[8] | 1 | !********************************************************************** |
---|
| 2 | ! Copyright 1998,1999,2000,2001,2002,2005,2007,2008,2009,2010 * |
---|
| 3 | ! Andreas Stohl, Petra Seibert, A. Frank, Gerhard Wotawa, * |
---|
| 4 | ! Caroline Forster, Sabine Eckhardt, John Burkhart, Harald Sodemann * |
---|
| 5 | ! * |
---|
| 6 | ! This file is part of FLEXPART. * |
---|
| 7 | ! * |
---|
| 8 | ! FLEXPART is free software: you can redistribute it and/or modify * |
---|
| 9 | ! it under the terms of the GNU General Public License as published by* |
---|
| 10 | ! the Free Software Foundation, either version 3 of the License, or * |
---|
| 11 | ! (at your option) any later version. * |
---|
| 12 | ! * |
---|
| 13 | ! FLEXPART is distributed in the hope that it will be useful, * |
---|
| 14 | ! but WITHOUT ANY WARRANTY; without even the implied warranty of * |
---|
| 15 | ! MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * |
---|
| 16 | ! GNU General Public License for more details. * |
---|
| 17 | ! * |
---|
| 18 | ! You should have received a copy of the GNU General Public License * |
---|
| 19 | ! along with FLEXPART. If not, see <http://www.gnu.org/licenses/>. * |
---|
| 20 | !********************************************************************** |
---|
| 21 | |
---|
| 22 | subroutine convmix(itime) |
---|
| 23 | ! i |
---|
| 24 | !************************************************************** |
---|
| 25 | !handles all the calculations related to convective mixing |
---|
| 26 | !Petra Seibert, Bernd C. Krueger, Feb 2001 |
---|
| 27 | !nested grids included, Bernd C. Krueger, May 2001 |
---|
| 28 | ! |
---|
| 29 | !Changes by Caroline Forster, April 2004 - February 2005: |
---|
| 30 | ! convmix called every lsynctime seconds |
---|
| 31 | !CHANGES by A. Stohl: |
---|
| 32 | ! various run-time optimizations - February 2005 |
---|
| 33 | !************************************************************** |
---|
| 34 | |
---|
| 35 | use par_mod, only: nxmax, maxpart, maxnests |
---|
| 36 | |
---|
| 37 | use com_mod, only: iflux, xresoln, xln, xrn, yln, yrn, tt2, ps, itra1, & |
---|
| 38 | nxn, nyn, xtra1, ytra1, itramem, ztra1, lage, qvh, tth, td2, qvhn, & |
---|
| 39 | tthn, td2n, tt2n, psn, path, yresoln, memind, memtime, nx, ny, numpart, & |
---|
| 40 | nageclass, nuvz, numbnests, lsynctime |
---|
| 41 | |
---|
| 42 | use conv_mod, only: tconv, qconv, pconv, nconvlev, psconv, & |
---|
| 43 | td2conv, tt2conv, cbaseflux, cbasefluxn,fmassfrac, nconvtop |
---|
| 44 | |
---|
| 45 | use random_mod, only: ran3, ran3_conv, idummy_ran3_conv |
---|
| 46 | |
---|
| 47 | use omp_lib, only: OMP_GET_THREAD_NUM |
---|
| 48 | |
---|
| 49 | implicit none |
---|
| 50 | |
---|
| 51 | integer :: igr,igrold, ipart, itime, ix, j, inest |
---|
| 52 | integer :: ipconv |
---|
| 53 | integer :: jy, kpart, ktop, ngrid,kz |
---|
| 54 | integer,allocatable,dimension(:) :: igrid, ipoint |
---|
| 55 | integer,allocatable,dimension(:,:) :: igridn |
---|
| 56 | ! itime [s] current time |
---|
