[12] | 1 | MODULE timeloop_gcm_mod |
---|
[133] | 2 | |
---|
| 3 | PRIVATE |
---|
[12] | 4 | |
---|
[133] | 5 | PUBLIC :: timeloop |
---|
| 6 | |
---|
| 7 | INTEGER, PARAMETER :: euler=1, rk4=2, mlf=3 |
---|
| 8 | |
---|
[12] | 9 | CONTAINS |
---|
| 10 | |
---|
| 11 | SUBROUTINE timeloop |
---|
[19] | 12 | USE icosa |
---|
[15] | 13 | USE dissip_gcm_mod |
---|
[17] | 14 | USE caldyn_mod |
---|
[12] | 15 | USE etat0_mod |
---|
[17] | 16 | USE guided_mod |
---|
| 17 | USE advect_tracer_mod |
---|
[81] | 18 | USE physics_mod |
---|
[131] | 19 | USE mpipara |
---|
[17] | 20 | |
---|
[12] | 21 | IMPLICIT NONE |
---|
| 22 | TYPE(t_field),POINTER :: f_phis(:) |
---|
[129] | 23 | ! TYPE(t_field),POINTER :: f_theta(:) |
---|
[17] | 24 | TYPE(t_field),POINTER :: f_q(:) |
---|
[132] | 25 | TYPE(t_field),POINTER :: f_dtheta(:), f_rhodz(:) |
---|
[12] | 26 | TYPE(t_field),POINTER :: f_ps(:),f_psm1(:), f_psm2(:) |
---|
| 27 | TYPE(t_field),POINTER :: f_u(:),f_um1(:),f_um2(:) |
---|
| 28 | TYPE(t_field),POINTER :: f_theta_rhodz(:),f_theta_rhodzm1(:),f_theta_rhodzm2(:) |
---|
| 29 | TYPE(t_field),POINTER :: f_dps(:),f_dpsm1(:), f_dpsm2(:) |
---|
| 30 | TYPE(t_field),POINTER :: f_du(:),f_dum1(:),f_dum2(:) |
---|
| 31 | TYPE(t_field),POINTER :: f_dtheta_rhodz(:),f_dtheta_rhodzm1(:),f_dtheta_rhodzm2(:) |
---|
[133] | 32 | TYPE(t_field),POINTER :: f_hflux(:), f_wflux(:), f_hfluxt(:), f_wfluxt(:) |
---|
[12] | 33 | |
---|
| 34 | REAL(rstd),POINTER :: phis(:) |
---|
[17] | 35 | REAL(rstd),POINTER :: q(:,:,:) |
---|
[12] | 36 | REAL(rstd),POINTER :: ps(:) ,psm1(:), psm2(:) |
---|
| 37 | REAL(rstd),POINTER :: u(:,:) , um1(:,:), um2(:,:) |
---|
[133] | 38 | REAL(rstd),POINTER :: rhodz(:,:), theta_rhodz(:,:) , theta_rhodzm1(:,:), theta_rhodzm2(:,:) |
---|
[12] | 39 | REAL(rstd),POINTER :: dps(:), dpsm1(:), dpsm2(:) |
---|
| 40 | REAL(rstd),POINTER :: du(:,:), dum1(:,:), dum2(:,:) |
---|
| 41 | REAL(rstd),POINTER :: dtheta_rhodz(:,:),dtheta_rhodzm1(:,:),dtheta_rhodzm2(:,:) |
---|
[133] | 42 | REAL(rstd),POINTER :: hflux(:,:),wflux(:,:),hfluxt(:,:),wfluxt(:,:) |
---|
[50] | 43 | |
---|
[98] | 44 | ! REAL(rstd) :: dt, run_length |
---|
[12] | 45 | INTEGER :: ind |
---|
[133] | 46 | INTEGER :: it,i,j,n, nb_stage, stage, matsuno_period, scheme |
---|
| 47 | CHARACTER(len=255) :: scheme_name |
---|
[138] | 48 | LOGICAL :: fluxt_zero(ndomain) ! set to .TRUE. to start accumulating fluxes in time |
---|
[98] | 49 | ! INTEGER :: itaumax |
---|
| 50 | ! REAL(rstd) ::write_period |
---|
| 51 | ! INTEGER :: itau_out |
