| 1 |
C $Header: /u/gcmpack/models/MITgcmUV/model/src/ini_spherical_polar_grid.F,v 1.17.2.3 2001/04/08 21:58:27 cnh Exp $ |
| 2 |
C $Name: $ |
| 3 |
|
| 4 |
#include "CPP_OPTIONS.h" |
| 5 |
|
| 6 |
#undef USE_BACKWARD_COMPATIBLE_GRID |
| 7 |
|
| 8 |
CStartOfInterface |
| 9 |
SUBROUTINE INI_SPHERICAL_POLAR_GRID( myThid ) |
| 10 |
C /==========================================================\ |
| 11 |
C | SUBROUTINE INI_SPHERICAL_POLAR_GRID | |
| 12 |
C | o Initialise model coordinate system | |
| 13 |
C |==========================================================| |
| 14 |
C | These arrays are used throughout the code in evaluating | |
| 15 |
C | gradients, integrals and spatial avarages. This routine | |
| 16 |
C | is called separately by each thread and initialise only | |
| 17 |
C | the region of the domain it is "responsible" for. | |
| 18 |
C | Notes: | |
| 19 |
C | Two examples are included. One illustrates the | |
| 20 |
C | initialisation of a cartesian grid. The other shows the | |
| 21 |
C | inialisation of a spherical polar grid. Other orthonormal| |
| 22 |
C | grids can be fitted into this design. In this case | |
| 23 |
C | custom metric terms also need adding to account for the | |
| 24 |
C | projections of velocity vectors onto these grids. | |
| 25 |
C | The structure used here also makes it possible to | |
| 26 |
C | implement less regular grid mappings. In particular | |
| 27 |
C | o Schemes which leave out blocks of the domain that are | |
| 28 |
C | all land could be supported. | |
| 29 |
C | o Multi-level schemes such as icosohedral or cubic | |
| 30 |
C | grid projections onto a sphere can also be fitted | |
| 31 |
C | within the strategy we use. | |
| 32 |
C | Both of the above also require modifying the support | |
| 33 |
C | routines that map computational blocks to simulation | |
| 34 |
C | domain blocks. | |
| 35 |
C | Under the spherical polar grid mode primitive distances | |
| 36 |
C | in X and Y are in degrees. Distance in Z are in m or Pa | |
| 37 |
C | depending on the vertical gridding mode. | |
| 38 |
C \==========================================================/ |
| 39 |
IMPLICIT NONE |
| 40 |
|
| 41 |
C === Global variables === |
| 42 |
#include "SIZE.h" |
| 43 |
#include "EEPARAMS.h" |
| 44 |
#include "PARAMS.h" |
| 45 |
#include "GRID.h" |
| 46 |
|
| 47 |
C == Routine arguments == |
| 48 |
