1 |
jmc |
1.6 |
C $Header: /u/gcmpack/MITgcm/pkg/thsice/thsice_advection.F,v 1.5 2007/05/23 23:40:19 jmc Exp $ |
2 |
jmc |
1.1 |
C $Name: $ |
3 |
|
|
|
4 |
|
|
#include "THSICE_OPTIONS.h" |
5 |
|
|
#ifdef ALLOW_GENERIC_ADVDIFF |
6 |
|
|
# include "GAD_OPTIONS.h" |
7 |
|
|
#endif |
8 |
|
|
|
9 |
|
|
C---+----1----+----2----+----3----+----4----+----5----+----6----+----7-|--+----| |
10 |
|
|
CBOP |
11 |
|
|
C !ROUTINE: THSICE_ADVECTION |
12 |
|
|
|
13 |
|
|
C !INTERFACE: ========================================================== |
14 |
|
|
SUBROUTINE THSICE_ADVECTION( |
15 |
|
|
I tracerIdentity, |
16 |
|
|
I advectionScheme, |
17 |
|
|
I useGridVolume, |
18 |
|
|
I uTrans, vTrans, maskOc, deltaTadvect, iceEps, |
19 |
|
|
U iceVol, iceFld, |
20 |
|
|
O afx, afy, |
21 |
|
|
I bi, bj, myTime, myIter, myThid) |
22 |
|
|
|
23 |
|
|
C !DESCRIPTION: |
24 |
|
|
C Calculates the tendency of a sea-ice field due to advection. |
25 |
|
|
C It uses the multi-dimensional method given in \ref{sect:multiDimAdvection} |
26 |
|
|
C and can only be used for the non-linear advection schemes such as the |
27 |
|
|
C direct-space-time method and flux-limiters. |
28 |
|
|
C |
29 |
|
|
C This routine is an adaption of the GAD_ADVECTION for 2D-fields. |
30 |
|
|
C for Area, effective thickness or other "extensive" sea-ice field, |
31 |
|
|
C the contribution iceFld*div(u) (that is present in gad_advection) |
32 |
|
|
C is not included here. |
33 |
|
|
C |
34 |
|
|
C The algorithm is as follows: |
35 |
|
|
C \begin{itemize} |
36 |
|
|
C \item{$\theta^{(n+1/2)} = \theta^{(n)} |
37 |
|
|
C - \Delta t \partial_x (u\theta^{(n)}) + \theta^{(n)} \partial_x u$} |
38 |
|
|
C \item{$\theta^{(n+2/2)} = \theta^{(n+1/2)} |
39 |
|
|
C - \Delta t \partial_y (v\theta^{(n+1/2)}) + \theta^{(n)} \partial_y v$} |
40 |
|
|
C \item{$G_\theta = ( \theta^{(n+2/2)} - \theta^{(n)} )/\Delta t$} |
41 |
|
|
C \end{itemize} |
42 |
|
|
C |
43 |
|
|
C The tendency (output) is over-written by this routine. |
44 |
|
|
|
45 |
|
|
C !USES: =============================================================== |
46 |
|
|
IMPLICIT NONE |
47 |
|
|
#include "SIZE.h" |
48 |
|
|
#include "EEPARAMS.h" |
49 |
|
|
#include "PARAMS.h" |
50 |
|
|
#include "GRID.h" |
51 |
|
|
#include "THSICE_SIZE.h" |
52 |
|
|
#ifdef ALLOW_GENERIC_ADVDIFF |
53 |
|
|
# include "GAD.h" |
54 |
|
|
#endif |
55 |
|
|
#ifdef ALLOW_EXCH2 |
56 |
|
|
#include "W2_EXCH2_TOPOLOGY.h" |
57 |
|
|
#include "W2_EXCH2_PARAMS.h" |
58 |
|
|
#endif /* ALLOW_EXCH2 */ |
59 |
heimbach |
1.4 |
#ifdef ALLOW_AUTODIFF_TAMC |
60 |
|
|
# include "THSICE_PARAMS.h" |
61 |
|
|
# include "tamc.h" |
62 |
|
|
# include "tamc_keys.h" |
63 |
|
|
#endif |
64 |
jmc |
1.1 |
|
65 |
|
|
C !INPUT PARAMETERS: =================================================== |
66 |
|
|