---|
[63] | 52 | |
---|
[98] | 53 | ! dt=90. |
---|
| 54 | ! CALL getin('dt',dt) |
---|
[32] | 55 | |
---|
[98] | 56 | ! itaumax=100 |
---|
| 57 | ! CALL getin('itaumax',itaumax) |
---|
[17] | 58 | |
---|
[98] | 59 | ! run_length=dt*itaumax |
---|
| 60 | ! CALL getin('run_length',run_length) |
---|
| 61 | ! itaumax=run_length/dt |
---|
| 62 | ! PRINT *,'itaumax=',itaumax |
---|
| 63 | ! dt=dt/scale_factor |
---|
[45] | 64 | |
---|
[129] | 65 | ! Trends |
---|
| 66 | CALL allocate_field(f_dps,field_t,type_real) |
---|
| 67 | CALL allocate_field(f_du,field_u,type_real,llm) |
---|
| 68 | CALL allocate_field(f_dtheta_rhodz,field_t,type_real,llm) |
---|
| 69 | ! Model state at current time step (RK/MLF/Euler) |
---|
| 70 | CALL allocate_field(f_phis,field_t,type_real) |
---|
| 71 | CALL allocate_field(f_ps,field_t,type_real) |
---|
| 72 | CALL allocate_field(f_u,field_u,type_real,llm) |
---|
| 73 | CALL allocate_field(f_theta_rhodz,field_t,type_real,llm) |
---|
| 74 | ! Model state at previous time step (RK/MLF) |
---|
| 75 | CALL allocate_field(f_psm1,field_t,type_real) |
---|
| 76 | CALL allocate_field(f_um1,field_u,type_real,llm) |
---|
| 77 | CALL allocate_field(f_theta_rhodzm1,field_t,type_real,llm) |
---|
| 78 | ! Tracers |
---|
| 79 | CALL allocate_field(f_q,field_t,type_real,llm,nqtot) |
---|
[132] | 80 | CALL allocate_field(f_rhodz,field_t,type_real,llm) |
---|
[134] | 81 | ! Mass fluxes |
---|
[133] | 82 | CALL allocate_field(f_hflux,field_u,type_real,llm) ! instantaneous mass fluxes |
---|
| 83 | CALL allocate_field(f_wflux,field_t,type_real,llm+1) ! computed by caldyn |
---|
| 84 | CALL allocate_field(f_hfluxt,field_u,type_real,llm) ! mass "fluxes" accumulated in time |
---|
| 85 | CALL allocate_field(f_wfluxt,field_t,type_real,llm+1) |
---|
[129] | 86 | |
---|
[133] | 87 | scheme_name='runge_kutta' |
---|
| 88 | CALL getin('scheme',scheme_name) |
---|
[129] | 89 | |
---|
[133] | 90 | SELECT CASE (TRIM(scheme_name)) |
---|
[129] | 91 | CASE('euler') |
---|
[133] | 92 | scheme=euler |
---|
[129] | 93 | nb_stage=1 |
---|
| 94 | CASE ('runge_kutta') |
---|
[133] | 95 | scheme=rk4 |
---|
[129] | 96 | nb_stage=4 |
---|
| 97 | CASE ('leapfrog_matsuno') |
---|
[133] | 98 | scheme=mlf |
---|
[129] | 99 | matsuno_period=5 |
---|
| 100 | CALL getin('matsuno_period',matsuno_period) |
---|
| 101 | nb_stage=matsuno_period+1 |
---|
| 102 | ! Model state 2 time steps ago (MLF) |
---|
| 103 | CALL allocate_field(f_psm2,field_t,type_real) |
---|
| 104 | CALL allocate_field(f_theta_rhodzm2,field_t,type_real,llm) |