C myThid - Number of this instance of INI_CARTESIAN_GRID |
| 49 |
INTEGER myThid |
| 50 |
CEndOfInterface |
| 51 |
|
| 52 |
C == Local variables == |
| 53 |
C xG, yG - Global coordinate location. |
| 54 |
C xBase - South-west corner location for process. |
| 55 |
C yBase |
| 56 |
C zUpper - Work arrays for upper and lower |
| 57 |
C zLower cell-face heights. |
| 58 |
C phi - Temporary scalar |
| 59 |
C iG, jG - Global coordinate index. Usually used to hold |
| 60 |
C the south-west global coordinate of a tile. |
| 61 |
C bi,bj - Loop counters |
| 62 |
C zUpper - Temporary arrays holding z coordinates of |
| 63 |
C zLower upper and lower faces. |
| 64 |
C xBase - Lower coordinate for this threads cells |
| 65 |
C yBase |
| 66 |
C lat, latN, - Temporary variables used to hold latitude |
| 67 |
C latS values. |
| 68 |
C I,J,K |
| 69 |
INTEGER iG, jG |
| 70 |
INTEGER bi, bj |
| 71 |
INTEGER I, J |
| 72 |
_RL lat, dlat, dlon, xG0, yG0 |
| 73 |
|
| 74 |
|
| 75 |
C "Long" real for temporary coordinate calculation |
| 76 |
C NOTICE the extended range of indices!! |
| 77 |
_RL xGloc(1-Olx:sNx+Olx+1,1-Oly:sNy+Oly+1) |
| 78 |
_RL yGloc(1-Olx:sNx+Olx+1,1-Oly:sNy+Oly+1) |
| 79 |
|
| 80 |
C The functions iGl, jGl return the "global" index with valid values beyond |
| 81 |
C halo regions |
| 82 |
C cnh wrote: |
| 83 |
C > I don't understand why we would ever have to multiply the |
| 84 |
C > overlap by the total domain size e.g |
| 85 |
C > OLx*Nx, OLy*Ny. |
| 86 |
C > Can anybody explain? Lines are in ini_spherical_polar_grid.F. |
| 87 |
C > Surprised the code works if its wrong, so I am puzzled. |
| 88 |
C jmc relied: |
| 89 |
C Yes, I can explain this since I put this modification to work |
| 90 |
C with small domain (where Oly > Ny, as for instance, zonal-average |
| 91 |
C case): |
| 92 |
C This has no effect on the acuracy of the evaluation of iGl(I,bi) |
| 93 |
C and jGl(J,bj) since we take mod(a+OLx*Nx,Nx) and mod(b+OLy*Ny,Ny). |
| 94 |
C But in case a or b is negative, then the FORTRAN function "mod" |
| 95 |
C does not return the matematical value of the "modulus" function, |
| 96 |
C and this is not good for your purpose. |
| 97 |
C This is why I add +OLx*Nx and +OLy*Ny to be sure that the 1rst |
| 98 |
C argument of the mod function is positive. |
| 99 |
INTEGER iGl,jGl |
| 100 |
iGl(I,bi) = 1+mod(myXGlobalLo-1+(bi-1)*sNx+I+Olx*Nx-1,Nx) |
| 101 |