C tracerIdentity :: tracer identifier (required only for OBCS) |
67 |
|
|
C advectionScheme :: advection scheme to use (Horizontal plane) |
68 |
|
|
C useGridVolume :: use grid-cell Area & Volume (instead of "iceVol" field) |
69 |
|
|
C uTrans,vTrans :: volume transports at U,V points |
70 |
|
|
C maskOc :: oceanic mask |
71 |
|
|
C iceVol :: sea-ice volume |
72 |
|
|
C iceFld :: sea-ice field |
73 |
|
|
C deltaTadvect :: time-step used for advection [s] |
74 |
|
|
C iceEps :: small volume (to avoid division by zero if iceVol==0) |
75 |
|
|
C bi,bj :: tile indices |
76 |
|
|
C myTime :: current time in simulation [s] |
77 |
|
|
C myIter :: current iteration number |
78 |
|
|
C myThid :: my thread Id. number |
79 |
|
|
INTEGER tracerIdentity |
80 |
|
|
INTEGER advectionScheme |
81 |
|
|
LOGICAL useGridVolume |
82 |
|
|
_RL uTrans(1-OLx:sNx+OLx,1-OLy:sNy+OLy) |
83 |
|
|
_RL vTrans(1-OLx:sNx+OLx,1-OLy:sNy+OLy) |
84 |
|
|
_RS maskOc(1-OLx:sNx+OLx,1-OLy:sNy+OLy) |
85 |
|
|
_RL iceFld(1-OLx:sNx+OLx,1-OLy:sNy+OLy) |
86 |
|
|
_RL iceVol(1-OLx:sNx+OLx,1-OLy:sNy+OLy) |
87 |
|
|
_RL deltaTadvect, iceEps |
88 |
|
|
INTEGER bi,bj |
89 |
|
|
_RL myTime |
90 |
|
|
INTEGER myIter |
91 |
|
|
INTEGER myThid |
92 |
|
|
|
93 |
|
|
C !OUTPUT PARAMETERS: ================================================== |
94 |
|
|
C iceVol (Updated):: sea-ice volume |
95 |
|
|
C iceFld (Updated):: sea-ice field |
96 |
|
|
C afx :: horizontal advective flux, x direction |
97 |
|
|
C afy :: horizontal advective flux, y direction |
98 |
|
|
_RL afx (1-OLx:sNx+OLx,1-OLy:sNy+OLy) |
99 |
|
|
_RL afy (1-OLx:sNx+OLx,1-OLy:sNy+OLy) |
100 |
|
|
|
101 |
jmc |
1.2 |
#ifdef ALLOW_GENERIC_ADVDIFF |
102 |
jmc |
1.1 |
C !LOCAL VARIABLES: ==================================================== |
103 |
|
|
C maskLocW :: 2-D array for mask at West points |
104 |
|
|
C maskLocS :: 2-D array for mask at South points |
105 |
|
|
C iMin,iMax, :: loop range for called routines |
106 |
|
|
C jMin,jMax :: loop range for called routines |
107 |
|
|
C [iMin,iMax]Upd :: loop range to update sea-ice field |
108 |
|
|
C [jMin,jMax]Upd :: loop range to update sea-ice field |
109 |
|
|
C i,j :: loop indices |
110 |
|
|
C uCfl :: CFL number, zonal direction |
111 |
|
|
C vCfl :: CFL number, meridional direction |
112 |
|
|
C af :: 2-D array for horizontal advective flux |
113 |
|
|
C calc_fluxes_X :: logical to indicate to calculate fluxes in X dir |
114 |
|
|
C calc_fluxes_Y :: logical to indicate to calculate fluxes in Y dir |
115 |
|
|
C interiorOnly :: only update the interior of myTile, but not the edges |
116 |
|
|
C overlapOnly :: only update the edges of myTile, but not the interior |
117 |
|
|
C nipass :: number of passes in multi-dimensional method |
118 |
|
|
C ipass :: number of the current pass being made |