---|
| 105 | CALL allocate_field(f_um2,field_u,type_real,llm) |
---|
| 106 | CASE default |
---|
[133] | 107 | PRINT*,'Bad selector for variable scheme : <', TRIM(scheme_name), & |
---|
[129] | 108 | ' > options are <euler>, <runge_kutta>, <leapfrog_matsuno>' |
---|
| 109 | STOP |
---|
| 110 | END SELECT |
---|
[12] | 111 | |
---|
[98] | 112 | ! write_period=0 |
---|
| 113 | ! CALL getin('write_period',write_period) |
---|
| 114 | ! write_period=write_period/scale_factor |
---|
| 115 | ! itau_out=FLOOR(.5+write_period/dt) |
---|
| 116 | ! PRINT *, 'Output frequency (scaled) set to ',write_period, ' : itau_out = ',itau_out |
---|
[63] | 117 | |
---|
[129] | 118 | ! Trends at previous time steps needed only by Adams-Bashforth |
---|
| 119 | ! CALL allocate_field(f_dpsm1,field_t,type_real) |
---|
| 120 | ! CALL allocate_field(f_dpsm2,field_t,type_real) |
---|
| 121 | ! CALL allocate_field(f_dum1,field_u,type_real,llm) |
---|
| 122 | ! CALL allocate_field(f_dum2,field_u,type_real,llm) |
---|
| 123 | ! CALL allocate_field(f_dtheta_rhodzm1,field_t,type_real,llm) |
---|
| 124 | ! CALL allocate_field(f_dtheta_rhodzm2,field_t,type_real,llm) |
---|
[12] | 125 | |
---|
[129] | 126 | ! CALL allocate_field(f_theta,field_t,type_real,llm) |
---|
| 127 | ! CALL allocate_field(f_dtheta,field_t,type_real,llm) |
---|
[12] | 128 | |
---|
[98] | 129 | CALL init_dissip |
---|
| 130 | CALL init_caldyn |
---|
| 131 | CALL init_guided |
---|
| 132 | CALL init_advect_tracer |
---|
| 133 | CALL init_physics |
---|
[12] | 134 | |
---|
[17] | 135 | CALL etat0(f_ps,f_phis,f_theta_rhodz,f_u, f_q) |
---|
[97] | 136 | CALL writefield("phis",f_phis,once=.TRUE.) |
---|
[81] | 137 | CALL transfert_request(f_q,req_i1) |
---|
[50] | 138 | |
---|
[133] | 139 | DO ind=1,ndomain |
---|
| 140 | CALL swap_dimensions(ind) |
---|
| 141 | CALL swap_geometry(ind) |
---|
| 142 | rhodz=f_rhodz(ind); ps=f_ps(ind) |
---|
[138] | 143 | CALL compute_rhodz(.TRUE., ps,rhodz) ! save rhodz for transport scheme before dynamics update ps |
---|
[133] | 144 | END DO |
---|
[138] | 145 | fluxt_zero=.TRUE. |
---|
[132] | 146 | |
---|
[138] | 147 | ! check that rhodz is consistent with ps |
---|
| 148 | CALL transfert_request(f_rhodz,req_i1) |
---|
| 149 | CALL transfert_request(f_ps,req_i1) |
---|
| 150 | DO ind=1,ndomain |
---|
| 151 | CALL swap_dimensions(ind) |
---|
| 152 | CALL swap_geometry(ind) |
---|
| 153 | rhodz=f_rhodz(ind); ps=f_ps(ind) |
---|
| 154 | CALL compute_rhodz(.FALSE., ps, rhodz) |
---|
| 155 | END DO |