| 57 | ! igrid(maxpart) horizontal grid position of each particle |
---|
| 58 | ! igridn(maxpart,maxnests) dto. for nested grids |
---|
| 59 | ! ipoint(maxpart) pointer to access particles according to grid position |
---|
| 60 | |
---|
| 61 | logical :: lconv |
---|
| 62 | real :: x, y, xtn,ytn, ztold, delt |
---|
| 63 | real :: dt1,dt2,dtt |
---|
| 64 | integer :: mind1,mind2 |
---|
| 65 | ! dt1,dt2,dtt,mind1,mind2 variables used for time interpolation |
---|
| 66 | integer :: itage,nage |
---|
| 67 | real,parameter :: eps=nxmax/3.e5 |
---|
| 68 | integer, allocatable, dimension(:) :: frst |
---|
| 69 | integer :: cnt, stat, kk |
---|
| 70 | ! for debugging, remove again |
---|
| 71 | ! character(len=255) :: fn |
---|
| 72 | ! integer :: thread |
---|
| 73 | ! integer :: sumconv |
---|
| 74 | |
---|
| 75 | ! allocate local variables on the heap |
---|
| 76 | allocate(igrid(maxpart),stat=stat) |
---|
| 77 | allocate(ipoint(maxpart),stat=stat) |
---|
| 78 | allocate(igridn(maxpart,maxnests),stat=stat) |
---|
| 79 | |
---|
| 80 | !monitoring variables |
---|
| 81 | !real sumconv,sumall |
---|
| 82 | |
---|
| 83 | |
---|
| 84 | ! Calculate auxiliary variables for time interpolation |
---|
| 85 | !***************************************************** |
---|
| 86 | |
---|
| 87 | dt1=real(itime-memtime(1)) |
---|
| 88 | dt2=real(memtime(2)-itime) |
---|
| 89 | dtt=1./(dt1+dt2) |
---|
| 90 | mind1=memind(1) |
---|
| 91 | mind2=memind(2) |
---|
| 92 | delt=real(abs(lsynctime)) |
---|
| 93 | |
---|
| 94 | |
---|
| 95 | lconv = .false. |
---|
| 96 | |
---|
| 97 | ! if no particles are present return after initialization |
---|
| 98 | !******************************************************** |
---|
| 99 | |
---|
| 100 | if (numpart.le.0) return |
---|
| 101 | |
---|
| 102 | ! Assign igrid and igridn, which are pseudo grid numbers indicating particles |
---|
| 103 | ! that are outside the part of the grid under consideration |
---|
| 104 | ! (e.g. particles near the poles or particles in other nests). |
---|
| 105 | ! Do this for all nests but use only the innermost nest; for all others |
---|
| 106 | ! igrid shall be -1 |
---|
| 107 | ! Also, initialize index vector ipoint |
---|
| 108 | !************************************************************************ |
---|
| 109 | |
---|
| 110 | !$OMP PARALLEL DEFAULT(none) & |
---|
| 111 | !$OMP PRIVATE(ipart, x, y, j, ngrid, xtn, ytn, ix, jy) & |
---|
| 112 | !$OMP SHARED(numpart, igrid, igridn, ipoint, itra1, itime, numbnests, & |
---|
| 113 | !$OMP xln, yln, xrn, yrn, xresoln, yresoln, xtra1, ytra1, nxn, nx) |
---|
| 114 | |
---|
| 115 | #if (defined STATIC_SCHED) |
---|
| 116 | !$OMP DO SCHEDULE(static) |
---|
| 117 | #else |
---|
| 118 | !$OMP DO SCHEDULE(dynamic, max(1,numpart/1000)) |
---|
| 119 | #endif |
---|
| 120 | |
---|
| 121 | do ipart=1,numpart |
---|
| 122 | igrid(ipart)=-1 |
---|
| 123 | do j=numbnests,1,-1 |
---|
| 124 | igridn(ipart,j)=-1 |
---|
| 125 | end do |
---|
| 126 | ipoint(ipart)=ipart |
---|