119 |
|
|
C myTile :: variables used to determine which cube face |
120 |
|
|
C nCFace :: owns a tile for cube grid runs using |
121 |
|
|
C :: multi-dim advection. |
122 |
|
|
C [N,S,E,W]_edge :: true if N,S,E,W edge of myTile is an Edge of the cube |
123 |
|
|
_RS maskLocW(1-OLx:sNx+OLx,1-OLy:sNy+OLy) |
124 |
|
|
_RS maskLocS(1-OLx:sNx+OLx,1-OLy:sNy+OLy) |
125 |
|
|
INTEGER iMin,iMax,jMin,jMax |
126 |
|
|
INTEGER iMinUpd,iMaxUpd,jMinUpd,jMaxUpd |
127 |
|
|
INTEGER i,j,k |
128 |
|
|
_RL uCfl (1-OLx:sNx+OLx,1-OLy:sNy+OLy) |
129 |
|
|
_RL vCfl (1-OLx:sNx+OLx,1-OLy:sNy+OLy) |
130 |
|
|
_RL af (1-OLx:sNx+OLx,1-OLy:sNy+OLy) |
131 |
|
|
LOGICAL calc_fluxes_X, calc_fluxes_Y, withSigns |
132 |
|
|
LOGICAL interiorOnly, overlapOnly |
133 |
|
|
INTEGER nipass,ipass |
134 |
|
|
INTEGER nCFace |
135 |
|
|
LOGICAL N_edge, S_edge, E_edge, W_edge |
136 |
|
|
#ifdef ALLOW_EXCH2 |
137 |
|
|
INTEGER myTile |
138 |
|
|
#endif |
139 |
|
|
LOGICAL dBug |
140 |
|
|
_RL tmpFac |
141 |
|
|
_RL tmpVol |
142 |
|
|
CEOP |
143 |
|
|
|
144 |
heimbach |
1.4 |
C---+----1----+----2----+----3----+----4----+----5----+----6----+----7-|--+----| |
145 |
|
|
|
146 |
|
|
#ifdef ALLOW_AUTODIFF_TAMC |
147 |
|
|
act1 = bi - myBxLo(myThid) |
148 |
|
|
max1 = myBxHi(myThid) - myBxLo(myThid) + 1 |
149 |
|
|
act2 = bj - myByLo(myThid) |
150 |
|
|
max2 = myByHi(myThid) - myByLo(myThid) + 1 |
151 |
|
|
act3 = myThid - 1 |
152 |
|
|
max3 = nTx*nTy |
153 |
|
|
act4 = ikey_dynamics - 1 |
154 |
|
|
iicekey = (act1 + 1) + act2*max1 |
155 |
|
|
& + act3*max1*max2 |
156 |
|
|
& + act4*max1*max2*max3 |
157 |
|
|
#endif /* ALLOW_AUTODIFF_TAMC */ |
158 |
|
|
|
159 |
jmc |
1.3 |
k = 1 |
160 |
jmc |
1.1 |
dBug = debugLevel.GE.debLevB |
161 |
|
|
& .AND. myIter.EQ.nIter0 |
162 |
|
|
& .AND. ( tracerIdentity.EQ.GAD_SI_HICE .OR. |
163 |
|
|
& tracerIdentity.EQ.GAD_SI_QICE2 ) |
164 |
|
|
c & .AND. tracerIdentity.EQ.GAD_SI_HICE |
165 |
|
|
|
166 |
|
|
C-- Set up work arrays with valid (i.e. not NaN) values |
167 |
|
|
C These inital values do not alter the numerical results. They |
168 |
|
|
C just ensure that all memory references are to valid floating |
169 |
|
|
C point numbers. This prevents spurious hardware signals due to |
170 |
|
|
C uninitialised but inert locations. |
171 |
|
|
|
172 |
|
|
C-- Set tile-specific parameters for horizontal fluxes |
173 |
|
|
IF (useCubedSphereExchange) THEN |
174 |
|
|
nipass=3 |
175 |
|
|
#ifdef ALLOW_EXCH2 |
176 |
|
|
myTile = W2_myTileList(bi) |
177 |
|
|
nCFace = exch2_myFace(myTile) |
178 |
|
|
N_edge = exch2_isNedge(myTile).EQ.1 |
179 |
|
|
S_edge = exch2_isSedge(myTile).EQ.1 |
180 |
|
|
E_edge = exch2_isEedge(myTile).EQ.1 |
181 |
|
|
W_edge = exch2_isWedge(myTile).EQ.1 |
182 |
|
|
#else |
183 |
|
|
nCFace = bi |
184 |
|
|
N_edge = .TRUE. |
