| 4 |
Notes on tags used in MITgcmUV |
Notes on tags used in MITgcmUV |
| 5 |
============================== |
============================== |
| 6 |
|
|
| 7 |
|
o pkg/exf (interpolation): |
| 8 |
|
- assume periodicity in X only if input field cover full longitude range. |
| 9 |
|
- improve search for lat. index (supposed to be faster, in ~log2(ny) steps, |
| 10 |
|
and should vectorise). |
| 11 |
|
- fix input lat of the 2 added row (in case we provide N.pole data). |
| 12 |
|
|
| 13 |
|
checkpoint63h (2011/12/30) |
| 14 |
|
o pkg/seaice: |
| 15 |
|
- change declaration of UVM from _RL to _RS in order to get B-grid |
| 16 |
|
uIce,vIce monitor when _RS = real*4 |
| 17 |
|
o pkg/offline: |
| 18 |
|
- new S/R OFFLINE_RESET_PARMS for resetting main-model params (switch off flags); |
| 19 |
|
- compute vertical velocity if not read from file (switch on exactConserv |
| 20 |
|
& staggerTimeStep); |
| 21 |
|
- fix output of ptracer monitor iter & time for default PTRACERS_monitorFreq. |
| 22 |
|
o verification: |
| 23 |
|
- update grid-files (in tutorial_held_suarez_cs/input and fizhi-cs-32x32x40/input) |
| 24 |
|
with grid-angle generated using "utils/matlab/cs_grid/cubeCalcAngle.m": |
| 25 |
|
fix Pb in yG definition @ 2 missing corners (previously yG was averaged); |
| 26 |
|
- update output of exp. global_ocean.cs32x15.viscA4 after changing |
| 27 |
|
grid-file angles (used here for coriolis-3d); |
| 28 |
|
- remove (or gzip) few files not needed to run fizhi-cs experiments. |
| 29 |
|
o pkg/exf (vector interpolation): |
| 30 |
|
- rotate vector component toward local grid-axis orientation using grid-angles |
| 31 |
|
(stored in GRID.h) instead of locally computed rotation angle; keep old code |
| 32 |
|
available with exf option EXF_USE_OLD_VEC_ROTATION (undef by default). |
| 33 |
|
o pkg/seaice: |
| 34 |
|
- allow to compile with exf options ALLOW_ATM_TEMP or ALLOW_DOWNWARD_RADIATION |
| 35 |
|
undefined. Fix and remove the stop when ALLOW_RUNOFF is undefined. |
| 36 |
|
o pkg/PKG_OPTIONS.h files: |
| 37 |
|
- move #include "CPP_OPTIONS.h" just after #include "PACKAGES_CONFIG.h", |
| 38 |
|
outside of the #ifdef ALLOW_THISPKG / #endif block. |
| 39 |
|
- update verification/*/code/PKG_OPTIONS.h files. |
| 40 |
|
o add a verification experiment (offline_exf_seaice/input.seaice) that test |
| 41 |
|
the B-grid LSOR solver |
| 42 |
|
o pkg/seaice: sort out emissivity and long wave radiation step 2: |
| 43 |
|
- add emissivity for snow and ice as new run-time parameters |
| 44 |
|
(unfortunately) that default to the corresponding exf values; this |
| 45 |
|
changes results in most verification experiments so that for backward |
| 46 |
|
compatibility these parameter have to be set explicitly to inconsistent |
| 47 |
|
values (ugly); |
| 48 |
|
- the exf-flag EXF_LWDOWN_WITH_EMISSIVITY then gets rid off the hard-wired |
| 49 |
|
emissivities of 0.97 associated with lwdownloc in S/R seaice_solve4temp |
| 50 |
|
until we agree how to handle this bug properly |
| 51 |
|
o pkg/exf (interpolation): |
| 52 |
|
- add a set of run-time param ({inputfield}_interpMethod), one for |
| 53 |
|
each interpolated input field, to select the interpolation method |
| 54 |
|
with a value of zero switching off the interpolation; |
| 55 |
|
- print interpolation parameters to standard output; |
| 56 |
|
- refine exf_check.F accordingly ; stop if USE_NO_INTERP_RUNOFF is defined |
| 57 |
|
(to use instead runoff_interpMethod=0 in data.exf). |
| 58 |
|
- simplify default setting of interpolation parameters (for all fields, |
| 59 |
|
set position to match model grid cell-center position, leading to trivial |
| 60 |
|
interpolation in case of regular Lat-Lon grid) and fix for the case of |
| 61 |
|
non-uniform delY. |
| 62 |
|
o pkg/exf/exf_radiation.F: add the factor ocean_emissivity to lwdown, for |
| 63 |
|
within #ifdef EXF_LWDOWN_WITH_EMISSIVITY/#endif. Unless ocean_emissivity=1. |
| 64 |
|
this factor is absolutely required, because the lw exitance (= out-going |
| 65 |
|
long wave radiation) is emissivity*stefanBoltzmann*T^4 + rho*lwdown, |
| 66 |
|
where the reflectivity rho = 1-emissivity for conservation reasons. Once |
| 67 |
|
this is established, we can remove the CPP-Flag (if we want to). |
| 68 |
|
o pkg/seaice: |
| 69 |
|
- qualify a stop statement in S/R seaice_check: EXF_READ_EVAP and |
| 70 |
|
ALLOW_BULKFORMULAE are only required if SEAICE_EXTERNAL_FLUXES is defined |
| 71 |
|
- sort out emissivity and long wave radiation, step 1: make SEAICE_emissivity |
| 72 |
|
what the name implies (and not emissivity*BoltzmannConstant), this means |
| 73 |
|
the old default value of 5.5e-8 now means |
| 74 |
|
5.5e-8/5.670e-8 = 0.97001763668430343479 as ocean_emissivity in |
| 75 |
|
S/R exf_readparms |
| 76 |
|
- FV-discretization of the B-grid LSOR solver, turn on with |
| 77 |
|
SEAICE_LSRBNEW. This new solver also involves a slight re-organization |
| 78 |
|
of the calling routine dynsolver, so that B and C-grid code look a |
| 79 |
|
little more alike. |
| 80 |
|
- fix index bugs in ostres.F (B-grid code) |
| 81 |
|
o pkg/seaice/seaice_ocean_budget.F: change (unused) code to make it look better |
| 82 |
|
o pkg/seaice/seaice_readparms.F: make ocean_emissivity*stefanBoltzmann the |
| 83 |
|
default for SEAICE_emissivity, if useEXF; requires setting a wrong |
| 84 |
|
parameter value in global_ocean.cs32x15/input.seaice/data.seaice for |
| 85 |
|
backward compatiblity |
| 86 |
|
o model/src: use sigmaR in calc_ivdc |
| 87 |
|
|
| 88 |
|
checkpoint63g (2011/12/16) |
| 89 |
|
o pkg/seaice/seaice_growth.F: |
| 90 |
|
fix heat conservation of the coupled ocean-seaice system. |
| 91 |
|
when SEAICE_HEAT_CONSERV_FIX is defined, the mismatch of |
| 92 |
|
solid vs liquid water treatment is resolved via QNET. |
| 93 |
|
o pkg/seaice/seaice_growth.F: |
| 94 |
|
add missing term for sublimation in saltFlux, which was breaking conservation. |
| 95 |
|
o pkg/seaice/seaice_model.F: |
| 96 |
|
Remove this patholigical case treatment in evolution branch, |
| 97 |
|
where it is un-necessary and breaks conservation. |
| 98 |
|
Update global_ocean.cs32x15/results accordingly for |
| 99 |
|
output.seaice.txt output_adm.seaice.txt and output_adm.seaice_dynmix.txt |
| 100 |
|
o model/src: |
| 101 |
|
useRealFreshWaterFlux: remove masking of EmPmP (using maskInC) in |
| 102 |
|
external_forcing_surf.F and apply mask directly in S/R SOLVE_FOR_PRESSURE. |
| 103 |
|
o pkg/obcs: |
| 104 |
|
modify Stevens boundary conditions so that S/R obcs_calc_stevens computes |
| 105 |
|
tendency for temperature and salinity that is then applied to update |
| 106 |
|
theta and salt in S/R obcs_apply_ts; this requires proper masking of all |
| 107 |
|
horizontal advection and diffusion contributions on the boundary (in |
| 108 |
|
S/Rs gad_advection, gad_calc_rhs, gmredi_rtransport, gmredi_calc_diff), and |
| 109 |
|
removing the mask in S/R timestep_tracer. The advantage of this approach is |
| 110 |
|
that the updated tracer values include contributions from the surface |
| 111 |
|
forcing and vertical diffusion. |
| 112 |
|
This modification changes exp4.stevens because of vertical diffusion. |
| 113 |
|
o model/src: |
| 114 |
|
change non-curvilinear grid setting to also work with exch2 + blank tiles: |
| 115 |
|
- move delX,delY to new header file (SET_GRID.h) and adjust length to 1rst |
| 116 |
|
face dimensions. |
| 117 |
|
- new S/R INI_LOCAL_GRID to define model tile-local grid (using same units |
| 118 |
|
as delX,delY) and called from S/R INI_CARTESIAN_GRID, INI_CYLINDER_GRID |
| 119 |
|
and INI_SPHERICAL_POLAR_GRID. |
| 120 |
|
change test-experiment global_ocean.90x40x15: |
| 121 |
|
- split the domain in many more tiles (9x4 tiles of 10x10 each); |
| 122 |
|
- switch to exch2 and remove 1 blank-tile (but only for MPI test). |
| 123 |
|
- update results of both test (std & dwnslp). |
| 124 |
|
o pkg/diagnostics/diagstat_calc.F: fix problem with non-initialised halos in |
| 125 |
|
tmp-fields for TARGET_NEC_SX at the cost of extra loops and less vectorisation |
| 126 |
|
o model/src: |
| 127 |
|
- take sigma-coords related code out of update_etah.F into new S/R UPDATE_ETAWS |
| 128 |
|
and fix initialisation when using sigma-coords. |
| 129 |
|
o model/src: |
| 130 |
|
- improve Gael's changes from Jan 14, 2011 "addition of an 'update' of hfacc etc. |
| 131 |
|
to their current value at the beginning of forward_step.F": add a new run-time |
| 132 |
|
switch "doResetHFactors" (default=false) and use it to do (or not) the resetting. |
| 133 |
|
- set doResetHFactors=T in the following test exp: |
| 134 |
|
global_ocean.90x40x15/input (rStar) & input.dwnslp (linFS), |
| 135 |
|
global_ocean.cs32x15/input.seaice (rStar), hs94.cs-32x32x5/input.impIGW (linFS) |
| 136 |
|
and tutorial_global_oce_in_p (nonLinFS in p) |
| 137 |
|
o pkg/ecco |
| 138 |
|
- added IES cost term |
| 139 |
|
o verification/advect_xz |
| 140 |
|
- scale down the depth by 1/100 (from 200 km to just 2.km) |
| 141 |
|
- add secondary test using divergent flow, r-star and implicit vertical advection. |
| 142 |
|
- add note (check_conserve_TS.txt) and matlab script to check global |
| 143 |
|
conservation of T & S with NonLin-FreeSurf and Adams-Bashforth. |
| 144 |
|
o pkg/rbcs: |
| 145 |
|
- add missing relaxUFile & relaxVFile in namelist (in rbcs_readparms.F). |
| 146 |
|
o pkg/obcs (Stevens BCs): |
| 147 |
|
- fix various problems in the phase speed computations of the Stevens |
| 148 |
|
boundary conditions: |
| 149 |
|
- sign error for western and southern boundaries, |
| 150 |
|
- time tendencies were underestimated by factor of dTracerLev and some just |
| 151 |
|
zero (for multidim advection schemes) -> now estimate time tendencies |
| 152 |
|
from difference between previous (n) time level and the one before (n-1); |
| 153 |
