| 4 |
Notes on tags used in MITgcmUV |
Notes on tags used in MITgcmUV |
| 5 |
============================== |
============================== |
| 6 |
|
|
| 7 |
|
o model & momentum pkgs: |
| 8 |
|
- implement fully implicit bottom friction combined with implicit surf. |
| 9 |
|
pressure, for now, only in hydrostic formulation: |
| 10 |
|
+ to use: compile with #define ALLOW_SOLVE4_PS_AND_DRAG in CPP_OPTIONS.h |
| 11 |
|
and run with selectImplicitDrag=2 ; |
| 12 |
|
+ currently missing diagnostics of bottom drag and vert. viscosity tendency; |
| 13 |
|
- partial implicit time-stepping (bot. friction + vert. visc first, then |
| 14 |
|
surf. press) corresponding to selectImplicitDrag=1 should be applicable |
| 15 |
|
to NH formulation but not available until we solve for surf. pressure |
| 16 |
|
increment (instead of full etaN). |
| 17 |
|
o pkg/seaice: add linear remapping scheme (Lipscomb 2001) to ITD model |
| 18 |
|
- this is a first version that includes many regularisations that may |
| 19 |
|
not all be necessary, to be sorted out, |
| 20 |
|
- turned on by runtime parameter SEAICEuseLinRemapITD (true by default) |
| 21 |
|
- change one verification experiment in seaice_itd to use and test new code |
| 22 |
|
- seaice_growth: catch a case where the argument of log can be zero |
| 23 |
|
or smaller (important for code stability) |
| 24 |
|
o model/src: |
| 25 |
|
- modify SOLVE_DIAGONAL_LOWMEMORY version of solve_tridiagonal.F |
| 26 |
|
& solve_pentadiagonal.F to enable to re-used inverse matrix for solving |
| 27 |
|
similar A.X'=Y' system (with same matrix A) in subsequent calls; |
| 28 |
|
switch based on errCode (In/Out) argument value; in calling routines, |
| 29 |
|
initialise errCode before first call. |
| 30 |
|
|
| 31 |
|
checkpoint66a (2016/10/18) |
| 32 |
|
o pkg/bling: |
| 33 |
|
- BLING package now loads in atmospheric pCO2 file using exf package routines |
| 34 |
|
Enable with #define USE_EXFCO2 in BLING_OPTIONS.h |
| 35 |
|
o model/src & pkg/ptracers: |
| 36 |
|
- with INCLUDE_IMPLVERTADV_CODE defined, also call MOM_U,V_IMPLICIT_R & |
| 37 |
|
GAD_IMPLICIT_R (instead of IMPLDIFF) when just implicitViscosity and |
| 38 |
|
implicitDiffusion (respectively) are used (even without momImplVertAdv |
| 39 |
|
or temp,salt,PTRACERS_ImplVertAdv). |
| 40 |
|
o model/src & generic_advdiff: |
| 41 |
|
- add parenthesis in main diagonal calculation of implicit vertical |
| 42 |
|
diffusion matrix (impldiff.F & gad_implicit_r.F). |
| 43 |
|
This affects results at machine truncation level; update output of many |
| 44 |
|
verification experiments (31 fwd ; 10 adm ; 8 tlm ; 3 oad). |
| 45 |
|
o pkg/mom_commom & generic_advdiff: |
| 46 |
|
- add deep atmosphere and anelastic scaling factor in implicit vertical |
| 47 |
|
advection routines (mom_u,v_implicit_r.F & gad*_impl*_r.F). |
| 48 |
|
o model/src: |
| 49 |
|
- remove few unnecessary temp array copies in solve_tridiagonal.F |
| 50 |
|
& solve_pentadiagonal.F |
| 51 |
|
|
| 52 |
|
checkpoint65z (2016/09/28) |
| 53 |
|
o pkg/seaice: fix multithreading for seaice_fgmres.F |
| 54 |
|
- remove unecessary variables from common blocks |
| 55 |
|
- use extra dimension MAX_NO_THREADS for all variables of the Krylov space |
| 56 |
|
basis, that are stored in commom blocks (including integer variable i1) |
| 57 |
|
- clean up code a little by removing commented lines that refer to |
| 58 |
|
original code |
| 59 |
|
- turn on multi-threaded test for experiment offline_exf_seaice.dyn_jfnk. |
| 60 |
|
o MITgcm manual: |
| 61 |
|
- add uptodate pkg/ecco documentation to the manual (section 8.1; ecco.tex), |
| 62 |
|
