| 1 | MODULE dynspg_flt_tam |
|---|
| 2 | !!---------------------------------------------------------------------- |
|---|
| 3 | !! This software is governed by the CeCILL licence (Version 2) |
|---|
| 4 | !!---------------------------------------------------------------------- |
|---|
| 5 | #if defined key_tam |
|---|
| 6 | !!====================================================================== |
|---|
| 7 | !! *** MODULE dynspg_flt_tam : TANGENT/ADJOINT OF MODULE dynspg_flt *** |
|---|
| 8 | !! |
|---|
| 9 | !! Ocean dynamics: surface pressure gradient trend |
|---|
| 10 | !! |
|---|
| 11 | !!====================================================================== |
|---|
| 12 | !! History of the direct module: |
|---|
| 13 | !! History OPA ! 1998-05 (G. Roullet) free surface |
|---|
| 14 | !! ! 1998-10 (G. Madec, M. Imbard) release 8.2 |
|---|
| 15 | !! NEMO O.1 ! 2002-08 (G. Madec) F90: Free form and module |
|---|
| 16 | !! - ! 2002-11 (C. Talandier, A-M Treguier) Open boundaries |
|---|
| 17 | !! 1.0 ! 2004-08 (C. Talandier) New trends organization |
|---|
| 18 | !! - ! 2005-11 (V. Garnier) Surface pressure gradient organization |
|---|
| 19 | !! 2.0 ! 2006-07 (S. Masson) distributed restart using iom |
|---|
| 20 | !! - ! 2006-08 (J.Chanut, A.Sellar) Calls to BDY routines. |
|---|
| 21 | !! 3.2 ! 2009-03 (G. Madec, M. Leclair, R. Benshila) introduce sshwzv module |
|---|
| 22 | !! History of the TAM module: |
|---|
| 23 | !! 9.0 ! 2008-08 (A. Vidard) skeleton |
|---|
| 24 | !! - ! 2008-08 (A. Weaver) original version |
|---|
| 25 | !! NEMO 3.2 ! 2010-04 (F. Vigilant) converison to 3.2 |
|---|
| 26 | !!---------------------------------------------------------------------- |
|---|
| 27 | # if defined key_dynspg_flt || defined key_esopa |
|---|
| 28 | !!---------------------------------------------------------------------- |
|---|
| 29 | !! 'key_dynspg_flt' filtered free surface |
|---|
| 30 | !!---------------------------------------------------------------------- |
|---|
| 31 | !!---------------------------------------------------------------------- |
|---|
| 32 | !! dyn_spg_flt_tan : update the momentum trend with the surface pressure |
|---|
| 33 | !! gradient in the filtered free surface case |
|---|
| 34 | !! (tangent routine) |
|---|
| 35 | !! dyn_spg_flt_adj : update the momentum trend with the surface pressure |
|---|
| 36 | !! gradient in the filtered free surface case |
|---|
| 37 | !! (adjoint routine) |
|---|
| 38 | !! dyn_spg_flt_adj_tst : Test of the adjoint routine |
|---|
| 39 | !!---------------------------------------------------------------------- |
|---|
| 40 | USE par_kind , ONLY: & ! Precision variables |
|---|
| 41 | & wp |
|---|
| 42 | USE prtctl ! Print control |
|---|
| 43 | USE par_oce , ONLY: & ! Ocean space and time domain variables |
|---|
| 44 | & jpi, & |
|---|
| 45 | & jpj, & |
|---|
| 46 | & jpk, & |
|---|
| 47 | & jpim1, & |
|---|
| 48 | & jpjm1, & |
|---|
| 49 | & jpkm1, & |
|---|
| 50 | & jpr2di, & |
|---|
| 51 | & jpr2dj, & |
|---|
| 52 | & jpij, & |
|---|
| 53 | & jpiglo, & |
|---|
| 54 | & lk_vopt_loop |
|---|
| 55 | USE in_out_manager, ONLY: & ! I/O manager |
|---|
| 56 | & lwp, & |
|---|
| 57 | & numout, & |
|---|
| 58 | & nit000, & |
|---|
| 59 | & nitend, & |
|---|
| 60 | & ctl_stop, & |
|---|
| 61 | & ln_ctl, & |
|---|
| 62 | & ctmp1 |
|---|
| 63 | USE phycst , ONLY: & ! Physical constants |
|---|
| 64 | & grav |
|---|
| 65 | USE lib_mpp , ONLY: & ! distributed memory computing |
|---|
| 66 | & lk_mpp, & |
|---|
| 67 | & mpp_sum |
|---|
| 68 | USE mppsum , ONLY: & ! Summation of arrays across processors |
|---|
| 69 | & mpp_sum_indep |
|---|
| 70 | USE dom_oce , ONLY: & ! Ocean space and time domain |
|---|
| 71 | & e1u, & |
|---|
| 72 | & e2u, & |
|---|
| 73 | & e1v, & |
|---|
| 74 | & e2v, & |
|---|
| 75 | & e1t, & |
|---|
| 76 | & e2t, & |
|---|
| 77 | # if defined key_zco |
|---|
| 78 | & e3t_0, & |
|---|
| 79 | # else |
|---|
| 80 | & e3u, & |
|---|
| 81 | & e3v, & |
|---|
| 82 | # endif |
|---|
| 83 | & tmask, & |
|---|
| 84 | & umask, & |
|---|
| 85 | & vmask, & |
|---|
| 86 | & bmask, & |
|---|
| 87 | & rdt, & |
|---|
| 88 | & neuler, & |
|---|
| 89 | & n_cla, & |
|---|
| 90 | & mig, & |
|---|
| 91 | & mjg, & |
|---|
| 92 | & nldi, & |
|---|
| 93 | & nldj, & |
|---|
| 94 | & nlei, & |
|---|
| 95 | & nlej, & |
|---|
| 96 | & atfp, & |
|---|
| 97 | & atfp1, & |
|---|
| 98 | & lk_vvl |
|---|
| 99 | USE solver , ONLY: & ! Solvers |
|---|
| 100 | & sol_mat |
|---|
| 101 | USE sol_oce , ONLY: & ! Solver variables in memory |
|---|
| 102 | & nn_solv, & |
|---|
| 103 | & nn_nmod, & |
|---|
| 104 | & ncut, & |
|---|
| 105 | & niter, & |
|---|
| 106 | & eps, & |
|---|
| 107 | & epsr, & |
|---|
| 108 | & rnorme, & |
|---|
| 109 | & res, & |
|---|
| 110 | & c_solver_pt, & |
|---|
| 111 | & gcdprc |
|---|
| 112 | USE oce_tam , ONLY: & ! Tangent-linear and adjoint model variables |
|---|
| 113 | & ub_tl, & |
|---|
| 114 | & vb_tl, & |
|---|
| 115 | & ua_tl, & |
|---|
| 116 | & va_tl, & |
|---|
| 117 | & ub_ad, & |
|---|
| 118 | & vb_ad, & |
|---|
| 119 | & ua_ad, & |
|---|
| 120 | & va_ad, & |
|---|
| 121 | & spgu_tl, & |
|---|
| 122 | & spgv_tl, & |
|---|
| 123 | & spgu_ad, & |
|---|
| 124 | & spgv_ad, & |
|---|
| 125 | & sshn_tl, & |
|---|
| 126 | & sshn_ad |
|---|
| 127 | USE sbc_oce_tam , ONLY: & ! Surface BCs for tangent and adjoint model |
|---|
| 128 | & emp_tl, & |
|---|
| 129 | & emp_ad |
|---|
| 130 | # if defined key_obc |
|---|
| 131 | # if defined key_pomme_r025 |
|---|
| 132 | USE obc_oce |
|---|
| 133 | USE obcdyn_tam |
|---|
| 134 | # else |
|---|
| 135 | Error, OBC not ready. |
|---|
| 136 | # endif |
|---|
| 137 | # endif |
|---|
| 138 | USE sol_oce , ONLY: & ! Solver arrays |
|---|
| 139 | & gcdmat |
|---|
| 140 | USE sol_oce_tam , ONLY: & ! Solver arrays for tangent and adjoint model |
|---|
| 141 | & gcx_tl, & |
|---|
| 142 | & gcxb_tl, & |
|---|
| 143 | & gcb_tl, & |
|---|
| 144 | & gcx_ad, & |
|---|
| 145 | & gcxb_ad, & |
|---|
| 146 | & gcb_ad |
|---|
| 147 | USE solsor_tam , ONLY: & ! SOR solver for tangent and adjoint model |
|---|
| 148 | & sol_sor_tan, & |
|---|
| 149 | & sol_sor_adj |
|---|
| 150 | USE solpcg_tam , ONLY: & ! PCG solver for tangent and adjoint model |
|---|
| 151 | & sol_pcg_tan, & |
|---|
| 152 | & sol_pcg_adj |
|---|
| 153 | USE lbclnk , ONLY: & ! Lateral boundary conditions |
|---|
| 154 | & lbc_lnk, & |
|---|
| 155 | & lbc_lnk_e |
|---|
| 156 | USE lbclnk_tam , ONLY: & ! TAM lateral boundary conditions |
|---|
| 157 | & lbc_lnk_adj, & |
|---|
| 158 | & lbc_lnk_e_adj |
|---|
| 159 | USE gridrandom , ONLY: & ! Random Gaussian noise on grids |
|---|
| 160 | & grid_random |
|---|
| 161 | USE dotprodfld , ONLY: & ! Computes dot product for 3D and 2D fields |
|---|
| 162 | & dot_product |
|---|
| 163 | USE paresp , ONLY: & ! Normalized energy weights |
|---|
| 164 | & wesp_emp, & |
|---|
| 165 | & wesp_ssh |
|---|
| 166 | USE dynadv, ONLY: & |
|---|
| 167 | & ln_dynadv_vec ! vector form flag |
|---|
| 168 | USE cla_dynspg_tam, ONLY: & ! Cross land advection on surface pressure |
|---|
| 169 | & dyn_spg_cla_tan, & |
|---|
| 170 | & dyn_spg_cla_adj |
|---|
| 171 | USE tstool_tam , ONLY: & |
|---|
| 172 | & stdu, & |
|---|
| 173 | & stdv, & |
|---|
| 174 | & stdw, & |
|---|
| 175 | & stdemp, & |
|---|
| 176 | & stdssh, & |
|---|
| 177 | & prntst_adj, & |
|---|
| 178 | & prntst_tlm |
|---|
| 179 | |
|---|
| 180 | IMPLICIT NONE |
|---|
| 181 | PRIVATE |
|---|
| 182 | |
|---|
| 183 | !! * Accessibility |
|---|
| 184 | PUBLIC dyn_spg_flt_tan, & ! routine called by step_tan.F90 |
|---|
| 185 | & dyn_spg_flt_adj, & ! routine called by step_adj.F90 |
|---|
| 186 | & dyn_spg_flt_adj_tst ! routine called by the tst.F90 |
|---|
| 187 | # if defined key_tst_tlm |
|---|
| 188 | PUBLIC dyn_spg_flt_tlm_tst |
|---|
| 189 | # endif |
|---|
| 190 | !! * Substitutions |
|---|
| 191 | # include "domzgr_substitute.h90" |
|---|
| 192 | # include "vectopt_loop_substitute.h90" |
|---|
| 193 | !!---------------------------------------------------------------------- |
|---|
| 194 | CONTAINS |
|---|
| 195 | |
|---|
| 196 | SUBROUTINE dyn_spg_flt_tan( kt, kindic ) |
|---|
| 197 | !!--------------------------------------------------------------------- |
|---|
| 198 | !! *** routine dyn_spg_flt_tan *** |
|---|
| 199 | !! |
|---|
| 200 | !! ** Purpose of the direct routine: |
|---|
| 201 | !! Compute the now trend due to the surface pressure |
|---|
| 202 | !! gradient in case of filtered free surface formulation and add |
|---|
| 203 | !! it to the general trend of momentum equation. |
|---|
| 204 | !! |
|---|
| 205 | !! ** Method of the direct routine: |
|---|
| 206 | !! Filtered free surface formulation. The surface |
|---|
| 207 | !! pressure gradient is given by: |
|---|
| 208 | !! spgu = 1/rau0 d/dx(ps) = 1/e1u di( sshn + btda ) |
|---|
| 209 | !! spgv = 1/rau0 d/dy(ps) = 1/e2v dj( sshn + btda ) |
|---|
| 210 | !! where sshn is the free surface elevation and btda is the after |
|---|
| 211 | !! time derivative of the free surface elevation |
|---|
| 212 | !! -1- evaluate the surface presure trend (including the addi- |
|---|
| 213 | !! tional force) in three steps: |
|---|
| 214 | !! a- compute the right hand side of the elliptic equation: |
|---|
| 215 | !! gcb = 1/(e1t e2t) [ di(e2u spgu) + dj(e1v spgv) ] |
|---|
| 216 | !! where (spgu,spgv) are given by: |
|---|
| 217 | !! spgu = vertical sum[ e3u (ub+ 2 rdt ua ) ] |
|---|
| 218 | !! - grav 2 rdt hu /e1u di[sshn + emp] |
|---|
| 219 | !! spgv = vertical sum[ e3v (vb+ 2 rdt va) ] |
|---|
| 220 | !! - grav 2 rdt hv /e2v dj[sshn + emp] |
|---|
| 221 | !! and define the first guess from previous computation : |
|---|
| 222 | !! zbtd = btda |
|---|
| 223 | !! btda = 2 zbtd - btdb |
|---|
| 224 | !! btdb = zbtd |
|---|
| 225 | !! b- compute the relative accuracy to be reached by the |
|---|
| 226 | !! iterative solver |
|---|
| 227 | !! c- apply the solver by a call to sol... routine |
|---|
| 228 | !! -2- compute and add the free surface pressure gradient inclu- |
|---|
| 229 | !! ding the additional force used to stabilize the equation. |