---|
| 156 | |
---|
[12] | 157 | DO it=0,itaumax |
---|
[81] | 158 | |
---|
[131] | 159 | IF (is_mpi_root) PRINT *,"It No :",It," t :",dt*It |
---|
[63] | 160 | IF (mod(it,itau_out)==0 ) THEN |
---|
| 161 | CALL writefield("q",f_q) |
---|
[81] | 162 | CALL update_time_counter(dt*it) |
---|
[63] | 163 | ENDIF |
---|
| 164 | |
---|
[139] | 165 | CALL guided(it*dt,f_ps,f_theta_rhodz,f_u,f_q) |
---|
[129] | 166 | |
---|
| 167 | DO stage=1,nb_stage |
---|
| 168 | CALL caldyn((stage==1) .AND. (MOD(it,itau_out)==0), & |
---|
[132] | 169 | f_phis,f_ps,f_theta_rhodz,f_u, f_q, & |
---|
[134] | 170 | f_hflux, f_wflux, f_dps, f_dtheta_rhodz, f_du) |
---|
[133] | 171 | SELECT CASE (scheme) |
---|
| 172 | CASE(euler) |
---|
| 173 | CALL euler_scheme(.TRUE.) |
---|
| 174 | CASE (rk4) |
---|
[129] | 175 | CALL rk_scheme(stage) |
---|
[133] | 176 | CASE (mlf) |
---|
[129] | 177 | CALL leapfrog_matsuno_scheme(stage) |
---|
| 178 | |
---|
| 179 | ! CASE ('leapfrog') |
---|
| 180 | ! CALL leapfrog_scheme |
---|
| 181 | ! |
---|
| 182 | ! CASE ('adam_bashforth') |
---|
| 183 | ! CALL dissip(f_u,f_du,f_ps,f_phis, f_theta_rhodz,f_dtheta_rhodz) |
---|
| 184 | ! CALL adam_bashforth_scheme |
---|
[133] | 185 | CASE DEFAULT |
---|
[129] | 186 | STOP |
---|
| 187 | END SELECT |
---|
| 188 | END DO |
---|
[130] | 189 | |
---|
[131] | 190 | CALL dissip(f_u,f_du,f_ps,f_phis, f_theta_rhodz,f_dtheta_rhodz) |
---|
| 191 | CALL euler_scheme(.FALSE.) |
---|
[130] | 192 | |
---|
[133] | 193 | IF(MOD(it+1,itau_adv)==0) THEN |
---|
[138] | 194 | CALL transfert_request(f_wfluxt,req_i1) ! FIXME |
---|
| 195 | ! CALL transfert_request(f_hfluxt,req_e1) ! FIXME |
---|
| 196 | |
---|
[135] | 197 | CALL advect_tracer(f_hfluxt,f_wfluxt,f_u, f_q,f_rhodz) ! update q and rhodz after RK step |
---|
[134] | 198 | fluxt_zero=.TRUE. |
---|
[138] | 199 | |
---|
| 200 | ! FIXME : check that rhodz is consistent with ps |
---|
| 201 | CALL transfert_request(f_rhodz,req_i1) |
---|
| 202 | CALL transfert_request(f_ps,req_i1) |
---|
| 203 | CALL transfert_request(f_dps,req_i1) ! FIXME |
---|
| 204 | CALL transfert_request(f_wflux,req_i1) ! FIXME |
---|
| 205 | DO ind=1,ndomain |
---|
| 206 | CALL swap_dimensions(ind) |
---|
| 207 | CALL swap_geometry(ind) |
---|
| 208 | rhodz=f_rhodz(ind); ps=f_ps(ind); dps=f_dps(ind); |
---|
| 209 | wflux=f_wflux(ind); wfluxt=f_wfluxt(ind) |
---|
| 210 | CALL compute_rhodz(.FALSE., ps, rhodz) |
---|
| 211 | END DO |
---|
| 212 | |
---|
[133] | 213 | END IF |
---|
| 214 | |
---|
