| 1 |
C $Header: /u/gcmpack/MITgcm/pkg/flt/flt_bilinear.F,v 1.1 2001/09/13 17:43:55 adcroft Exp $ |
| 2 |
C $Name: $ |
| 3 |
|
| 4 |
#include "FLT_CPPOPTIONS.h" |
| 5 |
|
| 6 |
subroutine flt_bilinear( |
| 7 |
I xp, |
| 8 |
I yp, |
| 9 |
O uu, |
| 10 |
I kp, |
| 11 |
I u, |
| 12 |
I nu, |
| 13 |
I bi, |
| 14 |
I bj |
| 15 |
& ) |
| 16 |
|
| 17 |
c ================================================================== |
| 18 |
c SUBROUTINE flt_bilinear |
| 19 |
c ================================================================== |
| 20 |
c |
| 21 |
c o Bilinear scheme to find u of particle at given xp,yp location |
| 22 |
c |
| 23 |
c ================================================================== |
| 24 |
c SUBROUTINE flt_bilinear |
| 25 |
c ================================================================== |
| 26 |
|
| 27 |
c == global variables == |
| 28 |
|
| 29 |
#include "SIZE.h" |
| 30 |
|
| 31 |
c == routine arguments == |
| 32 |
|
| 33 |
_RL xp, yp |
| 34 |
_RL uu |
| 35 |
integer nu, kp, bi, bj |
| 36 |
_RL u (1-OLx:sNx+OLx,1-OLy:sNy+OLy,Nr,nSx,nSy) |
| 37 |
|
| 38 |
c == local variables == |
| 39 |
|
| 40 |
INTEGER nnx, nny, nfx, nfy, nfxp, nfyp |
| 41 |
_RL dx, dy, ddx, ddy |
| 42 |
integer ip |
| 43 |
_RL xx, yy, phi, scalex, scaley |
| 44 |
_RL u11, u12, u22, u21 |
| 45 |
|
| 46 |
c == end of interface == |
| 47 |
|
| 48 |
nnx = int(xp) |
| 49 |
nny = int(yp) |
| 50 |
dx = xp - float(nnx) |
| 51 |
dy = yp - float(nny) |
| 52 |
c |
| 53 |
c to choose the u box in which the particle is found |
| 54 |
c nu=1 for T, S |
| 55 |
c nu=2 for u |
| 56 |
c nu=3 for v |
| 57 |
c nu=4 for w |
| 58 |
c |
| 59 |
if (nu.eq.1.or.nu.eq.4) then |
| 60 |
nfx = nnx |
| 61 |
nfy = nny |
| 62 |
ddx = dx |
| 63 |
ddy = dy |
| 64 |
endif |
| 65 |
c |
| 66 |
if (nu.eq.2) then |
| 67 |
if (dx.le.0.5) then |
| 68 |
nfx = nnx |
| 69 |
ddx = dx + 0.5 |
| 70 |
else |
| 71 |
nfx = nnx + 1 |
| 72 |
ddx = dx - 0.5 |
| 73 |
endif |
| 74 |
nfy = nny |
| 75 |
ddy = dy |
| 76 |
endif |
| 77 |
c |
| 78 |
if (nu.eq.3) then |
| 79 |
if (dy.le.0.5) then |
| 80 |
nfy = nny |
| 81 |
ddy = dy + 0.5 |
| 82 |
else |
| 83 |
nfy = nny + 1 |
| 84 |
ddy = dy - 0.5 |
| 85 |
endif |
| 86 |
nfx = nnx |
| 87 |
ddx = dx |
| 88 |
endif |
| 89 |
c |
| 90 |
c |
| 91 |
cab change start |
| 92 |
c was correct only for global? |
| 93 |
c if(nfx.gt.nx) nfx=nfx-nx |
| 94 |
if(nfx.gt.nx) nfx=nx |
| 95 |
cab change end |
| 96 |
if(nfy.gt.ny) nfy=ny |
| 97 |
nfxp = nfx + 1 |
| 98 |
nfyp = nfy + 1 |
| 99 |
cab change start |
| 100 |
c if (nfx.eq.nx) nfxp = 1 |
| 101 |
if (nfx.eq.nx) nfxp = nfx |
| 102 |
cab change end |
| 103 |
if (nfy.eq.ny) nfyp = nfy |
| 104 |
|
| 105 |
if (nu.lt.4) then |
| 106 |
u11 = u(nfx,nfy,kp,bi,bj) |
| 107 |
u21 = u(nfxp,nfy,kp,bi,bj) |
| 108 |
u22 = u(nfxp,nfyp,kp,bi,bj) |
| 109 |
u12 = u(nfx,nfyp,kp,bi,bj) |
| 110 |
endif |
| 111 |
if (nu.eq.4) then |
| 112 |