|---|
| 230 | !! |
|---|
| 231 | !! ** Action : - Update (ua,va) with the surf. pressure gradient trend |
|---|
| 232 | !! |
|---|
| 233 | !! References : Roullet and Madec 1999, JGR. |
|---|
| 234 | !!--------------------------------------------------------------------- |
|---|
| 235 | !! * Arguments |
|---|
| 236 | INTEGER, INTENT( IN ) :: & |
|---|
| 237 | & kt ! ocean time-step index |
|---|
| 238 | INTEGER, INTENT( OUT ) :: & |
|---|
| 239 | & kindic ! solver convergence flag (<0 if not converge) |
|---|
| 240 | |
|---|
| 241 | !! * Local declarations |
|---|
| 242 | INTEGER :: & |
|---|
| 243 | & ji, & ! dummy loop indices |
|---|
| 244 | & jj, & |
|---|
| 245 | & jk |
|---|
| 246 | REAL(wp) :: & |
|---|
| 247 | & z2dt, & ! temporary scalars |
|---|
| 248 | & z2dtg, & |
|---|
| 249 | & zgcb, & |
|---|
| 250 | & zbtd, & |
|---|
| 251 | & ztdgu, & |
|---|
| 252 | & ztdgv |
|---|
| 253 | !! Variable volume |
|---|
| 254 | REAL(wp), DIMENSION(jpi,jpj) :: & ! 2D workspace |
|---|
| 255 | & zsshub, & |
|---|
| 256 | & zsshua, & |
|---|
| 257 | & zsshvb, & |
|---|
| 258 | & zsshva, & |
|---|
| 259 | & zssha |
|---|
| 260 | REAL(wp), DIMENSION(jpi,jpj,jpk) :: & ! 3D workspace |
|---|
| 261 | & zfse3ub, & |
|---|
| 262 | & zfse3ua, & |
|---|
| 263 | & zfse3vb, & |
|---|
| 264 | & zfse3va |
|---|
| 265 | !!---------------------------------------------------------------------- |
|---|
| 266 | ! |
|---|
| 267 | IF( kt == nit000 ) THEN |
|---|
| 268 | |
|---|
| 269 | IF(lwp) WRITE(numout,*) |
|---|
| 270 | IF(lwp) WRITE(numout,*) 'dyn_spg_flt_tan : surface pressure gradient trend' |
|---|
| 271 | IF(lwp) WRITE(numout,*) '~~~~~~~~~~~~~~~ (free surface constant volume case)' |
|---|
| 272 | |
|---|
| 273 | ! set to zero free surface specific arrays |
|---|
| 274 | spgu_tl(:,:) = 0.0_wp ! surface pressure gradient (i-direction) |
|---|
| 275 | spgv_tl(:,:) = 0.0_wp ! surface pressure gradient (j-direction) |
|---|
| 276 | |
|---|
| 277 | ENDIF |
|---|
| 278 | |
|---|
| 279 | ! Local constant initialization |
|---|
| 280 | IF ( neuler == 0 .AND. kt == nit000 ) THEN |
|---|
| 281 | |
|---|
| 282 | z2dt = rdt ! time step: Euler if restart from rest |
|---|
| 283 | CALL sol_mat(kt) ! initialize matrix |
|---|
| 284 | |
|---|
| 285 | ELSEIF ( neuler == 0 .AND. kt == nit000 + 1 ) THEN |
|---|
| 286 | |
|---|
| 287 | z2dt = 2.0_wp * rdt ! time step: leap-frog |
|---|
| 288 | CALL sol_mat(kt) ! initialize matrix |
|---|
| 289 | |
|---|
| 290 | ELSEIF ( neuler /= 0 .AND. kt == nit000 ) THEN |
|---|
| 291 | |
|---|
| 292 | z2dt = 2.0_wp * rdt ! time step: leap-frog |
|---|
| 293 | CALL sol_mat(kt) ! initialize matrix |
|---|
| 294 | |
|---|
| 295 | ELSE |
|---|
| 296 | |
|---|
| 297 | z2dt = 2.0_wp * rdt ! time step: leap-frog |
|---|
| 298 | |
|---|
| 299 | ENDIF |
|---|
| 300 | |
|---|
| 301 | z2dtg = grav * z2dt |
|---|
| 302 | |
|---|
| 303 | ! Evaluate the masked next velocity (effect of the additional force not included) |
|---|
| 304 | |
|---|
| 305 | IF( lk_vvl ) THEN ! variable volume (surface pressure gradient already included in dyn_hpg) |
|---|
| 306 | ! |
|---|
| 307 | IF( ln_dynadv_vec ) THEN ! vector form : applied on velocity |
|---|
| 308 | DO jk = 1, jpkm1 |
|---|
| 309 | DO jj = 2, jpjm1 |
|---|
| 310 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 311 | ua_tl(ji,jj,jk) = ( ub_tl(ji,jj,jk) + z2dt * ua_tl(ji,jj,jk) ) * umask(ji,jj,jk) |
|---|
| 312 | va_tl(ji,jj,jk) = ( vb_tl(ji,jj,jk) + z2dt * va_tl(ji,jj,jk) ) * vmask(ji,jj,jk) |
|---|
| 313 | END DO |
|---|
| 314 | END DO |
|---|
| 315 | END DO |
|---|
| 316 | ! |
|---|
| 317 | ELSE ! flux form : applied on thickness weighted velocity |
|---|
| 318 | DO jk = 1, jpkm1 |
|---|
| 319 | DO jj = 2, jpjm1 |
|---|
| 320 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 321 | ua_tl(ji,jj,jk) = ( ub_tl(ji,jj,jk) * fse3u_b(ji,jj,jk) & |
|---|
| 322 | & + z2dt * ua_tl(ji,jj,jk) * fse3u_n(ji,jj,jk) ) & |
|---|
| 323 | & / fse3u_a(ji,jj,jk) * umask(ji,jj,jk) |
|---|
| 324 | va_tl(ji,jj,jk) = ( vb_tl(ji,jj,jk) * fse3v_b(ji,jj,jk) & |
|---|
| 325 | & + z2dt * va_tl(ji,jj,jk) * fse3v_n(ji,jj,jk) ) & |
|---|
| 326 | & / fse3v_a(ji,jj,jk) * vmask(ji,jj,jk) |
|---|
| 327 | END DO |
|---|
| 328 | END DO |
|---|
| 329 | END DO |
|---|
| 330 | ! |
|---|
| 331 | ENDIF |
|---|
| 332 | ! |
|---|
| 333 | ELSE ! fixed volume (add the surface pressure gradient + unweighted time stepping) |
|---|
| 334 | ! |
|---|
| 335 | DO jj = 2, jpjm1 ! Surface pressure gradient (now) |
|---|
| 336 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 337 | spgu_tl(ji,jj) = - grav * ( sshn_tl(ji+1,jj) - sshn_tl(ji,jj) ) / e1u(ji,jj) |
|---|
| 338 | spgv_tl(ji,jj) = - grav * ( sshn_tl(ji,jj+1) - sshn_tl(ji,jj) ) / e2v(ji,jj) |
|---|
| 339 | END DO |
|---|
| 340 | END DO |
|---|
| 341 | DO jk = 1, jpkm1 ! unweighted time stepping |
|---|
| 342 | DO jj = 2, jpjm1 |
|---|
| 343 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 344 | ua_tl(ji,jj,jk) = ( ub_tl(ji,jj,jk) + z2dt * ( ua_tl(ji,jj,jk) + spgu_tl(ji,jj) ) ) * umask(ji,jj,jk) |
|---|
| 345 | va_tl(ji,jj,jk) = ( vb_tl(ji,jj,jk) + z2dt * ( va_tl(ji,jj,jk) + spgv_tl(ji,jj) ) ) * vmask(ji,jj,jk) |
|---|
| 346 | END DO |
|---|
| 347 | END DO |
|---|
| 348 | END DO |
|---|
| 349 | ! |
|---|
| 350 | ENDIF |
|---|
| 351 | |
|---|
| 352 | # if defined key_obc |
|---|
| 353 | IF( lk_obc ) CALL obc_dyn_tan( kt ) ! Update velocities on each open boundary with the radiation algorithm |
|---|
| 354 | ! IF( lk_obc ) CALL obc_vol_tan( kt ) ! Correction of the barotropic componant velocity to control the volume of the system |
|---|
| 355 | # endif |
|---|
| 356 | # if defined key_bdy |
|---|
| 357 | CALL ctl_stop( 'dyn_spg_flt_tan: key_bdy is not available' ) |
|---|
| 358 | # endif |
|---|
| 359 | # if defined key_agrif |
|---|
| 360 | CALL ctl_stop( 'dyn_spg_flt_tan: key_agrif is not available' ) |
|---|
| 361 | # endif |
|---|
| 362 | # if defined key_orca_r2 |
|---|
| 363 | IF( n_cla == 1 ) CALL dyn_spg_cla_tan( kt ) ! Cross Land Advection (update (ua,va)) |
|---|
| 364 | # endif |
|---|
| 365 | |
|---|
| 366 | ! compute the next vertically averaged velocity (effect of the additional force not included) |
|---|
| 367 | ! --------------------------------------------- |
|---|
| 368 | DO jj = 2, jpjm1 |
|---|
| 369 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 370 | spgu_tl(ji,jj) = 0.0_wp |
|---|
| 371 | spgv_tl(ji,jj) = 0.0_wp |
|---|
| 372 | END DO |
|---|
| 373 | END DO |
|---|
| 374 | |
|---|
| 375 | ! vertical sum |
|---|
| 376 | !CDIR NOLOOPCHG |
|---|
| 377 | IF( lk_vopt_loop ) THEN ! vector opt., forced unroll |
|---|
| 378 | DO jk = 1, jpkm1 |
|---|
| 379 | DO ji = 1, jpij |
|---|
| 380 | spgu_tl(ji,1) = spgu_tl(ji,1) + fse3u(ji,1,jk) * ua_tl(ji,1,jk) |
|---|
| 381 | spgv_tl(ji,1) = spgv_tl(ji,1) + fse3v(ji,1,jk) * va_tl(ji,1,jk) |
|---|
| 382 | END DO |
|---|
| 383 | END DO |
|---|
| 384 | ELSE ! No vector opt. |
|---|
| 385 | DO jk = 1, jpkm1 |
|---|
| 386 | DO jj = 2, jpjm1 |
|---|
| 387 | DO ji = 2, jpim1 |
|---|
| 388 | spgu_tl(ji,jj) = spgu_tl(ji,jj) + fse3u(ji,jj,jk) * ua_tl(ji,jj,jk) |
|---|
| 389 | spgv_tl(ji,jj) = spgv_tl(ji,jj) + fse3v(ji,jj,jk) * va_tl(ji,jj,jk) |
|---|
| 390 | END DO |
|---|
| 391 | END DO |
|---|
| 392 | END DO |
|---|
| 393 | ENDIF |
|---|
| 394 | |
|---|
| 395 | ! transport: multiplied by the horizontal scale factor |
|---|
| 396 | DO jj = 2, jpjm1 |
|---|
| 397 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 398 | spgu_tl(ji,jj) = spgu_tl(ji,jj) * e2u(ji,jj) |
|---|
| 399 | spgv_tl(ji,jj) = spgv_tl(ji,jj) * e1v(ji,jj) |
|---|
| 400 | END DO |
|---|
| 401 | END DO |
|---|
| 402 | |
|---|
| 403 | CALL lbc_lnk( spgu_tl, 'U', -1.0_wp ) ! lateral boundary conditions |
|---|
| 404 | CALL lbc_lnk( spgv_tl, 'V', -1.0_wp ) |
|---|
| 405 | |
|---|
| 406 | IF( lk_vvl ) CALL ctl_stop( 'dyn_spg_flt_tan: lk_vvl is not available' ) |
|---|
| 407 | |
|---|
| 408 | ! Right hand side of the elliptic equation and first guess |
|---|
| 409 | ! ----------------------------------------------------------- |
|---|
| 410 | DO jj = 2, jpjm1 |
|---|
| 411 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 412 | ! Divergence of the after vertically averaged velocity |
|---|
| 413 | zgcb = spgu_tl(ji,jj) - spgu_tl(ji-1,jj) & |
|---|
| 414 | & + spgv_tl(ji,jj) - spgv_tl(ji,jj-1) |
|---|
| 415 | gcb_tl(ji,jj) = gcdprc(ji,jj) * zgcb |
|---|
| 416 | ! First guess of the after barotropic transport divergence |
|---|
| 417 | zbtd = gcx_tl(ji,jj) |
|---|
| 418 | gcx_tl (ji,jj) = 2.0_wp * zbtd - gcxb_tl(ji,jj) |
|---|
| 419 | gcxb_tl(ji,jj) = zbtd |
|---|
| 420 | END DO |
|---|
| 421 | END DO |
|---|
| 422 | ! apply the lateral boundary conditions |
|---|
| 423 | IF( nn_solv == 2 .AND. MAX( jpr2di, jpr2dj ) > 0 ) CALL lbc_lnk_e( gcb_tl, c_solver_pt, 1.0_wp ) |
|---|
| 424 | # if defined key_agrif |
|---|
| 425 | CALL ctl_stop( 'dyn_spg_flt_tan: key_agrif is not available' ) |
|---|
| 426 | # endif |
|---|
| 427 | |
|---|
| 428 | ! Relative precision (computation on one processor) |
|---|
| 429 | ! ------------------ |
|---|
| 430 | rnorme = SUM( gcb_tl(1:jpi,1:jpj) * gcdmat(1:jpi,1:jpj) * gcb_tl(1:jpi,1:jpj) * bmask(:,:) ) |
|---|
| 431 | IF( lk_mpp ) CALL mpp_sum( rnorme ) ! sum over the global domain |
|---|
| 432 | |
|---|
| 433 | epsr = eps * eps * rnorme |
|---|
| 434 | ncut = 0 |
|---|
| 435 | ! if rnorme is 0, the solution is 0, the solver is not called |
|---|
| 436 | IF( rnorme == 0.0_wp ) THEN |
|---|
| 437 | gcx_tl(:,:) = 0.0_wp |
|---|
| 438 | res = 0.0_wp |
|---|
| 439 | niter = 0 |
|---|
| 440 | ncut = 999 |
|---|
| 441 | ENDIF |
|---|
| 442 | |
|---|
| 443 | ! Evaluate the next transport divergence |
|---|
| 444 | ! -------------------------------------- |
|---|
| 445 | ! Iterarive solver for the elliptic equation (except IF sol.=0) |
|---|
| 446 | ! (output in gcx with boundary conditions applied) |
|---|
| 447 | kindic = 0 |
|---|
| 448 | IF( ncut == 0 ) THEN |
|---|
| 449 | IF ( nn_solv == 1 ) THEN ; CALL sol_pcg_tan( kt, kindic ) ! diagonal preconditioned conjuguate gradient |
|---|
