Opened 22 months ago

Last modified 3 months ago

#692 assigned defect

flux can be computed with a surface temperature different from the actual temperature in case of snow

Reported by: fcheruy Owned by: cottle
Priority: major Milestone: Not scheduled yet
Component: Physical processes Version: orchidee_2_0
Keywords: snow, energy conservation Cc:

Description

temp_sol_new is calculated in enerbil_surftemp, and the fluxes are calculated consistently with this temperature.

at the end of thermosoil (so after enerbil) one finds the following lines:

!! Surface temperature is forced to zero celcius if its value is larger than melting point, only for explicit snow scheme
IF (ok_explicitsnow) THEN

IF (SUM(snowdz(ji,:)) .GT. 0.0) THEN

IF (temp_sol_new(ji) .GE. tp_00) THEN

temp_sol_new(ji) = tp_00

ENDIF

END I

The energy corresponding to this later change of temperature is used to melt the snow (in hydrol). We just saw that now. However it remains that the surface temperature is modified (in thermosoil) after the computation of the fluxes (in enerbil) in the same time-step. In other words it seems that the flux are computed with a surface temperature which is different from the actual temperature.

Here is the description of the experiment we did (just for info)
The graph attached shows the seasonal cycle of the differences of the global mean over continent of the surface temperature between a control experiment and an experiment where the surface temperature is prescribed with the climatological value at each time step.
The differences are not zero (as wee expected for a climatology). They are significantly reduced when the above patch is commented. Part of differences might be due to the fact that the snow is not perfectly consistent with the prescribed temperature.

Attachments (1)

tsol_glo_ave_output_minus_input_noSummer.jpeg (150.7 KB) - added by fcheruy 22 months ago.

Download all attachments as: .zip

Change History (4)

comment:1 Changed 22 months ago by fcheruy

The other graph is the map of the differences. The runs are 10 years long.

comment:2 Changed 22 months ago by luyssaert

  • Milestone set to Not scheduled yet
  • Version set to orchidee_2_0

comment:3 Changed 3 months ago by jgipsl

  • Owner set to cottle
  • Status changed from new to assigned
Note: See TracTickets for help on using tickets.