[124] | 215 | ! CALL physics(it,f_phis, f_ps, f_theta_rhodz, f_u, f_q) |
---|
[129] | 216 | ENDDO |
---|
| 217 | |
---|
[12] | 218 | CONTAINS |
---|
| 219 | |
---|
[130] | 220 | SUBROUTINE Euler_scheme(with_dps) |
---|
[12] | 221 | IMPLICIT NONE |
---|
[130] | 222 | LOGICAL :: with_dps |
---|
| 223 | INTEGER :: ind |
---|
| 224 | DO ind=1,ndomain |
---|
[138] | 225 | CALL swap_dimensions(ind) |
---|
| 226 | CALL swap_geometry(ind) |
---|
[130] | 227 | IF(with_dps) THEN |
---|
| 228 | ps=f_ps(ind) ; dps=f_dps(ind) ; |
---|
| 229 | ps(:)=ps(:)+dt*dps(:) |
---|
[133] | 230 | hflux=f_hflux(ind); hfluxt=f_hfluxt(ind) |
---|
[138] | 231 | wflux=f_wflux(ind); wfluxt=f_wfluxt(ind) |
---|
| 232 | CALL accumulate_fluxes(hflux,wflux,hfluxt,wfluxt,dt,fluxt_zero(ind)) |
---|
[130] | 233 | END IF |
---|
| 234 | u=f_u(ind) ; theta_rhodz=f_theta_rhodz(ind) |
---|
| 235 | du=f_du(ind) ; dtheta_rhodz=f_dtheta_rhodz(ind) |
---|
| 236 | u(:,:)=u(:,:)+dt*du(:,:) |
---|
| 237 | theta_rhodz(:,:)=theta_rhodz(:,:)+dt*dtheta_rhodz(:,:) |
---|
| 238 | ENDDO |
---|
[133] | 239 | |
---|
[12] | 240 | END SUBROUTINE Euler_scheme |
---|
[120] | 241 | |
---|
[129] | 242 | SUBROUTINE RK_scheme(stage) |
---|
[120] | 243 | IMPLICIT NONE |
---|
| 244 | INTEGER :: ind, stage |
---|
[129] | 245 | REAL(rstd), DIMENSION(4), PARAMETER :: coef = (/ .25, 1./3., .5, 1. /) |
---|
[120] | 246 | REAL(rstd) :: tau |
---|
| 247 | |
---|
| 248 | tau = dt*coef(stage) |
---|
[129] | 249 | |
---|
[120] | 250 | DO ind=1,ndomain |
---|
[138] | 251 | CALL swap_dimensions(ind) |
---|
| 252 | CALL swap_geometry(ind) |
---|
[120] | 253 | ps=f_ps(ind) ; u=f_u(ind) ; theta_rhodz=f_theta_rhodz(ind) |
---|
| 254 | psm1=f_psm1(ind) ; um1=f_um1(ind) ; theta_rhodzm1=f_theta_rhodzm1(ind) |
---|
| 255 | dps=f_dps(ind) ; du=f_du(ind) ; dtheta_rhodz=f_dtheta_rhodz(ind) |
---|
[129] | 256 | |
---|
| 257 | IF (stage==1) THEN ! first stage : save model state in XXm1 |
---|
[120] | 258 | psm1(:)=ps(:) ; um1(:,:)=u(:,:) ; theta_rhodzm1(:,:)=theta_rhodz(:,:) |
---|
| 259 | END IF |
---|
[129] | 260 | ! updates are of the form x1 := x0 + tau*f(x1) |
---|
[120] | 261 | ps(:)=psm1(:)+tau*dps(:) |
---|
| 262 | u(:,:)=um1(:,:)+tau*du(:,:) |
---|
| 263 | theta_rhodz(:,:)=theta_rhodzm1(:,:)+tau*dtheta_rhodz(:,:) |
---|
[133] | 264 | IF(stage==nb_stage) THEN ! accumulate mass fluxes at last stage |
---|
| 265 | hflux=f_hflux(ind); hfluxt=f_hfluxt(ind) |
---|
[138] | 266 | wflux=f_wflux(ind); wfluxt=f_wfluxt(ind) |
---|
| 267 | CALL accumulate_fluxes(hflux,wflux, hfluxt,wfluxt, dt,fluxt_zero(ind)) |