| 450 | ELSEIF( nn_solv == 2 ) THEN ; CALL sol_sor_tan( kt, kindic ) ! successive-over-relaxation |
|---|
| 451 | ENDIF |
|---|
| 452 | ENDIF |
|---|
| 453 | |
|---|
| 454 | ! Transport divergence gradient multiplied by z2dt |
|---|
| 455 | ! --------------------------------------------==== |
|---|
| 456 | DO jj = 2, jpjm1 |
|---|
| 457 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 458 | ! trend of Transport divergence gradient |
|---|
| 459 | ztdgu = z2dtg * ( gcx_tl(ji+1,jj ) - gcx_tl(ji,jj) ) / e1u(ji,jj) |
|---|
| 460 | ztdgv = z2dtg * ( gcx_tl(ji ,jj+1) - gcx_tl(ji,jj) ) / e2v(ji,jj) |
|---|
| 461 | ! multiplied by z2dt |
|---|
| 462 | # if defined key_obc |
|---|
| 463 | ! caution : grad D = 0 along open boundaries |
|---|
| 464 | IF( Agrif_Root() ) THEN |
|---|
| 465 | spgu_tl(ji,jj) = z2dt * ztdgu * obcumask(ji,jj) |
|---|
| 466 | spgv_tl(ji,jj) = z2dt * ztdgv * obcvmask(ji,jj) |
|---|
| 467 | ELSE |
|---|
| 468 | spgu_tl(ji,jj) = z2dt * ztdgu |
|---|
| 469 | spgv_tl(ji,jj) = z2dt * ztdgv |
|---|
| 470 | ENDIF |
|---|
| 471 | # elif defined key_bdy |
|---|
| 472 | CALL ctl_stop( 'dyn_spg_flt_tan: key_bdy is not available' ) |
|---|
| 473 | # else |
|---|
| 474 | spgu_tl(ji,jj) = z2dt * ztdgu |
|---|
| 475 | spgv_tl(ji,jj) = z2dt * ztdgv |
|---|
| 476 | # endif |
|---|
| 477 | END DO |
|---|
| 478 | END DO |
|---|
| 479 | |
|---|
| 480 | # if defined key_agrif |
|---|
| 481 | CALL ctl_stop( 'dyn_spg_flt_tan: key_agrif is not available' ) |
|---|
| 482 | # endif |
|---|
| 483 | |
|---|
| 484 | ! Add the trends multiplied by z2dt to the after velocity |
|---|
| 485 | ! ------------------------------------------------------- |
|---|
| 486 | ! ( c a u t i o n : (ua_tl,va_tl) here are the after velocity not the |
|---|
| 487 | ! trend, the leap-frog time stepping will not |
|---|
| 488 | ! be done in dynnxt_tan.F90 routine) |
|---|
| 489 | DO jk = 1, jpkm1 |
|---|
| 490 | DO jj = 2, jpjm1 |
|---|
| 491 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 492 | ua_tl(ji,jj,jk) = ( ua_tl(ji,jj,jk) + spgu_tl(ji,jj) ) * umask(ji,jj,jk) |
|---|
| 493 | va_tl(ji,jj,jk) = ( va_tl(ji,jj,jk) + spgv_tl(ji,jj) ) * vmask(ji,jj,jk) |
|---|
| 494 | END DO |
|---|
| 495 | END DO |
|---|
| 496 | END DO |
|---|
| 497 | |
|---|
| 498 | ! |
|---|
| 499 | END SUBROUTINE dyn_spg_flt_tan |
|---|
| 500 | |
|---|
| 501 | SUBROUTINE dyn_spg_flt_adj( kt, kindic ) |
|---|
| 502 | !!---------------------------------------------------------------------- |
|---|
| 503 | !! *** routine dyn_spg_flt_adj *** |
|---|
| 504 | !! |
|---|
| 505 | !! ** Purpose of the direct routine: |
|---|
| 506 | !! Compute the now trend due to the surface pressure |
|---|
| 507 | !! gradient in case of filtered free surface formulation and add |
|---|
| 508 | !! it to the general trend of momentum equation. |
|---|
| 509 | !! |
|---|
| 510 | !! ** Method of the direct routine: |
|---|
| 511 | !! Filtered free surface formulation. The surface |
|---|
| 512 | !! pressure gradient is given by: |
|---|
| 513 | !! spgu = 1/rau0 d/dx(ps) = 1/e1u di( sshn + btda ) |
|---|
| 514 | !! spgv = 1/rau0 d/dy(ps) = 1/e2v dj( sshn + btda ) |
|---|
| 515 | !! where sshn is the free surface elevation and btda is the after |
|---|
| 516 | !! time derivative of the free surface elevation |
|---|
| 517 | !! -1- evaluate the surface presure trend (including the addi- |
|---|
| 518 | !! tional force) in three steps: |
|---|
| 519 | !! a- compute the right hand side of the elliptic equation: |
|---|
| 520 | !! gcb = 1/(e1t e2t) [ di(e2u spgu) + dj(e1v spgv) ] |
|---|
| 521 | !! where (spgu,spgv) are given by: |
|---|
| 522 | !! spgu = vertical sum[ e3u (ub+ 2 rdt ua ) ] |
|---|
| 523 | !! - grav 2 rdt hu /e1u di[sshn + emp] |
|---|
| 524 | !! spgv = vertical sum[ e3v (vb+ 2 rdt va) ] |
|---|
| 525 | !! - grav 2 rdt hv /e2v dj[sshn + emp] |
|---|
| 526 | !! and define the first guess from previous computation : |
|---|
| 527 | !! zbtd = btda |
|---|
| 528 | !! btda = 2 zbtd - btdb |
|---|
| 529 | !! btdb = zbtd |
|---|
| 530 | !! b- compute the relative accuracy to be reached by the |
|---|
| 531 | !! iterative solver |
|---|
| 532 | !! c- apply the solver by a call to sol... routine |
|---|
| 533 | !! -2- compute and add the free surface pressure gradient inclu- |
|---|
| 534 | !! ding the additional force used to stabilize the equation. |
|---|
| 535 | !! |
|---|
| 536 | !! ** Action : - Update (ua,va) with the surf. pressure gradient trend |
|---|
| 537 | !! |
|---|
| 538 | !! References : Roullet and Madec 1999, JGR. |
|---|
| 539 | !!--------------------------------------------------------------------- |
|---|
| 540 | !! * Arguments |
|---|
| 541 | INTEGER, INTENT( IN ) :: & |
|---|
| 542 | & kt ! ocean time-step index |
|---|
| 543 | INTEGER, INTENT( OUT ) :: & |
|---|
| 544 | & kindic ! solver convergence flag (<0 if not converge) |
|---|
| 545 | |
|---|
| 546 | !! * Local declarations |
|---|
| 547 | INTEGER :: & |
|---|
| 548 | & ji, & ! dummy loop indices |
|---|
| 549 | & jj, & |
|---|
| 550 | & jk |
|---|
| 551 | REAL(wp) :: & |
|---|
| 552 | & z2dt, & ! temporary scalars |
|---|
| 553 | & z2dtg, & |
|---|
| 554 | & zgcb, & |
|---|
| 555 | & zbtd, & |
|---|
| 556 | & ztdgu, & |
|---|
| 557 | & ztdgv |
|---|
| 558 | !!---------------------------------------------------------------------- |
|---|
| 559 | ! |
|---|
| 560 | IF( kt == nitend ) THEN |
|---|
| 561 | IF(lwp) WRITE(numout,*) |
|---|
| 562 | IF(lwp) WRITE(numout,*) 'dyn_spg_flt_adj : surface pressure gradient trend' |
|---|
| 563 | IF(lwp) WRITE(numout,*) '~~~~~~~~~~~~~~~ (free surface constant volume case)' |
|---|
| 564 | ENDIF |
|---|
| 565 | |
|---|
| 566 | ! Local constant initialization |
|---|
| 567 | IF ( neuler == 0 .AND. kt == nit000 ) THEN |
|---|
| 568 | |
|---|
| 569 | z2dt = rdt ! time step: Euler if restart from rest |
|---|
| 570 | CALL sol_mat(kt) ! initialize matrix |
|---|
| 571 | |
|---|
| 572 | ELSEIF ( neuler == 0 .AND. kt == nit000 + 1 ) THEN |
|---|
| 573 | |
|---|
| 574 | z2dt = 2.0_wp * rdt ! time step: leap-frog |
|---|
| 575 | CALL sol_mat(kt) ! reinitialize matrix |
|---|
| 576 | |
|---|
| 577 | ELSEIF ( kt == nitend ) THEN |
|---|
| 578 | |
|---|
| 579 | z2dt = 2.0_wp * rdt ! time step: leap-frog |
|---|
| 580 | CALL sol_mat(kt) ! reinitialize matrix |
|---|
| 581 | |
|---|
| 582 | ELSEIF ( neuler /= 0 .AND. kt == nit000 ) THEN |
|---|
| 583 | |
|---|
| 584 | z2dt = 2.0_wp * rdt ! time step: leap-frog |
|---|
| 585 | CALL sol_mat(kt) ! initialize matrix |
|---|
| 586 | |
|---|
| 587 | ELSE |
|---|
| 588 | |
|---|
| 589 | z2dt = 2.0_wp * rdt ! time step: leap-frog |
|---|
| 590 | |
|---|
| 591 | ENDIF |
|---|
| 592 | |
|---|
| 593 | z2dtg = grav * z2dt |
|---|
| 594 | |
|---|
| 595 | ! Add the trends multiplied by z2dt to the after velocity |
|---|
| 596 | ! ----------------------------------------------------------- |
|---|
| 597 | ! ( c a u t i o n : (ua_ad,va_ad) here are the after velocity not the |
|---|
| 598 | ! trend, the leap-frog time stepping will not |
|---|
| 599 | ! be done in dynnxt_adj.F90 routine) |
|---|
| 600 | DO jk = 1, jpkm1 |
|---|
| 601 | DO jj = 2, jpjm1 |
|---|
| 602 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 603 | ua_ad(ji,jj,jk) = ua_ad(ji,jj,jk) * umask(ji,jj,jk) |
|---|
| 604 | va_ad(ji,jj,jk) = va_ad(ji,jj,jk) * vmask(ji,jj,jk) |
|---|
| 605 | spgu_ad(ji,jj) = spgu_ad(ji,jj) + ua_ad(ji,jj,jk) |
|---|
| 606 | spgv_ad(ji,jj) = spgv_ad(ji,jj) + va_ad(ji,jj,jk) |
|---|
| 607 | END DO |
|---|
| 608 | END DO |
|---|
| 609 | END DO |
|---|
| 610 | |
|---|
| 611 | # if defined key_agrif |
|---|
| 612 | CALL ctl_stop( 'dyn_spg_flt_adj: key_agrif is not available' ) |
|---|
| 613 | # endif |
|---|
| 614 | |
|---|
| 615 | ! Transport divergence gradient multiplied by z2dt |
|---|
| 616 | ! --------------------------------------------==== |
|---|
| 617 | DO jj = 2, jpjm1 |
|---|
| 618 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 619 | ! multiplied by z2dt |
|---|
| 620 | # if defined key_obc |
|---|
| 621 | ! caution : grad D = 0 along open boundaries |
|---|
| 622 | IF( Agrif_Root() ) THEN |
|---|
| 623 | ztdgu = z2dt * spgu_ad(ji,jj) * obcumask(ji,jj) |
|---|
| 624 | ztdgv = z2dt * spgv_ad(ji,jj) * obcvmask(ji,jj) |
|---|
| 625 | spgu_ad(ji,jj) = 0.0_wp |
|---|
| 626 | spgv_ad(ji,jj) = 0.0_wp |
|---|
| 627 | ELSE |
|---|
| 628 | ztdgu = z2dt * spgu_ad(ji,jj) |
|---|
| 629 | ztdgv = z2dt * spgv_ad(ji,jj) |
|---|
| 630 | spgu_ad(ji,jj) = 0.0_wp |
|---|
| 631 | spgv_ad(ji,jj) = 0.0_wp |
|---|
| 632 | ENDIF |
|---|
| 633 | # elif defined key_bdy |
|---|
| 634 | CALL ctl_stop( 'dyn_spg_flt_adj: key_bdy is not available' ) |
|---|
| 635 | # else |
|---|
| 636 | ztdgu = z2dt * spgu_ad(ji,jj) |
|---|
| 637 | ztdgv = z2dt * spgv_ad(ji,jj) |
|---|
| 638 | spgu_ad(ji,jj) = 0.0_wp |
|---|
| 639 | spgv_ad(ji,jj) = 0.0_wp |
|---|
| 640 | # endif |
|---|
| 641 | ! trend of Transport divergence gradient |
|---|
| 642 | ztdgu = ztdgu * z2dtg / e1u(ji,jj) |
|---|
| 643 | ztdgv = ztdgv * z2dtg / e2v(ji,jj) |
|---|
| 644 | gcx_ad(ji ,jj ) = gcx_ad(ji ,jj ) - ztdgu - ztdgv |
|---|
| 645 | gcx_ad(ji ,jj+1) = gcx_ad(ji ,jj+1) + ztdgv |
|---|
| 646 | gcx_ad(ji+1,jj ) = gcx_ad(ji+1,jj ) + ztdgu |
|---|
| 647 | END DO |
|---|
| 648 | END DO |
|---|
| 649 | |
|---|
| 650 | ! Evaluate the next transport divergence |
|---|
| 651 | ! -------------------------------------- |
|---|
| 652 | ! Iterative solver for the elliptic equation (except IF sol.=0) |
|---|
| 653 | ! (output in gcx_ad with boundary conditions applied) |
|---|
| 654 | |
|---|
| 655 | kindic = 0 |
|---|
| 656 | ncut = 0 ! Force solver |
|---|
| 657 | IF( ncut == 0 ) THEN |
|---|
| 658 | IF ( nn_solv == 1 ) THEN ; CALL sol_pcg_adj( kt, kindic ) ! diagonal preconditioned conjuguate gradient |
|---|
| 659 | ELSEIF( nn_solv == 2 ) THEN ; CALL sol_sor_adj( kt, kindic ) ! successive-over-relaxation |
|---|
| 660 | ENDIF |
|---|
| 661 | ENDIF |
|---|
| 662 | # if defined key_agrif |
|---|
| 663 | CALL ctl_stop( 'dyn_spg_flt_adj: key_agrif is not available' ) |
|---|
| 664 | # endif |
|---|
| 665 | |
|---|