|
this requires storing a field per tracer per boundary, also in the pickup |
| 154 |
|
files. |
| 155 |
|
- replace deltaTmom in cflMer/Zon computation with dTracerLev |
| 156 |
|
- set default phase velocity to cflMer/Zon |
| 157 |
|
- update verification experiment exp4.stevens |
| 158 |
|
o pkg/generic_advdiff (implicit_r): |
| 159 |
|
- fix implicit vertical advection conservation with AB and/or NonLin-FreeSurf |
| 160 |
|
(remove Tr*d/dz(w) in gad_implicit_r.F and add it in gad_calc_rhs.F); |
| 161 |
|
- fix implicit vertical diffusion or advection with NonLin FreeSurf |
| 162 |
|
(was using current recip_hFacC instead of future recip_hFacC): set future |
| 163 |
|
recip_hFac in thermodynamics.F and pass & use it in all vertical implicit |
| 164 |
|
solver S/R ; |
| 165 |
|
- fill diagnostics of vertical advective fluxes (computed from tendency) |
| 166 |
|
when using implicit vert. advection; |
| 167 |
|
- update output of exp that uses NonLin-FreeSurf & implicit vertical diffusion: |
| 168 |
|
global_ocean.90x40x15 (fwd+AD), global_ocean.cs32x15 (4 fwd +4 AD), |
| 169 |
|
tutorial_global_oce_in_p (fwd), tutorial_tracer_adjsens (AD) |
| 170 |
|
and cpl_aim+ocn (used in ocean); to be fixed: bottom_ctrl_5x5. |
| 171 |
|
o pkg/diagnostics: |
| 172 |
|
- add diags for T & S tendency which goes through Adams-Bashforth. |
| 173 |
|
|
| 174 |
|
checkpoint63f (2011/11/30) |
| 175 |
|
o pkg/obcs/obcs_calc_stevens.F: |
| 176 |
|
- join some loops to reduce number of if-statements |
| 177 |
|
- fix a bug in code for northern and southern boundaries. |
| 178 |
|
o pkg/seaice/seaice_init_varia.F |
| 179 |
|
- undo resetting UVM=0 that has slipped in unnoticed in r1.49 and |
| 180 |
|
caused zero ice-velocities for the B-grid code |
| 181 |
|
o pkg/seaice/seaice_init_fixed.F |
| 182 |
|
- fix a bug in the initialisation of the second metric terms |
| 183 |
|
k2AtC and k2AtZ: k2AtC was set to the value of k2AtZ and k2AtZ was 0. |
| 184 |
|
- this affects all seaice configurations with useCurviLinearGrid = .true. |
| 185 |
|
and SEAICEuseDynamics = .true. (default) and SEAICEuseMetricTerms |
| 186 |
|
= .true. (default) |
| 187 |
|
- the only verification experiment affected is global_ocean.cs32x15: |
| 188 |
|
icedyn, seaice, and seaice_dynmix (AD) |
| 189 |
|
o pkg/thsice/thsice_diagnostics_state.F |
| 190 |
|
- only for TARGET_NEC_SX: initialise all of tmpFld, so that overlap does |
| 191 |
|
not contain garbage that can lead to surprises in diagstats_calc.F |
| 192 |
|
o model/src: |
| 193 |
|
- move bi,bj loops inside INTEGR_CONTINUITY ; |
| 194 |
|
- call INTEGR_CONTINUITY directly from forward_step.F |
| 195 |
|
(previously called from momentum_correction_step.F) |
| 196 |
|
- call UPDATE_ETAH from integr_continuity.F and update initialise_varia.F, |
| 197 |
|
forward_step.F and pkg/ecco/the_main_loop.F |
| 198 |
|
o model/src/do_oceanic_phys.F: |
| 199 |
|
- reset addMass to zero only if useICEFRONT ; |
| 200 |
|
- separated k-loop for rhoInSitu calculation ; |
| 201 |
|
- add 2 diagnostics for wVel*delta^k(Rho) at constant pressure (= WdRHO_P) |
| 202 |
|
and at constant T,S (= WdRHOdP); |
| 203 |
|
NOTE: change diagnostics WRHOMASS : this is now simply the Vertical transport |
| 204 |
|
of In-Situ Density (Anomaly) (=wVel*rho^bar-k) (previously was Vertical |
| 205 |
|
transport of locally-referenced potential density); |
| 206 |
|
to recover old diag: (WRHOMASS)_old = (WRHOMASS)_new -1/2 * (WdRHOdP) |
| 207 |
|
o gluemnc: |
| 208 |
|
- replace ncrcat with ncks when adding variables to *.glob.nc |
| 209 |
|
- change the way variables without record dimensions are handled so that |
| 210 |
|
files that have variables with and without record variables can be glued: |
| 211 |
|
add a record dimension with a new unique name to be able to differentiate |
| 212 |
|
between variable with and without a-prior record dimensions. Remove |
| 213 |
|
this record dimension before adding the variable to the *.glob.nc file. |
| 214 |
|
This procedure automatically works also for grid.*.nc files that do not |
| 215 |
|
have any record dimension. |
| 216 |
|
o verification/cpl_aim+ocn: |
| 217 |
|
- generate new reference output on baudelaire (gfortran, mpich2) (previously |
| 218 |
|
from old aces-grid cluster, ao + ifort) and update 'run_cpl_test' script. |
| 219 |
|
o pkg/obcs/obcs_calc_stevens.F: |
| 220 |
|
- add code for northern and southern boundaries. |
| 221 |
|
- reorder some loops to save some operations and so that drFBar can be scalar |
| 222 |
|
o model/src: |
| 223 |
|
- re-arrange extrapolation calculation in Adams-Bashforth S/R (affects machine |
| 224 |
|
truncation) and add diagnostics for Adams-Bashforth tendency increment. |
| 225 |
|
- update test-exp. AD output: |
| 226 |
|
1D_ocean_ice_column, bottom_ctrl_5x5, lab_sea.noseaicedyn, obcs_ctrl; |
| 227 |
|
and FWD output: exp4 (x3), fizhi-cs-32x32x40, fizhi-cs-aqualev20, |
| 228 |
|
front_relax (x3), global_ocean.90x40x15 (x2), global_ocean.cs32x15 (x5), |
| 229 |
|
global_ocean_ebm, ideal_2D_oce, internal_wave, isomip (x3), lab_sea (x4), |
| 230 |
|
matrix_example, MLAdjust (x5), natl_box (x2), tutorial_advection_in_gyre, |
| 231 |
|
tutorial_global_oce_in_p, tutorial_global_oce_latlon and |
| 232 |
|
tutorial_plume_on_slope. |
| 233 |
|
o pkg/diagnostics/diagstats_calc.F: add special code for TARGET_NEC_SX |
| 234 |
|
that vectorises. This code uses 3 extra 2D fields, F90 array assignments |
| 235 |
|
and F90 intrinsic routines SUM, MINVAL, MAXVAL, but since TARGET_NEC_SX |
| 236 |
|
does not even come with an F77 compiler anymore, this should not be |
| 237 |
|
a limitation |
| 238 |
|
|
| 239 |
|
checkpoint63e (2011/11/06) |
| 240 |
|
o pkg/seaice adjoint of LSR: |
| 241 |
|
- substantially rework STORE dirs. |
| 242 |
|
- add new flag SEAICE_LSR_ADJOINT_ITER and parameter SOLV_MAX_FIXED |
| 243 |
|
this flag will do SOLV_MAX_ITERS but only store SOLV_MAX_FIXED fields |
| 244 |
|
and only perform SOLV_MAX_FIXED adjoint iters. |
| 245 |
|
If flag SEAICE_LSR_ADJOINT_ITER is undef-ined, the store in the iteration |
| 246 |
|
always goes to same record, meaning that the last forward field will be |
| 247 |
|
the one used in the adjoint - an approximation. |
| 248 |
|
- now produces non-zero LSR-related gradients, and acceptable gradients for |
| 249 |
|
An's zonal (x-z) test case (to be checked in) |
| 250 |
|
- lab_sea: produces acceptable gradients for lab_sea with forward/backward LSR |
| 251 |
|
(but these are screwed up by the thermodynamics in the first place); |
| 252 |
|
reduce from 4 to 3 timesteps to avoid t.d. issue as temporary solution; |
| 253 |
|
output_adm.txt also reflects changes to seaice_lsr |
| 254 |
|
- global_ocean.cs32x15: |
| 255 |
|
* the "default" input_ad.seaice has been re-defined to always exclude dynamics |
| 256 |
|
(output_adm.seaice.txt was updated) |
| 257 |
|
* new input_ad.seaice_dynmix tests a forward LSR / backward free-drift |
| 258 |
|
(output_adm.seaice_dynmix.txt) |
| 259 |
|
o test-exp. exp4+stevens: |
| 260 |
|
- switch off pTracers and nonHydrostatic (restart problems with NH); update output |
| 261 |
|
o pkg/seaice (lsr solver): |
| 262 |
|
- put coriolis term into FORCEX,Y (as in seaice_evp.F; previously in rhsU,V only); |
| 263 |
|
update output (change truncation) of FWD test exp lab_sea & global_ocean.cs32x15 |
| 264 |
|
- fix SEAICEuseFREEDRIFT for new OBCS version (ndef OBCS_UVICE_OLD) |
| 265 |
|
- new parameter "LSR_mixIniGuess" allows (if=2,4) to mix free-drift solution |
| 266 |
|
into initial guess for linearised solver; use weight ~ (1/err)^2,4. |
| 267 |
|
- compute (with new S/R) and print initial (+ free-driff sol. if available) |
| 268 |
|
and final residual of linearised system. |
| 269 |
|
o pkg/pp81: add diagnostics for vertical eddy viscosity and diffusivity |
| 270 |
|
(requires new subroutines pp81_init_fixed.F/pp81_diagnostics_init.F) |
| 271 |
|
o pkg/obcs: Stevens BCs (still only for east and west) |
| 272 |
|
- fix Stevens BCs by using intermediate velocities of previous |
| 273 |
|
time step as "baroclinic" contribution rather velocity of one grid cell |
| 274 |
|
inside the boundaries; more consistent but makes the entire treatment |
| 275 |
|
more complicated, requires new pickup files and two new S/Rs: |
| 276 |
|
obcs_copy_uv_n and obcs_save_uv_n |
| 277 |
|
keep old version with flag OBCS_STEVENS_USE_INTERIOR_VELOCITY |
| 278 |
|
(undefined by default) |
| 279 |
|
- silently move computation of some grid factors into separate loops to |
| 280 |
|
save some recomputations |
| 281 |
|
- add a hack for non-linear free surface |
| 282 |
|
- add a verification experiment to exp4 that tests this code |
| 283 |
|
o make gluemnc/xplodemnc work with (moron) non-GNU shell tools |
| 284 |
|
|
| 285 |
|
checkpoint63d (2011/10/25) |
| 286 |
|
o pkg/obcs + seaice-dynamics: |
| 287 |
|
- new implementation calls OBCS_APPLY_UVICE (in seaice_dynsolver.F) before |
| 288 |
|
solving for uIce,vIce and prevent SEAICE_LSR or SEAICE_EVP to modify OB |
| 289 |
|
values (consistent with an implicit method). |
| 290 |
|
- add option OBCS_UVICE_OLD (in OBCS_OPTIONS.h, defined by default) to |
| 291 |
|
return to older implementation (same results as before); note that none of |
| 292 |
|
the "experimental" OBCS_SEAICE options are implemented in the new version. |
| 293 |
|
- switch test exp. seaice_obcs to use new version (update output). |
| 294 |
|
o pkg/seaice: |
| 295 |
|
- add few stop (B-grid & CS-exch, B-grid & obcs, AUTODIFF & SEAICE_maskRHS) |
| 296 |
|
- add parenthesis in Water-Ice drag calculations (evp & lsr); |
| 297 |
|
- in seaice_lsr.F, simplify and fix AA3 setting in case sNx or sNy=1; |
| 298 |
|
and fix case SEAICE_waterDrag=0 (but unstable if no or little seaice) |
| 299 |
|
- update output of test exp lab_sea that uses evp or lsr (fwd+ad) |