| 127 | ! do not consider particles that are (yet) not part of simulation |
---|
| 128 | if (itra1(ipart).ne.itime) cycle |
---|
| 129 | |
---|
| 130 | x = real(xtra1(ipart)) |
---|
| 131 | y = real(ytra1(ipart)) |
---|
| 132 | |
---|
| 133 | ! Determine which nesting level to be used |
---|
| 134 | !********************************************************** |
---|
| 135 | |
---|
| 136 | ngrid=0 |
---|
| 137 | do j=numbnests,1,-1 |
---|
| 138 | if ( x.gt.xln(j)+eps .and. x.lt.xrn(j)-eps .and. & |
---|
| 139 | y.gt.yln(j)+eps .and. y.lt.yrn(j)-eps ) then |
---|
| 140 | ngrid=j |
---|
| 141 | exit |
---|
| 142 | endif |
---|
| 143 | end do |
---|
| 144 | |
---|
| 145 | ! Determine nested grid coordinates |
---|
| 146 | !********************************** |
---|
| 147 | |
---|
| 148 | if (ngrid.gt.0) then |
---|
| 149 | ! nested grids |
---|
| 150 | xtn=(x-xln(ngrid))*xresoln(ngrid) |
---|
| 151 | ytn=(y-yln(ngrid))*yresoln(ngrid) |
---|
| 152 | ix=nint(xtn) |
---|
| 153 | jy=nint(ytn) |
---|
| 154 | igridn(ipart,ngrid) = 1 + jy*nxn(ngrid) + ix |
---|
| 155 | else if(ngrid.eq.0) then |
---|
| 156 | ! mother grid |
---|
| 157 | ix=nint(x) |
---|
| 158 | jy=nint(y) |
---|
| 159 | igrid(ipart) = 1 + jy*nx + ix |
---|
| 160 | endif |
---|
| 161 | |
---|
| 162 | end do |
---|
| 163 | !$OMP END DO |
---|
| 164 | !$OMP END PARALLEL |
---|
| 165 | |
---|
| 166 | !sumall = 0. |
---|
| 167 | !sumconv = 0. |
---|
| 168 | !***************************************************************************** |
---|
| 169 | ! 1. Now, do everything for the mother domain and, later, for all of the nested domains |
---|
| 170 | ! While all particles have to be considered for redistribution, the Emanuel convection |
---|
| 171 | ! scheme only needs to be called once for every grid column where particles are present. |
---|
| 172 | ! Therefore, particles are sorted according to their grid position. Whenever a new grid |
---|
| 173 | ! cell is encountered by looping through the sorted particles, the convection scheme is called. |
---|
| 174 | !***************************************************************************** |
---|
| 175 | |
---|
| 176 | ! sort particles according to horizontal position and calculate index vector IPOINT |
---|
| 177 | |
---|
| 178 | call sort2(numpart,igrid,ipoint) |
---|
| 179 | |
---|
| 180 | ! determine index where new grid cell starts |
---|
| 181 | ! by this also selecting grid cells where particles are present |
---|
| 182 | allocate(frst(nx*ny), stat=stat) |
---|
| 183 | if (stat.ne.0) write(*,*) "ERROR: could not allocate frst" |
---|
| 184 | |
---|
| 185 | cnt = 1 |
---|
| 186 | igrold = -1 |
---|
| 187 | do kpart=1,numpart |
---|
| 188 | |
---|
| 189 | if (igrold.ne.igrid(kpart)) then |
---|
| 190 | frst(cnt) = kpart |
---|
| 191 | igrold=igrid(kpart) |
---|
| 192 | cnt=cnt+1 |
---|
| 193 | endif |
---|
| 194 | enddo |
---|
| 195 | frst(cnt) = numpart+1 |
---|
| 196 | ! if all particles in nested grids cnt will be 1 otherwise >1 |
---|
| 197 | |
---|
| 198 | do kpart=1, numpart |
---|