| 666 | ! Right hand side of the elliptic equation and first guess |
|---|
| 667 | ! -------------------------------------------------------- |
|---|
| 668 | ! apply the lateral boundary conditions |
|---|
| 669 | IF( nn_solv == 2 .AND. MAX( jpr2di, jpr2dj ) > 0 ) & |
|---|
| 670 | & CALL lbc_lnk_e_adj( gcb_ad, c_solver_pt, 1.0_wp ) |
|---|
| 671 | |
|---|
| 672 | DO jj = 2, jpjm1 |
|---|
| 673 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 674 | ! First guess of the after barotropic transport divergence |
|---|
| 675 | zbtd = gcxb_ad(ji,jj) + 2.0_wp * gcx_ad(ji,jj) |
|---|
| 676 | gcxb_ad(ji,jj) = - gcx_ad(ji,jj) |
|---|
| 677 | gcx_ad (ji,jj) = zbtd |
|---|
| 678 | ! Divergence of the after vertically averaged velocity |
|---|
| 679 | zgcb = gcb_ad(ji,jj) * gcdprc(ji,jj) |
|---|
| 680 | gcb_ad(ji,jj) = 0.0_wp |
|---|
| 681 | spgu_ad(ji-1,jj ) = spgu_ad(ji-1,jj ) - zgcb |
|---|
| 682 | spgu_ad(ji ,jj ) = spgu_ad(ji ,jj ) + zgcb |
|---|
| 683 | spgv_ad(ji ,jj-1) = spgv_ad(ji ,jj-1) - zgcb |
|---|
| 684 | spgv_ad(ji ,jj ) = spgv_ad(ji ,jj ) + zgcb |
|---|
| 685 | END DO |
|---|
| 686 | END DO |
|---|
| 687 | |
|---|
| 688 | IF( lk_vvl ) CALL ctl_stop( 'dyn_spg_flt_adj: lk_vvl is not available' ) |
|---|
| 689 | |
|---|
| 690 | ! Boundary conditions on (spgu_ad,spgv_ad) |
|---|
| 691 | CALL lbc_lnk_adj( spgu_ad, 'U', -1.0_wp ) |
|---|
| 692 | CALL lbc_lnk_adj( spgv_ad, 'V', -1.0_wp ) |
|---|
| 693 | |
|---|
| 694 | ! transport: multiplied by the horizontal scale factor |
|---|
| 695 | DO jj = 2,jpjm1 |
|---|
| 696 | DO ji = fs_2,fs_jpim1 ! vector opt. |
|---|
| 697 | spgu_ad(ji,jj) = spgu_ad(ji,jj) * e2u(ji,jj) |
|---|
| 698 | spgv_ad(ji,jj) = spgv_ad(ji,jj) * e1v(ji,jj) |
|---|
| 699 | END DO |
|---|
| 700 | END DO |
|---|
| 701 | |
|---|
| 702 | ! compute the next vertically averaged velocity (effect of the additional force not included) |
|---|
| 703 | ! --------------------------------------------- |
|---|
| 704 | |
|---|
| 705 | ! vertical sum |
|---|
| 706 | !CDIR NOLOOPCHG |
|---|
| 707 | IF( lk_vopt_loop ) THEN ! vector opt., forced unroll |
|---|
| 708 | DO jk = 1, jpkm1 |
|---|
| 709 | DO ji = 1, jpij |
|---|
| 710 | ua_ad(ji,1,jk) = ua_ad(ji,1,jk) + fse3u(ji,1,jk) * spgu_ad(ji,1) |
|---|
| 711 | va_ad(ji,1,jk) = va_ad(ji,1,jk) + fse3v(ji,1,jk) * spgv_ad(ji,1) |
|---|
| 712 | END DO |
|---|
| 713 | END DO |
|---|
| 714 | ELSE ! No vector opt. |
|---|
| 715 | DO jk = 1, jpkm1 |
|---|
| 716 | DO jj = 2, jpjm1 |
|---|
| 717 | DO ji = 2, jpim1 |
|---|
| 718 | ua_ad(ji,jj,jk) = ua_ad(ji,jj,jk) + fse3u(ji,jj,jk) * spgu_ad(ji,jj) |
|---|
| 719 | va_ad(ji,jj,jk) = va_ad(ji,jj,jk) + fse3v(ji,jj,jk) * spgv_ad(ji,jj) |
|---|
| 720 | END DO |
|---|
| 721 | END DO |
|---|
| 722 | END DO |
|---|
| 723 | ENDIF |
|---|
| 724 | |
|---|
| 725 | DO jj = 2, jpjm1 |
|---|
| 726 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 727 | spgu_ad(ji,jj) = 0.0_wp |
|---|
| 728 | spgv_ad(ji,jj) = 0.0_wp |
|---|
| 729 | END DO |
|---|
| 730 | END DO |
|---|
| 731 | |
|---|
| 732 | # if defined key_orca_r2 |
|---|
| 733 | IF( n_cla == 1 ) CALL dyn_spg_cla_adj( kt ) ! Cross Land Advection (update (ua,va)) |
|---|
| 734 | # endif |
|---|
| 735 | # if defined key_agrif |
|---|
| 736 | CALL ctl_stop( 'dyn_spg_flt_adj: key_agrif is not available' ) |
|---|
| 737 | # endif |
|---|
| 738 | # if defined key_bdy |
|---|
| 739 | CALL ctl_stop( 'dyn_spg_flt_adj: key_bdy is not available' ) |
|---|
| 740 | # endif |
|---|
| 741 | # if defined key_obc |
|---|
| 742 | ! IF( lk_obc ) CALL obc_vol_adj( kt ) ! Correction of the barotropic componant velocity to control the volume of the system |
|---|
| 743 | IF( lk_obc ) CALL obc_dyn_adj( kt ) ! Update velocities on each open boundary with the radiation algorithm |
|---|
| 744 | # endif |
|---|
| 745 | |
|---|
| 746 | ! Evaluate the masked next velocity (effect of the additional force not included) |
|---|
| 747 | IF( lk_vvl ) THEN ! variable volume (surface pressure gradient already included in dyn_hpg) |
|---|
| 748 | ! |
|---|
| 749 | IF( ln_dynadv_vec ) THEN ! vector form : applied on velocity |
|---|
| 750 | DO jk = 1, jpkm1 |
|---|
| 751 | DO jj = 2, jpjm1 |
|---|
| 752 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 753 | ub_ad(ji,jj,jk) = ub_ad(ji,jj,jk) + z2dt * ua_ad(ji,jj,jk) * umask(ji,jj,jk) |
|---|
| 754 | ua_ad(ji,jj,jk) = z2dt * ua_ad(ji,jj,jk) * umask(ji,jj,jk) |
|---|
| 755 | vb_ad(ji,jj,jk) = vb_ad(ji,jj,jk) + z2dt * va_ad(ji,jj,jk) * vmask(ji,jj,jk) |
|---|
| 756 | va_ad(ji,jj,jk) = z2dt * va_ad(ji,jj,jk) * vmask(ji,jj,jk) |
|---|
| 757 | END DO |
|---|
| 758 | END DO |
|---|
| 759 | END DO |
|---|
| 760 | ! |
|---|
| 761 | ELSE ! flux form : applied on thickness weighted velocity |
|---|
| 762 | DO jk = 1, jpkm1 |
|---|
| 763 | DO jj = 2, jpjm1 |
|---|
| 764 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 765 | ua_ad(ji,jj,jk) = ua_ad(ji,jj,jk) / fse3u_a(ji,jj,jk) * umask(ji,jj,jk) |
|---|
| 766 | ub_ad(ji,jj,jk) = ub_ad(ji,jj,jk) + ua_ad(ji,jj,jk) * fse3u_b(ji,jj,jk) |
|---|
| 767 | ua_ad(ji,jj,jk) = ua_ad(ji,jj,jk) * z2dt * fse3u_n(ji,jj,jk) |
|---|
| 768 | va_ad(ji,jj,jk) = va_ad(ji,jj,jk) / fse3v_a(ji,jj,jk) * vmask(ji,jj,jk) |
|---|
| 769 | vb_ad(ji,jj,jk) = vb_ad(ji,jj,jk) + va_ad(ji,jj,jk) * fse3v_b(ji,jj,jk) |
|---|
| 770 | va_ad(ji,jj,jk) = va_ad(ji,jj,jk) * z2dt * fse3v_n(ji,jj,jk) |
|---|
| 771 | END DO |
|---|
| 772 | END DO |
|---|
| 773 | END DO |
|---|
| 774 | ! |
|---|
| 775 | ENDIF |
|---|
| 776 | ! |
|---|
| 777 | ELSE ! fixed volume (add the surface pressure gradient + unweighted time stepping) |
|---|
| 778 | ! |
|---|
| 779 | DO jk = 1, jpkm1 ! unweighted time stepping |
|---|
| 780 | DO jj = 2, jpjm1 |
|---|
| 781 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 782 | ua_ad( ji,jj,jk) = ua_ad(ji,jj,jk) * umask(ji,jj,jk) |
|---|
| 783 | ub_ad( ji,jj,jk) = ub_ad(ji,jj,jk) + ua_ad(ji,jj,jk) |
|---|
| 784 | spgu_ad(ji,jj ) = spgu_ad(ji,jj) + ua_ad(ji,jj,jk) * z2dt |
|---|
| 785 | ua_ad( ji,jj,jk) = ua_ad(ji,jj,jk) * z2dt |
|---|
| 786 | va_ad( ji,jj,jk) = va_ad(ji,jj,jk) * vmask(ji,jj,jk) |
|---|
| 787 | vb_ad( ji,jj,jk) = vb_ad(ji,jj,jk) + va_ad(ji,jj,jk) |
|---|
| 788 | spgv_ad(ji,jj ) = spgv_ad(ji,jj) + va_ad(ji,jj,jk) * z2dt |
|---|
| 789 | va_ad( ji,jj,jk) = va_ad(ji,jj,jk) * z2dt |
|---|
| 790 | END DO |
|---|
| 791 | END DO |
|---|
| 792 | END DO |
|---|
| 793 | DO jj = 2, jpjm1 ! Surface pressure gradient (now) |
|---|
| 794 | DO ji = fs_2, fs_jpim1 ! vector opt. |
|---|
| 795 | spgu_ad(ji ,jj ) = spgu_ad(ji ,jj ) * grav / e1u(ji,jj) |
|---|
| 796 | spgv_ad(ji ,jj ) = spgv_ad(ji ,jj ) * grav / e2v(ji,jj) |
|---|
| 797 | sshn_ad(ji ,jj ) = sshn_ad(ji ,jj ) + spgv_ad(ji,jj) |
|---|
| 798 | sshn_ad(ji ,jj+1) = sshn_ad(ji ,jj+1) - spgv_ad(ji,jj) |
|---|
| 799 | sshn_ad(ji ,jj ) = sshn_ad(ji ,jj ) + spgu_ad(ji,jj) |
|---|
| 800 | sshn_ad(ji+1,jj ) = sshn_ad(ji+1,jj ) - spgu_ad(ji,jj) |
|---|
| 801 | spgu_ad(ji ,jj ) = 0.0_wp |
|---|
| 802 | spgv_ad(ji ,jj ) = 0.0_wp |
|---|
| 803 | END DO |
|---|
| 804 | END DO |
|---|
| 805 | ENDIF |
|---|
| 806 | |
|---|
| 807 | IF( kt == nit000 ) THEN |
|---|
| 808 | ! set to zero free surface specific arrays |
|---|
| 809 | spgu_ad(:,:) = 0.0_wp ! surface pressure gradient (i-direction) |
|---|
| 810 | spgv_ad(:,:) = 0.0_wp ! surface pressure gradient (j-direction) |
|---|
| 811 | ENDIF |
|---|
| 812 | |
|---|
| 813 | END SUBROUTINE dyn_spg_flt_adj |
|---|
| 814 | |
|---|
| 815 | SUBROUTINE dyn_spg_flt_adj_tst( kumadt ) |
|---|
| 816 | !!----------------------------------------------------------------------- |
|---|
| 817 | !! |
|---|
| 818 | !! *** ROUTINE dyn_spg_flt_adj_tst *** |
|---|
| 819 | !! |
|---|
| 820 | !! ** Purpose : Test the adjoint routine. |
|---|
| 821 | !! |
|---|
| 822 | !! ** Method : Verify the scalar product |
|---|
| 823 | !! |
|---|
| 824 | !! ( L dx )^T W dy = dx^T L^T W dy |
|---|
| 825 | !! |
|---|
| 826 | !! where L = tangent routine |
|---|
| 827 | !! L^T = adjoint routine |
|---|
| 828 | !! W = diagonal matrix of scale factors |
|---|
| 829 | !! dx = input perturbation (random field) |
|---|
| 830 | !! dy = L dx |
|---|
| 831 | !! |
|---|
| 832 | !! ** Action : |
|---|
| 833 | !! |
|---|
| 834 | !! History : |
|---|
| 835 | !! ! 09-01 (A. Weaver) |
|---|
| 836 | !!----------------------------------------------------------------------- |
|---|
| 837 | !! * Modules used |
|---|
| 838 | |
|---|
| 839 | !! * Arguments |
|---|
| 840 | INTEGER, INTENT(IN) :: & |
|---|
| 841 | & kumadt ! Output unit |
|---|
| 842 | |
|---|
| 843 | !! * Local declarations |
|---|
| 844 | REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: & |
|---|
| 845 | & zua_tlin, & ! Tangent input: ua_tl |
|---|
| 846 | & zva_tlin, & ! Tangent input: va_tl |
|---|
| 847 | & zub_tlin, & ! Tangent input: ub_tl |
|---|
| 848 | & zvb_tlin, & ! Tangent input: vb_tl |
|---|
| 849 | & zua_tlout, & ! Tangent output: ua_tl |
|---|
| 850 | & zva_tlout, & ! Tangent output: va_tl |
|---|
| 851 | & zub_tlout, & ! Tangent output: ua_tl |
|---|
| 852 | & zvb_tlout, & ! Tangent output: va_tl |
|---|
| 853 | & zub_adin, & ! Tangent output: ua_ad |
|---|
| 854 | & zvb_adin, & ! Tangent output: va_ad |
|---|
| 855 | & zua_adin, & ! Adjoint input: ua_ad |
|---|
| 856 | & zva_adin, & ! Adjoint input: va_ad |
|---|
| 857 | & zua_adout, & ! Adjoint output: ua_ad |
|---|
| 858 | & zva_adout, & ! Adjoint output: va_ad |
|---|
| 859 | & zub_adout, & ! Adjoint oputput: ub_ad |
|---|
| 860 | & zvb_adout, & ! Adjoint output: vb_ad |
|---|
| 861 | & znu ! 3D random field for u |
|---|
| 862 | |
|---|
| 863 | REAL(wp), DIMENSION(:,:), ALLOCATABLE :: & |
|---|
| 864 | & zgcx_tlin, zgcxb_tlin, zgcb_tlin, zgcx_tlout, zgcxb_tlout, zgcb_tlout, & |
|---|
| 865 | & zgcx_adin, zgcxb_adin, zgcb_adin, zgcx_adout, zgcxb_adout, zgcb_adout, & |
|---|
| 866 | & zspgu_tlout, zspgv_tlout, zspgu_adin, zspgv_adin |
|---|
| 867 | |
|---|
| 868 | REAL(KIND=wp), DIMENSION(:,:), ALLOCATABLE :: & |
|---|
| 869 | & zsshn_tlin, & ! Tangent input: sshn_tl |
|---|
| 870 | & zsshn_adout,& ! Adjoint output: sshn_ad |
|---|
| 871 | & zemp_tlin, & ! Tangent input: emp_tl |
|---|