| 300 |
|
and global_ocean.cs32x15 (icedyn & seaice fwd+ad). |
| 301 |
|
o pkg/layers: |
| 302 |
|
- Include potential density as new coordinate (Thanks to David Munday) |
| 303 |
|
o pkg/dic: |
| 304 |
|
- Remove subroutine CALC_PCO2_APPROX_CO3 from carbon_chem.F and add |
| 305 |
|
carbonate computation/output to CALC_PCO2_APPROX |
| 306 |
|
o pkg/generic_advdiff |
| 307 |
|
- make TAF generate code that vectorizes: |
| 308 |
|
gad_dst3fl_adv_*.F, gad_u3c4_impl_r.F |
| 309 |
|
o model/src: |
| 310 |
|
- improve vectorization also for adjoint code of |
| 311 |
|
solve_tridiagonal and solve_pentadiagonal. As this requires up to |
| 312 |
|
5 new 3D-fields, this code is only turned on for the combination of |
| 313 |
|
ALLOW_AUTODIFF_TAMC and TARGET_NEC_SX. Also remove error message |
| 314 |
|
for singular matrix check (only for TARGET_NEC_SX). |
| 315 |
|
- remove hard-wiring of CPP-flag ALLOW_SOLVERS_KLOOPINSIDE to |
| 316 |
|
ALLOW_AUTODIFF. This flag now needs to be set outside of |
| 317 |
|
solve_tri/pentadiagonal.F |
| 318 |
|
o pkg/exf: |
| 319 |
|
- fix previous modif in exf_readparms.F (revision 1.66: treatment of |
| 320 |
|
optional namelist EXF_NML_OBCS). |
| 321 |
|
- add many debug-print (to trace down where pkg/cal error msg comes from) |
| 322 |
|
o pkg/mom_vecinv/mom_vi_hdissip.F |
| 323 |
|
- rearrange loops so that vectorization is improved and in particular |
| 324 |
|
the adjoint routine vectorizes. The code is a little longer now, but I |
| 325 |
|
think easier to read. |
| 326 |
|
Note: CPP-flag MOM_VI_ORIGINAL_VISCA4 is not tested. |
| 327 |
|
o pkg/seaice: |
| 328 |
|
- allow to specify initial seaice velocity from files (uIceFile,vIceFile) |
| 329 |
|
o pkg/exch2: |
| 330 |
|
- alternative code (without tile x tile matrix) for cumulated (global) sum; |
| 331 |
|
used as default; keep using matrix version in adjustment.cs-32x32x1. |
| 332 |
|
o tools/build_options: |
| 333 |
|
- after merging some MPI and/or multi-threads specific optfiles so that |
| 334 |
|
the modified default optfile support both built (using genmake2 -mpi |
| 335 |
|
for the MPI built, using genmake2 -omp for OpenMP built), |
| 336 |
|
move now obsolete optfiles to tools/build_options/unsupported/ |
| 337 |
|
(add copy of email sent to support list in doc/optfiles_changes.txt) |
| 338 |
|
|
| 339 |
|
checkpoint63c (2011/09/28) |
| 340 |
|
o pkg/ebm: |
| 341 |
|
- fix ebm_zonalmean.F for several procs in Ydir (nPy > 1); update output of |
| 342 |
|
test exp. global_ocean_ebm (affects machine truncation, 12 digits for cg2d) |
| 343 |
|
o pkg/seaice (SItracer pickup) |
| 344 |
|
* Simple code for pickup of SItracer fields (#define ALLOW_SITRACER) |
| 345 |
|
To be done: |
| 346 |
|
- thorough testing of missing fields in S/R seaice_read_pickup |
| 347 |
|
- thorough matching between generic tracers and specific fields used |
| 348 |
|
(e.g. volume-weighted age, area-weighted age, salinity, ...) |
| 349 |
|
o pkg/grdchk: |
| 350 |
|
- all procs calculate and print finite-difference gradient (for testreport) |
| 351 |
|
- change description output used by testreport; |
| 352 |
|
- AD testing: also compare and report Finite-difference gradient; |
| 353 |
|
=> update all output_adm*.txt |
| 354 |
|
o test exp. global_ocean_ebm: |
| 355 |
|
- switch back useRealFreshWaterFlux (no problem anymore with run-off); |
| 356 |
|
and turn on exactConserve; update (fwd) output. |
| 357 |
|
o pkg/seaice (LSR): |
| 358 |
|
- print to STDOUT number of solver iterations at SEAICE_monFreq interval. |
| 359 |
|
- modify solver to deal cleanly with u/v switch at face edges (CS-grid): |
| 360 |
|
* identical results with non-CS topology (but save up to half of EXCH calls) |
| 361 |
|
* no much changes in cs-32 verification tests since no ice at face edges |
| 362 |
|
(differences due to convergence criteria, since nb of iterations changed) |
| 363 |
|
* expect some differences on lat-lon-cap grid. |
| 364 |
|
update FWD results of global_ocean.cs32x15 test exp. (icedyn & seaice) |
| 365 |
|
o pkg/exch2: |
| 366 |
|
- add new CPP option to disable Cumul-Sum matrix (to save storage when using |
| 367 |
|
many tiles); until alternative coding is added, just add a stop. |
| 368 |
|
o pkg/kpp: |
| 369 |
|
- move ptracer/salt correction to KPP diffusivity from calc_3d_diffusivity.F |
| 370 |
|
to S/R KPP_CALC_DIFF_PTR (+ add tracer-index to argument list). |
| 371 |
|
o optfiles: |
| 372 |
|
- merged few MPI specific (*+mpi*) and multi-threads specific (*+mth) optfiles |
| 373 |
|
into corresponding standard one, which now support different build: |
| 374 |
|
(when using "genmake2 -mpi" for MPI built, "genmake2 -omp" for OpenMP built) |
| 375 |
|
- add "darwin_amd64_gfortran" optfile (initially from Constantinos, tested by Gael) |
| 376 |
|
o pkg/obcs (seaice): |
| 377 |
|
- add missing EXCH for uIce,vIce after applying OBCS; |
| 378 |
|
- remove S/R EXF_SWAPFFIELDS_X/Y , disable S/R EXF_SET_OBCS_X/Y by adding |
| 379 |
|
argument = number of levels to S/R EXF_SWAPFFIELDS_XZ/YZ and S/R |
| 380 |
|
EXF_SET_OBCS_XZ/YZ ; update obcs_exf_load.F; |
| 381 |
|
- fix EXF_SET_OBCS_XZ/YZ for multi-threaded (assuming loaded arrays are shared) |
| 382 |
|
- verification exp. seaice_obcs: split domain in 2 tiles ; add SIZE.h_mpi |
| 383 |
|
and eedata.mth for MPI and mutlti-threaded testing. |
| 384 |
|
o pkg/flt: |
| 385 |
|
- make FLT EXCH working with multi-threads (and MPI + multi-threads); |
| 386 |
|
- use buffer for floats IO (trajectories & profiles); size set in FLT_SIZE.h; |
| 387 |
|
- fix pickup and trajectories + profiles output with multi-threads; |
| 388 |
|
- add parameter to select group of fields to write along trajectory and |
| 389 |
|
profile (default unchanged); |
| 390 |
|
- only call PORT_RAND* function when flt_noise <> 0; |
| 391 |
|
- add multi-threaded test for verification exp. flt_example. |
| 392 |
|
o eesupp/src: |
| 393 |
|
- fix S/R ALL_PROC_DIE for MPI+multi-threads. |
| 394 |
|
- USE_OMP_THREADING: check that number of threads (from eedata) matches |
| 395 |
|
actual number (from env.var. setting). |
| 396 |
|
o pkg/ebm: |
| 397 |
|
- fix multi-threading in ebm_area_t.F ; clean-up EBM.h header file. |
| 398 |
|
- add multi-threaded test for verification exp. global_ocean_ebm |
| 399 |
|
|
| 400 |
|
checkpoint63b (2011/08/29) |
| 401 |
|
o pkg/ebm: |
| 402 |
|
- fix sign of runoff; |
| 403 |
|
- update directly forcing fields (Qnet,EmPmR) to use standard |
| 404 |
|
external_forcing_surf.F (remove ebm_forcing_surf.F); |
| 405 |
|
- avoid loading forcing fields a 2nd time (1rst time from ini_forcing.F); |
| 406 |
|
- set either tauThetaClimRelax (in data) or tauThetaZonRelax (in data.ebm) |
| 407 |
|
and apply relaxation towards corresponding SST field (2-D map or ZonAver). |
| 408 |
|
- update global_ocean_ebm parameters file data and update output (fwd+ad). |
| 409 |
|
o verification/matrix_example: |
| 410 |
|
- split domain in 8 tiles (2x4); add SIZE.h_mpi and eedata.mth for |
| 411 |
|
MPI and mutlti-threaded testing. |
| 412 |
|
o pkg/matrix: |
| 413 |
|
- change name of S/R MATRIX_INIT to S/R MATRIX_INIT_VARIA ; |
| 414 |
|
- fix for multi-tiles (add bi,bj indices to exp/impMatrixCounter); |
| 415 |
|
- fix for multi-threaded (remove DATA statement); |
| 416 |
|
- use standard DIFFERENT_MULTIPLE function; |
| 417 |
|
- add MATRIX_OPTIONS.h |
| 418 |
|
o verification/matrix_example: |
| 419 |
|
To allow to split the domain in several tiles, change the domain size |
| 420 |
|
from 31x31 to 32x32 by adding 1 row (on Southern side) + 1 column (on |
| 421 |
|
Western side) of land points. |
| 422 |
|
o (*_)write_pickup: |
| 423 |
|
- fix few bugs in mypackage_read/write_pickup.F |
| 424 |
|
- in MDS_WR_METAFILES calls, replace "myTime" by a local copy into |
| 425 |
|
array of length 1 (was caught by compiler syntax-check). |
| 426 |
|
o pkg/diagnostics (velocity potential solver): |
| 427 |
|
- fix initialisation of main diagonal matrix (for overlap values) |
| 428 |
|
o tools/build_options: |
| 429 |
|
- towards single optfile for MPI and non-MPI built: merge few +mpi optfile |
| 430 |
|
into previously non MPI one (linux_am64 pgf77,open64,pathf90,ifort11) |
| 431 |
|
- linux_amd64_gfortran: merge +mpi_generic version into standard optfile: |
| 432 |
|
* move IEEE flags to DEVEL section; take IEEE flags from +mpi_generic ; |
| 433 |
|
* FAST case: for now, keep -O3 with ini_masks_etc.F in NOOPTFILES list; |
| 434 |
|
and use "-devel" testreport option for standard testing on baudelaire. |
| 435 |
|
o pkg/grdchk: |
| 436 |
|
- Assume that all gradient-check position are within the same tile |
| 437 |
|
corresponding to global indices (iGloTile,jGloTile) and that |
| 438 |
|
positions (i,j indices or nbeg) are local to this tile; |
| 439 |
|
- set tile local-indices and proc. number according to tile global-indices; |
| 440 |
|
- remove grdchkwhichproc from namelist (retired param); |
| 441 |
|
- send adjoint gradient to all procs so that all procs print it to STDOUT; |
| 442 |
|
- switch on MPI test for global_ocean.90x40x15 AD test experiment. |
| 443 |
|
o adding package bbl (Bottom Boundary Layer) |
| 444 |
|
description in MITgcm/pkg/bbl/bbl_description.tex |
| 445 |
|
example/test experiment in MITgcm_contrib/bbl |
| 446 |
|
o adjoint verifications |
| 447 |
|
- changed 5 experiments to multi-tiles |
| 448 |
|
* global_ocean.90x40x15 |
| 449 |
|
* obcs_ctrl |
| 450 |
|
* offline_exf_seaice |
| 451 |
|
* tutorial_dic_adjoffline |
| 452 |
|
* tutorial_tracer_adjsens |
| 453 |
|
changed reference output for those that crossed tile boundaries |
| 454 |
|
(changed choices of gradient check positions) |
| 455 |
|
- with same tiling, switch on MPI test for the last 4 (add SIZE.h_mpi) |
| 456 |
|
- temporary bug fix in grdchk_loc.F for obcs |
