Incoherence in qsatcalc when temperature is too high

[maignan]

Sujet : 	Re: [orchidee-help] soil temperature exceed threshold (of 1M K) in grid w/o vegetation: model stops
Date : 	Thu, 24 Aug 2017 11:38:12 +0200
De : 	Fabienne Maignan <maignan@lsce.ipsl.fr>
Pour : 	Daniel Goll <dsgoll123@gmail.com>, orchidee-help@listes.ipsl.fr


Hello,

Le 11/08/2017 à 12:03, Daniel Goll a écrit :
> Hi
>
> my simulation stop because the soil temperature in a gridbox without vegetation growing since 1yr of the simulation exceeds the temperature of the core of the sun (1M K; threshold 'missing insrc_sechiba/enerbil.f90) after several hundred of years. (The gridbox is a coastal point (where there is actually ocean in real life); so no vegetation is okay.)
>
> 1. Are you serious about the threshold? What is the rational behind this threshold? 
The logic is the following one:
In qsat_moisture.f90, the temperature is tested against the threshold max_temp defined in the module as 370K.

In the old version (subroutine qsat_old) if the temperature exceeded the threshold and the flag diag_qsat was activated (TRUE is the default value), the return value for qsat was set to 999999.
This mechanism has disappeared in the new version (qsatcalc).

In enerbil_begin, qsatcalc is called, if the flag diag_qsat is activated qsol_sat is tested against missing = 999998 and then stops if qsol_sat is higher.

> 2. Any suggestion what to do?
To be coherent, the return value for qsat should be set again to 999999 if the temperature is too high, or another more elegant solution should be implemented.
However I'm not aware of the reasons why the qsat subroutines were modified.
I'll open a ticket to track this.

[aducharne]

Réunion tickets du 18/12/18:
soit on met un stop qd temp_sol devient trop grande (on veut que le code plante ici et pas plus loin), soit on prend une valeur plus raisonnable à la valeur max de qsol_sat

[fcheruy]

calcul de qsat dans LMDZ

[aducharne]

Proposition le 20/01/2020:

• prendre la fonction qsatcalc de LMDZ.
• mettre un warning si temp_sol > 100°C
• mettre un stop qd on dépasse une valeur plus grande, cf h_gardefou de LMDZ

=> PP will do it and commit will be done with JG