Changeset 4682 for trunk

Timestamp:

2014-06-21T11:47:21+02:00 (10 years ago)

Author:

gm

Message:

#1345 : correct a bug in CORE bulk formulae in the v3.6alpha (trunk)

File:

• trunk/NEMOGCM/NEMO/OPA_SRC/SBC/sbcblk_core.F90 (modified) (2 diffs)

trunk/NEMOGCM/NEMO/OPA_SRC/SBC/sbcblk_core.F90

@@ -820 +820 @@
            sqrt_Cd=kappa/((kappa/sqrt_Cd_n10) - zpsi_m) ; Cd=sqrt_Cd*sqrt_Cd;
          ELSE
-           !! Shifting the wind speed to 10m and neutral stability :
-           U_n10 = dU10*1./(1. + sqrt_Cd_n10/kappa*(log(zu/10.) - zpsi_m)) !  L & Y eq. (9a)
+           !! Shifting the wind speed to 10m and neutral stability :  L & Y eq. (9a)
+           !   In very rare low-wind conditions, the old way of estimating the
+           !   neutral wind speed at 10m leads to a negative value that causes the code
+           !   to crash. To prevent this a threshold of 0.25m/s is now imposed.
+           U_n10 = MAX( 0.25 , dU10/(1. + sqrt_Cd_n10/kappa*(log(zu/10.) - zpsi_m)) )
 
            !! Updating the neutral 10m transfer coefficients :
@@ -956 +959 @@
          zpsi_m  = psi_m(zeta_u)
          !!
-         !! Shifting the wind speed to 10m and neutral stability : (L & Y eq.(9a))
-!        U_n10 = dU10/(1. + sqrt_Cd_n10/kappa*(log(zu/10.) - psi_m(zeta_u)))
-         U_n10 = dU10/(1. + sqrt_Cd_n10/kappa*(log(zu/10.) - zpsi_m))
+         !! Shifting the wind speed to 10m and neutral stability : L & Y eq.(9a)
+         !   In very rare low-wind conditions, the old way of estimating the
+         !   neutral wind speed at 10m leads to a negative value that causes the code
+         !   to crash. To prevent this a threshold of 0.25m/s is now imposed.
+         U_n10 = MAX( 0.25 , dU10/(1. + sqrt_Cd_n10/kappa*(log(zu/10.) - zpsi_m)) )
          !!
          !! Shifting temperature and humidity at zu :          (L & Y eq. (9b-9c))