---|
[133] | 268 | END IF |
---|
[120] | 269 | END DO |
---|
| 270 | END SUBROUTINE RK_scheme |
---|
| 271 | |
---|
[12] | 272 | SUBROUTINE leapfrog_scheme |
---|
| 273 | IMPLICIT NONE |
---|
| 274 | INTEGER :: ind |
---|
| 275 | |
---|
| 276 | DO ind=1,ndomain |
---|
[138] | 277 | CALL swap_dimensions(ind) |
---|
| 278 | CALL swap_geometry(ind) |
---|
[12] | 279 | ps=f_ps(ind) ; u=f_u(ind) ; theta_rhodz=f_theta_rhodz(ind) |
---|
| 280 | psm1=f_psm1(ind) ; um1=f_um1(ind) ; theta_rhodzm1=f_theta_rhodzm1(ind) |
---|
| 281 | psm2=f_psm2(ind) ; um2=f_um2(ind) ; theta_rhodzm2=f_theta_rhodzm2(ind) |
---|
| 282 | dps=f_dps(ind) ; du=f_du(ind) ; dtheta_rhodz=f_dtheta_rhodz(ind) |
---|
| 283 | |
---|
| 284 | IF (it==0) THEN |
---|
| 285 | psm2(:)=ps(:) ; theta_rhodzm2(:,:)=theta_rhodz(:,:) ; um2(:,:)=u(:,:) |
---|
| 286 | |
---|
| 287 | ps(:)=ps(:)+dt*dps(:) |
---|
| 288 | u(:,:)=u(:,:)+dt*du(:,:) |
---|
| 289 | theta_rhodz(:,:)=theta_rhodz(:,:)+dt*dtheta_rhodz(:,:) |
---|
| 290 | |
---|
| 291 | psm1(:)=ps(:) ; theta_rhodzm1(:,:)=theta_rhodz(:,:) ; um1(:,:)=u(:,:) |
---|
| 292 | ELSE |
---|
| 293 | |
---|
| 294 | ps(:)=psm2(:)+2*dt*dps(:) |
---|
| 295 | u(:,:)=um2(:,:)+2*dt*du(:,:) |
---|
| 296 | theta_rhodz(:,:)=theta_rhodzm2(:,:)+2*dt*dtheta_rhodz(:,:) |
---|
| 297 | |
---|
| 298 | psm2(:)=psm1(:) ; theta_rhodzm2(:,:)=theta_rhodzm1(:,:) ; um2(:,:)=um1(:,:) |
---|
| 299 | psm1(:)=ps(:) ; theta_rhodzm1(:,:)=theta_rhodz(:,:) ; um1(:,:)=u(:,:) |
---|
| 300 | ENDIF |
---|
| 301 | |
---|
| 302 | ENDDO |
---|
| 303 | END SUBROUTINE leapfrog_scheme |
---|
| 304 | |
---|
[129] | 305 | SUBROUTINE leapfrog_matsuno_scheme(stage) |
---|
[12] | 306 | IMPLICIT NONE |
---|
[129] | 307 | INTEGER :: ind, stage |
---|
| 308 | REAL :: tau |
---|
| 309 | tau = dt/nb_stage |
---|
[12] | 310 | DO ind=1,ndomain |
---|
[138] | 311 | CALL swap_dimensions(ind) |
---|
| 312 | CALL swap_geometry(ind) |
---|
| 313 | |
---|
[12] | 314 | ps=f_ps(ind) ; u=f_u(ind) ; theta_rhodz=f_theta_rhodz(ind) |
---|
| 315 | psm1=f_psm1(ind) ; um1=f_um1(ind) ; theta_rhodzm1=f_theta_rhodzm1(ind) |
---|
| 316 | psm2=f_psm2(ind) ; um2=f_um2(ind) ; theta_rhodzm2=f_theta_rhodzm2(ind) |
---|
| 317 | dps=f_dps(ind) ; du=f_du(ind) ; dtheta_rhodz=f_dtheta_rhodz(ind) |
---|
| 318 | |
---|
| 319 | |
---|
[129] | 320 | ! IF (MOD(it,matsuno_period+1)==0) THEN |
---|
| 321 | IF (stage==1) THEN |
---|
[12] | 322 | psm1(:)=ps(:) ; um1(:,:)=u(:,:) ; theta_rhodzm1(:,:)=theta_rhodz(:,:) |
---|
[129] | 323 | ps(:)=psm1(:)+tau*dps(:) |
---|
