Consistency in fluxes and setting radiative properties of land

[kmdeck]

OKR2.3.8:

"Prepare for coupled runs: ensure that we follow the flux and unit conventions set by the coupler. Compute bulk albedo and temperature of surface, set emissivity to 1."

Units

• all energy fluxes are in J/m^2/s
• we store evaporation as a liquid water volume flux. The coupler converts the mass flux computed by atmos for us: https://github.com/CliMA/ClimaCoupler.jl/blob/59abc31b66ccb6d62a06fb86be9f0340fae4d4c0/experiments/ClimaEarth/components/land/climaland_bucket.jl#L211
• we store precipitation as a liquid water volume flux. The coupler converts a mass flux to volume flux for us right now: https://github.com/CliMA/ClimaCoupler.jl/blob/59abc31b66ccb6d62a06fb86be9f0340fae4d4c0/experiments/ClimaEarth/components/land/climaland_bucket.jl#L197
  Land to do: convert internally to volume fluxes
• What units will carbon fluxes be in? (mass flux of C)

Flux sign convention

We have followed the convention for fluxes that positive = towards atmos and negative = towards land. This applies for precipitation, SHF, LHF, evap, and everywhere in the land model, except radiation.

To do: change precip to be defined to be positive

Radiation internals for Land

For radiation, we have "LW_d" and "SW_d" in prescribed atmosphere runs. These are scalars, and indicate the magnitude of the downwelling radiation. They are always positive. We then compute "LW_u" and SW_u", which are the magnitude of the upwelling radiation, and hence are always positive.

To do: Change all instances of in/out to down/up.

2. Internally, we also compute the net radiation per land component "R_n = downwelling - upwelling". Hence this is positive if the land component is absorbing energy. Note that this is opposite the flux convention we use everywhere else, with negative indicating "towards land".

• ChosencConvention: R_n > 0 implies land absorbs energy. The net flux at the surface would then be SHF + LHF - R_n.

To do: make sure this convention is followed throughout

Coupled atmosphere runs - radiation:

• we will need SW_d in PAR and NIR wavelength bands currently, eventually we may have a more spectrally resolved model. Can the coupler carry out the conversion between the bands of the atmosphere model and of the land model?
• What sign convention will the downwelling SW fluxes follow? Answer: same as ours
• we will need LW_d. What sign convention will this follow? Answer: same as ours
  Atmos wants a broadband albedo, emissivity, and a single temperature (see below)
  Land will hand splitting SW into PAR and NIR by dividing by 2

Bulk surface properties

RRTMGP also requires the bulk properties of the land surface. We compute the net LW_u and SW_u (broadband) from all land components. there is therefore not a way to separate LW_u into emissivity and Temperature uniquely, so we set:
epsilon_sfc = 1 [note epsilon_soil and epsilon_canopy are not 1]
T_sfc = [LW_u/(epsilon_sfc*sigma)]^(1/4) [effective temperature, not the skin temperature]
alpha_sfc = SW_u/SW_d

To do: implement. To do this properly, we need to store the fraction reflected, and not this number multiplied by SW_d. This is because at night SW_d is zero, but the albedo should not be zero

Tasks

Beta Give feedback

1. Upwell/downwelling notation consistency Consistent naming for radiation, land properties for radiation #776
2. Bulk land properties needed by atmos Consistent naming for radiation, land properties for radiation #776
3. Fix albedo to not be zero at night
4. Sign convention for R_n Sign consistency with R_n #827
5. Precip and evaporation as mass fluxes
6. SIgn of precip
7. Documentation page Sign consistency with R_n #827