jGl(J,bj) = 1+mod(myYGlobalLo-1+(bj-1)*sNy+J+Oly*Ny-1,Ny) |
| 102 |
|
| 103 |
|
| 104 |
C For each tile ... |
| 105 |
DO bj = myByLo(myThid), myByHi(myThid) |
| 106 |
DO bi = myBxLo(myThid), myBxHi(myThid) |
| 107 |
|
| 108 |
C-- "Global" index (place holder) |
| 109 |
jG = myYGlobalLo + (bj-1)*sNy |
| 110 |
iG = myXGlobalLo + (bi-1)*sNx |
| 111 |
|
| 112 |
C-- First find coordinate of tile corner (meaning outer corner of halo) |
| 113 |
xG0 = thetaMin |
| 114 |
C Find the X-coordinate of the outer grid-line of the "real" tile |
| 115 |
DO i=1, iG-1 |
| 116 |
xG0 = xG0 + delX(i) |
| 117 |
ENDDO |
| 118 |
C Back-step to the outer grid-line of the "halo" region |
| 119 |
DO i=1, Olx |
| 120 |
xG0 = xG0 - delX( 1+mod(Olx*Nx-1+iG-i,Nx) ) |
| 121 |
ENDDO |
| 122 |
C Find the Y-coordinate of the outer grid-line of the "real" tile |
| 123 |
yG0 = phiMin |
| 124 |
DO j=1, jG-1 |
| 125 |
yG0 = yG0 + delY(j) |
| 126 |
ENDDO |
| 127 |
C Back-step to the outer grid-line of the "halo" region |
| 128 |
DO j=1, Oly |
| 129 |
yG0 = yG0 - delY( 1+mod(Oly*Ny-1+jG-j,Ny) ) |
| 130 |
ENDDO |
| 131 |
|
| 132 |
C-- Calculate coordinates of cell corners for N+1 grid-lines |
| 133 |
DO J=1-Oly,sNy+Oly +1 |
| 134 |
xGloc(1-Olx,J) = xG0 |
| 135 |
DO I=1-Olx,sNx+Olx |
| 136 |
c xGloc(I+1,J) = xGloc(I,J) + delX(1+mod(Nx-1+iG-1+i,Nx)) |
| 137 |
xGloc(I+1,J) = xGloc(I,J) + delX( iGl(I,bi) ) |
| 138 |
ENDDO |
| 139 |
ENDDO |
| 140 |
DO I=1-Olx,sNx+Olx +1 |
| 141 |
yGloc(I,1-Oly) = yG0 |
| 142 |
DO J=1-Oly,sNy+Oly |
| 143 |
c yGloc(I,J+1) = yGloc(I,J) + delY(1+mod(Ny-1+jG-1+j,Ny)) |
| 144 |
yGloc(I,J+1) = yGloc(I,J) + delY( jGl(J,bj) ) |
| 145 |
ENDDO |
| 146 |
ENDDO |
| 147 |
|
| 148 |
C-- Make a permanent copy of [xGloc,yGloc] in [xG,yG] |
| 149 |
DO J=1-Oly,sNy+Oly |
| 150 |
DO I=1-Olx,sNx+Olx |
| 151 |
xG(I,J,bi,bj) = xGloc(I,J) |
| 152 |
yG(I,J,bi,bj) = yGloc(I,J) |
| 153 |
ENDDO |
| 154 |
ENDDO |
| 155 |
|
| 156 |
C-- Calculate [xC,yC], coordinates of cell centers |
| 157 |
DO J=1-Oly,sNy+Oly |
| 158 |
DO I=1-Olx,sNx+Olx |
| 159 |
C by averaging |
| 160 |
xC(I,J,bi,bj) = 0.25*( |
| 161 |
& xGloc(I,J)+xGloc(I+1,J)+xGloc(I,J+1)+xGloc(I+1,J+1) ) |
| 162 |
yC(I,J,bi,bj) = 0.25*( |
| 163 |
& yGloc(I,J)+yGloc(I+1,J)+yGloc(I,J+1)+yGloc(I+1,J+1) ) |
| 164 |
ENDDO |
| 165 |
ENDDO |
| 166 |
|
| 167 |
C-- Calculate [dxF,dyF], lengths between cell faces (through center) |
| 168 |
DO J=1-Oly,sNy+Oly |
