Opened 7 years ago

Last modified 8 months ago

#350 assigned defect

Improper use of vbeta in enerbil

Reported by: aducharne Owned by: aducharne
Priority: major Milestone: Work on sft and multitiling
Component: Physical processes Version:
Keywords: Cc:


This ticket has links with #331 and #60.

In diffuco, 5 vbeta are defined for 5 evaporation fluxes (subli, inetrception loss, transpiration, soil evap, and flood plain evap). These vbeta account for both the "specific beta" of each flux (reduction factor compared to evapot, thus conveying information of the stresses), but they also account for the effect of the grid-cell fraction assigned to each flux on the grid-cell average (the larger the fraction, the larger the contribution of the subflux).

In doing so, it is assumed that each sub-flux originates from an independent fraction of the grid-cell: a particular fraction cannot contribute to two different sub-fluxes.

Two different problems are identified:

1) in diffuco_trans_co2, the fraction used for vbeta3 is veget_max*(un - zqsvegrap), which is not consistent with assuming that evapnu proceeds from (veget_max-veget). For comparaison, in diffuco_trans, the fraction used for vbeta3 is veget*(un - zqsvegrap).
=> The proposed solution is to weight vbeta3 in diffuco_trans_co2 like in diffuco_trans (preliminary tests over 10 yrs with WFDEIv1 show that it reduces the transpiration of 4% on global average, and the total ET of 1.5%)

2) in diffuco, the vbeta for transpir, interception loss, and evapnu, are calculated as if there was no snow nor floodplains (while the vbeta1 and vbeta5 for subli and vevapflo respectively depend on the fraction covered by snow and floodplains).

The reduction of transpir, interception loss, and evapnu, as a function of the fraction occupied by snow and floodplains, is performed in enerbil, with conceptual errors: basically, vbeta1 and vbeta5 are used correctly to define subli and vevapflo, but they are also used as the effective snow and floodplain fraction to calculate transpir, interception loss, and evapnu, and this is wrong. More specifically, I see a problem :

  • with this use of vbeta1 as an effective snow fraction if frac_snownobio is not 1,
  • with this use of vbeta5 as a floodplain fraction because evapot_penm/evapot is smaller than 1

I think this does not induce conservation problems, but just incorrect estimations of the fluxes. Yet, I may be wrong...

=> There are two ways to solve this:

  • either by fully accounting for the fractions in diffuco (implying to make vbeta2, vbeta3, and vbeta4 dependent on the snow and floodplain fractions); then, enerbil becomes very easy, since we just add the different vbeta_i * evapot
  • else, we can choose to correct the weighting scheme used in enerbil and replace some uses of vbeta1 and vbeta5 by the correct fractions.

Change History (9)

comment:1 Changed 7 years ago by jgipsl

  • Milestone set to IPSLCM6.v1

comment:2 Changed 6 years ago by jgipsl

  • Milestone changed from IPSLCM6.v1 to ORCHIDEE 3.0
  • Owner changed from somebody to maignan
  • Status changed from new to assigned

comment:3 Changed 6 years ago by jgipsl

  • Owner changed from maignan to aducharne

comment:4 Changed 6 years ago by aducharne

  • Component changed from Anthropogenic processes to Physical processes

comment:5 Changed 5 years ago by aducharne

  • Milestone changed from ORCHIDEE 3.0 to ORCHIDEE 4.0

comment:6 Changed 4 years ago by aducharne

Linked to ticket #60

comment:7 Changed 3 years ago by luyssaert

  • Milestone changed from ORCHIDEE 4.0 to Not scheduled yet

comment:8 Changed 3 years ago by aducharne

Might be worth to process with ticket #467

comment:9 Changed 8 months ago by jgipsl

  • Milestone changed from Not scheduled yet to sft_and_multitiling
Note: See TracTickets for help on using tickets.