| 457 |
|
o genmake2 & testreport |
| 458 |
|
- add target "obj" (just for OBJFILES files) to Makefile; |
| 459 |
|
- add option to specify FEXTRAFLAGS on command line. |
| 460 |
|
o pkg/mypackage: |
| 461 |
|
- remove bi,bj arguments from argument list of S/R MYPACKAGE_CALC_RHS |
| 462 |
|
|
| 463 |
|
checkpoint63a (2011/08/02) |
| 464 |
|
o pkg/diagnostics (velocity potential): |
| 465 |
|
- fix velocity-potential calculation when using OBCS: balance net inflow |
| 466 |
|
at each level (new S/R OBCS_DIAG_BALANCE). |
| 467 |
|
- fix some of the problems with OBCS in stream-function calculation. |
| 468 |
|
o pkg/exf/exf_readparms.F |
| 469 |
|
- make treatment of optional namelists (EXF_NML_04, EXF_NML_SGRUNOFF, |
| 470 |
|
EXF_NML_OBCS) more consistent, so that you need not have these lists |
| 471 |
|
in data.exf when they are not used. |
| 472 |
|
o optim/optim_readdata.F: |
| 473 |
|
- add some output about obcs variables that might help debugging |
| 474 |
|
- add a check about nvarlength |
| 475 |
|
o pkg/ctrl: once again, fix computation of endrec in s/r ctrl_init_rec |
| 476 |
|
o pkg/diagnostics (velocity potential): |
| 477 |
|
- fix mask for OBCS (still problems in stream-function with OBCS); |
| 478 |
|
- add specific parameter (default = main code CG2D params) for solver; |
| 479 |
|
- in case of poor convergence, use solution corresponding to lowest residual. |
| 480 |
|
o pkg/seaice/seaice_growth.F: small changes that permit vectorizable |
| 481 |
|
adjoint code. In general statements like tmpscal=fld(i,j) are problematic |
| 482 |
|
and should be avoided where possible. |
| 483 |
|
o model/src (rStar): |
| 484 |
|
- include r* effect on reference-state (PhiRef) in totPhiHyd |
| 485 |
|
(so that it's really the potential anomaly at the cell center); |
| 486 |
|
affect solution only if z* with full pressure in EOS. |
| 487 |
|
- test-exp global_ocean.90x40x15: reduce viscA4 (was not stable) and |
| 488 |
|
update output (because of z* in totPhiHyd and viscA4 changes); |
| 489 |
|
- keep the previous diagnostics as "PHIHYDcR" (closer to potential anomaly |
| 490 |
|
@ constant r) until interpolation at constant r is implemented. |
| 491 |
|
o pkg/diagnostics |
| 492 |
|
- fix RCENTER diagnostics in case cell center is not at interface middle |
| 493 |
|
o add code for TEOS-10 (www.teos-10.org, McDougall et al. 2011). Use |
| 494 |
|
this eos with eosType = 'TEOS10', in data (PARM01). This eos implies |
| 495 |
|
that THETA and SALT are "conservative temperature" and "absolute |
| 496 |
|
salinity" |
| 497 |
|
o pkg/ctrl |
| 498 |
|
- fix computation of endrec in s/r ctrl_init_rec |
| 499 |
|
- ctrl_pack/unpack_*: remove unused variables and some obsolete code |
| 500 |
|
o pkg/seaice/seaice_growth.F |
| 501 |
|
- fix recomputation with FENTY_AREA_EXPANSION_CONTRACTION code, |
| 502 |
|
remove then obsolete STORE directives |
| 503 |
|
- fix STORE directives for tice and tices, so that only the |
| 504 |
|
appropriate part of the fields are stored (and not the entire |
| 505 |
|
field); update global_ocean.cs32x15.seaice AD output. |
| 506 |
|
- cosmetic changes: adjust CPP flags and indentation for better |
| 507 |
|
legibilty |
| 508 |
|
o pkg/autodiff: add mnc-output of adBottomDragFld (code contributed by |
| 509 |
|
Chris Wilson) |
| 510 |
|
o pkg/gmredi: |
| 511 |
|
Add Sub-Meso Eddies parameterisation from Baylor (for now, disabled with TAF), |
| 512 |
|
and test it in verification exp. "front_relax.bvp" (update output.bvp.txt) |
| 513 |
|
|
| 514 |
|
checkpoint63 (2011/07/11) |
| 515 |
|
o pkg/exch2: |
| 516 |
|
- rename+move: nTiles in W2_EXCH2_PARAMS.h --> exch2_nTiles in W2_EXCH2_TOPOLOGY.h |
| 517 |
|
- exch2 version of Cumul-Sum routine: S/R W2_SET_MAP_CUMSUM set up the |
| 518 |
|
(2*nTiles)x(nTiles) Cumul-Sum matrix and Master Proc collects all tiles X,Y |
| 519 |
|
increments, calculates Cumul-Sum @ tile origin and sends results to each tile. |
| 520 |
|
o pkg/diagnostics: |
| 521 |
|
- deal with pairs of post-processed diags which are computed together; in |
| 522 |
|
particular, re-order list of diags within 1 file to save 2nd computation. |
| 523 |
|
- add S/R to compute velocity stream-function (called after velocity-potential |
| 524 |
|
calculation) from divergence free transport -> new diagnostics "PsiVEL". |
| 525 |
|
- add parameter to select grid-point location where PsiVEL == 0. |
| 526 |
|
o eesupp/src: |
| 527 |
|
new S/R to calculate cumulated sum on tiled array, corner grid-cell location |
| 528 |
|
(specific S/R needed with exch2 topology - for now, just print an error msg) |
| 529 |
|
o pkg/shelfice |
| 530 |
|
- Implement friction velocity-dependent transfer coefficients following |
| 531 |
|
Holland and Jenkins, JPO, 1999 |
| 532 |
|
Original code by M. Losch with small modifs. |
| 533 |
|
Enable with CPP option SHI_ALLOW_GAMMAFRICT |
| 534 |
|
and runtime flag SHELFICEuseGammaFrict |
| 535 |
|
o verification lab_sea & global_ocean.cs32x15: |
| 536 |
|
- use tools/xmakedepend (instead of cyrus_makedepend, in genmake_local): |
| 537 |
|
does not miss some dependencies (but is slower) and also works with many files |
| 538 |
|
(default makedepend limitation for AD built). |
| 539 |
|
- use cyrus_makedepend in adjustment.128x64x1 (so that it still tested). |
| 540 |
|
o pkg/diagnostics: |
| 541 |
|
- add run-time parameter "useMissingValue" (def=False) to fill land-point |
| 542 |
|
(i.e., where mask=0) with MissingValue ; used only in MNC output file. |
| 543 |
|
This replaces CPP-option DIAGNOSTICS_MISSING_VALUE. |
| 544 |
|
- Extend the use of "missing_value" setting from data.diagnostics for the case |
| 545 |
|
of counter-diagnostics (where fraction-weight is zero). This applies both to |
| 546 |
|
MNC and MDS output files. |
| 547 |
|
o pkg/seaice/seaice_growth.F: |
| 548 |
|
- add 'a priori' cap of latent heat flux based on actual mass considerations |
| 549 |
|
when SEAICE_ADD_SUBLIMATION_TO_FWBUDGET is used in conjunction with the |
| 550 |
|
non-legacy seaice_solve4temp. |
| 551 |
|
- add diagnostic of the residual sublimation freshwater flux (the freshwater |
| 552 |
|
flux taken from the ocean if the total actual latent heat fluxes exceed |
| 553 |
|
the maximum) |
| 554 |
|
- remove the CPP switch around the *_FWbySublim unit change and move diagnostic |
| 555 |
|
fill for SIatmQnt, SIfwSubl, SIatmFW to the end of seaice_growth.F. |
| 556 |
|
- add diagnostic of the actual sublimation freshwater flux (that is 0. ifndef |
| 557 |
|
SEAICE_ADD_SUBLIMATION_TO_FWBUDGET) and of the latent heat flux (evap+sublim). |
| 558 |
|
- bug fix : frWtrAtm needs a_FWbySublim ifdef SEAICE_ADD_SUBLIMATION_TO_FWBUDGET. |
| 559 |
|
- bug fix : missing part to Martin's r_FWbySublim pathological case fix. |
| 560 |
|
o verification 1D_ocean_ice_column: |
| 561 |
|
made model domain truly 1x1x23 instead of 5x5x23 with only center cell (3,3,:) wet |
| 562 |
|
update results |
| 563 |
|
o pkg/diagnostics: |
| 564 |
|
- Implement setting of "Post-Processed" diagnostics (corresponding to gdiag(5)='P') |
| 565 |
|
which are not filled-up but computed from other diags ; In this case, the mate |
| 566 |
|
diag points to the primary (filled-up) diag to used for post processing. |
| 567 |
|
- Update Velocity-potential computation using post-processed diag framework. |
| 568 |
|
==> Change the ordering of fields and time-periods in MDS output file |
| 569 |
|
when using periodic averaging: now writes one time record with the full set |
| 570 |
|
of fields for this time period, then the next time reccord (it used to be |
| 571 |
|
one field with the full set of time records then the next field). |
| 572 |
|
Note: this is now similar to NetCDF (MNC) output file. <== |
| 573 |
|
|
| 574 |
|
checkpoint62z (2011/06/21) |
| 575 |
|
o adjoint monitor: |
| 576 |
|
- Add adjoint monitor output for packages exf, seaice |
| 577 |
|
and update relevant testreport output |
| 578 |
|
o pkg/seaice: |
| 579 |
|
- re-order diags inits so that they appear in order in available_diagnostics.log |
| 580 |
|
- added diags: hsnow thermo. rates of change, area preceeding ridiging process, |
| 581 |
|
area/heff/hsnow preceeding thermodynamic growth/melt. |
| 582 |
|
- modified diag: SIdA; output actual effect, as the description states. |
| 583 |
|
- modified description: SIqneti, SIqneto, SIuheff, SIvheff, SIdAb*, HSNOW. |
| 584 |
|
- retired redundant diags: SIyneg, SIfice, SIsnwice, SIthdgrh. |
| 585 |
|
(SIyneg->SIdHbOCN; SIfice->SIdHbATC; SIsnwice->SIdHbFLO; |
| 586 |
|
and SIthdgrh->SIdHbOCN+SIdHbATC+SIdHbATO) |
| 587 |
|
- modified diags (2) : switch all SIdA*, SIdH*, SIdS* from delta to rates. |
| 588 |
|
- added diagnostics for seaice generic tracers (SItr*). |
| 589 |
|
- added SItrUnit and SItrNameLong run time param (for SItr* diags). |
| 590 |
|
- in diag names, replaced 'PrTh' abbrev. of 'preceeding thermo' with 'PT'. |
| 591 |
|
I. Fenty: 18-Jun-2011: |
| 592 |
|
Minor changes to seaice package: |
| 593 |
|
1) Retired old variables (A22, SEAICE_lhsublim, areaMax, areaMin, hiceMin) and |
| 594 |
|
added some new ones (SEAICE_area_reg, SEAICE_hice_reg, SEAICE_area_floor) |
| 595 |
|
- Differentiated "regularization variables" from "floor variables" |
| 596 |
|
* areaMin became SEAICE_area_reg (old A22) and SEAICE_area_floor |
| 597 |
|
* hiceMin became SEAICE_hice_reg (old hiceMin) |
| 598 |
|
(with _reg meaning regularization variable) |
| 599 |
|
- SEAICE_lhSublim becomes lhSublim, the sum of SEAICE_lhEvap and SEAICE_lhFusion |
| 600 |
|
so as to ensure energy conservation when going between phases |
| 601 |
|
- A22 was not used anywhere |
| 602 |
|
2) Changed regularization procedure for heffActual and hsnowActual to ensure |
| 603 |
|
well-boundedness and smooth adjoint in seaice_growth.F |
| 604 |
|