| 324 | u(:,:)=um1(:,:)+tau*du(:,:) |
---|
| 325 | theta_rhodz(:,:)=theta_rhodzm1(:,:)+tau*dtheta_rhodz(:,:) |
---|
[12] | 326 | |
---|
[129] | 327 | ! ELSE IF (MOD(it,matsuno_period+1)==1) THEN |
---|
| 328 | ELSE IF (stage==2) THEN |
---|
[12] | 329 | |
---|
[129] | 330 | ps(:)=psm1(:)+tau*dps(:) |
---|
| 331 | u(:,:)=um1(:,:)+tau*du(:,:) |
---|
| 332 | theta_rhodz(:,:)=theta_rhodzm1(:,:)+tau*dtheta_rhodz(:,:) |
---|
[12] | 333 | |
---|
| 334 | psm2(:)=psm1(:) ; theta_rhodzm2(:,:)=theta_rhodzm1(:,:) ; um2(:,:)=um1(:,:) |
---|
| 335 | psm1(:)=ps(:) ; theta_rhodzm1(:,:)=theta_rhodz(:,:) ; um1(:,:)=u(:,:) |
---|
| 336 | |
---|
| 337 | ELSE |
---|
| 338 | |
---|
[129] | 339 | ps(:)=psm2(:)+2*tau*dps(:) |
---|
| 340 | u(:,:)=um2(:,:)+2*tau*du(:,:) |
---|
| 341 | theta_rhodz(:,:)=theta_rhodzm2(:,:)+2*tau*dtheta_rhodz(:,:) |
---|
[12] | 342 | |
---|
| 343 | psm2(:)=psm1(:) ; theta_rhodzm2(:,:)=theta_rhodzm1(:,:) ; um2(:,:)=um1(:,:) |
---|
| 344 | psm1(:)=ps(:) ; theta_rhodzm1(:,:)=theta_rhodz(:,:) ; um1(:,:)=u(:,:) |
---|
| 345 | |
---|
| 346 | ENDIF |
---|
| 347 | |
---|
| 348 | ENDDO |
---|
| 349 | |
---|
| 350 | END SUBROUTINE leapfrog_matsuno_scheme |
---|
| 351 | |
---|
| 352 | SUBROUTINE adam_bashforth_scheme |
---|
| 353 | IMPLICIT NONE |
---|
| 354 | INTEGER :: ind |
---|
| 355 | |
---|
| 356 | DO ind=1,ndomain |
---|
[138] | 357 | CALL swap_dimensions(ind) |
---|
| 358 | CALL swap_geometry(ind) |
---|
[12] | 359 | ps=f_ps(ind) ; u=f_u(ind) ; theta_rhodz=f_theta_rhodz(ind) |
---|
| 360 | dps=f_dps(ind) ; du=f_du(ind) ; dtheta_rhodz=f_dtheta_rhodz(ind) |
---|
| 361 | dpsm1=f_dpsm1(ind) ; dum1=f_dum1(ind) ; dtheta_rhodzm1=f_dtheta_rhodzm1(ind) |
---|
| 362 | dpsm2=f_dpsm2(ind) ; dum2=f_dum2(ind) ; dtheta_rhodzm2=f_dtheta_rhodzm2(ind) |
---|
| 363 | |
---|
| 364 | IF (it==0) THEN |
---|
| 365 | dpsm1(:)=dps(:) ; dum1(:,:)=du(:,:) ; dtheta_rhodzm1(:,:)=dtheta_rhodz(:,:) |
---|
| 366 | dpsm2(:)=dpsm1(:) ; dum2(:,:)=dum1(:,:) ; dtheta_rhodzm2(:,:)=dtheta_rhodzm1(:,:) |
---|
| 367 | ENDIF |
---|
| 368 | |
---|
| 369 | ps(:)=ps(:)+dt*(23*dps(:)-16*dpsm1(:)+5*dpsm2(:))/12 |
---|
| 370 | u(:,:)=u(:,:)+dt*(23*du(:,:)-16*dum1(:,:)+5*dum2(:,:))/12 |
---|
| 371 | theta_rhodz(:,:)=theta_rhodz(:,:)+dt*(23*dtheta_rhodz(:,:)-16*dtheta_rhodzm1(:,:)+5*dtheta_rhodzm2(:,:))/12 |
---|
| 372 | |
---|
| 373 | dpsm2(:)=dpsm1(:) ; dum2(:,:)=dum1(:,:) ; dtheta_rhodzm2(:,:)=dtheta_rhodzm1(:,:) |
---|
| 374 | dpsm1(:)=dps(:) ; dum1(:,:)=du(:,:) ; dtheta_rhodzm1(:,:)=dtheta_rhodz(:,:) |
---|
| 375 | |
---|