185 |
|
|
S_edge = .TRUE. |
186 |
|
|
E_edge = .TRUE. |
187 |
|
|
W_edge = .TRUE. |
188 |
|
|
#endif |
189 |
|
|
ELSE |
190 |
|
|
nipass=2 |
191 |
|
|
nCFace = bi |
192 |
|
|
N_edge = .FALSE. |
193 |
|
|
S_edge = .FALSE. |
194 |
|
|
E_edge = .FALSE. |
195 |
|
|
W_edge = .FALSE. |
196 |
|
|
ENDIF |
197 |
|
|
|
198 |
|
|
iMin = 1-OLx |
199 |
|
|
iMax = sNx+OLx |
200 |
|
|
jMin = 1-OLy |
201 |
|
|
jMax = sNy+OLy |
202 |
|
|
|
203 |
|
|
C-- Start horizontal fluxes |
204 |
|
|
|
205 |
|
|
C-- set mask West & South |
206 |
|
|
DO j=1-OLy,sNy+OLy |
207 |
|
|
maskLocW(1-Olx,j) = 0. |
208 |
|
|
DO i=2-OLx,sNx+OLx |
209 |
|
|
maskLocW(i,j) = MIN( maskOc(i-1,j), maskOc(i,j) ) |
210 |
|
|
ENDDO |
211 |
|
|
ENDDO |
212 |
|
|
DO i=1-OLx,sNx+OLx |
213 |
|
|
maskLocS(i,1-Oly) = 0. |
214 |
|
|
ENDDO |
215 |
|
|
DO j=2-OLy,sNy+OLy |
216 |
|
|
DO i=1-OLx,sNx+OLx |
217 |
|
|
maskLocS(i,j) = MIN( maskOc(i,j-1), maskOc(i,j) ) |
218 |
|
|
ENDDO |
219 |
|
|
ENDDO |
220 |
|
|
|
221 |
heimbach |
1.4 |
#ifndef ALLOW_AUTODIFF_TAMC |
222 |
jmc |
1.1 |
IF (useCubedSphereExchange) THEN |
223 |
|
|
withSigns = .FALSE. |
224 |
|
|
CALL FILL_CS_CORNER_UV_RS( |
225 |
|
|
& withSigns, maskLocW,maskLocS, bi,bj, myThid ) |
226 |
|
|
ENDIF |
227 |
heimbach |
1.4 |
#endif |
228 |
jmc |
1.1 |
|
229 |
|
|
C-- Multiple passes for different directions on different tiles |
230 |
|
|
C-- For cube need one pass for each of red, green and blue axes. |
231 |
|
|
DO ipass=1,nipass |
232 |
heimbach |
1.4 |
#ifdef ALLOW_AUTODIFF_TAMC |
233 |
|
|
ikey_4 = ipass |
234 |
|
|
& + nipass*act1 |
235 |
|
|
& + nipass*max1*act2 |
236 |
|
|
& + nipass*max1*max2*act3 |
237 |
|
|
& + nipass*max1*max2*max3*act4 |
238 |
|
|
#endif |
239 |
|
|
|
240 |
|
|
#ifdef ALLOW_AUTODIFF_TAMC |
241 |
|
|
CADJ STORE iceFld(:,:) = comlev1_thsice_4, key=ikey_4, byte=isbyte |
242 |
|
|
CADJ STORE iceVol(:,:) = comlev1_thsice_4, key=ikey_4, byte=isbyte |
243 |
|
|
CADJ STORE af(:,:) = comlev1_thsice_4, key=ikey_4, byte=isbyte |
244 |
|
|
#endif |
245 |
jmc |
1.1 |
|
246 |
|
|
interiorOnly = .FALSE. |
247 |
|
|
overlapOnly = .FALSE. |
248 |
|
|
IF (useCubedSphereExchange) THEN |
249 |
|
|
C-- CubedSphere : pass 3 times, with partial update of local seaice field |
250 |
|
|
IF (ipass.EQ.1) THEN |
251 |
|
|
overlapOnly = MOD(nCFace,3).EQ.0 |
252 |
|
|
interiorOnly = MOD(nCFace,3).NE.0 |
253 |
|
|
calc_fluxes_X = nCFace.EQ.6 .OR. nCFace.EQ.1 .OR. nCFace.EQ.2 |
254 |
|
|
calc_fluxes_Y = nCFace.EQ.3 .OR. nCFace.EQ.4 .OR. nCFace.EQ.5 |
255 |
|
|
ELSEIF (ipass.EQ.2) THEN |
256 |
|
|
overlapOnly = MOD(nCFace,3).EQ.2 |
257 |
|
|
calc_fluxes_X = nCFace.EQ.2 .OR. nCFace.EQ.3 .OR. nCFace.EQ.4 |
258 |
|
|
calc_fluxes_Y = nCFace.EQ.5 .OR. nCFace.EQ.6 .OR. nCFace.EQ.1 |
259 |
|
|
ELSE |
260 |
jmc |
1.5 |
interiorOnly = MOD(nCFace,3).EQ.0 |