| 872 | & zemp_adout, & ! Adjoint output: emp_ad |
|---|
| 873 | & znssh ! 2D random field for SSH |
|---|
| 874 | REAL(wp) :: & |
|---|
| 875 | & zsp1, & ! scalar product involving the tangent routine |
|---|
| 876 | & zsp2 ! scalar product involving the adjoint routine |
|---|
| 877 | INTEGER, DIMENSION(jpi,jpj) :: & |
|---|
| 878 | & iseed_2d ! 2D seed for the random number generator |
|---|
| 879 | INTEGER :: & |
|---|
| 880 | & indic, & |
|---|
| 881 | & istp |
|---|
| 882 | INTEGER :: & |
|---|
| 883 | & ji, & |
|---|
| 884 | & jj, & |
|---|
| 885 | & jk, & |
|---|
| 886 | & kmod, & |
|---|
| 887 | & jstp |
|---|
| 888 | CHARACTER (LEN=14) :: & |
|---|
| 889 | & cl_name |
|---|
| 890 | INTEGER :: & |
|---|
| 891 | & jpert |
|---|
| 892 | INTEGER, PARAMETER :: jpertmax = 7 |
|---|
| 893 | |
|---|
| 894 | ! Allocate memory |
|---|
| 895 | |
|---|
| 896 | ALLOCATE( & |
|---|
| 897 | & zua_tlin(jpi,jpj,jpk), & |
|---|
| 898 | & zva_tlin(jpi,jpj,jpk), & |
|---|
| 899 | & zub_tlin(jpi,jpj,jpk), & |
|---|
| 900 | & zvb_tlin(jpi,jpj,jpk), & |
|---|
| 901 | & zua_tlout(jpi,jpj,jpk), & |
|---|
| 902 | & zva_tlout(jpi,jpj,jpk), & |
|---|
| 903 | & zua_adin(jpi,jpj,jpk), & |
|---|
| 904 | & zva_adin(jpi,jpj,jpk), & |
|---|
| 905 | & zua_adout(jpi,jpj,jpk), & |
|---|
| 906 | & zva_adout(jpi,jpj,jpk), & |
|---|
| 907 | & zub_adout(jpi,jpj,jpk), & |
|---|
| 908 | & zvb_adout(jpi,jpj,jpk), & |
|---|
| 909 | & znu(jpi,jpj,jpk) & |
|---|
| 910 | & ) |
|---|
| 911 | ALLOCATE( & |
|---|
| 912 | & zsshn_tlin(jpi,jpj), & |
|---|
| 913 | & zsshn_adout(jpi,jpj),& |
|---|
| 914 | & zemp_tlin(jpi,jpj), & |
|---|
| 915 | & zemp_adout(jpi,jpj), & |
|---|
| 916 | & znssh(jpi,jpj) & |
|---|
| 917 | & ) |
|---|
| 918 | |
|---|
| 919 | # if defined key_obc |
|---|
| 920 | ALLOCATE( zemp_tlout(jpi,jpj), zemp_adin (jpi,jpj), & |
|---|
| 921 | zwn_tlout (jpi,jpj), zwn_adin (jpi,jpj) ) |
|---|
| 922 | # endif |
|---|
| 923 | |
|---|
| 924 | ALLOCATE( zgcx_tlin (jpi,jpj), zgcx_tlout (jpi,jpj), zgcx_adin (jpi,jpj), zgcx_adout (jpi,jpj), & |
|---|
| 925 | zgcxb_tlin(jpi,jpj), zgcxb_tlout(jpi,jpj), zgcxb_adin(jpi,jpj), zgcxb_adout(jpi,jpj), & |
|---|
| 926 | zgcb_tlin (jpi,jpj), zgcb_tlout (jpi,jpj), zgcb_adin (jpi,jpj), zgcb_adout (jpi,jpj) & |
|---|
| 927 | & ) |
|---|
| 928 | |
|---|
| 929 | ALLOCATE ( zub_tlout(jpi,jpj,jpk), zvb_tlout(jpi,jpj,jpk), & |
|---|
| 930 | zub_adin (jpi,jpj,jpk), zvb_adin (jpi,jpj,jpk) ) |
|---|
| 931 | |
|---|
| 932 | ALLOCATE( zspgu_tlout (jpi,jpj), zspgv_tlout (jpi,jpj), zspgu_adin (jpi,jpj), zspgv_adin (jpi,jpj)) |
|---|
| 933 | |
|---|
| 934 | !========================================================================= |
|---|
| 935 | ! dx = ( sshb_tl, sshn_tl, ub_tl, ua_tl, vb_tl, va_tl, wn_tl, emp_tl ) |
|---|
| 936 | ! and dy = ( sshb_tl, sshn_tl, ua_tl, va_tl ) |
|---|
| 937 | !========================================================================= |
|---|
| 938 | |
|---|
| 939 | ! Test for time steps nit000 and nit000 + 1 (the matrix changes) |
|---|
| 940 | |
|---|
| 941 | DO jstp = nit000, nit000 + 1 |
|---|
| 942 | DO jpert = 1, jpertmax |
|---|
| 943 | istp = jstp |
|---|
| 944 | |
|---|
| 945 | !-------------------------------------------------------------------- |
|---|
| 946 | ! Reset the tangent and adjoint variables |
|---|
| 947 | !-------------------------------------------------------------------- |
|---|
| 948 | |
|---|
| 949 | zub_tlin (:,:,:) = 0.0_wp |
|---|
| 950 | zvb_tlin (:,:,:) = 0.0_wp |
|---|
| 951 | zua_tlin (:,:,:) = 0.0_wp |
|---|
| 952 | zva_tlin (:,:,:) = 0.0_wp |
|---|
| 953 | zua_tlout(:,:,:) = 0.0_wp |
|---|
| 954 | zva_tlout(:,:,:) = 0.0_wp |
|---|
| 955 | zua_adin (:,:,:) = 0.0_wp |
|---|
| 956 | zva_adin (:,:,:) = 0.0_wp |
|---|
| 957 | zub_adout(:,:,:) = 0.0_wp |
|---|
| 958 | zvb_adout(:,:,:) = 0.0_wp |
|---|
| 959 | zua_adout(:,:,:) = 0.0_wp |
|---|
| 960 | zva_adout(:,:,:) = 0.0_wp |
|---|
| 961 | |
|---|
| 962 | zsshn_tlin (:,:) = 0.0_wp |
|---|
| 963 | zemp_tlin (:,:) = 0.0_wp |
|---|
| 964 | zsshn_adout(:,:) = 0.0_wp |
|---|
| 965 | zemp_adout (:,:) = 0.0_wp |
|---|
| 966 | zspgu_adin (:,:) = 0.0_wp |
|---|
| 967 | zspgv_adin (:,:) = 0.0_wp |
|---|
| 968 | zspgu_tlout(:,:) = 0.0_wp |
|---|
| 969 | zspgv_tlout(:,:) = 0.0_wp |
|---|
| 970 | |
|---|
| 971 | zgcx_tlout (:,:) = 0.0_wp ; zgcx_adin (:,:) = 0.0_wp ; zgcx_adout (:,:) = 0.0_wp |
|---|
| 972 | zgcxb_tlout(:,:) = 0.0_wp ; zgcxb_adin(:,:) = 0.0_wp ; zgcxb_adout(:,:) = 0.0_wp |
|---|
| 973 | zgcb_tlout (:,:) = 0.0_wp ; zgcb_adin (:,:) = 0.0_wp ; zgcb_adout (:,:) = 0.0_wp |
|---|
| 974 | |
|---|
| 975 | ub_tl(:,:,:) = 0.0_wp |
|---|
| 976 | vb_tl(:,:,:) = 0.0_wp |
|---|
| 977 | ua_tl(:,:,:) = 0.0_wp |
|---|
| 978 | va_tl(:,:,:) = 0.0_wp |
|---|
| 979 | sshn_tl(:,:) = 0.0_wp |
|---|
| 980 | emp_tl(:,:) = 0.0_wp |
|---|
| 981 | gcb_tl(:,:) = 0.0_wp |
|---|
| 982 | gcx_tl(:,:) = 0.0_wp |
|---|
| 983 | gcxb_tl(:,:) = 0.0_wp |
|---|
| 984 | spgu_tl(:,:) = 0.0_wp |
|---|
| 985 | spgv_tl(:,:) = 0.0_wp |
|---|
| 986 | ub_ad(:,:,:) = 0.0_wp |
|---|
| 987 | vb_ad(:,:,:) = 0.0_wp |
|---|
| 988 | ua_ad(:,:,:) = 0.0_wp |
|---|
| 989 | va_ad(:,:,:) = 0.0_wp |
|---|
| 990 | sshn_ad(:,:) = 0.0_wp |
|---|
| 991 | emp_ad(:,:) = 0.0_wp |
|---|
| 992 | gcb_ad(:,:) = 0.0_wp |
|---|
| 993 | gcx_ad(:,:) = 0.0_wp |
|---|
| 994 | gcxb_ad(:,:) = 0.0_wp |
|---|
| 995 | spgu_ad(:,:) = 0.0_wp |
|---|
| 996 | spgv_ad(:,:) = 0.0_wp |
|---|
| 997 | !-------------------------------------------------------------------- |
|---|
| 998 | ! Initialize the tangent input with random noise: dx |
|---|
| 999 | !-------------------------------------------------------------------- |
|---|
| 1000 | IF ( (jpert == 1) .OR. (jpert == jpertmax) ) THEN |
|---|
| 1001 | |
|---|
| 1002 | DO jj = 1, jpj |
|---|
| 1003 | DO ji = 1, jpi |
|---|
| 1004 | iseed_2d(ji,jj) = - ( 456953 + & |
|---|
| 1005 | & mig(ji) + ( mjg(jj) - 1 ) * jpiglo ) |
|---|
| 1006 | END DO |
|---|
| 1007 | END DO |
|---|
| 1008 | CALL grid_random( iseed_2d, znu, 'U', 0.0_wp, stdu ) |
|---|
| 1009 | |
|---|
| 1010 | DO jk = 1, jpk |
|---|
| 1011 | DO jj = nldj, nlej |
|---|
| 1012 | DO ji = nldi, nlei |
|---|
| 1013 | zua_tlin(ji,jj,jk) = znu(ji,jj,jk) |
|---|
| 1014 | END DO |
|---|
| 1015 | END DO |
|---|
| 1016 | END DO |
|---|
| 1017 | |
|---|
| 1018 | ENDIF |
|---|
| 1019 | IF ( (jpert == 2) .OR. (jpert == jpertmax) ) THEN |
|---|
| 1020 | DO jj = 1, jpj |
|---|
| 1021 | DO ji = 1, jpi |
|---|
| 1022 | iseed_2d(ji,jj) = - ( 3434334 + & |
|---|
| 1023 | & mig(ji) + ( mjg(jj) - 1 ) * jpiglo ) |
|---|
| 1024 | END DO |
|---|
| 1025 | END DO |
|---|
| 1026 | CALL grid_random( iseed_2d, znu, 'V', 0.0_wp, stdv ) |
|---|
| 1027 | |
|---|
| 1028 | DO jk = 1, jpk |
|---|
| 1029 | DO jj = nldj, nlej |
|---|
| 1030 | DO ji = nldi, nlei |
|---|
| 1031 | zva_tlin(ji,jj,jk) = znu(ji,jj,jk) |
|---|
| 1032 | END DO |
|---|
| 1033 | END DO |
|---|
| 1034 | END DO |
|---|
| 1035 | |
|---|
| 1036 | ENDIF |
|---|
| 1037 | IF ( (jpert == 3) .OR. (jpert == jpertmax) ) THEN |
|---|
| 1038 | DO jj = 1, jpj |
|---|
| 1039 | DO ji = 1, jpi |
|---|
| 1040 | iseed_2d(ji,jj) = - ( 935678 + & |
|---|
| 1041 | & mig(ji) + ( mjg(jj) - 1 ) * jpiglo ) |
|---|
| 1042 | END DO |
|---|
| 1043 | END DO |
|---|
| 1044 | CALL grid_random( iseed_2d, znu, 'U', 0.0_wp, stdu ) |
|---|
| 1045 | |
|---|
| 1046 | DO jk = 1, jpk |
|---|
| 1047 | DO jj = nldj, nlej |
|---|
| 1048 | DO ji = nldi, nlei |
|---|
| 1049 | zub_tlin(ji,jj,jk) = znu(ji,jj,jk) |
|---|
| 1050 | END DO |
|---|
| 1051 | END DO |
|---|
| 1052 | END DO |
|---|
| 1053 | |
|---|
| 1054 | ENDIF |
|---|
| 1055 | IF ( (jpert == 4) .OR. (jpert == jpertmax) ) THEN |
|---|
| 1056 | DO jj = 1, jpj |
|---|
| 1057 | DO ji = 1, jpi |
|---|
| 1058 | iseed_2d(ji,jj) = - ( 1462578 + & |
|---|
| 1059 | & mig(ji) + ( mjg(jj) - 1 ) * jpiglo ) |
|---|
| 1060 | END DO |
|---|
| 1061 | END DO |
|---|
| 1062 | CALL grid_random( iseed_2d, znu, 'V', 0.0_wp, stdv ) |
|---|
| 1063 | |
|---|
| 1064 | DO jk = 1, jpk |
|---|
| 1065 | DO jj = nldj, nlej |
|---|
| 1066 | DO ji = nldi, nlei |
|---|
| 1067 | zvb_tlin(ji,jj,jk) = znu(ji,jj,jk) |
|---|
| 1068 | END DO |
|---|
| 1069 | END DO |
|---|
| 1070 | END DO |
|---|
| 1071 | |
|---|
| 1072 | ENDIF |
|---|
| 1073 | IF ( (jpert == 5) .OR. (jpert == jpertmax) ) THEN |
|---|
| 1074 | DO jj = 1, jpj |
|---|
| 1075 | DO ji = 1, jpi |
|---|
| 1076 | iseed_2d(ji,jj) = - ( 25674910 + & |
|---|
| 1077 | & mig(ji) + ( mjg(jj) - 1 ) * jpiglo ) |
|---|
| 1078 | END DO |
|---|
| 1079 | END DO |
|---|
| 1080 | CALL grid_random( iseed_2d, znssh, 'T', 0.0_wp, stdemp ) |
|---|
| 1081 | |
|---|
| 1082 | DO jj = nldj, nlej |
|---|
| 1083 | DO ji = nldi, nlei |
|---|
| 1084 | zemp_tlin(ji,jj) = znssh(ji,jj) |
|---|
| 1085 | END DO |
|---|
| 1086 | END DO |
|---|
| 1087 | |
|---|
| 1088 | ENDIF |
|---|
| 1089 | IF ( (jpert == 6) .OR. (jpert == jpertmax) ) THEN |
|---|
| 1090 | DO jj = 1, jpj |
|---|
| 1091 | DO ji = 1, jpi |
|---|
| 1092 | iseed_2d(ji,jj) = - ( 12672456 + & |
|---|
| 1093 | & mig(ji) + ( mjg(jj) - 1 ) * jpiglo ) |
|---|
| 1094 | END DO |
|---|
| 1095 | END DO |
|---|
| 1096 | CALL grid_random( iseed_2d, znssh, 'T', 0.0_wp, stdssh ) |
|---|
| 1097 | DO jj = nldj, nlej |
|---|
| 1098 | DO ji = nldi, nlei |
|---|
| 1099 | zsshn_tlin(ji,jj) = znssh(ji,jj) |
|---|
| 1100 | END DO |
|---|
| 1101 | END DO |
|---|
| 1102 | END IF |
|---|
| 1103 | # if defined key_obc |
|---|
| 1104 | zgcx_tlin (:,:) = ( zua_tlin(:,:,1) + zub_tlin(:,:,1) ) / 10. |
|---|
| 1105 | zgcxb_tlin (:,:) = ( zua_tlin(:,:,2) + zub_tlin(:,:,2) ) / 10. |
|---|
| 1106 | zgcb_tlin (:,:) = ( zua_tlin(:,:,3) + zub_tlin(:,:,3) ) / 10. |
|---|
| 1107 | # endif |
|---|
| 1108 | zgcx_tlin (:,:) = ( zua_tlin(:,:,1) + zub_tlin(:,:,1) ) / 10. |
|---|
| 1109 | zgcxb_tlin (:,:) = ( zua_tlin(:,:,2) + zub_tlin(:,:,2) ) / 10. |
|---|
| 1110 | !-------------------------------------------------------------------- |
|---|
| 1111 | ! Call the tangent routine: dy = L dx |
|---|
| 1112 | !-------------------------------------------------------------------- |
|---|
| 1113 | |
|---|
| 1114 | ua_tl(:,:,:) = zua_tlin(:,:,:) |