caw This may be incorrect. |
| 113 |
u11 = u(nfx,nfy,kp,bi,bj)+u(nfx,nfy,kp-1,bi,bj) |
| 114 |
u21 = u(nfxp,nfy,kp,bi,bj)+u(nfxp,nfy,kp-1,bi,bj) |
| 115 |
u22 = u(nfxp,nfyp,kp,bi,bj)+u(nfxp,nfyp,kp-1,bi,bj) |
| 116 |
u12 = u(nfx,nfyp,kp,bi,bj)+u(nfx,nfyp,kp-1,bi,bj) |
| 117 |
endif |
| 118 |
|
| 119 |
c |
| 120 |
c |
| 121 |
c bilinear interpolation (from numerical recipes) |
| 122 |
uu = (1-ddx)*(1-ddy)*u11 + ddx*(1-ddy)*u21 + ddx*ddy*u22 |
| 123 |
. + (1-ddx)*ddy*u12 |
| 124 |
c |
| 125 |
c |
| 126 |
return |
| 127 |
end |
| 128 |
|
| 129 |
subroutine flt_trilinear( |
| 130 |
I xp, |
| 131 |
I yp, |
| 132 |
I zp, |
| 133 |
O uu, |
| 134 |
I u, |
| 135 |
I nu, |
| 136 |
I bi, |
| 137 |
I bj |
| 138 |
& ) |
| 139 |
|
| 140 |
c ================================================================== |
| 141 |
c SUBROUTINE flt_trilinear |
| 142 |
c ================================================================== |
| 143 |
c |
| 144 |
c o Trilinear scheme to find u of particle at a given xp,yp,zp |
| 145 |
c location. This routine is a straight forward generalization of the |
| 146 |
c bilinear interpolation scheme. |
| 147 |
c |
| 148 |
c started: 2004.05.28 Antti Westerlund (antti.westerlund@fimr.fi) |
| 149 |
c and Sergio Jaramillo (sju@eos.ubc.ca). |
| 150 |
c (adopted from subroutine bilinear by Arne Biastoch) |
| 151 |
c |
| 152 |
c ================================================================== |
| 153 |
c SUBROUTINE flt_trilinear |
| 154 |
c ================================================================== |
| 155 |
|
| 156 |
c == global variables == |
| 157 |
|
| 158 |
#include "SIZE.h" |
| 159 |
|
| 160 |
c == routine arguments == |
| 161 |
|
| 162 |
_RL xp, yp, zp |
| 163 |
_RL uu |
| 164 |
integer nu, bi, bj |
| 165 |
_RL u (1-OLx:sNx+OLx,1-OLy:sNy+OLy,Nr,nSx,nSy) |
| 166 |
|
| 167 |
c == local variables == |
| 168 |
|
| 169 |
INTEGER nnx, nny, nnz, nfx, nfy, nfz, nfxp, nfyp, nfzp |
| 170 |
_RL dx, dy, dz, ddx, ddy, ddz |
| 171 |
integer ip |
| 172 |
_RL xx, yy, zz, phi, scalex, scaley, scalez |
| 173 |
_RL u111, u121, u221, u211, u112, u122, u222, u212 |
| 174 |
|
| 175 |
c == end of interface == |
| 176 |
|
| 177 |
c Round xp,yp,zp down to find a grid point. |
| 178 |
nnx = int(xp) |
| 179 |
nny = int(yp) |
| 180 |
nnz = int(zp) |
| 181 |
|
| 182 |
c Find out the distance from the gridpoint. |
| 183 |
dx = xp - float(nnx) |
| 184 |
dy = yp - float(nny) |
| 185 |
dz = zp - float(nnz) |
| 186 |
c |
| 187 |
c to choose the u box in which the particle is found |
| 188 |
c nu=1 for T, S |
| 189 |
c nu=2 for u |
| 190 |
c nu=3 for v |
| 191 |
c nu=4 for w |
| 192 |
c |
| 193 |
c Velocities are face quantities and must therefore be treated differently |
| 194 |
c |
| 195 |
c if the variable is T,S |
| 196 |
if (nu.eq.1) then |
| 197 |
nfx = nnx |
| 198 |
ddx = dx |
| 199 |
nfy = nny |
| 200 |
ddy = dy |
| 201 |
nfz = nnz |
| 202 |
ddz = dz |
| 203 |