261 |
jmc |
1.1 |
calc_fluxes_X = nCFace.EQ.5 .OR. nCFace.EQ.6 |
262 |
|
|
calc_fluxes_Y = nCFace.EQ.2 .OR. nCFace.EQ.3 |
263 |
|
|
ENDIF |
264 |
|
|
ELSE |
265 |
|
|
C-- not CubedSphere |
266 |
|
|
calc_fluxes_X = MOD(ipass,2).EQ.1 |
267 |
|
|
calc_fluxes_Y = .NOT.calc_fluxes_X |
268 |
|
|
ENDIF |
269 |
|
|
IF (dBug.AND.bi.EQ.3 ) WRITE(6,*) 'ICE_adv:',tracerIdentity, |
270 |
|
|
& ipass,calc_fluxes_X,calc_fluxes_Y,overlapOnly,interiorOnly |
271 |
|
|
|
272 |
|
|
C---+----1----+----2----+----3----+----4----+----5----+----6----+----7-|--+----| |
273 |
|
|
C-- X direction |
274 |
|
|
|
275 |
|
|
IF (calc_fluxes_X) THEN |
276 |
|
|
|
277 |
|
|
C- Do not compute fluxes if |
278 |
|
|
C a) needed in overlap only |
279 |
|
|
C and b) the overlap of myTile are not cube-face Edges |
280 |
heimbach |
1.4 |
#ifdef ALLOW_AUTODIFF_TAMC |
281 |
|
|
CADJ STORE iceFld(:,:) = comlev1_thsice_4, key=ikey_4, byte=isbyte |
282 |
|
|
CADJ STORE iceVol(:,:) = comlev1_thsice_4, key=ikey_4, byte=isbyte |
283 |
|
|
#endif |
284 |
jmc |
1.1 |
IF ( .NOT.overlapOnly .OR. N_edge .OR. S_edge ) THEN |
285 |
|
|
|
286 |
|
|
C- Advective flux in X |
287 |
|
|
DO j=1-Oly,sNy+Oly |
288 |
|
|
DO i=1-Olx,sNx+Olx |
289 |
|
|
af(i,j) = 0. |
290 |
|
|
ENDDO |
291 |
|
|
ENDDO |
292 |
|
|
|
293 |
heimbach |
1.4 |
#ifndef ALLOW_AUTODIFF_TAMC |
294 |
jmc |
1.1 |
C- Internal exchange for calculations in X |
295 |
|
|
IF ( useCubedSphereExchange .AND. |
296 |
|
|
& ( overlapOnly .OR. ipass.EQ.1 ) ) THEN |
297 |
jmc |
1.6 |
CALL FILL_CS_CORNER_TR_RL( .TRUE., .FALSE., |
298 |
|
|
& iceFld, bi,bj, myThid ) |
299 |
jmc |
1.1 |
IF ( .NOT.useGridVolume ) |
300 |
jmc |
1.6 |
& CALL FILL_CS_CORNER_TR_RL( .TRUE., .FALSE., |
301 |
|
|
& iceVol, bi,bj, myThid ) |
302 |
jmc |
1.1 |
ENDIF |
303 |
heimbach |
1.4 |
#endif |
304 |
jmc |
1.1 |
|
305 |
|
|
C- Compute CFL number |
306 |
|
|
IF ( useGridVolume ) THEN |
307 |
|
|
DO j=1-Oly,sNy+Oly |
308 |
|
|
DO i=2-Olx,sNx+Olx |
309 |
|
|
uCfl(i,j) = deltaTadvect*( |
310 |
|
|
& MAX( uTrans(i,j), 0. _d 0 )*recip_rA(i-1,j,bi,bj) |
311 |
|
|
& +MAX(-uTrans(i,j), 0. _d 0 )*recip_rA( i ,j,bi,bj) |
312 |
|
|
& ) |
313 |
|
|
ENDDO |
314 |
|
|
ENDDO |
315 |
|
|
ELSE |
316 |
|
|
DO j=1-Oly,sNy+Oly |
317 |
|
|
DO i=2-Olx,sNx+Olx |
318 |
|
|
uCfl(i,j) = deltaTadvect*( |
319 |
|
|
& MAX( uTrans(i,j), 0. _d 0 )/MAX( iceVol(i-1,j), iceEps ) |
320 |
|
|
& +MAX(-uTrans(i,j), 0. _d 0 )/MAX( iceVol( i ,j), iceEps ) |
321 |
|
|
& ) |
322 |
|
|
ENDDO |
323 |
|
|
ENDDO |
324 |
|
|
ENDIF |
325 |
|
|
|
326 |
|
|
IF ( advectionScheme.EQ.ENUM_UPWIND_1RST |
327 |
|
|
& .OR. advectionScheme.EQ.ENUM_DST2 ) THEN |
328 |
|
|
CALL GAD_DST2U1_ADV_X( bi,bj,k, advectionScheme, .FALSE., |
329 |
|
|
I deltaTadvect, uTrans, uCfl, iceFld, |
330 |
|
|
O af, myThid ) |
331 |
|