|---|
| 1115 | va_tl(:,:,:) = zva_tlin(:,:,:) |
|---|
| 1116 | ub_tl(:,:,:) = zub_tlin(:,:,:) |
|---|
| 1117 | vb_tl(:,:,:) = zvb_tlin(:,:,:) |
|---|
| 1118 | emp_tl (:,:) = zemp_tlin (:,:) |
|---|
| 1119 | sshn_tl(:,:) = zsshn_tlin(:,:) |
|---|
| 1120 | |
|---|
| 1121 | # if defined key_obc |
|---|
| 1122 | gcb_tl (:,:) = zgcb_tlin (:,:) |
|---|
| 1123 | # else |
|---|
| 1124 | gcx_tl (:,:) = 0.e0 ; gcxb_tl(:,:) = 0.e0 |
|---|
| 1125 | gcb_tl (:,:) = 0.e0 |
|---|
| 1126 | gcx_tl (:,:) = zgcx_tlin (:,:) ; gcxb_tl(:,:) = zgcxb_tlin(:,:) |
|---|
| 1127 | # endif |
|---|
| 1128 | |
|---|
| 1129 | CALL dyn_spg_flt_tan( istp, indic ) |
|---|
| 1130 | |
|---|
| 1131 | zua_tlout(:,:,:) = ua_tl(:,:,:) ; zva_tlout(:,:,:) = va_tl(:,:,:) |
|---|
| 1132 | # if defined key_obc |
|---|
| 1133 | zgcx_tlout (:,:) = gcx_tl (:,:) ; zgcxb_tlout(:,:) = gcxb_tl(:,:) |
|---|
| 1134 | # endif |
|---|
| 1135 | zspgu_tlout(:,:) = spgu_tl(:,:) ; zspgv_tlout(:,:) = spgv_tl(:,:) |
|---|
| 1136 | zgcb_tlout (:,:) = gcb_tl (:,:) |
|---|
| 1137 | |
|---|
| 1138 | !-------------------------------------------------------------------- |
|---|
| 1139 | ! Initialize the adjoint variables: dy^* = W dy |
|---|
| 1140 | !-------------------------------------------------------------------- |
|---|
| 1141 | |
|---|
| 1142 | DO jk = 1, jpk |
|---|
| 1143 | DO jj = nldj, nlej |
|---|
| 1144 | DO ji = nldi, nlei |
|---|
| 1145 | zua_adin(ji,jj,jk) = zua_tlout(ji,jj,jk) & |
|---|
| 1146 | & * e1u(ji,jj) * e2u(ji,jj) * fse3u(ji,jj,jk) & |
|---|
| 1147 | & * umask(ji,jj,jk) |
|---|
| 1148 | zva_adin(ji,jj,jk) = zva_tlout(ji,jj,jk) & |
|---|
| 1149 | & * e1v(ji,jj) * e2v(ji,jj) * fse3v(ji,jj,jk) & |
|---|
| 1150 | & * vmask(ji,jj,jk) |
|---|
| 1151 | END DO |
|---|
| 1152 | END DO |
|---|
| 1153 | END DO |
|---|
| 1154 | # if defined key_obc |
|---|
| 1155 | DO jj = nldj, nlej |
|---|
| 1156 | DO ji = nldi, nlei |
|---|
| 1157 | zgcx_adin (ji,jj) = zgcx_tlout (ji,jj) & |
|---|
| 1158 | & * e1t(ji,jj) * e2t(ji,jj) * fse3u(ji,jj,1) * tmask(ji,jj,1) |
|---|
| 1159 | zgcxb_adin(ji,jj) = zgcxb_tlout(ji,jj) & |
|---|
| 1160 | & * e1t(ji,jj) * e2t(ji,jj) * fse3u(ji,jj,1) * tmask(ji,jj,1) |
|---|
| 1161 | END DO |
|---|
| 1162 | END DO |
|---|
| 1163 | # endif |
|---|
| 1164 | DO jj = nldj, nlej |
|---|
| 1165 | DO ji = nldi, nlei |
|---|
| 1166 | zgcb_adin (ji,jj) = zgcb_tlout (ji,jj) & |
|---|
| 1167 | & * e1t(ji,jj) * e2t(ji,jj) * fse3u(ji,jj,1) * tmask(ji,jj,1) |
|---|
| 1168 | zspgu_adin (ji,jj) = zspgu_tlout (ji,jj) & |
|---|
| 1169 | & * e1u(ji,jj) * e2u(ji,jj) * fse3u(ji,jj,1) * umask(ji,jj,1) |
|---|
| 1170 | zspgv_adin(ji,jj) = zspgv_tlout(ji,jj) & |
|---|
| 1171 | & * e1v(ji,jj) * e2v(ji,jj) * fse3v(ji,jj,1) * vmask(ji,jj,1) |
|---|
| 1172 | END DO |
|---|
| 1173 | END DO |
|---|
| 1174 | |
|---|
| 1175 | !-------------------------------------------------------------------- |
|---|
| 1176 | ! Compute the scalar product: ( L dx )^T W dy |
|---|
| 1177 | !-------------------------------------------------------------------- |
|---|
| 1178 | |
|---|
| 1179 | zsp1 = DOT_PRODUCT( zua_tlout , zua_adin ) & |
|---|
| 1180 | # if defined key_obc |
|---|
| 1181 | & + DOT_PRODUCT( zgcx_tlout , zgcx_adin ) & |
|---|
| 1182 | & + DOT_PRODUCT( zgcxb_tlout, zgcxb_adin ) & |
|---|
| 1183 | # endif |
|---|
| 1184 | & + DOT_PRODUCT( zgcb_tlout , zgcb_adin ) & |
|---|
| 1185 | & + DOT_PRODUCT( zspgu_tlout , zspgu_adin ) & |
|---|
| 1186 | & + DOT_PRODUCT( zspgv_tlout , zspgv_adin ) & |
|---|
| 1187 | & + DOT_PRODUCT( zva_tlout , zva_adin ) |
|---|
| 1188 | |
|---|
| 1189 | |
|---|
| 1190 | !-------------------------------------------------------------------- |
|---|
| 1191 | ! Call the adjoint routine: dx^* = L^T dy^* |
|---|
| 1192 | !-------------------------------------------------------------------- |
|---|
| 1193 | |
|---|
| 1194 | ua_ad(:,:,:) = zua_adin(:,:,:) |
|---|
| 1195 | va_ad(:,:,:) = zva_adin(:,:,:) |
|---|
| 1196 | |
|---|
| 1197 | # if defined key_obc |
|---|
| 1198 | gcx_ad (:,:) = zgcx_adin (:,:) ; gcxb_ad(:,:) = zgcxb_adin(:,:) |
|---|
| 1199 | # else |
|---|
| 1200 | gcx_ad (:,:) = 0.0_wp ; gcxb_ad(:,:) = 0.0_wp |
|---|
| 1201 | gcb_ad (:,:) = zgcb_adin (:,:) |
|---|
| 1202 | spgu_ad(:,:) = zspgu_adin(:,:) |
|---|
| 1203 | spgv_ad(:,:) = zspgv_adin(:,:) |
|---|
| 1204 | ub_ad (:,:,:) = zub_adin (:,:,:) ; vb_ad (:,:,:) = zvb_adin (:,:,:) |
|---|
| 1205 | # endif |
|---|
| 1206 | |
|---|
| 1207 | CALL dyn_spg_flt_adj( istp, indic ) |
|---|
| 1208 | |
|---|
| 1209 | zua_adout(:,:,:) = ua_ad(:,:,:) |
|---|
| 1210 | zva_adout(:,:,:) = va_ad(:,:,:) |
|---|
| 1211 | zub_adout(:,:,:) = ub_ad(:,:,:) |
|---|
| 1212 | zvb_adout(:,:,:) = vb_ad(:,:,:) |
|---|
| 1213 | zemp_adout (:,:) = emp_ad (:,:) |
|---|
| 1214 | zsshn_adout(:,:) = sshn_ad(:,:) |
|---|
| 1215 | zgcx_adout (:,:) = gcx_ad (:,:) |
|---|
| 1216 | zgcxb_adout(:,:) = gcxb_ad(:,:) |
|---|
| 1217 | # if defined key_obc |
|---|
| 1218 | zgcb_adout (:,:) = gcb_ad (:,:) |
|---|
| 1219 | # endif |
|---|
| 1220 | |
|---|
| 1221 | !-------------------------------------------------------------------- |
|---|
| 1222 | ! Compute the scalar product: dx^T L^T W dy |
|---|
| 1223 | !-------------------------------------------------------------------- |
|---|
| 1224 | |
|---|
| 1225 | zsp2 = DOT_PRODUCT( zua_tlin , zua_adout ) & |
|---|
| 1226 | & + DOT_PRODUCT( zva_tlin , zva_adout ) & |
|---|
| 1227 | & + DOT_PRODUCT( zub_tlin , zub_adout ) & |
|---|
| 1228 | & + DOT_PRODUCT( zvb_tlin , zvb_adout ) & |
|---|
| 1229 | & + DOT_PRODUCT( zgcx_tlin , zgcx_adout ) & |
|---|
| 1230 | & + DOT_PRODUCT( zgcxb_tlin, zgcxb_adout ) & |
|---|
| 1231 | # if defined key_obc |
|---|
| 1232 | & + DOT_PRODUCT( zgcb_tlin , zgcb_adout ) & |
|---|
| 1233 | # endif |
|---|
| 1234 | & + DOT_PRODUCT( zsshn_tlin, zsshn_adout ) & |
|---|
| 1235 | & + DOT_PRODUCT( zemp_tlin , zemp_adout ) |
|---|
| 1236 | |
|---|
| 1237 | ! Compare the scalar products |
|---|
| 1238 | |
|---|
| 1239 | ! 14 char:'12345678901234' |
|---|
| 1240 | IF ( istp == nit000 ) THEN |
|---|
| 1241 | SELECT CASE (jpert) |
|---|
| 1242 | CASE(1) |
|---|
| 1243 | cl_name = 'spg_flt Ua T1' |
|---|
| 1244 | CASE(2) |
|---|
| 1245 | cl_name = 'spg_flt Va T1' |
|---|
| 1246 | CASE(3) |
|---|
| 1247 | cl_name = 'spg_flt Ub T1' |
|---|
| 1248 | CASE(4) |
|---|
| 1249 | cl_name = 'spg_flt Vb T1' |
|---|
| 1250 | CASE(5) |
|---|
| 1251 | cl_name = 'spg_flt emp T1' |
|---|
| 1252 | CASE(6) |
|---|
| 1253 | cl_name = 'spg_flt ssh T1' |
|---|
| 1254 | CASE(jpertmax) |
|---|
| 1255 | cl_name = 'dyn_spg_flt T1' |
|---|
| 1256 | END SELECT |
|---|
| 1257 | ELSEIF ( istp == nit000 + 1 ) THEN |
|---|
| 1258 | SELECT CASE (jpert) |
|---|
| 1259 | CASE(1) |
|---|
| 1260 | cl_name = 'spg_flt Ua T2' |
|---|
| 1261 | CASE(2) |
|---|
| 1262 | cl_name = 'spg_flt Va T2' |
|---|
| 1263 | CASE(3) |
|---|
| 1264 | cl_name = 'spg_flt Ub T2' |
|---|
| 1265 | CASE(4) |
|---|
| 1266 | cl_name = 'spg_flt Vb T2' |
|---|
| 1267 | CASE(5) |
|---|
| 1268 | cl_name = 'spg_flt emp T2' |
|---|
| 1269 | CASE(6) |
|---|
| 1270 | cl_name = 'spg_flt ssh T2' |
|---|
| 1271 | CASE(jpertmax) |
|---|
| 1272 | cl_name = 'dyn_spg_flt T2' |
|---|
| 1273 | END SELECT |
|---|
| 1274 | END IF |
|---|
| 1275 | CALL prntst_adj( cl_name, kumadt, zsp1, zsp2 ) |
|---|
| 1276 | |
|---|
| 1277 | END DO |
|---|
| 1278 | END DO |
|---|
| 1279 | |
|---|
| 1280 | !nn_nmod = kmod ! restore initial frequency of test for the SOR solver |
|---|
| 1281 | |
|---|
| 1282 | ! Deallocate memory |
|---|
| 1283 | |
|---|
| 1284 | DEALLOCATE( & |
|---|
| 1285 | & zua_tlin, & |
|---|
| 1286 | & zva_tlin, & |
|---|
| 1287 | & zub_tlin, & |
|---|
| 1288 | & zvb_tlin, & |
|---|
| 1289 | & zua_tlout, & |
|---|
| 1290 | & zva_tlout, & |
|---|
| 1291 | & zua_adin, & |
|---|
| 1292 | & zva_adin, & |
|---|
| 1293 | & zua_adout, & |
|---|
| 1294 | & zva_adout, & |
|---|
| 1295 | & zub_adout, & |
|---|
| 1296 | & zvb_adout, & |
|---|
| 1297 | & znu & |
|---|
| 1298 | & ) |
|---|
| 1299 | DEALLOCATE( & |
|---|
| 1300 | & zsshn_tlin, & |
|---|
| 1301 | & zemp_tlin, & |
|---|
| 1302 | & zsshn_adout,& |
|---|
| 1303 | & zemp_adout, & |
|---|
| 1304 | & znssh & |
|---|
| 1305 | & ) |
|---|
| 1306 | DEALLOCATE( zgcx_tlin , zgcx_tlout , zgcx_adin , zgcx_adout, & |
|---|
| 1307 | & zgcxb_tlin, zgcxb_tlout, zgcxb_adin, zgcxb_adout, & |
|---|
| 1308 | & zgcb_tlin , zgcb_tlout , zgcb_adin , zgcb_adout & |
|---|
| 1309 | & ) |
|---|
| 1310 | DEALLOCATE ( zub_tlout, zvb_tlout, zub_adin , zvb_adin ) |
|---|
| 1311 | # if defined key_obc |
|---|
| 1312 | DEALLOCATE( zemp_tlout, zemp_adin , & |
|---|
| 1313 | zwn_tlout , zwn_adin ) |
|---|
| 1314 | # endif |
|---|
| 1315 | END SUBROUTINE dyn_spg_flt_adj_tst |
|---|
| 1316 | |
|---|
| 1317 | # if defined key_tst_tlm |
|---|
| 1318 | SUBROUTINE dyn_spg_flt_tlm_tst( kumadt ) |
|---|
| 1319 | !!----------------------------------------------------------------------- |
|---|
| 1320 | !! |
|---|
| 1321 | !! *** ROUTINE dyn_spg_flt_tlm_tst *** |
|---|
| 1322 | !! |
|---|
| 1323 | !! ** Purpose : Test the tangent routine. |
|---|
| 1324 | !! |
|---|
| 1325 | !! ** Method : Verify the tangent with Taylor expansion |
|---|
| 1326 | !! |
|---|
| 1327 | !! M(x+hdx) = M(x) + L(hdx) + O(h^2) |
|---|
| 1328 | !! |
|---|
| 1329 | !! where L = tangent routine |
|---|
| 1330 | !! M = direct routine |
|---|
| 1331 | !! dx = input perturbation (random field) |
|---|
| 1332 | !! h = ration on perturbation |
|---|
| 1333 | !! |
|---|
| 1334 | !! In the tangent test we verify that: |
|---|
| 1335 | !! M(x+h*dx) - M(x) |
|---|
| 1336 | !! g(h) = ------------------ ---> 1 as h ---> 0 |
|---|
| 1337 | !! L(h*dx) |
|---|
| 1338 | !! and |
|---|
| 1339 | !! g(h) - 1 |
|---|
| 1340 | !! f(h) = ---------- ---> k (costant) as h ---> 0 |
|---|
| 1341 | !! p |
|---|
| 1342 | !! |
|---|
| 1343 | !! History : |
|---|
| 1344 | !! ! 09-08 (A. Vigilant) |
|---|
| 1345 | !!----------------------------------------------------------------------- |
|---|
| 1346 | !! * Modules used |
|---|
| 1347 | USE dynspg_flt |
|---|
| 1348 | USE tamtrj ! writing out state trajectory |