preliminary documentation for generic controls (8.3; ctrl.tex), |
| 63 |
|
a placeholder for smooth (8.4; smooth.tex), and comment out the deprecated |
| 64 |
|
ecco_costfunction.tex |
| 65 |
|
o pkg/ecco: |
| 66 |
|
- update and cleanup list of CCP options |
| 67 |
|
- re-assign gencost_mask to specify masks rather than gencost_errfile; 'c' |
| 68 |
|
is now always passed to cost_generic; gencost_mask is now MAX_LEN_FNAM |
| 69 |
|
- add checks that retired run-time options are not used, and in relation |
| 70 |
|
to gencost_mask and gencost_errfile |
| 71 |
|
- finish deprecating gencost_scalefile, gencost_smooth2Ddiffnbt, and |
| 72 |
|
gencost_nrecperiod |
| 73 |
|
- fix m_eta computation for virtual salt flux case |
| 74 |
|
- add m_horflux_vol case (selected via gencost_barfile) that computes |
| 75 |
|
transports via cost_gencost_boxmean (gencost_flag=-3) |
| 76 |
|
- use gencost_flag more systematically (i.e., avoid using gencost_name) |
| 77 |
|
for code switches (aside from physical variable selection) |
| 78 |
|
o pkg/seaice: make seaice_fgmres.F fit for multithreading; this first |
| 79 |
|
attempt appears to be a bit of a hack and my need some repair down the road |
| 80 |
|
o pkg/exf: |
| 81 |
|
- interpolation near South Pole: fix criteria on input field latitude. |
| 82 |
|
o pkg/icefront & pkg/exf: |
| 83 |
|
- move setting of icefront Sub-Glacial RunOff forcing (currently unused) |
| 84 |
|
from pkg/exf (read from data.exf) to pkg/icefront (read from data.icefront) |
| 85 |
|
o pkg/exf: |
| 86 |
|
- replace CPP option ALLOW_EXF_OUTPUT_INTERP with run-time parameter |
| 87 |
|
"exf_output_interp" (in namelist: EXF_NML_04); change output file name |
| 88 |
|
(add suffix "_out" instead of a prefix); keep same record number as in |
| 89 |
|
input file but always write record 1 first (to avoid a STOP in I/O S/R). |
| 90 |
|
Change output file content by writing interpolated field before applying |
| 91 |
|
land-sea mask (helpful for debug). |
| 92 |
|
- 2 compon. vector interpolation with rotation: apply scaling factor on |
| 93 |
|
each compon. (Eastward and Northward) just before rotation. |
| 94 |
|
o pkg/bling: |
| 95 |
|
- Cleaned package. Removed not needed code and comments. |
| 96 |
|
- adjoint example now works |
| 97 |
|
(MITgcm_contrib/verification_other/global_oce_biogeo_bling/code_ad) |
| 98 |
|
o pkg/ptracers: |
| 99 |
|
- add Bruno Deremble's code to complete the linear surface correction |
| 100 |
|
for tracer conservation; |
| 101 |
|
- enabled for individual ptracers by PTRACERS_linFSConserve(iTr)=.TRUE. |
| 102 |
|
|
| 103 |
checkpoint65y (2016/07/29) |
checkpoint65y (2016/07/29) |
| 104 |
o pkg/debug (stats): |
o pkg/debug (stats): |
| 105 |
- change GLOBAL_SUM to GLOBAL_SUM_TILE (result independent of tile to |
- change GLOBAL_SUM to GLOBAL_SUM_TILE (result independent of tile to |
| 113 |
writing delayed until next monitor output. |
writing delayed until next monitor output. |
| 114 |
o pkg/obcs: |
o pkg/obcs: |
| 115 |
- fix bug in OBCS_BALANCE_FLOW (OBCS_balanceFacN misspelled OBCS_balanceFacW) |
- fix bug in OBCS_BALANCE_FLOW (OBCS_balanceFacN misspelled OBCS_balanceFacW) |
| 116 |
|
o pkg/cal: |
| 117 |
|
- calandar-dump: treat case myTime < timePhase in the same way as in |
| 118 |
|
eesupp/src/diff_phase_multiple.F (when not using calandar-dump). |
| 119 |
o pkg/dic & seaice: |
o pkg/dic & seaice: |
| 120 |
- fix compilation of ocean component for coupled atm+ocn model with |
- fix compilation of ocean component for coupled atm+ocn model with |
| 121 |
seaice dynamics & dic. |
seaice dynamics & dic. |
Parent Directory
| 
Revision Log
| 
Revision Graph
| 
Patch