Error in the diagnostic bottom pressure

[Robin_Waldman]

Context

Within DIA/diaar5.F90, the routine dia_ar5 computes wrongly the bottom pressure (variable zbotpres)

Analysis

The bottom pressure formulation in L.185 of diaar5.F90 has a dimensional error. Indeed, the hydrostatic part is multiplied twice by density, once in the zbotpres computation of L.164-179, and a second time within the constant coefficient zztmp computed in L.184.

Recommendation

1. For already existing model outputs with the variable zbotpres (pbo in the CMIP nomenclature): the bottom pressure can be retrieved offline by combining zbtotpres and sshn as follows:

   true_zbotpres(:,:) = zbotpres(:,:)/rau0 + (sshn(:,:) + thick0(:,:)) * g * (rau0 - 1)
   

2. The correction to diaar5.F90 to retrieve the true bottom pressure is:

   184            zztmp = grav * 1.e-4_wp               ! recover pressure from pressure anomaly and cover to dbar = 1.e4 Pa
   185            zbotpres(:,:) = zztmp * ( zbotpres(:,:) + rho0 * (ssh(:,:,Kmm) + thick0(:,:)) )
   

...

[Robin_Waldman]

rau0 instead of rho0 in L.185.