|---|
| 1349 | USE par_tlm, ONLY: & |
|---|
| 1350 | & tlm_bch, & |
|---|
| 1351 | & cur_loop, & |
|---|
| 1352 | & h_ratio |
|---|
| 1353 | USE istate_mod |
|---|
| 1354 | USE gridrandom, ONLY: & |
|---|
| 1355 | & grid_rd_sd |
|---|
| 1356 | USE trj_tam |
|---|
| 1357 | USE oce , ONLY: & ! ocean dynamics and tracers variables |
|---|
| 1358 | & ua, va, ub, vb, & |
|---|
| 1359 | & sshn |
|---|
| 1360 | USE sbc_oce , ONLY: & |
|---|
| 1361 | & emp |
|---|
| 1362 | USE tamctl, ONLY: & ! Control parameters |
|---|
| 1363 | & numtan, numtan_sc |
|---|
| 1364 | !! * Arguments |
|---|
| 1365 | INTEGER, INTENT(IN) :: & |
|---|
| 1366 | & kumadt ! Output unit |
|---|
| 1367 | |
|---|
| 1368 | !! * Local declarations |
|---|
| 1369 | REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: & |
|---|
| 1370 | & zua_tlin, & ! Tangent input: ua_tl |
|---|
| 1371 | & zva_tlin, & ! Tangent input: va_tl |
|---|
| 1372 | & zub_tlin, & ! Tangent input: ub_tl |
|---|
| 1373 | & zvb_tlin, & ! |
|---|
| 1374 | & zua_out, & ! |
|---|
| 1375 | & zva_out, & ! |
|---|
| 1376 | & zua_wop, & ! |
|---|
| 1377 | & zva_wop, & |
|---|
| 1378 | & z3r ! 3D field |
|---|
| 1379 | |
|---|
| 1380 | REAL(wp), DIMENSION(:,:), ALLOCATABLE :: & |
|---|
| 1381 | & zsshn_tlin, & ! Tangent input: sshn_tl |
|---|
| 1382 | & zemp_tlin, & ! Tangent input: emp_tl |
|---|
| 1383 | & zsshn_out, & ! |
|---|
| 1384 | & zsshn_wop, & |
|---|
| 1385 | & z2r ! 2D field |
|---|
| 1386 | REAL(wp) :: & |
|---|
| 1387 | & zsp1, zsp1_1, zsp1_2, zsp1_3, zsp1_4, & ! |
|---|
| 1388 | & zsp2, zsp2_1, zsp2_2, zsp2_3, zsp2_4, & ! |
|---|
| 1389 | & zsp3, zsp3_1, zsp3_2, zsp3_3, zsp3_4, & ! |
|---|
| 1390 | & zzsp, zzsp_1, zzsp_2, zzsp_3, zzsp_4, & |
|---|
| 1391 | & gamma, & |
|---|
| 1392 | & zgsp1, zgsp2, zgsp3, zgsp4, zgsp5, & |
|---|
| 1393 | & zgsp6, zgsp7 |
|---|
| 1394 | INTEGER :: & |
|---|
| 1395 | & indic, & |
|---|
| 1396 | & istp |
|---|
| 1397 | INTEGER :: & |
|---|
| 1398 | & ji, & |
|---|
| 1399 | & jj, & |
|---|
| 1400 | & jk |
|---|
| 1401 | CHARACTER(LEN=14) :: cl_name |
|---|
| 1402 | CHARACTER (LEN=128) :: file_out, file_wop, file_xdx |
|---|
| 1403 | CHARACTER (LEN=90) :: FMT |
|---|
| 1404 | REAL(KIND=wp), DIMENSION(100):: & |
|---|
| 1405 | & zscua, zscva, & |
|---|
| 1406 | & zscerrua, zscerrva, & |
|---|
| 1407 | & zscsshb, zscsshn, & |
|---|
| 1408 | & zscerrsshb, zscerrsshn |
|---|
| 1409 | INTEGER, DIMENSION(100):: & |
|---|
| 1410 | & iipossshb, iipossshn, iiposua, iiposva, & |
|---|
| 1411 | & ijpossshb, ijpossshn, ijposua, ijposva, & |
|---|
| 1412 | & ikposua, ikposva |
|---|
| 1413 | INTEGER:: & |
|---|
| 1414 | & ii, & |
|---|
| 1415 | & isamp=40, & |
|---|
| 1416 | & jsamp=40, & |
|---|
| 1417 | & ksamp=40, & |
|---|
| 1418 | & numsctlm |
|---|
| 1419 | REAL(KIND=wp), DIMENSION(jpi,jpj,jpk) :: & |
|---|
| 1420 | & zerrua, zerrva |
|---|
| 1421 | REAL(KIND=wp), DIMENSION(jpi,jpj) :: & |
|---|
| 1422 | & zerrsshb, zerrsshn |
|---|
| 1423 | ! Allocate memory |
|---|
| 1424 | |
|---|
| 1425 | ALLOCATE( & |
|---|
| 1426 | & zua_tlin(jpi,jpj,jpk), & |
|---|
| 1427 | & zva_tlin(jpi,jpj,jpk), & |
|---|
| 1428 | & zub_tlin(jpi,jpj,jpk), & |
|---|
| 1429 | & zvb_tlin(jpi,jpj,jpk), & |
|---|
| 1430 | & zua_out( jpi,jpj,jpk), & |
|---|
| 1431 | & zva_out( jpi,jpj,jpk), & |
|---|
| 1432 | & zua_wop( jpi,jpj,jpk), & |
|---|
| 1433 | & zva_wop( jpi,jpj,jpk), & |
|---|
| 1434 | & z3r(jpi,jpj,jpk) & |
|---|
| 1435 | & ) |
|---|
| 1436 | ALLOCATE( & |
|---|
| 1437 | & zsshn_tlin(jpi,jpj), & |
|---|
| 1438 | & zemp_tlin(jpi,jpj), & |
|---|
| 1439 | & zsshn_out(jpi,jpj), & |
|---|
| 1440 | & zsshn_wop(jpi,jpj), & |
|---|
| 1441 | & z2r(jpi,jpj) & |
|---|
| 1442 | & ) |
|---|
| 1443 | !-------------------------------------------------------------------- |
|---|
| 1444 | ! Reset variables |
|---|
| 1445 | !-------------------------------------------------------------------- |
|---|
| 1446 | zua_tlin( :,:,:) = 0.0_wp |
|---|
| 1447 | zva_tlin( :,:,:) = 0.0_wp |
|---|
| 1448 | zub_tlin( :,:,:) = 0.0_wp |
|---|
| 1449 | zvb_tlin( :,:,:) = 0.0_wp |
|---|
| 1450 | zsshn_tlin( :,:) = 0.0_wp |
|---|
| 1451 | zemp_tlin( :,:) = 0.0_wp |
|---|
| 1452 | zua_out ( :,:,:) = 0.0_wp |
|---|
| 1453 | zva_out ( :,:,:) = 0.0_wp |
|---|
| 1454 | zsshn_out ( :,:) = 0.0_wp |
|---|
| 1455 | zua_wop ( :,:,:) = 0.0_wp |
|---|
| 1456 | zva_wop ( :,:,:) = 0.0_wp |
|---|
| 1457 | zsshn_wop( :,:) = 0.0_wp |
|---|
| 1458 | |
|---|
| 1459 | zscua(:) = 0.0_wp |
|---|
| 1460 | zscva(:) = 0.0_wp |
|---|
| 1461 | zscerrua(:) = 0.0_wp |
|---|
| 1462 | zscerrva(:) = 0.0_wp |
|---|
| 1463 | zerrua(:,:,:) = 0.0_wp |
|---|
| 1464 | zerrva(:,:,:) = 0.0_wp |
|---|
| 1465 | zscsshn(:) = 0.0_wp |
|---|
| 1466 | zscerrsshn(:) = 0.0_wp |
|---|
| 1467 | zerrsshn(:,:) = 0.0_wp |
|---|
| 1468 | !-------------------------------------------------------------------- |
|---|
| 1469 | ! Output filename Xn=F(X0) |
|---|
| 1470 | !-------------------------------------------------------------------- |
|---|
| 1471 | !!!! CALL tlm_namrd |
|---|
| 1472 | gamma = h_ratio |
|---|
| 1473 | file_wop='trj_wop_dynspg' |
|---|
| 1474 | file_xdx='trj_xdx_dynspg' |
|---|
| 1475 | !-------------------------------------------------------------------- |
|---|
| 1476 | ! Initialize the tangent input with random noise: dx |
|---|
| 1477 | !-------------------------------------------------------------------- |
|---|
| 1478 | IF (( cur_loop .NE. 0) .OR. ( gamma .NE. 0.0_wp) )THEN |
|---|
| 1479 | CALL grid_rd_sd( 456953, z3r, 'U', 0.0_wp, stdu) |
|---|
| 1480 | DO jk = 1, jpk |
|---|
| 1481 | DO jj = nldj, nlej |
|---|
| 1482 | DO ji = nldi, nlei |
|---|
| 1483 | zua_tlin(ji,jj,jk) = z3r(ji,jj,jk) |
|---|
| 1484 | END DO |
|---|
| 1485 | END DO |
|---|
| 1486 | END DO |
|---|
| 1487 | CALL grid_rd_sd( 3434334, z3r, 'V', 0.0_wp, stdv) |
|---|
| 1488 | DO jk = 1, jpk |
|---|
| 1489 | DO jj = nldj, nlej |
|---|
| 1490 | DO ji = nldi, nlei |
|---|
| 1491 | zva_tlin(ji,jj,jk) = z3r(ji,jj,jk) |
|---|
| 1492 | END DO |
|---|
| 1493 | END DO |
|---|
| 1494 | END DO |
|---|
| 1495 | CALL grid_rd_sd( 935678, z3r, 'U', 0.0_wp, stdu) |
|---|
| 1496 | DO jk = 1, jpk |
|---|
| 1497 | DO jj = nldj, nlej |
|---|
| 1498 | DO ji = nldi, nlei |
|---|
| 1499 | zub_tlin(ji,jj,jk) = z3r(ji,jj,jk) |
|---|
| 1500 | END DO |
|---|
| 1501 | END DO |
|---|
| 1502 | END DO |
|---|
| 1503 | CALL grid_rd_sd( 1462578, z3r, 'V', 0.0_wp, stdv) |
|---|
| 1504 | DO jk = 1, jpk |
|---|
| 1505 | DO jj = nldj, nlej |
|---|
| 1506 | DO ji = nldi, nlei |
|---|
| 1507 | zvb_tlin(ji,jj,jk) = z3r(ji,jj,jk) |
|---|
| 1508 | END DO |
|---|
| 1509 | END DO |
|---|
| 1510 | END DO |
|---|
| 1511 | CALL grid_rd_sd( 25674910, z2r, 'T', 0.0_wp, stdemp) |
|---|
| 1512 | DO jj = nldj, nlej |
|---|
| 1513 | DO ji = nldi, nlei |
|---|
| 1514 | zemp_tlin(ji,jj) = z2r(ji,jj) |
|---|
| 1515 | END DO |
|---|
| 1516 | END DO |
|---|
| 1517 | CALL grid_rd_sd( 12672456, z2r,'T', 0.0_wp, stdssh) |
|---|
| 1518 | DO jj = nldj, nlej |
|---|
| 1519 | DO ji = nldi, nlei |
|---|
| 1520 | zsshn_tlin(ji,jj) = z2r(ji,jj) |
|---|
| 1521 | END DO |
|---|
| 1522 | END DO |
|---|
| 1523 | ENDIF |
|---|
| 1524 | |
|---|
| 1525 | !-------------------------------------------------------------------- |
|---|
| 1526 | ! Complete Init for Direct |
|---|
| 1527 | !------------------------------------------------------------------- |
|---|
| 1528 | CALL istate_p |
|---|
| 1529 | ! *** initialize the reference trajectory |
|---|
| 1530 | ! ------------ |
|---|
| 1531 | CALL trj_rea( nit000-1, 1 ) |
|---|
| 1532 | CALL trj_rea( nit000, 1 ) |
|---|
| 1533 | |
|---|
| 1534 | |
|---|
| 1535 | IF (( cur_loop .NE. 0) .OR. ( gamma .NE. 0.0_wp) )THEN |
|---|
| 1536 | zua_tlin(:,:,:) = gamma * zua_tlin(:,:,:) |
|---|
| 1537 | ua(:,:,:) = ua(:,:,:) + zua_tlin(:,:,:) |
|---|
| 1538 | |
|---|
| 1539 | zva_tlin(:,:,:) = gamma * zva_tlin(:,:,:) |
|---|
| 1540 | va(:,:,:) = va(:,:,:) + zva_tlin(:,:,:) |
|---|
| 1541 | |
|---|
| 1542 | zub_tlin(:,:,:) = gamma * zub_tlin(:,:,:) |
|---|
| 1543 | ub(:,:,:) = ub(:,:,:) + zub_tlin(:,:,:) |
|---|
| 1544 | |
|---|
| 1545 | zvb_tlin(:,:,:) = gamma * zvb_tlin(:,:,:) |
|---|
| 1546 | vb(:,:,:) = vb(:,:,:) + zvb_tlin(:,:,:) |
|---|
| 1547 | |
|---|
| 1548 | zemp_tlin(:,:) = gamma * zemp_tlin(:,:) |
|---|
| 1549 | emp(:,:) = emp(:,:) + zemp_tlin(:,:) |
|---|
| 1550 | |
|---|
| 1551 | zsshn_tlin(:,:) = gamma * zsshn_tlin(:,:) |
|---|
| 1552 | sshn(:,:) = sshn(:,:) + zsshn_tlin(:,:) |
|---|
| 1553 | ENDIF |
|---|
| 1554 | |
|---|
| 1555 | !-------------------------------------------------------------------- |
|---|
| 1556 | ! Compute the direct model F(X0,t=n) = Xn |
|---|
| 1557 | !-------------------------------------------------------------------- |
|---|
| 1558 | IF ( tlm_bch /= 2 ) CALL dyn_spg_flt(nit000, indic) |
|---|
| 1559 | IF ( tlm_bch == 0 ) CALL trj_wri_spl(file_wop) |
|---|
| 1560 | IF ( tlm_bch == 1 ) CALL trj_wri_spl(file_xdx) |
|---|
| 1561 | !-------------------------------------------------------------------- |
|---|
| 1562 | ! Compute the Tangent |
|---|
| 1563 | !-------------------------------------------------------------------- |
|---|
| 1564 | IF ( tlm_bch == 2 ) THEN |
|---|
| 1565 | !-------------------------------------------------------------------- |
|---|
| 1566 | ! Initialize the tangent variables |
|---|
| 1567 | !-------------------------------------------------------------------- |
|---|
| 1568 | CALL trj_rea( nit000-1, 1 ) |
|---|
| 1569 | CALL trj_rea( nit000, 1 ) |
|---|
| 1570 | ua_tl (:,:,:) = zua_tlin (:,:,:) |
|---|
| 1571 | va_tl (:,:,:) = zva_tlin (:,:,:) |
|---|
| 1572 | ub_tl (:,:,:) = zub_tlin (:,:,:) |
|---|
| 1573 | vb_tl (:,:,:) = zvb_tlin (:,:,:) |
|---|
| 1574 | emp_tl (:,: ) = zemp_tlin (:,: ) |
|---|
| 1575 | sshn_tl(:,: ) = zsshn_tlin(:,: ) |
|---|
| 1576 | |
|---|
| 1577 | CALL dyn_spg_flt_tan(nit000, indic) |
|---|
| 1578 | !-------------------------------------------------------------------- |
|---|
| 1579 | ! Compute the scalar product: ( L(t0,tn) gamma dx0 ) ) |
|---|
| 1580 | !-------------------------------------------------------------------- |
