fix to help numerical stability of the soil hydrology particularly relevant for soils with low bulk density

[eleanorgb]

I propose we add this under the jules_soil namelist as a fix for stability when have soils of very low bulk density. This helps along with shorter timestep.

Remi has a the fix in version 7.9. Noah developed the fix. Im not sure Noah is yet properly connected to the Met Office git hub.

https://github.com/eleanorgb/jules_wiemip/tree/fix_Noah_l_soilsat_down

[caroduro]

Is the soilsat_method variable will replace the l_soil_sat_down logical?

[noahdsmith]

Yes!

[caroduro]

OK.
Is there any equivalence between l_soil_sat_down=true,false with soilsat_method 1,2,3?
Can we said that:
l_soil_sat_down=T corresponds to soilsat_method=2?
l_soil_sat_down=F corresponds to soilsat_method=1?

[noahdsmith]

Yes, that's right:

soilsat_method = 1
      ! Method for handling supersaturated and negative soil layers
      ! 1: Adjust flux from layer above (old l_soilsat_down=.FALSE.)
      ! 2: Adjust flux to layer below   (old l_soilsat_down=.TRUE.)
      ! 3: Limit fluxes based on direction of flow

[caroduro]

Which output variables will be affected by this development? Please let me know which other varaibles to look at.
I did a quick comparison of smc_avail_top for the 3 options of soilsat_method for the LBA_BAN (Brazil) site, found no big differences between 1 and 2, but 3.

soilsat_method=1: LBA_BAN_JULES_vn8.1-presc0_smc_avail_top_page1.pdf

soilsat_method=2: LBA_BAN_JULES_vn8.1-presc0_smc_avail_top_page1.pdf

soilsat_method=3: LBA_BAN_JULES_vn8.1-presc0_smc_avail_top_page1.pdf

Thanks

[noahdsmith]

It affects the moisture fluxes between layers, which will affect soil moisture (sthu and sthf).

I've added an output variable for w_flux. It should mostly have an effect in permafrost environments on freeze-up really, so sort of surprising that it's changing things so much in Brazil.

I'll have a look too - I was going to test whether using an EPSILON condition works instead of the equality_tolarance currently in there.