|
IF (dBug.AND. bi.EQ.3) WRITE(6,'(A,1P4E14.6)') |
332 |
|
|
& 'ICE_adv: xFx=',af(13,11),iceFld(12,11),uTrans(13,11), |
333 |
|
|
& af(13,11)/uTrans(13,11) |
334 |
|
|
ELSEIF (advectionScheme.EQ.ENUM_FLUX_LIMIT) THEN |
335 |
|
|
CALL GAD_FLUXLIMIT_ADV_X( bi,bj,k, .FALSE., deltaTadvect, |
336 |
|
|
I uTrans, uCfl, maskLocW, iceFld, |
337 |
|
|
O af, myThid ) |
338 |
|
|
ELSEIF (advectionScheme.EQ.ENUM_DST3 ) THEN |
339 |
|
|
CALL GAD_DST3_ADV_X( bi,bj,k, .FALSE., deltaTadvect, |
340 |
|
|
I uTrans, uCfl, maskLocW, iceFld, |
341 |
|
|
O af, myThid ) |
342 |
|
|
ELSEIF (advectionScheme.EQ.ENUM_DST3_FLUX_LIMIT ) THEN |
343 |
|
|
CALL GAD_DST3FL_ADV_X( bi,bj,k, .FALSE., deltaTadvect, |
344 |
|
|
I uTrans, uCfl, maskLocW, iceFld, |
345 |
|
|
O af, myThid ) |
346 |
|
|
ELSE |
347 |
|
|
STOP |
348 |
|
|
& 'THSICE_ADVECTION: adv. scheme incompatibale with multi-dim' |
349 |
|
|
ENDIF |
350 |
|
|
|
351 |
|
|
C-- Advective flux in X : done |
352 |
|
|
ENDIF |
353 |
|
|
|
354 |
heimbach |
1.4 |
#ifndef ALLOW_AUTODIFF_TAMC |
355 |
jmc |
1.1 |
C-- Internal exchange for next calculations in Y |
356 |
|
|
IF ( overlapOnly .AND. ipass.EQ.1 ) THEN |
357 |
jmc |
1.6 |
CALL FILL_CS_CORNER_TR_RL(.FALSE., .FALSE., |
358 |
|
|
& iceFld, bi,bj, myThid ) |
359 |
jmc |
1.1 |
IF ( .NOT.useGridVolume ) |
360 |
jmc |
1.6 |
& CALL FILL_CS_CORNER_TR_RL(.FALSE., .FALSE., |
361 |
|
|
& iceVol, bi,bj, myThid ) |
362 |
jmc |
1.1 |
ENDIF |
363 |
heimbach |
1.4 |
#endif |
364 |
jmc |
1.1 |
|
365 |
|
|
C- Update the local seaice field where needed: |
366 |
|
|
|
367 |
|
|
C update in overlap-Only |
368 |
|
|
IF ( overlapOnly ) THEN |
369 |
|
|
iMinUpd = 1-OLx+1 |
370 |
|
|
iMaxUpd = sNx+OLx-1 |
371 |
|
|
C-- notes: these 2 lines below have no real effect (because recip_hFac=0 |
372 |
|
|
C in corner region) but safer to keep them. |
373 |
|
|
IF ( W_edge ) iMinUpd = 1 |
374 |
|
|
IF ( E_edge ) iMaxUpd = sNx |
375 |
|
|
|
376 |
heimbach |
1.4 |
#ifdef ALLOW_AUTODIFF_TAMC |
377 |
|
|
CADJ STORE iceFld(:,:) = comlev1_thsice_4, key=ikey_4, byte=isbyte |
378 |
|
|
CADJ STORE iceVol(:,:) = comlev1_thsice_4, key=ikey_4, byte=isbyte |
379 |
|
|
#endif |
380 |
jmc |
1.1 |
IF ( S_edge .AND. useGridVolume ) THEN |
381 |
|
|
DO j=1-OLy,0 |
382 |
|
|
DO i=iMinUpd,iMaxUpd |
383 |
|
|
iceFld(i,j) = iceFld(i,j) |
384 |
|
|
& -deltaTadvect*maskOc(i,j) |
385 |
|
|
& *recip_rA(i,j,bi,bj) |
386 |
|
|
& *( af(i+1,j)-af(i,j) ) |
387 |
|
|
ENDDO |
388 |
|
|
ENDDO |
389 |
|
|
ELSEIF ( S_edge ) THEN |
390 |
|
|
DO j=1-OLy,0 |
391 |
|
|
DO i=iMinUpd,iMaxUpd |
392 |
|
|
tmpVol = iceVol(i,j) |
393 |
|
|
iceVol(i,j) = iceVol(i,j) |
394 |
|
|
& -deltaTadvect*maskOc(i,j) |
395 |
|
|
& *( uTrans(i+1,j)-uTrans(i,j) ) |
396 |
|
|
IF ( iceVol(i,j).GT.iceEps ) |
397 |
|
|
& iceFld(i,j) = ( iceFld(i,j)*tmpVol |