|---|
| 1581 | |
|---|
| 1582 | zsp2_1 = DOT_PRODUCT( ua_tl, ua_tl ) |
|---|
| 1583 | zsp2_2 = DOT_PRODUCT( va_tl, va_tl ) |
|---|
| 1584 | zsp2_4 = DOT_PRODUCT( sshn_tl, sshn_tl ) |
|---|
| 1585 | zsp2 = zsp2_1 + zsp2_2 + zsp2_4 |
|---|
| 1586 | !-------------------------------------------------------------------- |
|---|
| 1587 | ! Storing data |
|---|
| 1588 | !-------------------------------------------------------------------- |
|---|
| 1589 | CALL trj_rd_spl(file_wop) |
|---|
| 1590 | zua_wop (:,:,:) = ua (:,:,:) |
|---|
| 1591 | zva_wop (:,:,:) = va (:,:,:) |
|---|
| 1592 | zsshn_wop(:,:) = sshn(:,:) |
|---|
| 1593 | CALL trj_rd_spl(file_xdx) |
|---|
| 1594 | zua_out (:,:,:) = ua (:,:,:) |
|---|
| 1595 | zva_out (:,:,:) = va (:,:,:) |
|---|
| 1596 | zsshn_out(:,: ) = sshn (:,: ) |
|---|
| 1597 | !-------------------------------------------------------------------- |
|---|
| 1598 | ! Compute the Linearization Error |
|---|
| 1599 | ! Nn = M( X0+gamma.dX0, t0,tn) - M(X0, t0,tn) |
|---|
| 1600 | ! and |
|---|
| 1601 | ! Compute the Linearization Error |
|---|
| 1602 | ! En = Nn -TL(gamma.dX0, t0,tn) |
|---|
| 1603 | !-------------------------------------------------------------------- |
|---|
| 1604 | ! Warning: Here we re-use local variables z()_out and z()_wop |
|---|
| 1605 | ii=0 |
|---|
| 1606 | DO jk = 1, jpk |
|---|
| 1607 | DO jj = 1, jpj |
|---|
| 1608 | DO ji = 1, jpi |
|---|
| 1609 | zua_out (ji,jj,jk) = zua_out (ji,jj,jk) - zua_wop (ji,jj,jk) |
|---|
| 1610 | zua_wop (ji,jj,jk) = zua_out (ji,jj,jk) - ua_tl (ji,jj,jk) |
|---|
| 1611 | IF ( ua_tl(ji,jj,jk) .NE. 0.0_wp ) & |
|---|
| 1612 | & zerrua(ji,jj,jk) = zua_out(ji,jj,jk)/ua_tl(ji,jj,jk) |
|---|
| 1613 | IF( (MOD(ji, isamp) .EQ. 0) .AND. & |
|---|
| 1614 | & (MOD(jj, jsamp) .EQ. 0) .AND. & |
|---|
| 1615 | & (MOD(jk, ksamp) .EQ. 0) ) THEN |
|---|
| 1616 | ii = ii+1 |
|---|
| 1617 | iiposua(ii) = ji |
|---|
| 1618 | ijposua(ii) = jj |
|---|
| 1619 | ikposua(ii) = jk |
|---|
| 1620 | IF ( INT(umask(ji,jj,jk)) .NE. 0) THEN |
|---|
| 1621 | zscua (ii) = zua_wop(ji,jj,jk) |
|---|
| 1622 | zscerrua (ii) = ( zerrua(ji,jj,jk) - 1.0_wp ) /gamma |
|---|
| 1623 | ENDIF |
|---|
| 1624 | ENDIF |
|---|
| 1625 | END DO |
|---|
| 1626 | END DO |
|---|
| 1627 | END DO |
|---|
| 1628 | ii=0 |
|---|
| 1629 | DO jk = 1, jpk |
|---|
| 1630 | DO jj = 1, jpj |
|---|
| 1631 | DO ji = 1, jpi |
|---|
| 1632 | zva_out (ji,jj,jk) = zva_out (ji,jj,jk) - zva_wop (ji,jj,jk) |
|---|
| 1633 | zva_wop (ji,jj,jk) = zva_out (ji,jj,jk) - va_tl (ji,jj,jk) |
|---|
| 1634 | IF ( va_tl(ji,jj,jk) .NE. 0.0_wp ) & |
|---|
| 1635 | & zerrva(ji,jj,jk) = zva_out(ji,jj,jk)/va_tl(ji,jj,jk) |
|---|
| 1636 | IF( (MOD(ji, isamp) .EQ. 0) .AND. & |
|---|
| 1637 | & (MOD(jj, jsamp) .EQ. 0) .AND. & |
|---|
| 1638 | & (MOD(jk, ksamp) .EQ. 0) ) THEN |
|---|
| 1639 | ii = ii+1 |
|---|
| 1640 | iiposva(ii) = ji |
|---|
| 1641 | ijposva(ii) = jj |
|---|
| 1642 | ikposva(ii) = jk |
|---|
| 1643 | IF ( INT(vmask(ji,jj,jk)) .NE. 0) THEN |
|---|
| 1644 | zscva (ii) = zua_wop(ji,jj,jk) |
|---|
| 1645 | zscerrva (ii) = ( zerrva(ji,jj,jk) - 1.0_wp ) /gamma |
|---|
| 1646 | ENDIF |
|---|
| 1647 | ENDIF |
|---|
| 1648 | END DO |
|---|
| 1649 | END DO |
|---|
| 1650 | END DO |
|---|
| 1651 | ii=0 |
|---|
| 1652 | DO jj = 1, jpj |
|---|
| 1653 | DO ji = 1, jpi |
|---|
| 1654 | zsshn_out (ji,jj) = zsshn_out (ji,jj) - zsshn_wop (ji,jj) |
|---|
| 1655 | zsshn_wop (ji,jj) = zsshn_out (ji,jj) - sshn_tl (ji,jj) |
|---|
| 1656 | IF ( sshn_tl(ji,jj) .NE. 0.0_wp ) & |
|---|
| 1657 | & zerrsshn(ji,jj) = zsshn_out(ji,jj)/sshn_tl(ji,jj) |
|---|
| 1658 | IF( (MOD(ji, isamp) .EQ. 0) .AND. & |
|---|
| 1659 | & (MOD(jj, jsamp) .EQ. 0) ) THEN |
|---|
| 1660 | ii = ii+1 |
|---|
| 1661 | iipossshn(ii) = ji |
|---|
| 1662 | ijpossshn(ii) = jj |
|---|
| 1663 | IF ( INT(tmask(ji,jj,1)) .NE. 0) THEN |
|---|
| 1664 | zscsshn (ii) = zsshn_wop(ji,jj) |
|---|
| 1665 | zscerrsshn (ii) = ( zerrsshn(ji,jj) - 1.0_wp ) /gamma |
|---|
| 1666 | ENDIF |
|---|
| 1667 | ENDIF |
|---|
| 1668 | END DO |
|---|
| 1669 | END DO |
|---|
| 1670 | zsp1_1 = DOT_PRODUCT( zua_out, zua_out ) |
|---|
| 1671 | zsp1_2 = DOT_PRODUCT( zva_out, zva_out ) |
|---|
| 1672 | zsp1_4 = DOT_PRODUCT( zsshn_out, zsshn_out ) |
|---|
| 1673 | zsp1 = zsp1_1 + zsp1_2 + zsp1_4 |
|---|
| 1674 | zsp3_1 = DOT_PRODUCT( zua_wop, zua_wop ) |
|---|
| 1675 | zsp3_2 = DOT_PRODUCT( zva_wop, zva_wop ) |
|---|
| 1676 | zsp3_4 = DOT_PRODUCT( zsshn_wop, zsshn_wop ) |
|---|
| 1677 | zsp3 = zsp3_1 + zsp3_2 + zsp3_4 |
|---|
| 1678 | |
|---|
| 1679 | !-------------------------------------------------------------------- |
|---|
| 1680 | ! Print the linearization error En - norme 2 |
|---|
| 1681 | !-------------------------------------------------------------------- |
|---|
| 1682 | ! 14 char:'12345678901234' |
|---|
| 1683 | cl_name = 'dynspg_tam:En ' |
|---|
| 1684 | zzsp = SQRT(zsp3) |
|---|
| 1685 | zzsp_1 = SQRT(zsp3_1) |
|---|
| 1686 | zzsp_2 = SQRT(zsp3_2) |
|---|
| 1687 | zzsp_4 = SQRT(zsp3_4) |
|---|
| 1688 | zgsp5 = zzsp |
|---|
| 1689 | CALL prntst_tlm( cl_name, kumadt, zzsp, h_ratio ) |
|---|
| 1690 | !-------------------------------------------------------------------- |
|---|
| 1691 | ! Compute TLM norm2 |
|---|
| 1692 | !-------------------------------------------------------------------- |
|---|
| 1693 | zzsp = SQRT(zsp2) |
|---|
| 1694 | zzsp_1 = SQRT(zsp2_1) |
|---|
| 1695 | zzsp_2 = SQRT(zsp2_2) |
|---|
| 1696 | zzsp_4 = SQRT(zsp2_4) |
|---|
| 1697 | zgsp4 = zzsp |
|---|
| 1698 | cl_name = 'dynspg_tam:Ln2' |
|---|
| 1699 | CALL prntst_tlm( cl_name, kumadt, zzsp, h_ratio ) |
|---|
| 1700 | !-------------------------------------------------------------------- |
|---|
| 1701 | ! Print the linearization error Nn - norme 2 |
|---|
| 1702 | !-------------------------------------------------------------------- |
|---|
| 1703 | zzsp = SQRT(zsp1) |
|---|
| 1704 | zzsp_1 = SQRT(zsp1_1) |
|---|
| 1705 | zzsp_2 = SQRT(zsp1_2) |
|---|
| 1706 | zzsp_4 = SQRT(zsp1_4) |
|---|
| 1707 | cl_name = 'dynspg:Mhdx-Mx' |
|---|
| 1708 | CALL prntst_tlm( cl_name, kumadt, zzsp, h_ratio ) |
|---|
| 1709 | |
|---|
| 1710 | |
|---|
| 1711 | zgsp3 = SQRT( zsp3/zsp2 ) |
|---|
| 1712 | zgsp7 = zgsp3/gamma |
|---|
| 1713 | zgsp1 = zzsp |
|---|
| 1714 | zgsp2 = zgsp1 / zgsp4 |
|---|
| 1715 | zgsp6 = (zgsp2 - 1.0_wp)/gamma |
|---|
| 1716 | |
|---|
| 1717 | FMT = "(A8,2X,I4.4,2X,E6.1,2X,E20.13,2X,E20.13,2X,E20.13,2X,E20.13,2X,E20.13,2X,E20.13,2X,E20.13)" |
|---|
| 1718 | WRITE(numtan,FMT) 'dspgflt ', cur_loop, h_ratio, zgsp1, zgsp2, zgsp3, zgsp4, zgsp5, zgsp6, zgsp7 |
|---|
| 1719 | !-------------------------------------------------------------------- |
|---|
| 1720 | ! Unitary calculus |
|---|
| 1721 | !-------------------------------------------------------------------- |
|---|
| 1722 | FMT = "(A8,2X,A8,2X,I4.4,2X,E6.1,2X,I4.4,2X,I4.4,2X,I4.4,2X,E20.13,1X)" |
|---|
| 1723 | cl_name = 'dspgflt ' |
|---|
| 1724 | IF(lwp) THEN |
|---|
| 1725 | DO ii=1, 100, 1 |
|---|
| 1726 | IF ( zscua(ii) .NE. 0.0_wp ) WRITE(numtan_sc,FMT) cl_name, 'zscua ', & |
|---|
| 1727 | & cur_loop, h_ratio, ii, iiposua(ii), ijposua(ii), zscua(ii) |
|---|
| 1728 | ENDDO |
|---|
| 1729 | DO ii=1, 100, 1 |
|---|
| 1730 | IF ( zscva(ii) .NE. 0.0_wp ) WRITE(numtan_sc,FMT) cl_name, 'zscva ', & |
|---|
| 1731 | & cur_loop, h_ratio, ii, iiposva(ii), ijposva(ii), zscva(ii) |
|---|
| 1732 | ENDDO |
|---|
| 1733 | DO ii=1, 100, 1 |
|---|
| 1734 | IF ( zscsshn(ii) .NE. 0.0_wp ) WRITE(numtan_sc,FMT) cl_name, 'zscsshn ', & |
|---|
| 1735 | & cur_loop, h_ratio, ii, iipossshn(ii), ijpossshn(ii), zscsshn(ii) |
|---|
| 1736 | ENDDO |
|---|
| 1737 | DO ii=1, 100, 1 |
|---|
| 1738 | IF ( zscerrua(ii) .NE. 0.0_wp ) WRITE(numsctlm,FMT) cl_name, 'zscerrua ', & |
|---|
| 1739 | & cur_loop, h_ratio, ii, iiposua(ii), ijposua(ii), zscerrua(ii) |
|---|
| 1740 | ENDDO |
|---|
| 1741 | DO ii=1, 100, 1 |
|---|
| 1742 | IF ( zscerrva(ii) .NE. 0.0_wp ) WRITE(numtan_sc,FMT) cl_name, 'zscerrva ', & |
|---|
| 1743 | & cur_loop, h_ratio, ii, iiposva(ii), ijposva(ii), zscerrva(ii) |
|---|
| 1744 | ENDDO |
|---|
| 1745 | DO ii=1, 100, 1 |
|---|
| 1746 | IF ( zscerrsshn(ii) .NE. 0.0_wp ) WRITE(numtan_sc,FMT) cl_name, 'zscerrsshn ', & |
|---|
| 1747 | & cur_loop, h_ratio, ii, iipossshn(ii), ijpossshn(ii), zscerrsshn(ii) |
|---|
| 1748 | ENDDO |
|---|
| 1749 | ! write separator |
|---|
| 1750 | WRITE(numtan_sc,"(A4)") '====' |
|---|
| 1751 | ENDIF |
|---|
| 1752 | ENDIF |
|---|
| 1753 | |
|---|
| 1754 | DEALLOCATE( & |
|---|
| 1755 | & zemp_tlin, zsshn_tlin, & |
|---|
| 1756 | & zua_tlin, zva_tlin, & |
|---|
| 1757 | & zua_out, zva_out, & |
|---|
| 1758 | & zua_wop, zva_wop, & |
|---|
| 1759 | & zsshn_out, zsshn_wop, & |
|---|
| 1760 | & z3r, z2r & |
|---|
| 1761 | & ) |
|---|
| 1762 | END SUBROUTINE dyn_spg_flt_tlm_tst |
|---|
| 1763 | # endif |
|---|
| 1764 | # else |
|---|
| 1765 | !!---------------------------------------------------------------------- |
|---|
| 1766 | !! Default case : Empty module No standart explicit free surface |
|---|
| 1767 | !!---------------------------------------------------------------------- |
|---|
| 1768 | CONTAINS |
|---|
| 1769 | SUBROUTINE dyn_spg_flt_tan( kt, kindic ) ! Empty routine |
|---|
| 1770 | WRITE(*,*) 'dyn_spg_flt: You should not have seen this print! error?', kt |
|---|
| 1771 | END SUBROUTINE dyn_spg_flt_tan |
|---|
| 1772 | SUBROUTINE dyn_spg_flt_adj( kt, kindic ) ! Empty routine |
|---|
| 1773 | WRITE(*,*) 'dyn_spg_flt: You should not have seen this print! error?', kt |
|---|
| 1774 | END SUBROUTINE dyn_spg_flt_adj |
|---|
| 1775 | SUBROUTINE dyn_spg_flt_adj_tst( kt ) ! Empty routine |
|---|
| 1776 | WRITE(*,*) 'dyn_spg_flt: You should not have seen this print! error?', kt |
|---|
| 1777 | END SUBROUTINE dyn_spg_flt_adj_tst |
|---|
| 1778 | SUBROUTINE dyn_spg_flt_tlm_tst( kt ) ! Empty routine |
|---|
| 1779 | WRITE(*,*) 'dyn_spg_flt: You should not have seen this print! error?', kt |
|---|
| 1780 | END SUBROUTINE dyn_spg_flt_tlm_tst |
|---|
| 1781 | # endif |
|---|
| 1782 | #endif |
|---|
| 1783 | END MODULE dynspg_flt_tam |
|---|
