at some point but the CLM4 and CLM4.5 documentation had already been released though.
The CLM4 simulations documented in the Lawrence et al. (2019) CLM5 paper also used the squared formulation.
Unfortunately, I don't seem to have any record of why that change was made.