3) Fixed a bug where seaice_solve4temp would not recognize ice-free grid cells |
| 605 |
|
because the old regularization always set heffActual >= 0.05 cm |
| 606 |
|
4) Changed the model so that the default behavior is to put a small (10^-5) "floor" |
| 607 |
|
on AREA when HEFF > 0. |
| 608 |
|
- went from requiring ALLOW_PRECLUDE_INFINITESIMAL_AREA to be defined to |
| 609 |
|
requiring that DISABLE_AREA_FLOOR *not* be defined |
| 610 |
|
Affected verifications: |
| 611 |
|
- 1D_ocean_ice_column (output.txt output_adm.txt) |
| 612 |
|
- global_ocean.cs32x15 (output.seaice.txt output_adm.seaice.txt) |
| 613 |
|
o pkg/diagnostics: |
| 614 |
|
- move MNC code from diagnostics_out.F into new diagnostics_mnc_out.F (2 S/R). |
| 615 |
|
- rename S/R GETDIAG to DIAGNOSTICS_GET_DIAG and change type of 1rst argument |
| 616 |
|
(was _RL, now integer) with option = 0 to retrieve all levels. |
| 617 |
|
- add pieces of code to compute velocity potential at each level (new diagnostics |
| 618 |
|
PhiVEL, works only when UVELMASS & VVELMASS are also turned on); |
| 619 |
|
for now, using same params as CG2D (target-resisual, max Nb of iter ...). |
| 620 |
|
- fix few issues with counter diagnostics in diagnostics_setdiag.F & |
| 621 |
|
diagstats_setdiag.F (if diagnostic counter is itself a counter diag) |
| 622 |
|
and in diagnostics_fract_fill.F (do not need a valid pointer for the |
| 623 |
|
diagnostics counter-mate); check that definition of counter-diagnostic |
| 624 |
|
includes a valid counter-mate (in diagnostics_set_levels.F). |
| 625 |
|
o model/src (rStar): |
| 626 |
|
- add rStar scaling factor in diags_phi_hyd.F (similar to calc_grad_phi_hyd.F); |
| 627 |
|
affect solution only if z* with full pressure in EOS; |
| 628 |
|
update output of test-exp global_ocean.90x40x15. |
| 629 |
|
o pkg/seaice: |
| 630 |
|
- introduce siEps=1e-5 parameter |
| 631 |
|
- use areaMax in AREA update (part 4), consistent with ridging step (part 2.5). |
| 632 |
|
- re-organize SEAICE_ADD_SUBLIMATION_TO_FWBUDGET code. |
| 633 |
|
- introduce ALLOW_SITRACER and seaice_tracer_phys.F to handle generic seaice tracer. |
| 634 |
|
For now it covers, and was tested for, ice salinity and age (attached to HEFF), |
| 635 |
|
as well as ice cover age and ridging index (attached to AREA). |
| 636 |
|
For now the run time parameters (in data.seaice/SEAICE_PARM03) are |
| 637 |
|
SItrName (e.g. 'age') and SItrMate ('HEFF', which is the default, or 'AREA'). |
| 638 |
|
Up to SItrMaxNum=3 tracers can be used simultaneously. |
| 639 |
|
o pkg/salt_plume: |
| 640 |
|
-introduce SPsalFRAC run time parameter that sets the fraction of the |
| 641 |
|
salt by-product of ice growth (SPsalFRAC*(1-SIsalFRAC)*salt typically) |
| 642 |
|
that gets vertically redistributed (fluxed down) by pkg/salt_plume. |
| 643 |
|
o model debug options: |
| 644 |
|
- Switch reading of debugMode from S/R INI_PARMS (parameter file "data", stored |
| 645 |
|
in PARAMS.h) to S/R EESET_PARMS (parameter file "eedata", stored in EEPARAMS.h). |
| 646 |
|
- Refine debug-msg printing selection (3 new levels): |
| 647 |
|
debugMode : controls printing of debug messages (sequence of S/R calls). |
| 648 |
|
debugLevel : controls printing of algorithm-relevant auxiliary/debug messages |
| 649 |
|
debLevZero = 0 ; /* level of disabled aux. msg printing */ |
| 650 |
|
debLevA = 1 ; /* level of minimum aux. msg printing */ |
| 651 |
|
debLevB = 2 ; /* level of low aux. print (report read-file opening)*/ |
| 652 |
|
debLevC = 3 ; /* level of moderate debug prt (most pkgs debug msg) */ |
| 653 |
|
debLevD = 4 ; /* level of enhanced debug prt (add DEBUG_STATS prt) */ |
| 654 |
|
debLevE = 5 ; /* level of extensive debug printing */ |
| 655 |
|
New debugLevel default: with default debugMode value (=F), |
| 656 |
|
= 2 (debLevB) or = 1 (debLevA) in case ALLOW_AUTODIFF is defined.; |
| 657 |
|
and if debugMode=T, default debugLevel = 4 (debLevD). |
| 658 |
|
Relation with previous settings: |
| 659 |
|
debLevB (=2) corresponds to former debLevA (=1) |
| 660 |
|
debLevD (=4) corresponds to former debLevB (=2) |
| 661 |
|
Note: |
| 662 |
|
a) if debugLevel=0, skip WRITE_GRID call when nIter0 > 0 (same as before); |
| 663 |
|
b) default debugLevel=1 is convenient in AD runs (avoid many file opening msgs); |
| 664 |
|
to recover previous default, set debugLevel=2 (e.g. AD-test hs94.1x64x5) |
| 665 |
|
- Add new parameter (printResidualFreq, read from PARM02 namelist) to print |
| 666 |
|
detailed CG solvers convergence; off (=0) by default, except if debugLevel=5. |
| 667 |
|
o pkg/obcs: |
| 668 |
|
- change OBCSfixTopo default from True to False (should not change the results); |
| 669 |
|
keep OBCSfixTopo=T in verification-exp. seaice_obcs. |
| 670 |
|
o pkg/thsice: |
| 671 |
|
- fix bug in thsice_calc_thickn.F (was introduced in version 1.18) in |
| 672 |
|
2nd layer seaice sublimation (unlikely to happen) |
| 673 |
|
- use Interior mask for OBCS in thsice_advection.F (similar to what is |
| 674 |
|
done in gad_advection & seaice_advection). |
| 675 |
|
o adjoint: |
| 676 |
|
bug fix spotted by Gael Forget: |
| 677 |
|
seaice_advection and thsice_advection still had |
| 678 |
|
legacy CPP options to avoid CS corner updates for adjoint |
| 679 |
|
o pkg/seaice: |
| 680 |
|
=> Change some of the defaults (2011/05/29): |
| 681 |
|
- change SEAICEuseFluxForm default (from False to True) |
| 682 |
|
- change DIFF1 default (from 0.004 m/s to 0. m/s) |
| 683 |
|
And print warnings if DIFF1 is not set in data.seaice and centered |
| 684 |
|
avection scheme is used without any diffusion. |
| 685 |
|
also: |
| 686 |
|
- fix seaice_diffusion calls in default AdvScheme part of seaice_advdiff.F; |
| 687 |
|
- set SEAICEdiffKh[] default in the same way as SEAICEadvSch[] |
| 688 |
|
- stop if DIFF1<>0 with SEAICEadvScheme <> 2, if DIFF1<>0 with any |
| 689 |
|
other non-zero SEAICEdiffKh[] diffusivity, if SEAICEadvSalt=T with |
| 690 |
|
#undef SEAICE_VARIABLE_SALINITY, if SEAICEadvAge=T with #undef SEAICE_AGE. |
| 691 |
|
update results of test exp.: |
| 692 |
|
lab_sea: output.fd (switch to AdvScheme=33 with diffKh=20) |
| 693 |
|
output.hb87, output_adm.evp (SEAICEuseFluxForm=T and diffKh=200) |
| 694 |
|
output_adm.noseaicedyn (SEAICEuseFluxForm=T) |
| 695 |
|
offline_exf_seaice: output.seaicetd, output_adm (SEAICEuseFluxForm=T) |
| 696 |
|
seaice_obcs: output (SEAICEuseFluxForm=T) |
| 697 |
|
and prescribe old default for SEAICEuseFluxForm (=F) and DIFF1 (=0.004) |
| 698 |
|
for standard fwd & AD lab_sea tests (to keep the same output). |
| 699 |
|
o pkg/seaice/seaice_growth.F: |
| 700 |
|
- use Ian Fenty's capping formulas for actual ice |
| 701 |
|
thickness thoughout EVOLUTION branch. |
| 702 |
|
- put treatment pathological case #2) in CPP brackets |
| 703 |
|
(ALLOW_AVOID_INFINITESIMAL_AREA, undef by default). |
| 704 |
|
- update global_ocean.cs32x15 and 1D_ocean_ice_colum results accordingly. |
| 705 |
|
o pkg/seaice: |
| 706 |
|
- seaice_diffusion.F : remove the min(DX,DY) factor permanently; |
| 707 |
|
pass the Laplacian diffusivity (in m^2/s) diffKh as a parameter; |
| 708 |
|
also pass fac as a param, which can be 1. (to update the tendency |
| 709 |
|
field) or the time step (to update the field itself). |
| 710 |
|
- seaice_advdiff.F: for multidim, call seaice_diffusion with |
| 711 |
|
e.g. SEAICEdiffKhHeff and with fac=1; for single-dim, call |
| 712 |
|
seaice_diffusion with e.g. SEAICEdiffKhHeff and with fac=time step. |
| 713 |
|
- SEAICE_PARAMS.h, seaice_readparms.F, seaice_check.F : |
| 714 |
|
hooks to handle SEAICEdiffKhHeff, SEAICEdiffKhArea, etc. |
| 715 |
|
o pkg/seaice: |
| 716 |
|
diffusion with centered advection scheme (diffus.F & advect.F): |
| 717 |
|
- simplify S/R DIFFUS (put deltaT & DIFF1 factor outside, move EXCH outside) |
| 718 |
|
to just compute the laplacian of input-field. |
| 719 |
|
- unroll k loop (harmonic,biharmonic diff) in advect.F to save 1 exch call. |
| 720 |
|
- fix flux-form version in diffus.F to conserve tracer. |
| 721 |
|
update output of lab_sea.fd & lab_sea.hb87 (1rst set of modif affects |
| 722 |
|
machine truncation) and lab_sea.salt_plume (conservative flux-form modif) |
| 723 |
|
|
| 724 |
|
checkpoint62y (2011/05/26) |
| 725 |
|
o pkg/seaice: |
| 726 |
|
- seaice_growth.F : in EVOLUTION branch, always include |
| 727 |
|
a_QSWbyATM_cover in qnet (as it should be). |
| 728 |
|
- verification/global_ocean.cs32x15 : update results accordingly |
| 729 |
|
(output.seaice.txt output_adm.seaice.txt). |
| 730 |
|
o pkg/seaice + obcs: |
| 731 |
|
- use maskInW & maskInS in multi-dim advection and biharmonic diffusion: |
| 732 |
|
equivalent of assuming no tracer gradient outside the OB region. |
| 733 |
|
update output of seaice_obcs (uses DIFF1 <>0 => biharmonic diffusion) |
| 734 |
|
- remove calls to OBCS_COPY_TRACER after EXCH calls (no longer needed |
| 735 |
|
after changes above); |
| 736 |
|
- obcs+seaice cleaning: apply seaice-tracer OB values to tracer field |
| 737 |
|
only at OB location (instead of over a band of OLx width). |
| 738 |
|
o pkg/seaice: |
| 739 |
|
- fix wrong indices in old (non Flux-Form) diffusion default S/R diffus.F |
| 740 |
|
(dxG(I+1, ...) changed to dxG(I, ...), no impact on Lat-Lon grid) |
| 741 |
|
o pkg/obcs: |
| 742 |
|
- split header file "OBCS.h" into 4 separated files: |
| 743 |
|
OBCS_PARAMS.h, OBCS_GRID.h, OBCS_FIELDS.h & OBCS_SEAICE.h |
| 744 |
|
Note: OBCS_PARAMS.h & OBCS_GRID.h can be included without OBCS_OPTIONS.h |
| 745 |
|
(contain no conditional statement which depends on OBCS options). |
| 746 |
|
- move "useOBCSYearlyFields" from "data.obcs" (hold in OBCS_PARAMS.h) |
| 747 |
|
to parameter file "data.exf", namelist EXF_NML_OBCS (in EXF_PARAM.h). |
| 748 |
|
o ctrl, cost, grdchk & autodiff pkgs: |
| 749 |
|
- make options file more standard (like other {PKG}_OPTIONS.h file but |