endif |
| 204 |
c if the variable is u |
| 205 |
if (nu.eq.2) then |
| 206 |
if (dx.le.0.5) then |
| 207 |
nfx = nnx |
| 208 |
ddx = dx + 0.5 |
| 209 |
else |
| 210 |
nfx = nnx + 1 |
| 211 |
ddx = dx - 0.5 |
| 212 |
endif |
| 213 |
nfy = nny |
| 214 |
ddy = dy |
| 215 |
nfz = nnz |
| 216 |
ddz = dz |
| 217 |
endif |
| 218 |
c if the variable is v |
| 219 |
if (nu.eq.3) then |
| 220 |
nfx = nnx |
| 221 |
ddx = dx |
| 222 |
if (dy.le.0.5) then |
| 223 |
nfy = nny |
| 224 |
ddy = dy + 0.5 |
| 225 |
else |
| 226 |
nfy = nny + 1 |
| 227 |
ddy = dy - 0.5 |
| 228 |
endif |
| 229 |
nfz = nnz |
| 230 |
ddz = dz |
| 231 |
endif |
| 232 |
c if the variable is w |
| 233 |
if (nu.eq.4) then |
| 234 |
nfx = nnx |
| 235 |
ddx = dx |
| 236 |
nfy = nny |
| 237 |
ddy = dy |
| 238 |
if (dz.le.0.5) then |
| 239 |
nfz = nnz |
| 240 |
ddz = dz + 0.5 |
| 241 |
else |
| 242 |
nfz = nnz + 1 |
| 243 |
ddz = dz - 0.5 |
| 244 |
endif |
| 245 |
endif |
| 246 |
c |
| 247 |
c if we are near or over the edge, limit nfx/y/z |
| 248 |
if(nfx.gt.nx) nfx=nx |
| 249 |
if(nfy.gt.ny) nfy=ny |
| 250 |
if(nfz.gt.nr) nfz=nr |
| 251 |
if(nfz.le.1) nfz=1 |
| 252 |
c We should possibly check something else too... |
| 253 |
c |
| 254 |
c the coordinates for the other grid points |
| 255 |
nfxp = nfx + 1 |
| 256 |
nfyp = nfy + 1 |
| 257 |
nfzp = nfz + 1 |
| 258 |
c if we are near the edge, also limit nf?p |
| 259 |
if (nfx.eq.nx) nfxp = nfx |
| 260 |
if (nfy.eq.ny) nfyp = nfy |
| 261 |
if (nfz.eq.nr) nfzp = nfz |
| 262 |
|
| 263 |
c Values of the field at relevant grid points |
| 264 |
u111 = u(nfx,nfy,nfz,bi,bj) |
| 265 |
u211 = u(nfxp,nfy,nfz,bi,bj) |
| 266 |
u221 = u(nfxp,nfyp,nfz,bi,bj) |
| 267 |
u121 = u(nfx,nfyp,nfz,bi,bj) |
| 268 |
u112 = u(nfx,nfy,nfzp,bi,bj) |
| 269 |
u212 = u(nfxp,nfy,nfzp,bi,bj) |
| 270 |
u222 = u(nfxp,nfyp,nfzp,bi,bj) |
| 271 |
u122 = u(nfx,nfyp,nfzp,bi,bj) |
| 272 |
|
| 273 |
c Trilinear interpolation, a straight forward generalization |
| 274 |
c of the bilinear interpolation scheme. |
| 275 |
uu = (1-ddx)*(1-ddy)*(1-ddz)*u111 + ddx*(1-ddy)*(1-ddz)*u211 |
| 276 |
& + ddx*ddy*(1-ddz)*u221 + (1-ddx)*ddy*(1-ddz)*u121 |
| 277 |
& + (1-ddx)*(1-ddy)*ddz*u112 + ddx*(1-ddy)*ddz*u212 |
| 278 |
& + ddx*ddy*ddz*u222 + (1-ddx)*ddy*ddz*u122 |
| 279 |
c |
| 280 |
c |
| 281 |
return |
| 282 |
end |
| 283 |
|
| 284 |
subroutine flt_bilinear2d( |
| 285 |
I xp, |
| 286 |
I yp, |
| 287 |
O uu, |
| 288 |
I u, |
| 289 |
I nu, |
| 290 |
I bi, |
| 291 |
I bj |
| 292 |
& ) |
| 293 |
|
| 294 |
c ================================================================== |
| 295 |
c SUBROUTINE flt_bilinear2d |
| 296 |
c ================================================================== |
| 297 |
c |
| 298 |
c o Bilinear scheme to find u of particle at given xp,yp location |
| 299 |
c o For 2D fields |
| 300 |
c |
| 301 |
c started: Arne Biastoch abiastoch@ucsd.edu 13-Jan-2000 |
| 302 |