398 |
|
|
& -deltaTadvect*maskOc(i,j) |
399 |
|
|
& *( af(i+1,j)-af(i,j) ) |
400 |
|
|
& )/iceVol(i,j) |
401 |
|
|
ENDDO |
402 |
|
|
ENDDO |
403 |
|
|
ENDIF |
404 |
|
|
IF ( N_edge .AND. useGridVolume ) THEN |
405 |
|
|
DO j=sNy+1,sNy+OLy |
406 |
|
|
DO i=iMinUpd,iMaxUpd |
407 |
|
|
iceFld(i,j) = iceFld(i,j) |
408 |
|
|
& -deltaTadvect*maskOc(i,j) |
409 |
|
|
& *recip_rA(i,j,bi,bj) |
410 |
|
|
& *( af(i+1,j)-af(i,j) ) |
411 |
|
|
ENDDO |
412 |
|
|
ENDDO |
413 |
|
|
ELSEIF ( N_edge ) THEN |
414 |
|
|
DO j=sNy+1,sNy+OLy |
415 |
|
|
DO i=iMinUpd,iMaxUpd |
416 |
|
|
tmpVol = iceVol(i,j) |
417 |
|
|
iceVol(i,j) = iceVol(i,j) |
418 |
|
|
& -deltaTadvect*maskOc(i,j) |
419 |
|
|
& *( uTrans(i+1,j)-uTrans(i,j) ) |
420 |
|
|
IF ( iceVol(i,j).GT.iceEps ) |
421 |
|
|
& iceFld(i,j) = ( iceFld(i,j)*tmpVol |
422 |
|
|
& -deltaTadvect*maskOc(i,j) |
423 |
|
|
& *( af(i+1,j)-af(i,j) ) |
424 |
|
|
& )/iceVol(i,j) |
425 |
|
|
ENDDO |
426 |
|
|
ENDDO |
427 |
|
|
ENDIF |
428 |
|
|
C-- keep advective flux (for diagnostics) |
429 |
|
|
IF ( S_edge ) THEN |
430 |
|
|
DO j=1-OLy,0 |
431 |
|
|
DO i=1-OLx+1,sNx+OLx |
432 |
|
|
afx(i,j) = af(i,j) |
433 |
|
|
ENDDO |
434 |
|
|
ENDDO |
435 |
|
|
ENDIF |
436 |
|
|
IF ( N_edge ) THEN |
437 |
|
|
DO j=sNy+1,sNy+OLy |
438 |
|
|
DO i=1-OLx+1,sNx+OLx |
439 |
|
|
afx(i,j) = af(i,j) |
440 |
|
|
ENDDO |
441 |
|
|
ENDDO |
442 |
|
|
ENDIF |
443 |
|
|
|
444 |
|
|
ELSE |
445 |
|
|
C do not only update the overlap |
446 |
|
|
jMinUpd = 1-OLy |
447 |
|
|
jMaxUpd = sNy+OLy |
448 |
|
|
IF ( interiorOnly .AND. S_edge ) jMinUpd = 1 |
449 |
|
|
IF ( interiorOnly .AND. N_edge ) jMaxUpd = sNy |
450 |
|
|
IF ( useGridVolume ) THEN |
451 |
|
|
DO j=jMinUpd,jMaxUpd |
452 |
|
|
DO i=1-OLx+1,sNx+OLx-1 |
453 |
|
|
iceFld(i,j) = iceFld(i,j) |
454 |
|
|
& -deltaTadvect*maskOc(i,j) |
455 |
|
|
& *recip_rA(i,j,bi,bj) |
456 |
|
|
& *( af(i+1,j)-af(i,j) ) |
457 |
|
|
ENDDO |
458 |
|
|
ENDDO |
459 |
|
|
ELSE |
460 |
|
|
DO j=jMinUpd,jMaxUpd |
461 |
|
|
DO i=1-OLx+1,sNx+OLx-1 |
462 |
|
|
tmpVol = iceVol(i,j) |
463 |
|
|
iceVol(i,j) = iceVol(i,j) |
464 |
|
|
& -deltaTadvect*maskOc(i,j) |
465 |
|
|
& *( uTrans(i+1,j)-uTrans(i,j) ) |
466 |
|
|
IF ( iceVol(i,j).GT.iceEps ) |
467 |
|
|
& iceFld(i,j) = ( iceFld(i,j)*tmpVol |
468 |
|
|
& -deltaTadvect*maskOc(i,j) |
469 |
|
|
& *( af(i+1,j)-af(i,j) ) |
470 |
|
|
& )/iceVol(i,j) |
471 |
|
|
ENDDO |
472 |
|
|
ENDDO |
473 |
|
|
ENDIF |
474 |
|
|
C-- keep advective flux (for diagnostics) |
475 |
|
|
DO j=jMinUpd,jMaxUpd |
476 |
|
|
DO i=1-OLx+1,sNx+OLx |
477 |
|
|
afx(i,j) = af(i,j) |
478 |
|
|
ENDDO |
479 |
|
|
ENDDO |
480 |
|
|
|
481 |
|
|
C- end if/else update overlap-Only |
482 |
|
|
ENDIF |
483 |
|
|
|
484 |
|
|
C-- End of X direction |
485 |
|
|
ENDIF |
486 |
|
|
|
487 |
|
|