| 199 | ! get random number outside parallel region |
---|
| 200 | ! (not necessary to do again for nested domains, because each particle is only treated once) |
---|
| 201 | if (itra1(kpart).eq.itime) then |
---|
| 202 | ran3_conv(kpart) = ran3(idummy_ran3_conv) |
---|
| 203 | endif |
---|
| 204 | end do |
---|
| 205 | |
---|
| 206 | ! Loop over grid columns with particles |
---|
| 207 | if (cnt.gt.1) then ! any particles in non-nested areas? |
---|
| 208 | |
---|
| 209 | |
---|
| 210 | !$OMP PARALLEL DEFAULT (none) & |
---|
| 211 | !$OMP PRIVATE(kk,jy,ix,kz, ktop, lconv, kpart, ipart, ztold, & |
---|
| 212 | !$OMP nage, ipconv, itage) & |
---|
| 213 | !$OMP SHARED(cnt, cbaseflux, frst, igrid, nx, mind1, mind2, ps, & |
---|
| 214 | !$OMP tt2, td2, tth, qvh, dt1, dt2, dtt, nuvz, delt, ipoint, & |
---|
| 215 | !$OMP iflux, itra1, itramem, nageclass, lage, xtra1, ytra1, ztra1) |
---|
| 216 | |
---|
| 217 | ! #if (defined _OPENMP) |
---|
| 218 | ! thread = OMP_GET_THREAD_NUM() |
---|
| 219 | ! #endif |
---|
| 220 | |
---|
| 221 | #if (defined STATIC_SCHED) |
---|
| 222 | !$OMP DO SCHEDULE(static) |
---|
| 223 | #else |
---|
| 224 | !$OMP DO SCHEDULE(dynamic) ! using default chunck sizes |
---|
| 225 | #endif |
---|
| 226 | do kk=1,cnt-1 ! loop through grid columns |
---|
| 227 | |
---|
| 228 | ! mother grid cell in nested domain, don't calculate convection |
---|
| 229 | if (igrid(frst(kk)) .eq. -1) cycle |
---|
| 230 | |
---|
| 231 | jy = (igrid(frst(kk))-1)/nx |
---|
| 232 | ix = igrid(frst(kk)) - jy*nx - 1 |
---|
| 233 | |
---|
| 234 | ! Interpolate all meteorological data needed for the convection scheme |
---|
| 235 | psconv=(ps(ix,jy,1,mind1)*dt2+ps(ix,jy,1,mind2)*dt1)*dtt |
---|
| 236 | tt2conv=(tt2(ix,jy,1,mind1)*dt2+tt2(ix,jy,1,mind2)*dt1)*dtt |
---|
| 237 | td2conv=(td2(ix,jy,1,mind1)*dt2+td2(ix,jy,1,mind2)*dt1)*dtt |
---|
| 238 | !!$ do kz=1,nconvlev+1 !old |
---|
| 239 | do kz=1,nuvz-1 !bugfix |
---|
| 240 | tconv(kz)=(tth(ix,jy,kz+1,mind1)*dt2+ & |
---|
| 241 | tth(ix,jy,kz+1,mind2)*dt1)*dtt |
---|
| 242 | qconv(kz)=(qvh(ix,jy,kz+1,mind1)*dt2+ & |
---|
| 243 | qvh(ix,jy,kz+1,mind2)*dt1)*dtt |
---|
| 244 | end do |
---|
| 245 | |
---|
| 246 | ! Calculate translocation matrix |
---|
| 247 | call calcmatrix(lconv,delt,cbaseflux(ix,jy)) |
---|
| 248 | |
---|
| 249 | ! treat column only if column has convection |
---|
| 250 | if (lconv) then |
---|
| 251 | ktop = 0 |
---|
| 252 | ! sumconv = 0 |
---|
| 253 | ! assign new vertical position to particle |
---|
| 254 | do kpart=frst(kk),frst(kk+1)-1 |
---|
| 255 | ipart = ipoint(kpart) |
---|
| 256 | ztold=ztra1(ipart) |
---|
| 257 | call redist(ipart,ktop,ipconv) |
---|
| 258 | |
---|
| 259 | ! if (ipconv.le.0) sumconv = sumconv+1 |
---|
| 260 | |
---|
| 261 | ! Calculate the gross fluxes across layer interfaces |
---|
| 262 | !*************************************************** |
---|
| 263 | |
---|
| 264 | if (iflux.eq.1) then |
---|
| 265 | itage=abs(itra1(ipart)-itramem(ipart)) |
---|
| 266 | do nage=1,nageclass |
---|
| 267 | if (itage.lt.lage(nage)) exit |
