Custom Query (964 matches)

1) Line 1419 step2d_lf_am3.h is:

          DiaU2rhs(i,j,M2hadv)=DiaU2rhs(i,j,M2yadv)+fac1

and should be

          DiaU2rhs(i,j,M2hadv)=DiaU2rhs(i,j,M2hadv)+fac1

2) Line 561 step3d_uv.F is:

            Dwrk(i,M2yadv)=Dwrk(i,M2hadv)+                              &
     &                     DiaU3wrk(i,j,k,M3yadv)*Hzk(i,k)

and should be

            Dwrk(i,M2yadv)=Dwrk(i,M2yadv)+                              &
     &                     DiaU3wrk(i,j,k,M3yadv)*Hzk(i,k)

the diagnotics results still seem questionable...

Corrected an adjoint bug in tl_npzd_iron.h, rp_npzd_iron.h, and ad_npzd_iron.h when the option NPZD_IRON is activated. We were using the incorrect basic state during the linearization of the Iron uptake. Many thanks to Andy Moore for correcting this bug.

I also corrected a lowercase/uppercase typo in w4dpsas_ocean.h and w4dvar_ocean.h. We were using both rec and Rec dummy variables. The variable rec was renamed to irec. Many thanks to Brian Powell for bringing this to my attention.

There was a typo in obs_sen_is4dvar.h. A underscore was missing in line 348 for the object initialize_ocean:

       USE mod_ocean, ONLY : initialize_ocean

Also, the assignment for BLKname needed to be before the call to ad_initial and not after:

#if defined BULK_FLUXES && defined NL_BULK_FLUXES
!
!  Set file name containing the nonlinear model bulk fluxes to be read
!  and processed by other algorithms.
!
        BLKname(ng)=HISname(ng)
#endif
!
!  Initialize adjoint model and define sensitivity functional.
!
        Lstiffness=.FALSE.
        CALL ad_initial (ng)
        IF (exit_flag.ne.NoError) RETURN

Many thanks to Jerome Fiechter and Andy Moore for bringing this to my attention