| 169 |
DO I=1-Olx,sNx+Olx |
| 170 |
C by averaging |
| 171 |
c dXF(I,J,bi,bj) = 0.5*(dXG(I,J,bi,bj)+dXG(I,J+1,bi,bj)) |
| 172 |
c dYF(I,J,bi,bj) = 0.5*(dYG(I,J,bi,bj)+dYG(I+1,J,bi,bj)) |
| 173 |
C by formula |
| 174 |
lat = yC(I,J,bi,bj) |
| 175 |
dlon = delX( iGl(I,bi) ) |
| 176 |
dlat = delY( jGl(J,bj) ) |
| 177 |
dXF(I,J,bi,bj) = rSphere*COS(deg2rad*lat)*dlon*deg2rad |
| 178 |
#ifdef USE_BACKWARD_COMPATIBLE_GRID |
| 179 |
dXF(I,J,bi,bj) = delX(iGl(I,bi))*deg2rad*rSphere* |
| 180 |
& COS(yc(I,J,bi,bj)*deg2rad) |
| 181 |
#endif /* USE_BACKWARD_COMPATIBLE_GRID */ |
| 182 |
dYF(I,J,bi,bj) = rSphere*dlat*deg2rad |
| 183 |
ENDDO |
| 184 |
ENDDO |
| 185 |
|
| 186 |
C-- Calculate [dxG,dyG], lengths along cell boundaries |
| 187 |
DO J=1-Oly,sNy+Oly |
| 188 |
DO I=1-Olx,sNx+Olx |
| 189 |
C by averaging |
| 190 |
c dXG(I,J,bi,bj) = 0.5*(dXF(I,J,bi,bj)+dXF(I,J-1,bi,bj)) |
| 191 |
c dYG(I,J,bi,bj) = 0.5*(dYF(I,J,bi,bj)+dYF(I-1,J,bi,bj)) |
| 192 |
C by formula |
| 193 |
lat = 0.5*(yGloc(I,J)+yGloc(I+1,J)) |
| 194 |
dlon = delX( iGl(I,bi) ) |
| 195 |
dlat = delY( jGl(J,bj) ) |
| 196 |
dXG(I,J,bi,bj) = rSphere*COS(deg2rad*lat)*dlon*deg2rad |
| 197 |
if (dXG(I,J,bi,bj).LT.1.) dXG(I,J,bi,bj)=0. |
| 198 |
dYG(I,J,bi,bj) = rSphere*dlat*deg2rad |
| 199 |
ENDDO |
| 200 |
ENDDO |
| 201 |
|
| 202 |
C-- The following arrays are not defined in some parts of the halo |
| 203 |
C region. We set them to zero here for safety. If they are ever |
| 204 |
C referred to, especially in the denominator then it is a mistake! |
| 205 |
DO J=1-Oly,sNy+Oly |
| 206 |
DO I=1-Olx,sNx+Olx |
| 207 |
dXC(I,J,bi,bj) = 0. |
| 208 |
dYC(I,J,bi,bj) = 0. |
| 209 |
dXV(I,J,bi,bj) = 0. |
| 210 |
dYU(I,J,bi,bj) = 0. |
| 211 |
rAw(I,J,bi,bj) = 0. |
| 212 |
rAs(I,J,bi,bj) = 0. |
| 213 |
ENDDO |
| 214 |
ENDDO |
| 215 |
|
| 216 |
C-- Calculate [dxC], zonal length between cell centers |
| 217 |
DO J=1-Oly,sNy+Oly |
| 218 |
DO I=1-Olx+1,sNx+Olx ! NOTE range |
| 219 |
C by averaging |
| 220 |
dXC(I,J,bi,bj) = 0.5D0*(dXF(I,J,bi,bj)+dXF(I-1,J,bi,bj)) |
| 221 |
C by formula |
| 222 |
c lat = 0.5*(yC(I,J,bi,bj)+yC(I-1,J,bi,bj)) |
| 223 |
c dlon = 0.5*(delX( iGl(I,bi) ) + delX( iGl(I-1,bi) )) |
| 224 |
c dXC(I,J,bi,bj) = rSphere*COS(deg2rad*lat)*dlon*deg2rad |
| 225 |
C by difference |
| 226 |
c lat = 0.5*(yC(I,J,bi,bj)+yC(I-1,J,bi,bj)) |
| 227 |
c dlon = (xC(I,J,bi,bj)+xC(I-1,J,bi,bj)) |
| 228 |
