| 4 |
Notes on tags used in MITgcmUV |
Notes on tags used in MITgcmUV |
| 5 |
============================== |
============================== |
| 6 |
|
|
| 7 |
|
o model/src: |
| 8 |
|
useRealFreshWaterFlux: remove masking of EmPmP (using maskInC) in |
| 9 |
|
external_forcing_surf.F and apply mask directly in S/R SOLVE_FOR_PRESSURE. |
| 10 |
|
o pkg/obcs: |
| 11 |
|
modify Stevens boundary conditions so that S/R obcs_calc_stevens computes |
| 12 |
|
tendency for temperature and salinity that is then applied to update |
| 13 |
|
theta and salt in S/R obcs_apply_ts; this requires proper masking of all |
| 14 |
|
horizontal advection and diffusion contributions on the boundary (in |
| 15 |
|
S/Rs gad_advection, gad_calc_rhs, gmredi_rtransport, gmredi_calc_diff), and |
| 16 |
|
removing the mask in S/R timestep_tracer. The advantage of this approach is |
| 17 |
|
that the updated tracer values include contributions from the surface |
| 18 |
|
forcing and vertical diffusion. |
| 19 |
|
This modification changes exp4.stevens because of vertical diffusion. |
| 20 |
|
o model/src: |
| 21 |
|
change non-curvilinear grid setting to also work with exch2 + blank tiles: |
| 22 |
|
- move delX,delY to new header file (SET_GRID.h) and adjust length to 1rst |
| 23 |
|
face dimensions. |
| 24 |
|
- new S/R INI_LOCAL_GRID to define model tile-local grid (using same units |
| 25 |
|
as delX,delY) and called from S/R INI_CARTESIAN_GRID, INI_CYLINDER_GRID |
| 26 |
|
and INI_SPHERICAL_POLAR_GRID. |
| 27 |
|
change test-experiment global_ocean.90x40x15: |
| 28 |
|
- split the domain in many more tiles (9x4 tiles of 10x10 each); |
| 29 |
|
- switch to exch2 and remove 1 blank-tile (but only for MPI test). |
| 30 |
|
- update results of both test (std & dwnslp). |
| 31 |
|
o pkg/diagnostics/diagstat_calc.F: fix problem with non-initialised halos in |
| 32 |
|
tmp-fields for TARGET_NEC_SX at the cost of extra loops and less vectorisation |
| 33 |
|
o model/src: |
| 34 |
|
- take sigma-coords related code out of update_etah.F into new S/R UPDATE_ETAWS |
| 35 |
|
and fix initialisation when using sigma-coords. |
| 36 |
|
o model/src: |
| 37 |
|
- improve Gael's changes from Jan 14, 2011 "addition of an 'update' of hfacc etc. |
| 38 |
|
to their current value at the beginning of forward_step.F": add a new run-time |
| 39 |
|
switch "doResetHFactors" (default=false) and use it to do (or not) the resetting. |
| 40 |
|
- set doResetHFactors=T in the following test exp: |
| 41 |
|
global_ocean.90x40x15/input (rStar) & input.dwnslp (linFS), |
| 42 |
|
global_ocean.cs32x15/input.seaice (rStar), hs94.cs-32x32x5/input.impIGW (linFS) |
| 43 |
|
and tutorial_global_oce_in_p (nonLinFS in p) |
| 44 |
|
o pkg/ecco |
| 45 |
|
- added IES cost term |
| 46 |
|
o verification/advect_xz |
| 47 |
|
- scale down the depth by 1/100 (from 200 km to just 2.km) |
| 48 |
|
- add secondary test using divergent flow, r-star and implicit vertical advection. |
| 49 |
|
- add note (check_conserve_TS.txt) and matlab script to check global |
| 50 |
|
conservation of T & S with NonLin-FreeSurf and Adams-Bashforth. |
| 51 |
|
o pkg/rbcs: |
| 52 |
|
- add missing relaxUFile & relaxVFile in namelist (in rbcs_readparms.F). |
| 53 |
|
o pkg/obcs (Stevens BCs): |
| 54 |
|
- fix various problems in the phase speed computations of the Stevens |
| 55 |
|
boundary conditions: |
| 56 |
|
- sign error for western and southern boundaries, |
| 57 |
|
- time tendencies were underestimated by factor of dTracerLev and some just |
| 58 |
|
zero (for multidim advection schemes) -> now estimate time tendencies |
| 59 |
|
from difference between previous (n) time level and the one before (n-1); |