[50] | 376 | ENDDO |
---|
[12] | 377 | |
---|
| 378 | END SUBROUTINE adam_bashforth_scheme |
---|
| 379 | |
---|
[50] | 380 | END SUBROUTINE timeloop |
---|
[133] | 381 | |
---|
[138] | 382 | SUBROUTINE compute_rhodz(comp, ps, rhodz) |
---|
[133] | 383 | USE icosa |
---|
| 384 | USE disvert_mod |
---|
[138] | 385 | LOGICAL, INTENT(IN) :: comp ! .TRUE. to compute, .FALSE. to check |
---|
[133] | 386 | REAL(rstd), INTENT(IN) :: ps(iim*jjm) |
---|
| 387 | REAL(rstd), INTENT(OUT) :: rhodz(iim*jjm,llm) |
---|
[138] | 388 | REAL(rstd) :: m, err |
---|
| 389 | INTEGER :: l,i,j,ij,dd |
---|
| 390 | err=0. |
---|
| 391 | IF(comp) THEN |
---|
| 392 | dd=1 |
---|
| 393 | ELSE |
---|
| 394 | ! dd=-1 |
---|
| 395 | dd=0 |
---|
| 396 | END IF |
---|
| 397 | |
---|
[133] | 398 | DO l = 1, llm |
---|
[138] | 399 | DO j=jj_begin-dd,jj_end+dd |
---|
| 400 | DO i=ii_begin-dd,ii_end+dd |
---|
[133] | 401 | ij=(j-1)*iim+i |
---|
[138] | 402 | m = ( ap(l) - ap(l+1) + (bp(l)-bp(l+1))*ps(ij) )/g |
---|
| 403 | IF(comp) THEN |
---|
| 404 | rhodz(ij,l) = m |
---|
| 405 | ELSE |
---|
| 406 | err = MAX(err,abs(m-rhodz(ij,l))) |
---|
| 407 | END IF |
---|
[133] | 408 | ENDDO |
---|
| 409 | ENDDO |
---|
| 410 | ENDDO |
---|
[138] | 411 | |
---|
| 412 | IF(.NOT. comp) THEN |
---|
| 413 | IF(err>1e-10) THEN |
---|
| 414 | PRINT *, 'Discrepancy between ps and rhodz detected', err |
---|
| 415 | STOP |
---|
| 416 | ELSE |
---|
| 417 | ! PRINT *, 'No discrepancy between ps and rhodz detected' |
---|
| 418 | END IF |
---|
| 419 | END IF |
---|
| 420 | |
---|
[133] | 421 | END SUBROUTINE compute_rhodz |
---|
| 422 | |
---|
[138] | 423 | SUBROUTINE accumulate_fluxes(hflux,wflux, hfluxt,wfluxt, tau,fluxt_zero) |
---|
[133] | 424 | USE icosa |
---|
| 425 | REAL(rstd), INTENT(IN) :: hflux(3*iim*jjm,llm), wflux(iim*jjm,llm+1) |
---|
| 426 | REAL(rstd), INTENT(INOUT) :: hfluxt(3*iim*jjm,llm), wfluxt(iim*jjm,llm+1) |
---|
| 427 | REAL(rstd), INTENT(IN) :: tau |
---|
[134] | 428 | LOGICAL, INTENT(INOUT) :: fluxt_zero |
---|
| 429 | IF(fluxt_zero) THEN |
---|
[138] | 430 | ! PRINT *, 'Accumulating fluxes (first)' |
---|
[134] | 431 | fluxt_zero=.FALSE. |
---|
| 432 | hfluxt = tau*hflux |
---|
| 433 | wfluxt = tau*wflux |
---|
| 434 | ELSE |
---|
[138] | 435 | ! PRINT *, 'Accumulating fluxes (next)' |
---|
[134] | 436 | hfluxt = hfluxt + tau*hflux |
---|
| 437 | wfluxt = wfluxt + tau*wflux |
---|
| 438 | END IF |
---|
[133] | 439 | END SUBROUTINE accumulate_fluxes |
---|
[12] | 440 | |
---|
| 441 | END MODULE timeloop_gcm_mod |
---|