c dXC(I,J,bi,bj) = rSphere*COS(deg2rad*lat)*dlon*deg2rad |
| 229 |
ENDDO |
| 230 |
ENDDO |
| 231 |
|
| 232 |
C-- Calculate [dyC], meridional length between cell centers |
| 233 |
DO J=1-Oly+1,sNy+Oly ! NOTE range |
| 234 |
DO I=1-Olx,sNx+Olx |
| 235 |
C by averaging |
| 236 |
dYC(I,J,bi,bj) = 0.5*(dYF(I,J,bi,bj)+dYF(I,J-1,bi,bj)) |
| 237 |
C by formula |
| 238 |
c dlat = 0.5*(delY( jGl(J,bj) ) + delY( jGl(J-1,bj) )) |
| 239 |
c dYC(I,J,bi,bj) = rSphere*dlat*deg2rad |
| 240 |
C by difference |
| 241 |
c dlat = (yC(I,J,bi,bj)+yC(I,J-1,bi,bj)) |
| 242 |
c dYC(I,J,bi,bj) = rSphere*dlat*deg2rad |
| 243 |
ENDDO |
| 244 |
ENDDO |
| 245 |
|
| 246 |
C-- Calculate [dxV,dyU], length between velocity points (through corners) |
| 247 |
DO J=1-Oly+1,sNy+Oly ! NOTE range |
| 248 |
DO I=1-Olx+1,sNx+Olx ! NOTE range |
| 249 |
C by averaging (method I) |
| 250 |
dXV(I,J,bi,bj) = 0.5*(dXG(I,J,bi,bj)+dXG(I-1,J,bi,bj)) |
| 251 |
dYU(I,J,bi,bj) = 0.5*(dYG(I,J,bi,bj)+dYG(I,J-1,bi,bj)) |
| 252 |
C by averaging (method II) |
| 253 |
c dXV(I,J,bi,bj) = 0.5*(dXG(I,J,bi,bj)+dXG(I-1,J,bi,bj)) |
| 254 |
c dYU(I,J,bi,bj) = 0.5*(dYC(I,J,bi,bj)+dYC(I-1,J,bi,bj)) |
| 255 |
ENDDO |
| 256 |
ENDDO |
| 257 |
|
| 258 |
C-- Calculate vertical face area (tracer cells) |
| 259 |
DO J=1-Oly,sNy+Oly |
| 260 |
DO I=1-Olx,sNx+Olx |
| 261 |
lat=0.5*(yGloc(I,J)+yGloc(I+1,J)) |
| 262 |
dlon=delX( iGl(I,bi) ) |
| 263 |
dlat=delY( jGl(J,bj) ) |
| 264 |
rA(I,J,bi,bj) = rSphere*rSphere*dlon*deg2rad |
| 265 |
& *abs( sin((lat+dlat)*deg2rad)-sin(lat*deg2rad) ) |
| 266 |
#ifdef USE_BACKWARD_COMPATIBLE_GRID |
| 267 |
lat=yC(I,J,bi,bj)-delY( jGl(J,bj) )*0.5 _d 0 |
| 268 |
lon=yC(I,J,bi,bj)+delY( jGl(J,bj) )*0.5 _d 0 |
| 269 |
rA(I,J,bi,bj) = dyF(I,J,bi,bj) |
| 270 |
& *rSphere*(SIN(lon*deg2rad)-SIN(lat*deg2rad)) |
| 271 |
#endif /* USE_BACKWARD_COMPATIBLE_GRID */ |
| 272 |
ENDDO |
| 273 |
ENDDO |
| 274 |
|
| 275 |
C-- Calculate vertical face area (u cells) |
| 276 |
DO J=1-Oly,sNy+Oly |
| 277 |
DO I=1-Olx+1,sNx+Olx ! NOTE range |
| 278 |
C by averaging |
| 279 |
rAw(I,J,bi,bj) = 0.5*(rA(I,J,bi,bj)+rA(I-1,J,bi,bj)) |
| 280 |
C by formula |
| 281 |
c lat=yGloc(I,J) |
| 282 |
c dlon=0.5*( delX( iGl(I,bi) ) + delX( iGl(I-1,bi) ) ) |
| 283 |
c dlat=delY( jGl(J,bj) ) |
| 284 |
c rAw(I,J,bi,bj) = rSphere*rSphere*dlon*deg2rad |
| 285 |
c & *abs( sin((lat+dlat)*deg2rad)-sin(lat*deg2rad) ) |
| 286 |
ENDDO |
| 287 |