| 750 |
|
with AD_CONFIG.h); clean-up some of the included header files. |
| 751 |
|
o pkg/obcs: |
| 752 |
|
- do not mask tracer when applying OB (tracer == 0 could generate problems) |
| 753 |
|
- apply tracer OB values to tracer field only at OB location (instead of |
| 754 |
|
over a band of OLx width) |
| 755 |
|
- clean-up obcs_apply_w.F & obcs_apply_eta.F (since less calls to these S/R) |
| 756 |
|
- remove OBCS_MULTIDIM_OLD_VERSION code in gad_advection.F and remove |
| 757 |
|
obcs_apply_tloc.F & obcs_apply_sloc.F |
| 758 |
|
o obcs in model/src: |
| 759 |
|
- call OBCS_APPLY_W only if NonHydrostatic. |
| 760 |
|
- call OBCS_APPLY_ETA only if NonLinFreeSurf. |
| 761 |
|
- reset dEtaHdt and wVel to zero outside OB interior region ; |
| 762 |
|
- leave tracer unchanged outside OB interior region: This has no effect |
| 763 |
|
on the solution but just to prevent unrealistic tracer value outside OB. |
| 764 |
|
- zero out velocity outside OB interior region: This has no effect on |
| 765 |
|
the solution but just to prevent unrealistic velocity outside OB. |
| 766 |
|
o obcs and non-lin free-surf: |
| 767 |
|
move call to OBCS_APPLY_R_STAR from update_r_star.F to calc_r_star.F |
| 768 |
|
and call to OBCS_APPLY_SURF_DR from update_surf_dr.F to calc_surf_dr.F |
| 769 |
|
+ add argument "etaFld" to OBCS_APPLY_R_STAR & OBCS_APPLY_SURF_DR: |
| 770 |
|
this fixes missing EXCH + get consistent eta OB value (+ fix restart). |
| 771 |
|
update results of test-exp. exp4(nlfs). |
| 772 |
|
o exp4 (nlfs) test-exp: |
| 773 |
|
use OBC for Eta at Western & Eastern OB (to test prescribing Eta |
| 774 |
|
with NonLin-FreeSurf); update output.nlfs.txt |
| 775 |
|
o obcs and 2d/3d solver norm: |
| 776 |
|
use interior masks (maskInC) to cancel out cg2d & cg3d matrix coeff at OB |
| 777 |
|
and outside OB interior region (will allow to place OB further inside the |
| 778 |
|
domain). Depending on the domain, can affects cg2d-Norm (previous |
| 779 |
|
cancellation outside OB was only partial), e.g. in exp4 & obcs_ctrl, |
| 780 |
|
and cg3d-Norm (previous Norm was computed before canceling OB terms), |
| 781 |
|
e.g. in tutorial_plume_on_slope. |
| 782 |
|
- update output of exp.: exp4 (x2), obcs_ctrl and tutorial_plume_on_slope. |
| 783 |
|
o pkg rbcs: |
| 784 |
|
- split RBCS.h into 3 files: RBCS_SIZE.h, RBCS_PARAMS.h & RBCS_FIELDS.h |
| 785 |
|
- add capability to apply relaxation to horizontal velocity uVel & vVel. |
| 786 |
|
o pkg/grdchk: add gradient check for control variable xx_shifwflx |
| 787 |
|
o pkg/shelfice/shelfice_init_fixed.F |
| 788 |
|
fix CPP-flag combinations |
| 789 |
|
o pkg/seaice: Merge iceage code of Pierre Rampal |
| 790 |
|
* update/change/correct ice age calculations |
| 791 |
|
* add age updates for pathological cases |
| 792 |
|
Probably needs more testing... |
| 793 |
|
|
| 794 |
|
checkpoint62x (2011/05/13) |
| 795 |
|
o pkg/mom_vecinv: |
| 796 |
|
high-order vorticity scheme: use interior masks (maskInW & S) |
| 797 |
|
to zero-out (outside OB) gradient of vorticity. |
| 798 |
|
o verification exp.: |
| 799 |
|
turn on "useEnergyConservingCoriolis" (better discretisation) in exp. |
| 800 |
|
hs94.128x64x5 and seaice_obcs ; update results. |
| 801 |
|
o pkg/obcs: |
| 802 |
|
fix bug in OBeta for YZ sections in obcs_exf_load.F |
| 803 |
|
o pkg/ctrl, pkg/shelfice, and optim: |
| 804 |
|
- add new control variable xx_shifwflx (fresh water flux underneath ice |
| 805 |
|
shelves). This is almost as tedious as obcs-ctrl, because the |
| 806 |
|
variables needs its own mask. |
| 807 |
|
- add a cost function term that penalizes the deviation of xx_shifwflx |
| 808 |
|
from the first guess (zero) in analogy to other control variables |
| 809 |
|
o pkg/shelfice: |
| 810 |
|
- start basic adjoint |
| 811 |
|
o pkg/ctrl: |
| 812 |
|
- remove extra 2D field "for good measure" when initializing in |
| 813 |
|
ctrl_set_globfld_[xy/xz/yz].F |
| 814 |
|
- replace hard coded precision 32 (a bug) in ctrl_map_ini_gen by ctrlprec |
| 815 |
|
- improve error messages in ctrl_set_unpack_* |
| 816 |
|
o OBC implementation in momentum: |
| 817 |
|
use interior masks (maskInC,W & S) in momentum S/R: |
| 818 |
|
- multiply divergence, tension and some del^i.u or del^j.v by maskInC |
| 819 |
|
- multiply del2u,del2v by maskInW/S (biharmonic viscosity) |
| 820 |
|
- multiply horiz. grad of vorticity by maskInW/S (in Leith scheme) |
| 821 |
|
- multiply gradient of wVel by maskInW/S in del2w calculation (NonHyd) |
| 822 |
|
Motivation: fix/clean implementation of OBC with biharmonic-viscosity |
| 823 |
|
(and also with Leith/Smagorinsky harmonic-viscosity). |
| 824 |
|
Update output of experiment dome and exp4 (+nlfs). |
| 825 |
|
o exp4 test-exp: |
| 826 |
|
- turn on: nonHydrostatic, exactConserv, viscA4 (with momDissip_In_AB=F) |
| 827 |
|
and prescribe time-varying OB (without pkg/exf) for zonal-velocity |
| 828 |
|
at Western & Eastern Boundaries (+/- 1 per-cent); Also prescribe |
| 829 |
|
time-varying OB at Western-side for wVel (NonHyd); update output. |
| 830 |
|
- add secondary set-up (.nlfs) using NonLin FreeSurf (z*) + mom_vecinv |
| 831 |
|
and biharmonic-viscosity ; also use prescribed time varying OB. |
| 832 |
|
o pkg/obcs: |
| 833 |
|
fix bug in OBWeta,OBWeta indices in obcs_apply_r_star.F |
| 834 |
|
o model/src: |
| 835 |
|
- fix startTime setting when baseTime<>0 & nIter0=0 are specified ; |
| 836 |
|
- initialise startTime & nIter0 with unlikely values: allows to set one |
| 837 |
|
from the other without ambiguity. |
| 838 |
|
o seaice_obcs test-exp.: |
| 839 |
|
add small diffK4T,S to test biharmonic diffusivity with OBCS ; update output. |
| 840 |
|
o pkg/obcs: |
| 841 |
|
- remove calls to OBCS_COPY_TRACER after EXCH for T,S & pTracer: |
| 842 |
|
no longer needed with maskInC,W,S in pkg/generic_advdiff. |
| 843 |
|
o pkg/seaice: |
| 844 |
|
- retired SEAICE_SALINITY and SEAICE_salinity in favor of |
| 845 |
|
SEAICE_VARIABLE_SALINITY and SIsalFRAC, respectively. |
| 846 |
|
- fixed bug in the initialization of sea ice salinity variable HSALT |
| 847 |
|
in seaice_init_varia.F (removed dependence on ICE2WATR) |
| 848 |
|
o fizhi-cs test exp.: |
| 849 |
|
switch reference output convention for these 2 non-standard test-exp, |
| 850 |
|
from: pgi 6.1-5 on faulks (FC.6, -of linux_ia32_pgf77+authors_fc5) |
| 851 |
|
to: ifort 11.1 on baudelaire (FC.13, -of linux_amd64_ifort11) |
| 852 |
|
reason: faulks might get retired soon and pkg/fizhi has still Pb with gfortran |
| 853 |
|
o model/inc: |
| 854 |
|
- change type of PmEpR array (from _RS to _RL) to fix restart (with "-use_r4") |
| 855 |
|
of exp. global_ocean.90x40x15 (Suncronous time-stepping + Real-FW + NonLin-FS) |
| 856 |
|
o pkg/seaice: |
| 857 |
|
- fix calculation of saltFlux and saltPlumeFlux when using SEAICE_SALINITY |
| 858 |
|
o verification/lab_sea, 1D_ocean_ice_column, seaice_obcs, global_ocean.cs32x15 |
| 859 |
|
- updated output*.txt files due to bug fixes in salt, saltPlumeFlux, |
| 860 |
|
and initial HSALT |
| 861 |
|
|
| 862 |
|
checkpoint62w (2011/04/26) |
| 863 |
|
o pkg/mom_fluxform: |
| 864 |
|
- fix calculation of del2u & del2v (no-slip BC) when using Non-Lin Free-Surf. |
| 865 |
|
- use parameter "sideDragFactor" in no-slip BC side-drag part. |
| 866 |
|
- update output of exp. global_ocean.90x40x15 |
| 867 |
|
o pkg/mom_vecinv: |
| 868 |
|
mask del2u & del2v (no effect on results, but cleaner & closer to mom_fluxform) |
| 869 |
|
o update seaice exp.: |
| 870 |
|
use more consistent parameter values (remove "backward compatability only": |
| 871 |
|
ICE2WATR, SEAICE_lhSublim, SEAICE_cpAir, SEAICE_rhoAir + set rhoConstFresh) |
| 872 |
|
related to thermodynamics seaice. Update output (fwd+AD) of exp.: |
| 873 |
|
1D_ocean_ice_column (x2), lab_sea (x8), ioffline_exf_seaice (x2) & seaice_obcs |
| 874 |
|
o update test-exp global_ocean.90x40x15: |
| 875 |
|
o use default CD_CODE_OPTIONS file (Adams-Bashforth on CD Mom & Coriolis). |
| 876 |
|
o increase Overlap size from 2 to 3 (to allow to test bi-harmonic Visc.) |
| 877 |
|
o new pickup-files from a 100.yr run with pkg/down_slope (input.dwnslp params) |
| 878 |
|
o turn on zStar & viscA4 in standard set-up (dir input/) |
| 879 |
|
and update output files: output.txt & output.dwnslp.txt |
| 880 |
|
o new AD-test experiment: obcs_ctrl |
| 881 |
|
(From Matt MITgcm_Contrib/SOSE/BoxAdj) to test OBCS as control vector. |
| 882 |
|
o pkg/dic: |
| 883 |
|
fix two bugs in handling water vapour pressure in carbonate chemistry |
| 884 |
|
and air-sea fluxes. Found by Val Bennington and Galen McKinley |
| 885 |
|
(add DIC Option = WATERVAP_BUG to recover old results) |
| 886 |
|
update output of tutorial_global_oce_biogeo (fwd+AD) & tutorial_dic_adjoffline |
| 887 |
|
o adding allowInteriorFreezing option to check for water below freezing point |
| 888 |
|
at depth and bring the negative heat anomaly to the surface level |
| 889 |
|
o pkgs _fields_load: |
| 890 |
|
- extend GET_PERIODIC_INTERVAL to deal with non-periodic case (from pkg/rbcs) |
| 891 |
|
- use new S/R GET_PERIODIC_INTERVAL in pkgs: |
| 892 |
|
bulk_force, cheapaml, cfc, dic, obcs (+ ctrl/ctrl_get_gen_rec), offline and |
| 893 |
|
rbcs to get interpolation weights and time reccord number. |
| 894 |
|
- fix multi-threading in aim_fields_load.F |
| 895 |
|
- apply EXCH to temp arrays [0,1] after reading in dic_fields_load.F |
| 896 |
|
o external_fields_load: |
| 897 |
|
new S/R GET_PERIODIC_INTERVAL to calculate which time-record (and linear |
| 898 |
|
interpolation weights) to use in a periodic time-series ; do not divide by |
| 899 |
|
time-step to convert to integer: |
| 900 |
|
- this fixes case where externForcingPeriod is not multiple of time-step |
| 901 |
|
(Thanks to Christopher L. Wolfe) |
| 902 |
|