| 60 |
|
this requires storing a field per tracer per boundary, also in the pickup |
| 61 |
|
files. |
| 62 |
|
- replace deltaTmom in cflMer/Zon computation with dTracerLev |
| 63 |
|
- set default phase velocity to cflMer/Zon |
| 64 |
|
- update verification experiment exp4.stevens |
| 65 |
|
o pkg/generic_advdiff (implicit_r): |
| 66 |
|
- fix implicit vertical advection conservation with AB and/or NonLin-FreeSurf |
| 67 |
|
(remove Tr*d/dz(w) in gad_implicit_r.F and add it in gad_calc_rhs.F); |
| 68 |
|
- fix implicit vertical diffusion or advection with NonLin FreeSurf |
| 69 |
|
(was using current recip_hFacC instead of future recip_hFacC): set future |
| 70 |
|
recip_hFac in thermodynamics.F and pass & use it in all vertical implicit |
| 71 |
|
solver S/R ; |
| 72 |
|
- fill diagnostics of vertical advective fluxes (computed from tendency) |
| 73 |
|
when using implicit vert. advection; |
| 74 |
|
- update output of exp that uses NonLin-FreeSurf & implicit vertical diffusion: |
| 75 |
|
global_ocean.90x40x15 (fwd+AD), global_ocean.cs32x15 (4 fwd +4 AD), |
| 76 |
|
tutorial_global_oce_in_p (fwd), tutorial_tracer_adjsens (AD) |
| 77 |
|
and cpl_aim+ocn (used in ocean); to be fixed: bottom_ctrl_5x5. |
| 78 |
|
o pkg/diagnostics: |
| 79 |
|
- add diags for T & S tendency which goes through Adams-Bashforth. |
| 80 |
|
|
| 81 |
|
checkpoint63f (2011/11/30) |
| 82 |
|
o pkg/obcs/obcs_calc_stevens.F: |
| 83 |
|
- join some loops to reduce number of if-statements |
| 84 |
|
- fix a bug in code for northern and southern boundaries. |
| 85 |
|
o pkg/seaice/seaice_init_varia.F |
| 86 |
|
- undo resetting UVM=0 that has slipped in unnoticed in r1.49 and |
| 87 |
|
caused zero ice-velocities for the B-grid code |
| 88 |
|
o pkg/seaice/seaice_init_fixed.F |
| 89 |
|
- fix a bug in the initialisation of the second metric terms |
| 90 |
|
k2AtC and k2AtZ: k2AtC was set to the value of k2AtZ and k2AtZ was 0. |
| 91 |
|
- this affects all seaice configurations with useCurviLinearGrid = .true. |
| 92 |
|
and SEAICEuseDynamics = .true. (default) and SEAICEuseMetricTerms |
| 93 |
|
= .true. (default) |
| 94 |
|
- the only verification experiment affected is global_ocean.cs32x15: |
| 95 |
|
icedyn, seaice, and seaice_dynmix (AD) |
| 96 |
|
o pkg/thsice/thsice_diagnostics_state.F |
| 97 |
|
- only for TARGET_NEC_SX: initialise all of tmpFld, so that overlap does |
| 98 |
|
not contain garbage that can lead to surprises in diagstats_calc.F |
| 99 |
|
o model/src: |
| 100 |
|
- move bi,bj loops inside INTEGR_CONTINUITY ; |
| 101 |
|
- call INTEGR_CONTINUITY directly from forward_step.F |
| 102 |
|
(previously called from momentum_correction_step.F) |
| 103 |
|
- call UPDATE_ETAH from integr_continuity.F and update initialise_varia.F, |
| 104 |
|
forward_step.F and pkg/ecco/the_main_loop.F |
| 105 |
|
o model/src/do_oceanic_phys.F: |
| 106 |
|
- reset addMass to zero only if useICEFRONT ; |
| 107 |
|
- separated k-loop for rhoInSitu calculation ; |
| 108 |
|
- add 2 diagnostics for wVel*delta^k(Rho) at constant pressure (= WdRHO_P) |
| 109 |
|
and at constant T,S (= WdRHOdP); |
| 110 |
|
NOTE: change diagnostics WRHOMASS : this is now simply the Vertical transport |
| 111 |
|
of In-Situ Density (Anomaly) (=wVel*rho^bar-k) (previously was Vertical |
| 112 |
|
transport of locally-referenced potential density); |
| 113 |
|
to recover old diag: (WRHOMASS)_old = (WRHOMASS)_new -1/2 * (WdRHOdP) |
| 114 |
|
o gluemnc: |
| 115 |
|
- replace ncrcat with ncks when adding variables to *.glob.nc |
| 116 |
|
- change the way variables without record dimensions are handled so that |
| 117 |
|