ENDDO |
| 288 |
|
| 289 |
C-- Calculate vertical face area (v cells) |
| 290 |
DO J=1-Oly,sNy+Oly |
| 291 |
DO I=1-Olx,sNx+Olx |
| 292 |
C by formula |
| 293 |
lat=yC(I,J,bi,bj) |
| 294 |
dlon=delX( iGl(I,bi) ) |
| 295 |
dlat=0.5*( delY( jGl(J,bj) ) + delY( jGl(J-1,bj) ) ) |
| 296 |
rAs(I,J,bi,bj) = rSphere*rSphere*dlon*deg2rad |
| 297 |
& *abs( sin(lat*deg2rad)-sin((lat-dlat)*deg2rad) ) |
| 298 |
#ifdef USE_BACKWARD_COMPATIBLE_GRID |
| 299 |
lon=yC(I,J,bi,bj)-delY( jGl(J,bj) ) |
| 300 |
lat=yC(I,J,bi,bj) |
| 301 |
rAs(I,J,bi,bj) = rSphere*delX(iGl(I,bi))*deg2rad |
| 302 |
& *rSphere*(SIN(lat*deg2rad)-SIN(lon*deg2rad)) |
| 303 |
#endif /* USE_BACKWARD_COMPATIBLE_GRID */ |
| 304 |
IF (abs(lat).GT.90..OR.abs(lat-dlat).GT.90.) rAs(I,J,bi,bj)=0. |
| 305 |
ENDDO |
| 306 |
ENDDO |
| 307 |
|
| 308 |
C-- Calculate vertical face area (vorticity points) |
| 309 |
DO J=1-Oly,sNy+Oly |
| 310 |
DO I=1-Olx,sNx+Olx |
| 311 |
C by formula |
| 312 |
lat=yC(I,J,bi,bj) |
| 313 |
dlon=delX( iGl(I,bi) ) |
| 314 |
dlat=0.5*( delY( jGl(J,bj) ) + delY( jGl(J-1,bj) ) ) |
| 315 |
rAz(I,J,bi,bj) = rSphere*rSphere*dlon*deg2rad |
| 316 |
& *abs( sin(lat*deg2rad)-sin((lat-dlat)*deg2rad) ) |
| 317 |
IF (abs(lat).GT.90..OR.abs(lat-dlat).GT.90.) rAz(I,J,bi,bj)=0. |
| 318 |
ENDDO |
| 319 |
ENDDO |
| 320 |
|
| 321 |
C-- Calculate trigonometric terms |
| 322 |
DO J=1-Oly,sNy+Oly |
| 323 |
DO I=1-Olx,sNx+Olx |
| 324 |
lat=0.5*(yGloc(I,J)+yGloc(I,J+1)) |
| 325 |
tanPhiAtU(i,j,bi,bj)=tan(lat*deg2rad) |
| 326 |
lat=0.5*(yGloc(I,J)+yGloc(I+1,J)) |
| 327 |
tanPhiAtV(i,j,bi,bj)=tan(lat*deg2rad) |
| 328 |
ENDDO |
| 329 |
ENDDO |
| 330 |
|
| 331 |
C-- Cosine(lat) scaling |
| 332 |
DO J=1-OLy,sNy+OLy |
| 333 |
jG = myYGlobalLo + (bj-1)*sNy + J-1 |
| 334 |
jG = min(max(1,jG),Ny) |
| 335 |
IF (cosPower.NE.0.) THEN |
| 336 |
cosFacU(J,bi,bj)=COS(yC(1,J,bi,bj)*deg2rad) |
| 337 |
& **cosPower |
| 338 |
cosFacV(J,bi,bj)=COS((yC(1,J,bi,bj)-0.5*delY(jG))*deg2rad) |
| 339 |
& **cosPower |
| 340 |
sqcosFacU(J,bi,bj)=sqrt(cosFacU(J,bi,bj)) |
| 341 |
sqcosFacV(J,bi,bj)=sqrt(cosFacV(J,bi,bj)) |
| 342 |
ELSE |
| 343 |
cosFacU(J,bi,bj)=1. |
| 344 |
cosFacV(J,bi,bj)=1. |
| 345 |
sqcosFacU(J,bi,bj)=1. |
| 346 |
sqcosFacV(J,bi,bj)=1. |
| 347 |
ENDIF |
| 348 |
ENDDO |
| 349 |
|
| 350 |
ENDDO ! bi |
| 351 |
ENDDO ! bj |
| 352 |
|
| 353 |
RETURN |
| 354 |
END |
Parent Directory
| 
Revision Log
| 
Revision Graph