---|
| 268 | end do |
---|
| 269 | |
---|
| 270 | if (nage.le.nageclass) then |
---|
| 271 | call calcfluxes(nage,ipart,real(xtra1(ipart)), & |
---|
| 272 | real(ytra1(ipart)),ztold) |
---|
| 273 | end if |
---|
| 274 | end if ! flux calculation |
---|
| 275 | end do ! particles in column |
---|
| 276 | ! write(*,*) "convmix:", thread, igrid(frst(kk)), cbaseflux(ix,jy), sumconv |
---|
| 277 | end if ! convection present |
---|
| 278 | end do ! grid cell loop |
---|
| 279 | !$OMP END DO |
---|
| 280 | !$OMP END PARALLEL |
---|
| 281 | end if ! particles in non-nested areas |
---|
| 282 | |
---|
| 283 | deallocate(frst) |
---|
| 284 | |
---|
| 285 | ! write(fn, '(A,A,I6.6)') trim(path(2)), 'cbaseflux_', itime |
---|
| 286 | ! call dump_field(fn, cbaseflux(0:nx-1, 0:ny-1), nx, ny, 1) |
---|
| 287 | |
---|
| 288 | |
---|
| 289 | !***************************************************************************** |
---|
| 290 | ! 2. Nested domains |
---|
| 291 | !***************************************************************************** |
---|
| 292 | |
---|
| 293 | nestloop : do inest=1,numbnests |
---|
| 294 | |
---|
| 295 | ! sort particles according to horizontal position and calculate index vector IPOINT |
---|
| 296 | do ipart=1,numpart |
---|
| 297 | ipoint(ipart)=ipart |
---|
| 298 | igrid(ipart) = igridn(ipart,inest) |
---|
| 299 | enddo |
---|
| 300 | call sort2(numpart,igrid,ipoint) |
---|
| 301 | |
---|
| 302 | ! determine particle index where new grid cell starts |
---|
| 303 | ! by this also selecting grid cells where particles are present |
---|
| 304 | allocate(frst(nxn(inest)*nyn(inest)), stat=stat) |
---|
| 305 | if (stat.ne.0) write(*,*) "ERROR: could not allocate frst" |
---|
| 306 | |
---|
| 307 | cnt = 1 |
---|
| 308 | igrold = -1 |
---|
| 309 | do kpart=1,numpart |
---|
| 310 | if (igrold.ne.igrid(kpart)) then |
---|
| 311 | frst(cnt) = kpart |
---|
| 312 | igrold=igrid(kpart) |
---|
| 313 | cnt=cnt+1 |
---|
| 314 | endif |
---|
| 315 | enddo |
---|
| 316 | frst(cnt) = numpart+1 |
---|
| 317 | ! if all particles in nested grids cnt will be 1 otherwise >1 |
---|
| 318 | |
---|
| 319 | if (cnt.gt.1) then ! any particles in nested areas? |
---|
| 320 | |
---|
| 321 | !$OMP PARALLEL DEFAULT(none) & |
---|
| 322 | !$OMP PRIVATE(kk,jy,ix,kz, ktop, lconv, kpart, ipart, ztold, & |
---|
| 323 | !$OMP nage, ipconv, itage) & |
---|
| 324 | !$OMP SHARED(cnt, cbasefluxn, frst, igrid, nxn, mind1, mind2, psn, & |
---|
| 325 | !$OMP tt2n, td2n, tthn, qvhn, dt1, dt2, dtt, nuvz, delt, ipoint, & |
---|
| 326 | !$OMP iflux, itra1, itramem, nageclass, lage, xtra1, ytra1, ztra1, & |
---|
| 327 | !$OMP inest) |
---|
| 328 | |
---|
| 329 | #if (defined STATIC_SCHED) |
---|
| 330 | !$OMP DO SCHEDULE(static) |
---|
| 331 | #else |
---|
| 332 | !$OMP DO SCHEDULE(dynamic) ! using default chunck sizes |
---|
| 333 | #endif |
---|
| 334 | do kk=1,cnt-1 ! loop through grid columns |
---|
| 335 | |
---|
| 336 | jy = (igrid(frst(kk))-1)/nxn(inest) |
---|
| 337 | ix = igrid(frst(kk)) - jy*nxn(inest) - 1 |