files that have variables with and without record variables can be glued: |
| 118 |
|
add a record dimension with a new unique name to be able to differentiate |
| 119 |
|
between variable with and without a-prior record dimensions. Remove |
| 120 |
|
this record dimension before adding the variable to the *.glob.nc file. |
| 121 |
|
This procedure automatically works also for grid.*.nc files that do not |
| 122 |
|
have any record dimension. |
| 123 |
|
o verification/cpl_aim+ocn: |
| 124 |
|
- generate new reference output on baudelaire (gfortran, mpich2) (previously |
| 125 |
|
from old aces-grid cluster, ao + ifort) and update 'run_cpl_test' script. |
| 126 |
|
o pkg/obcs/obcs_calc_stevens.F: |
| 127 |
|
- add code for northern and southern boundaries. |
| 128 |
|
- reorder some loops to save some operations and so that drFBar can be scalar |
| 129 |
|
o model/src: |
| 130 |
|
- re-arrange extrapolation calculation in Adams-Bashforth S/R (affects machine |
| 131 |
|
truncation) and add diagnostics for Adams-Bashforth tendency increment. |
| 132 |
|
- update test-exp. AD output: |
| 133 |
|
1D_ocean_ice_column, bottom_ctrl_5x5, lab_sea.noseaicedyn, obcs_ctrl; |
| 134 |
|
and FWD output: exp4 (x3), fizhi-cs-32x32x40, fizhi-cs-aqualev20, |
| 135 |
|
front_relax (x3), global_ocean.90x40x15 (x2), global_ocean.cs32x15 (x5), |
| 136 |
|
global_ocean_ebm, ideal_2D_oce, internal_wave, isomip (x3), lab_sea (x4), |
| 137 |
|
matrix_example, MLAdjust (x5), natl_box (x2), tutorial_advection_in_gyre, |
| 138 |
|
tutorial_global_oce_in_p, tutorial_global_oce_latlon and |
| 139 |
|
tutorial_plume_on_slope. |
| 140 |
|
o pkg/diagnostics/diagstats_calc.F: add special code for TARGET_NEC_SX |
| 141 |
|
that vectorises. This code uses 3 extra 2D fields, F90 array assignments |
| 142 |
|
and F90 intrinsic routines SUM, MINVAL, MAXVAL, but since TARGET_NEC_SX |
| 143 |
|
does not even come with an F77 compiler anymore, this should not be |
| 144 |
|
a limitation |
| 145 |
|
|
| 146 |
|
checkpoint63e (2011/11/06) |
| 147 |
|
o pkg/seaice adjoint of LSR: |
| 148 |
|
- substantially rework STORE dirs. |
| 149 |
|
- add new flag SEAICE_LSR_ADJOINT_ITER and parameter SOLV_MAX_FIXED |
| 150 |
|
this flag will do SOLV_MAX_ITERS but only store SOLV_MAX_FIXED fields |
| 151 |
|
and only perform SOLV_MAX_FIXED adjoint iters. |
| 152 |
|
If flag SEAICE_LSR_ADJOINT_ITER is undef-ined, the store in the iteration |
| 153 |
|
always goes to same record, meaning that the last forward field will be |
| 154 |
|
the one used in the adjoint - an approximation. |
| 155 |
|
- now produces non-zero LSR-related gradients, and acceptable gradients for |
| 156 |
|
An's zonal (x-z) test case (to be checked in) |
| 157 |
|
- lab_sea: produces acceptable gradients for lab_sea with forward/backward LSR |
| 158 |
|
(but these are screwed up by the thermodynamics in the first place); |
| 159 |
|
reduce from 4 to 3 timesteps to avoid t.d. issue as temporary solution; |
| 160 |
|
output_adm.txt also reflects changes to seaice_lsr |
| 161 |
|
- global_ocean.cs32x15: |
| 162 |
|
* the "default" input_ad.seaice has been re-defined to always exclude dynamics |
| 163 |
|
(output_adm.seaice.txt was updated) |
| 164 |
|
* new input_ad.seaice_dynmix tests a forward LSR / backward free-drift |
| 165 |
|
(output_adm.seaice_dynmix.txt) |
| 166 |
o test-exp. exp4+stevens: |
o test-exp. exp4+stevens: |
| 167 |
- switch off pTracers and nonHydrostatic (restart problems with NH); update output |
- switch off pTracers and nonHydrostatic (restart problems with NH); update output |
| 168 |
o pkg/seaice (lsr solver): |
o pkg/seaice (lsr solver): |
Parent Directory
| 
Revision Log
| 
Revision Graph
| 
Patch