- this fixes case where externForcingPeriod is an odd multiple of time-step |
| 903 |
|
(was wrong by half of time-step); update results of tutorial_global_oce_in_p. |
| 904 |
|
store index of time-record which is currently loaded: |
| 905 |
|
- works without any assumption on sequence of calls to EXTERNAL_FIELDS_LOAD ; |
| 906 |
|
- needed for the adjoint but not yet used because of extensive recomputations; |
| 907 |
|
(to be fixed soon) |
| 908 |
|
|
| 909 |
|
checkpoint62v (2011/04/02) |
| 910 |
|
o verification/1D_ocean_ice_column added to test Ian Fenty's |
| 911 |
|
pkg/seaice_growth.F modifications for improving sea ice code differentiability |
| 912 |
|
o update output of exp4, internal_wave, seaice_obcs, tutorial_plume_on_slope |
| 913 |
|
(new velocity statistics after changing maskInW & maskInS definition) |
| 914 |
|
o pkg/generic_advdiff with OBCS: |
| 915 |
|
use maksInW & maskInS in advective flux and in biharmonic diffusion calculation: |
| 916 |
|
equivalent to assumption of no tracer gradient outside the OB region. |
| 917 |
|
This will affect the results if using obcs+biharmonic diffusion. |
| 918 |
|
update output_adm.txt of experiment bottom_ctrl_5x5 (not expected since no obc) |
| 919 |
|
o pkg/obcs: |
| 920 |
|
change definition of maskInW & maskInS: leave OB edges inside (maskIn=1) |
| 921 |
|
(e.g., maskInW(OB_Ie)=1 ) so that velocity normal-component at OB is still |
| 922 |
|
in Interior region. No impact on solution but affects Velocity monitor stats. |
| 923 |
|
o pkg/ptracers (adjoint) |
| 924 |
|
New option AUTODIFF_PTRACERS_SPLIT_FILES for writing separate |
| 925 |
|
tape file for each tracer |
| 926 |
|
o model/inc/src and pkg/autodiff: add parameters that allow to turn |
| 927 |
|
of parts of the adjoint code at run time. For now, there are only |
| 928 |
|
useKPPinAdMode, useGMrediInAdMode, and useSEAICEinAdMode. These |
| 929 |
|
parameters are set in data.pkg. |
| 930 |
|
This is the lazy coward's version of doing this, but still useful for |
| 931 |
|
debugging. In order to do it right, everything needs to be moved to |
| 932 |
|
pkg/autodiff. |
| 933 |
|
o pkg/autodiff: fix global_sum_tile_ad.F by resetting adsumPhi at |
| 934 |
|
the end; this fixes the adjoint balancing code for open boundaries. |
| 935 |
|
o pkg/ctrl: |
| 936 |
|
- change behavior of xx_${varname}period to be analogous to exf-conventions: |
| 937 |
|
xx_${varname}period = 0., means time-independent control part |
| 938 |
|
xx_${varname}period = -12., means one-year-periodic monthly control part |
| 939 |
|
- introduce helper routine ctrl_init_rec (part of ctrl_init.F) to |
| 940 |
|
clean up ctrl_init a little, remove unused variables |
| 941 |
|
- ctrl_init: fixed the handling of the start/endrec for obcs-variables |
| 942 |
|
- move "call cal_GetMonthsRec" from s/r ctrl_get_gen to s/r ctrl_get_gen_rec |
| 943 |
|
|
| 944 |
|
checkpoint62u (2011/03/15) |
| 945 |
|
o model/src: |
| 946 |
|
- for the adjoint, change NonLin-FreeSurf initialisation sequence: fill-in |
| 947 |
|
h0Fac from INITIALISE_FIXED and reset hFac to h0Fac in INITIALISE_VARIA ; |
| 948 |
|
- in initialise_varia.F (except for ALLOW_DEPTH_CONTROL) remove 2nd call to |
| 949 |
|
INI_DEPTHS, INI_MASKS_ETC & INI_LINEAR_PHISURF. |
| 950 |
|
o pkg/ctrl/obcs/ecco: |
| 951 |
|
- remove obsolete, unused, and partially broken code: |
| 952 |
|
ctrl_obcsvol/ctrl_obcsbal/ctrl_volflux and code within |
| 953 |
|
ALLOW_CTRL_OBCS_BALANCE |
| 954 |
|
- put a stop statement into ecco/cost_obcsvol.F, but also add |
| 955 |
|
cost_obcsvol to ecco_ad_diff.list |
| 956 |
|
o pkg/ecco: fix counting cost function contributions for obcs (code |
| 957 |
|
by Matt M.) and diffkr for diagnostics |
| 958 |
|
o pkg/ecco, pkg/ctrl, pkg/grdchk: |
| 959 |
|
add cost penalty terms for initial uvel, vvel, eta controls |
| 960 |
|
and I/O for prior error files |
| 961 |
|
o pkg/mdsio: |
| 962 |
|
move mds_check4file.F from eesupp/src to mdsio_check4file.F ; add option |
| 963 |
|
(new argument) to check for tiled file in mdsioLocalDir : this fix the |
| 964 |
|
restart from pickup_som , pickup_dic (& dic_co2atm) with mdsioLocalDir. |
| 965 |
|
o pkg/ctrl: |
| 966 |
|
- ctrl_obcsbal.F, ctrl_obcsvol.F: replace last calls of ctrl_getrec by |
| 967 |
|
ctrl_get_gen_rec, put a stop statement into ctrl_getrec. Remove |
| 968 |
|
unused and superfluous variables |
| 969 |
|
- ctrl_get_gen_rec.F: catch the case of xx_genperiod=0. |
| 970 |
|
o pkg/ecco |
| 971 |
|
- enable adjoint of Stevens OB |
| 972 |
|
o pkg/ctrl/ecco/grdchk: |
| 973 |
|
- remove global declaration of tmpfldx/yz and tmpfldx/yz2. |
| 974 |
|
Fortunately, we do not have any tests for the numerous cpp-flagged |
| 975 |
|
option of "obcs as control parameters" so we will never know |
| 976 |
|
if this is an improvement (but at least now things compile). |
| 977 |
|
- remove an obsolete initialization from ecco_cost_init_barfiles.F |
| 978 |
|
- fix some problems within never used ifdefs to make the code compile |
| 979 |
|
o pkg/obcs: |
| 980 |
|
- Refine Interior mask definition to work with more complex OB (e.g. |
| 981 |
|
if OB_Ie < OB_Iw); In case default setting is not sufficient, allows to |
| 982 |
|
read in Interior mask from file. Check if Interior mask is consistent |
| 983 |
|
with OB list of indices. |
| 984 |
|
o pkg/seaice |
| 985 |
|
- beginning code for a seaice tracer package, |
| 986 |
|
but currently only used for seaice age: SEAICE_AGE |
| 987 |
|
SEAICE_num = 4 |
| 988 |
|
iTracer = 1: area-weighted ice age |
| 989 |
|
= 2: volume-weighted ice age |
| 990 |
|
= 3, 4: template for one-year vs. multi-year ice |
| 991 |
|
data.seaice has new namelist SEAICE_PARM03 for arrays related to |
| 992 |
|
SEAICE_num |
| 993 |
|
|
| 994 |
|
checkpoint62t (2011/03/05) |
| 995 |
|
o pkg/obcs: |
| 996 |
|
- obcs_apply_uv.F: change to a 2 steps method, first tangential component; |
| 997 |
|
then normal component. This ensures that the normal component is set |
| 998 |
|
correctly even when it conficts with tangential setting from an other OB. |
| 999 |
|
Affects results of test-experiment seaice_obcs: update the output.txt |
| 1000 |
|
o pkg/obcs: |
| 1001 |
|
- add code to monitor statistics of dynamical fields at OB. |
| 1002 |
|
- take OBCS_BALANCE code out of obcs_calc.F into new S/R OBCS_BALANCE_FLOW : |
| 1003 |
|
- fix to OBCS_BALANCE code (right hFac ; deal with simple OB intersection). |
| 1004 |
|
- change default: balance the Net inflow through all OB (instead of balancing |
| 1005 |
|
each OB separately); keep the old method if setting OBC_balanceFacX < 0. |
| 1006 |
|
- switch on useOBCSbalance in dome experiment: only correct Western OB outflow |
| 1007 |
|
to match Nortern OB inflow (OBCS_balanceFacN=0). |
| 1008 |
|
o pkg/seaice: |
| 1009 |
|
- redefined actual sea ice age to be ICEAGE/HEFF ifdef SEAICE_AGE_VOL |
| 1010 |
|
o pkg/seaice (seaice_growth v 1.113): |
| 1011 |
|
- reverting seaice_growth to v 1.111 in order to have code with exact bit |
| 1012 |
|
compatibility with MITgcm_contrib/ifenty/Fenty_seaice_thermo_code_merge |
| 1013 |
|
CPP option SEAICE_OCN_MELT_ACT_ON_AREA removed in v 1.112 is reintroduced |
| 1014 |
|
runtime flag SEAICEareaFormula is kept but is inactive |
| 1015 |
|
o model/src: |
| 1016 |
|
- add new param to scale curvilinear grid distance and area when using |
| 1017 |
|
a different sphere radius from the grid-files. |
| 1018 |
|
- test/check multi-threads EXCH applied to local variable (cg2d_x) |
| 1019 |
|
o adjoint: |
| 1020 |
|
- enable adjoint of Stevens OB |
| 1021 |
|
- add dummy_in_dynamics to enable extraction of ADJrhoinsitu |
| 1022 |
|
(a bit delicate because of potentially modified adj. common block structure; |
| 1023 |
|
current switch is ALLOW_AUTODIFF_MONITOR_DIAG, could be changed later) |
| 1024 |
|
|
| 1025 |
|
checkpoint62s (2011/02/15) |
| 1026 |
|
o pkg/seaice (seaice_growth v 1.112): |
| 1027 |
|
- introduce SEAICEareaFormula run time parameter, which selects the recipe for |
| 1028 |
|
computing the ice cover tendency from effective thickness tendencies. Include |
| 1029 |
|
Ian Fenty's formula. |
| 1030 |
|
- remove 'undef SEAICE_OCN_MELT_ACT_ON_AREA' which may now be achieved using |
| 1031 |
|
SEAICEareaFormula. |
| 1032 |
|
- simplify notations for the various d_HEFF/d_AREA terms. |
| 1033 |
|
o pkg/gmredi: |
| 1034 |
|
- add Boundary-Value-Problem bolus transport (Ferrari etal, OM 2010) code |
| 1035 |
|
from Christopher Wolfe (UCSD); add partial cell factor in vertical derivative; |
| 1036 |
|
add a verification test-run (front_relax.bvp) to test GM_useBVP. |
| 1037 |
|
o pkg/seaice (seaice_growth v 1.111): |
| 1038 |
|
- for SEAICE_DO_OPEN_WATER_GROWTH, replace Gael's implementation with Ian's. |
| 1039 |
|
- the new SEAICE_DO_OPEN_WATER_MELT allows open-water air-sea heat fluxes to |
| 1040 |
|
also melt ice. |
| 1041 |
|
o pkg/seaice: |
| 1042 |
|
- McPhee formula for Ice-Ocean fluxes (provided by Ian Fenty). |
| 1043 |
|
- can be used by #define MCPHEE_OCEAN_ICE_HEAT_FLUX in SEAICE_OPTIONS.h |
| 1044 |
|
o pkg/cheapaml: |
| 1045 |
|
- from Nico Wienders: New version of this package. |
| 1046 |
|
o pkg/flt: |
| 1047 |
|
- add CPP options to prevent floats to re-enter the opposite side of |
| 1048 |
|
a periodic domain (default: keep periodicity) |
| 1049 |
|
o pkg/gmredi: |
| 1050 |
|
- apply exchange to GM_PsiX,GM_PsiY to get valid values near CS-corners |
| 1051 |
|
(needed for multi-dim advection). |
| 1052 |
|
- switch on GM Advective form in global_ocean.cs32x15 and update output.txt |
| 1053 |
|
o model/src: |
| 1054 |
|
- fix wrong re-setting of addMass array (ALLOW_ADDFLUID) for atmospheric run. |
| 1055 |
|
o tutorial_tracer_adjsens |
| 1056 |
|
- extend to NLFS and replace maskC by hFacC |
| 1057 |
|
(change reference output accordingly) |
| 1058 |
|
|
| 1059 |
|
