Custom Query (986 matches)

Turn off the processing of stations and floats if not running long enough to create such NetCDF files. That is, LdefSTA(ng) is false because nSTA(ng) < ntimes or nSTA(ng) = 0 when nrrec = 0. Therefore, we also need Lstations(ng) to be false.

The following patch to inp_par.F fix the ​problem:

diff --git a/ROMS/Utility/inp_par.F b/ROMS/Utility/inp_par.F
index dc13796..9a6fa88 100644
--- a/ROMS/Utility/inp_par.F
+++ b/ROMS/Utility/inp_par.F
@@ -11100,6 +11100,10 @@
 !  Report input parameters.
 !-----------------------------------------------------------------------
 !
+      DO ng=1,Ngrids
+        IF (.not. LdefFLT(ng) .and. Lfloats(ng))                        &
+     &                 Lfloats(ng) = .FALSE.
+      END DO
       IF (Lwrite) THEN
         DO ng=1,Ngrids
           IF (ncount(ng).ne.Nfloats(ng)) THEN
@@ -11534,6 +11538,10 @@
 !  Report input parameters.
 !-----------------------------------------------------------------------
 !
+      DO ng=1,Ngrids
+        IF (.not. LdefSTA(ng) .and. Lstations(ng))                      &
+     &                 Lstations(ng) = .FALSE.
+      END DO
       IF (Lwrite) THEN
         DO ng=1,Ngrids
           IF (Lstations(ng)) THEN

Corrected how the RHS of elliptic equation is computed in the balance operator. The metrics in the divergence operator were incorrect.

Many thanks to Andy Moore for his help in testing this option.

Corrected several bugs in the Matlab scripts used to compute the standard deviation for the unbalanced error covariance matrix. Added a new script, ssh_reference.m, to compute the balance operator reference sea surface height.

Many thanks to Andy Moore for his help in testing these complex scripts.