Opened 15 years ago
Closed 15 years ago
#399 closed defect (Fixed)
Corrected vertical integration in balance operator
Reported by: | arango | Owned by: | arango |
---|---|---|---|
Priority: | major | Milestone: | Adjoint Based Algorithms |
Component: | Adjoint | Version: | 3.4 |
Keywords: | Cc: |
Description
Ticket corrects the following:
- An update was made in src:ticket:373 to use the level of no motion when computing the balanced free-surface contribution in the balance operator. This strategy is only needed when integrating the non-hydrostatic relationship from the level of no motion (LNM_depth) to the surface.
This strategy is not needed when ZETA_ELLIPTIC is activated since an elliptical equation is solved (Fukumori et al., 1998; Weaver et al. 2005) to avoid the need for a level of no motion in the vertical integration. Therefore, ad_banlance.F, tl_balance.F, and zeta_balanced.F were changed to remove this integration option.
- Changed the Version variable to 3.4 in mod_ncparam.F:
character (len=5 ) :: version = '3.4 '
- Allowed integer overflow in ran_state.F when using the gfortran compiler (Linux-gfortran.mk, CYGWIN-gfortran.mk, MINGW-gfortran.mk):
$(SCRATCH_DIR)/ran_state.o: FFLAGS += -fno-strict-overflow
Many thanks to Alexander Barth for reporting and solving this problem.
Note:
See TracTickets
for help on using tickets.