checkpoint62r (2011/01/24) |
| 1060 |
|
o testreport: |
| 1061 |
|
- add option -deloutp (-do) to remove output file after successful run |
| 1062 |
|
- add option -mfile (-mf) for MPI machile-file and for each test-exp, |
| 1063 |
|
extract exact list of procs from this file ; adjust COMMAND by |
| 1064 |
|
replacing TR_MFILE & TR_NPROC with local machile-file & Nb of MPI procs. |
| 1065 |
|
o pkg/ecco: replace old ineffective way of penalizing deviations from |
| 1066 |
|
first guess (of obcs) with code by Matt Mazloff |
| 1067 |
|
o pkg/ctrl: add code that allows the use of ALLOW_OBCS?_CONTROL without |
| 1068 |
|
exf/cal |
| 1069 |
|
- ctrl_init: fix initialisation (computation of diffrec) for obcs-ctrl |
| 1070 |
|
and handle case of obcs?period == 0 |
| 1071 |
|
Not affected by these changes: There is still an inconsistency in |
| 1072 |
|
the computation of start/endrec for the obcs variables for ALLOW_CAL |
| 1073 |
|
defined (see comment by Menemenlis, 7-Mar-2003) |
| 1074 |
|
- make S/R ctrl_get_gen_rec return something meaningful even if |
| 1075 |
|
ALLOW_CAL is not defined |
| 1076 |
|
- ctrl_getobcs?: simplify code so that it does not need to use |
| 1077 |
|
S/R exf_swapfields |
| 1078 |
|
o testreport: |
| 1079 |
|
allow to test MPI with different number of processors for each experiment: |
| 1080 |
|
create file "SIZE.h.mpi" in each build dir. |
| 1081 |
|
Note: update of "-command" argument not yet implemented. |
| 1082 |
|
o model/src, pkg/autodiff, pkg/ecco, pkg/seaice: |
| 1083 |
|
alleviate the need for additional 3D tapes when using |
| 1084 |
|
the non-inear free surface in the adjoint. This is done |
| 1085 |
|
by adding an 'update' of hfacc etc. to their current |
| 1086 |
|
value at the beginning of forward_step.F. |
| 1087 |
|
o pkg/seaice |
| 1088 |
|
seaice_lsr: move stress-divergence computations out of outer pseudo- |
| 1089 |
|
timestepping loop |
| 1090 |
|
o pkg/obcs/obcs_prescribe_read.F: move 'CALL OBCS_EXTERNAL_FIELDS_LOAD' |
| 1091 |
|
up, so that the calls of the ctrl_* subroutines are really at the end |
| 1092 |
|
of the routine. |
| 1093 |
|
o pkg/gmredi: |
| 1094 |
|
allow to specify a 2-D horizontal map and a 1-D vertical profil |
| 1095 |
|
as scaling factor for Isopycnal diffusivity (Redi) and GM diffusivity. |
| 1096 |
|
o pkg/seaice: |
| 1097 |
|
move lsr iterations from seaice_dynsolver to seaice_lsr. |
| 1098 |
|
o pkg/seaice: |
| 1099 |
|
reduce memory footprint of seaice mutldim advection tapes, |
| 1100 |
|
by using a 2D tape independent of the generic_advdiff ones, |
| 1101 |
|
and setting GAD_AREA=1 etc. rather than GAD_AREA=101 etc. |
| 1102 |
|
|
| 1103 |
|
checkpoint62q (2010/12/29) |
| 1104 |
|
o model/src: |
| 1105 |
|
- check horizontal grid-spacing for unset or zero/negative values. |
| 1106 |
|
- change AD test hs94.1x64x5 to be truly a 2-D set-up. |
| 1107 |
|
o pkg/flt: |
| 1108 |
|
- use (recent) mdsio S/R MDS_READVEC_LOC instead of local FLT_MDSREADVECTOR |
| 1109 |
|
- precision of input and output files set according to readBinaryPrec |
| 1110 |
|
and writeBinaryPrec (previously always 64.b), pickup remains at 64 bits. |
| 1111 |
|
- initial float position from tiled file was not converted to index space |
| 1112 |
|
(looks like a bug): fixed to make it similar to initial pos. global file. |
| 1113 |
|
o global_ocean.cs32x15: |
| 1114 |
|
change SIZE.h of AD set-up to get same tiling as with MPI. |
| 1115 |
|
Update output_adm, output_adm.thsice (pass@14) & output_adm.seaice (fail@11) |
| 1116 |
|
o model/src: |
| 1117 |
|
add flag to control printing of domain fields (bathymetry, hFac ...) in |
| 1118 |
|
ini_mask_etc (call to plot_fieds_xy*). |
| 1119 |
|
o pkg/thsice: |
| 1120 |
|
if useEXF, some forcing fields are not valid in overlap: fixed by computing |
| 1121 |
|
over the interior only + apply EXCH at the end (like with SEAICE advection). |
| 1122 |
|
o pkg/mdsio: |
| 1123 |
|
- add 2 arguments to S/R MDS_PASS_R4/8toRL/S for overlap size of buffer |
| 1124 |
|
to copy to/from, so that it can be used to replace MDS_PASS_WH_R8toRL/RS. |
| 1125 |
|
- change barrier call for safe multi-threads access to 3-D shared buffer. |
| 1126 |
|
- MDS_READ/WRITEVEC_LOC use buffers from MDSIO_BUFF_3D.h (instead of local ones) |
| 1127 |
|
- replace call to MDS_PASS_WH_R4/8toRL with call to MDS_PASS_R4/8toRL and |
| 1128 |
|
remove the 4 S/R MDS_PASS_WH_R4/8toRL/S. |
| 1129 |
|
o pkg/flt: |
| 1130 |
|
- add fourth-order Runge-Kutta integration |
| 1131 |
|
- add basic exch2 support (only 1 facet supported so far!) |
| 1132 |
|
o pkg/mdsio+rw: |
| 1133 |
|
new S/R to read local (or tile independent) vector (read counterpart of |
| 1134 |
|
mds_writevec_loc, and front-end write_glvec_rl/rs). |
| 1135 |
|
o global_ocean.cs32x15: additional input_ad.seaice experiment |
| 1136 |
|
that uses pkg/seaice rather than pkg/thsice. |
| 1137 |
|
(using multidim advection for seaice) |
| 1138 |
|
o towards more F95 syntax for array specification in namelist: |
| 1139 |
|
- change "data.diagnostics" (data.ptracers was changed before checkpoint62p) |
| 1140 |
|
of all verification experiments. |
| 1141 |
|
e.g.: line: fields(1,3)='ETAN ','ETANSQ ', |
| 1142 |
|
was changed to: fields(1:2,3)='ETAN ','ETANSQ ', |
| 1143 |
|
- finish adding "-DNML_EXTENDED_F77" to DEFINES list in pgf77 and g77 optfiles |
| 1144 |
|
o global_ocean.cs32x15: additional input.seaice experiment |
| 1145 |
|
that uses pkg/seaice rather than pkg/thsice. |
| 1146 |
|
|
| 1147 |
|
checkpoint62p (2010/12/17) |
| 1148 |
|
o pkg/thsice: rename (ad tape index) iicekey as ticekey, and move it |
| 1149 |
|
to THSICE_SIZE.h to avoid conflict with pkg/seaice. |
| 1150 |
|
o pkg/seaice: add seaice_diffusion.f seaice_map_thsice.f to adjoint. |
| 1151 |
|
allow for use of multim. advection in seaice adjoint. |
| 1152 |
|
o pkg/seaice: add diagnostics for snow precipitation and sublimation flux |
| 1153 |
|
- first attempt to close budgets by taking into account the freshwater |
| 1154 |
|
flux implied by the sublimation of snow and ice. So far only the |
| 1155 |
|
volume budget is closed, turn on with soon-to-be-retired-again CPP-Flag |
| 1156 |
|
SEAICE_ADD_SUBLIMATION_TO_FWBUDGET |
| 1157 |
|
o eesupp/src: |
| 1158 |
|
new S/R with new option (NML_EXTENDED_F77) to change array specification |
| 1159 |
|
from F95 standard to commonly accepted extended F77 form. |
| 1160 |
|
This would allow to use F95 syntax in namelist (needed with gfortran 4.4.5) |
| 1161 |
|
and to revert back to the original syntax (with -DNML_EXTENDED_F77) |
| 1162 |
|
for less tolerant f77 compiler (e.g., pgf77, g77). |
| 1163 |
|
o globalFiles: |
| 1164 |
|
commented out in parameter file "data" since it's not safe with MPI |
| 1165 |
|
(e.g., wrong on aces); keep it if useSingleCpuIO=T (but not doing much |
| 1166 |
|
in this case); no longer commented out in restart test (tst_2+2). |
| 1167 |
|
o add adTapeDir to specify a special directory where adjoint tapes are |
| 1168 |
|
written/read (adread_adwrite.F). adTapeDir is specified in data/PARM05 |
| 1169 |
|
and cannot be used together with mdsioLocalDir. |
| 1170 |
|
o nesting pkgs: |
| 1171 |
|
- packages_init_variables : put call to OBCS_INIT_VARIABLES in last position |
| 1172 |
|
- update Gianmaria's Nesting pkgs (from Contrib/nesting_sannino) |
| 1173 |
|
- add parameter files + utils dir + output from short run on beagle. |
| 1174 |
|
o pkg/seaice: |
| 1175 |
|
fix seaice_growth.F store directives; update results of exp. lab_sea.ad |
| 1176 |
|
o pkg/seaice: explicitly store tsurfloc in seaice_solve4temp.F to avoid |
| 1177 |
|
unnecessary copying in adjoint code, requires new parameter NMAX_ITER |
| 1178 |
|
|
| 1179 |
|
checkpoint62o (2010/11/25) |
| 1180 |
|
o pkg/exf: |
| 1181 |
|
- change CPP option EXF_VERBOSE to run-time logical param exf_verbose. |
| 1182 |
|
o pkg/seaice: |
| 1183 |
|
- seaice_ocean_stress: split a loop for better vectorization of the |
| 1184 |
|
adjoint code |
| 1185 |
|
- seaice_growth.F: fix bug in conservation of freshwater connected to |
| 1186 |
|
capping of heff, put contribution into d_HEFFbyNEG |
| 1187 |
|
- replace irritating parameters SEAICE_latentWater/Ice and SEAICE_sensHeat |
| 1188 |
|
by something more sensible (parameters that are what their name implies) |
| 1189 |
|
- change some defaults, so that by default exf-parameters are used for |
| 1190 |
|
things like rhoAir, cpAir,latent/sensible heat parameters, in order |
| 1191 |
|
to reproduce the old results, many data.seaice now contain funny values |
| 1192 |
|
for cpAir and latent heat of sublimation |
| 1193 |
|
o packages_readparms.F: move exf_readparms up in line so that |
| 1194 |
|
exf-parameters are visible for other packages (e.g. seaice) |
| 1195 |
|
o debugMode & debugLevel=2: |
| 1196 |
|
- differentiate between debugMode & debugLevel=2: |
| 1197 |
|
debugMode: controls printing of debug messages (sequence of S/R calls). |
| 1198 |
|
debugLevel: controls printing of algorithm intermediate results and stats. |
| 1199 |
|
- remove "if debugLevel=2 => debugMode" but keep "debugMode => debugLevel=2"; |
| 1200 |
|
=> allows to print debug-stats without all sequence of call messages. |
| 1201 |
|
- switch control of debug print from debugLevel to debugMode in many S/R. |
| 1202 |
|
o pkg/obcs (obcs-sponge): |
| 1203 |
|
fix bug that was introduced on Oct 25 (in dynamics.F v1.151): put back the |
| 1204 |
|
call to OBCS_PRESCRIBE_EXCHANGES (need to exch gU,gV when using obcs-sponge) |
| 1205 |
o pkg/generic_advdiff: |
o pkg/generic_advdiff: |
| 1206 |
- record overlap minimum size for each advection scheme; |
- record overlap minimum size for each advection scheme; |
| 1207 |
- refine criteria on overlap minimum size; |
- refine criteria on overlap minimum size; |
Parent Directory
| 
Revision Log
| 
Revision Graph
| 
Patch