c (adopted from subroutine bilinear) |
| 303 |
c |
| 304 |
c ================================================================== |
| 305 |
c SUBROUTINE flt_bilinear2d |
| 306 |
c ================================================================== |
| 307 |
|
| 308 |
c == global variables == |
| 309 |
|
| 310 |
#include "SIZE.h" |
| 311 |
|
| 312 |
c == routine arguments == |
| 313 |
|
| 314 |
_RL xp, yp |
| 315 |
_RL uu |
| 316 |
integer nu, bi, bj |
| 317 |
_RL u (1-OLx:sNx+OLx,1-OLy:sNy+OLy,nSx,nSy) |
| 318 |
|
| 319 |
c == local variables == |
| 320 |
|
| 321 |
INTEGER nnx, nny, nfx, nfy, nfxp, nfyp |
| 322 |
_RL dx, dy, ddx, ddy |
| 323 |
integer ip |
| 324 |
_RL xx, yy, phi, scalex, scaley |
| 325 |
_RL u11, u12, u22, u21 |
| 326 |
|
| 327 |
c == end of interface == |
| 328 |
|
| 329 |
nnx = int(xp) |
| 330 |
nny = int(yp) |
| 331 |
dx = xp - float(nnx) |
| 332 |
dy = yp - float(nny) |
| 333 |
c |
| 334 |
c to choose the u box in which the particle is found |
| 335 |
c nu=1 for T, S |
| 336 |
c nu=2 for u |
| 337 |
c nu=3 for v |
| 338 |
c nu=4 for w |
| 339 |
c |
| 340 |
if (nu.eq.1.or.nu.eq.4) then |
| 341 |
nfx = nnx |
| 342 |
nfy = nny |
| 343 |
ddx = dx |
| 344 |
ddy = dy |
| 345 |
endif |
| 346 |
c |
| 347 |
if (nu.eq.2) then |
| 348 |
if (dx.le.0.5) then |
| 349 |
nfx = nnx |
| 350 |
ddx = dx + 0.5 |
| 351 |
else |
| 352 |
nfx = nnx + 1 |
| 353 |
ddx = dx - 0.5 |
| 354 |
endif |
| 355 |
nfy = nny |
| 356 |
ddy = dy |
| 357 |
endif |
| 358 |
c |
| 359 |
if (nu.eq.3) then |
| 360 |
if (dy.le.0.5) then |
| 361 |
nfy = nny |
| 362 |
ddy = dy + 0.5 |
| 363 |
else |
| 364 |
nfy = nny + 1 |
| 365 |
ddy = dy - 0.5 |
| 366 |
endif |
| 367 |
nfx = nnx |
| 368 |
ddx = dx |
| 369 |
endif |
| 370 |
c |
| 371 |
cab change start |
| 372 |
c was correct only for global? |
| 373 |
c if(nfx.gt.nx) nfx=nfx-nx |
| 374 |
if(nfx.gt.nx) nfx=nx |
| 375 |
cab change end |
| 376 |
if(nfy.gt.ny) nfy=ny |
| 377 |
nfxp = nfx + 1 |
| 378 |
nfyp = nfy + 1 |
| 379 |
cab change start |
| 380 |
c if (nfx.eq.nx) nfxp = 1 |
| 381 |
if (nfx.eq.nx) nfxp = nfx |
| 382 |
cab change end |
| 383 |
if (nfy.eq.ny) nfyp = nfy |
| 384 |
|
| 385 |
if (nu.lt.4) then |
| 386 |
u11 = u(nfx,nfy,bi,bj) |
| 387 |
u21 = u(nfxp,nfy,bi,bj) |
| 388 |
u22 = u(nfxp,nfyp,bi,bj) |
| 389 |
u12 = u(nfx,nfyp,bi,bj) |
| 390 |
endif |
| 391 |
if (nu.eq.4) then |
| 392 |
caw This may be incorrect. |
| 393 |
u11 = u(nfx,nfy,bi,bj)+u(nfx,nfy,bi,bj) |
| 394 |
u21 = u(nfxp,nfy,bi,bj)+u(nfxp,nfy,bi,bj) |
| 395 |
u22 = u(nfxp,nfyp,bi,bj)+u(nfxp,nfyp,bi,bj) |
| 396 |
u12 = u(nfx,nfyp,bi,bj)+u(nfx,nfyp,bi,bj) |
| 397 |
endif |
| 398 |
c |
| 399 |
c |
| 400 |
c bilinear interpolation (from numerical recipes) |
| 401 |
uu = (1-ddx)*(1-ddy)*u11 + ddx*(1-ddy)*u21 + ddx*ddy*u22 |
| 402 |
. + (1-ddx)*ddy*u12 |
| 403 |
c |
| 404 |
c |
| 405 |
return |
| 406 |
end |
| 407 |
|
Parent Directory
| 
Revision Log
| 
Revision Graph