---|
| 338 | |
---|
| 339 | ! Interpolate all meteorological data needed for the convection scheme |
---|
| 340 | psconv=(psn(ix,jy,1,mind1,inest)*dt2+ & |
---|
| 341 | psn(ix,jy,1,mind2,inest)*dt1)*dtt |
---|
| 342 | tt2conv=(tt2n(ix,jy,1,mind1,inest)*dt2+ & |
---|
| 343 | tt2n(ix,jy,1,mind2,inest)*dt1)*dtt |
---|
| 344 | td2conv=(td2n(ix,jy,1,mind1,inest)*dt2+ & |
---|
| 345 | td2n(ix,jy,1,mind2,inest)*dt1)*dtt |
---|
| 346 | !!$ do kz=1,nconvlev+1 !old |
---|
| 347 | do kz=1,nuvz-1 !bugfix |
---|
| 348 | tconv(kz)=(tthn(ix,jy,kz+1,mind1,inest)*dt2+ & |
---|
| 349 | tthn(ix,jy,kz+1,mind2,inest)*dt1)*dtt |
---|
| 350 | qconv(kz)=(qvhn(ix,jy,kz+1,mind1,inest)*dt2+ & |
---|
| 351 | qvhn(ix,jy,kz+1,mind2,inest)*dt1)*dtt |
---|
| 352 | end do |
---|
| 353 | |
---|
| 354 | ! calculate translocation matrix |
---|
| 355 | !******************************* |
---|
| 356 | call calcmatrix(lconv,delt,cbasefluxn(ix,jy,inest)) |
---|
| 357 | |
---|
| 358 | ! treat particle only if column has convection |
---|
| 359 | if (lconv) then |
---|
| 360 | ktop = 0 |
---|
| 361 | do kpart=frst(kk), frst(kk+1)-1 |
---|
| 362 | ! assign new vertical position to particle |
---|
| 363 | ipart = ipoint(kpart) |
---|
| 364 | ztold=ztra1(ipart) |
---|
| 365 | call redist(ipart,ktop,ipconv) |
---|
| 366 | ! if (ipconv.le.0) sumconv = sumconv+1 |
---|
| 367 | |
---|
| 368 | ! Calculate the gross fluxes across layer interfaces |
---|
| 369 | !*************************************************** |
---|
| 370 | |
---|
| 371 | if (iflux.eq.1) then |
---|
| 372 | itage=abs(itra1(ipart)-itramem(ipart)) |
---|
| 373 | do nage=1,nageclass |
---|
| 374 | if (itage.lt.lage(nage)) exit |
---|
| 375 | end do |
---|
| 376 | |
---|
| 377 | if (nage.le.nageclass) then |
---|
| 378 | call calcfluxes(nage,ipart,real(xtra1(ipart)), & |
---|
| 379 | real(ytra1(ipart)),ztold) |
---|
| 380 | end if |
---|
| 381 | end if ! flux calculation |
---|
| 382 | |
---|
| 383 | end do ! particles in colum |
---|
| 384 | end if !(lconv .eqv. .true.) |
---|
| 385 | end do ! grid cell loop |
---|
| 386 | !$OMP END DO |
---|
| 387 | !$OMP END PARALLEL |
---|
| 388 | end if ! any particles in nest |
---|
| 389 | deallocate(frst) |
---|
| 390 | end do nestloop |
---|
| 391 | |
---|
| 392 | !-------------------------------------------------------------------------- |
---|
| 393 | !write(*,*)'############################################' |
---|
| 394 | !write(*,*)'TIME=', |
---|
| 395 | ! & itime |
---|
| 396 | !write(*,*)'fraction of particles under convection', |
---|
| 397 | ! & sumconv/(sumall+0.001) |
---|
| 398 | !write(*,*)'total number of particles', |
---|
| 399 | ! & sumall |
---|
| 400 | !write(*,*)'number of particles under convection', |
---|
| 401 | ! & sumconv |
---|
| 402 | !write(*,*)'############################################' |
---|
| 403 | |
---|
| 404 | deallocate(igrid) |
---|
| 405 | deallocate(ipoint) |
---|
| 406 | deallocate(igridn) |
---|
| 407 | |
---|
| 408 | return |
---|
| 409 | end subroutine convmix |
---|