C---+----1----+----2----+----3----+----4----+----5----+----6----+----7-|--+----| |
488 |
|
|
C-- Y direction |
489 |
|
|
|
490 |
|
|
IF (calc_fluxes_Y) THEN |
491 |
|
|
|
492 |
|
|
C- Do not compute fluxes if |
493 |
|
|
C a) needed in overlap only |
494 |
|
|
C and b) the overlap of myTile are not cube-face edges |
495 |
heimbach |
1.4 |
#ifdef ALLOW_AUTODIFF_TAMC |
496 |
|
|
CADJ STORE iceFld(:,:) = comlev1_thsice_4, key=ikey_4, byte=isbyte |
497 |
|
|
CADJ STORE iceVol(:,:) = comlev1_thsice_4, key=ikey_4, byte=isbyte |
498 |
|
|
#endif |
499 |
jmc |
1.1 |
IF ( .NOT.overlapOnly .OR. E_edge .OR. W_edge ) THEN |
500 |
|
|
|
501 |
|
|
C- Advective flux in Y |
502 |
|
|
DO j=1-OLy,sNy+OLy |
503 |
|
|
DO i=1-OLx,sNx+OLx |
504 |
|
|
af(i,j) = 0. |
505 |
|
|
ENDDO |
506 |
|
|
ENDDO |
507 |
|
|
|
508 |
heimbach |
1.4 |
#ifndef ALLOW_AUTODIFF_TAMC |
509 |
jmc |
1.1 |
C- Internal exchange for calculations in Y |
510 |
|
|
IF ( useCubedSphereExchange .AND. |
511 |
|
|
& ( overlapOnly .OR. ipass.EQ.1 ) ) THEN |
512 |
jmc |
1.6 |
CALL FILL_CS_CORNER_TR_RL(.FALSE., .FALSE., |
513 |
|
|
& iceFld, bi,bj, myThid ) |
514 |
jmc |
1.1 |
IF ( .NOT.useGridVolume ) |
515 |
jmc |
1.6 |
& CALL FILL_CS_CORNER_TR_RL(.FALSE., .FALSE., |
516 |
|
|
& iceVol, bi,bj, myThid ) |
517 |
jmc |
1.1 |
ENDIF |
518 |
heimbach |
1.4 |
#endif |
519 |
jmc |
1.1 |
|
520 |
|
|
C- Compute CFL number |
521 |
|
|
IF ( useGridVolume ) THEN |
522 |
|
|
DO j=2-Oly,sNy+Oly |
523 |
|
|
DO i=1-Olx,sNx+Olx |
524 |
|
|
vCfl(i,j) = deltaTadvect*( |
525 |
|
|
& MAX( vTrans(i,j), 0. _d 0 )*recip_rA(i,j-1,bi,bj) |
526 |
|
|
& +MAX(-vTrans(i,j), 0. _d 0 )*recip_rA(i, j ,bi,bj) |
527 |
|
|
& ) |
528 |
|
|
ENDDO |
529 |
|
|
ENDDO |
530 |
|
|
ELSE |
531 |
|
|
DO j=2-Oly,sNy+Oly |
532 |
|
|
DO i=1-Olx,sNx+Olx |
533 |
|
|
vCfl(i,j) = deltaTadvect*( |
534 |
|
|
& MAX( vTrans(i,j), 0. _d 0 )/MAX( iceVol(i,j-1), iceEps ) |
535 |
|
|
& +MAX(-vTrans(i,j), 0. _d 0 )/MAX( iceVol(i, j ), iceEps ) |
536 |
|
|
& ) |
537 |
|
|
ENDDO |
538 |
|
|
ENDDO |
539 |
|
|
ENDIF |
540 |
|
|
|
541 |
|
|
IF ( advectionScheme.EQ.ENUM_UPWIND_1RST |
542 |
|
|
& .OR. advectionScheme.EQ.ENUM_DST2 ) THEN |
543 |
|
|
CALL GAD_DST2U1_ADV_Y( bi,bj,k, advectionScheme, .FALSE., |
544 |
|
|
I deltaTadvect, vTrans, vCfl, iceFld, |
545 |
|
|
O af, myThid ) |
546 |
|
|
IF (dBug.AND. bi.EQ.3) WRITE(6,'(A,1P4E14.6)') |
547 |
|
|
& 'ICE_adv: yFx=',af(12,12),iceFld(12,11),vTrans(12,12), |
548 |
|
|
& af(12,12)/vTrans(12,12) |
549 |
|
|
ELSEIF (advectionScheme.EQ.ENUM_FLUX_LIMIT) THEN |
550 |
|
|
CALL GAD_FLUXLIMIT_ADV_Y( bi,bj,k, .FALSE., deltaTadvect, |
551 |
|
|
I vTrans, vCfl, maskLocS, iceFld, |
552 |
|
|
O af, myThid ) |
553 |
|
|
ELSEIF (advectionScheme.EQ.ENUM_DST3 ) THEN |
554 |
|
|
CALL GAD_DST3_ADV_Y( bi,bj,k, .FALSE., deltaTadvect, |
555 |
|
|
I vTrans, vCfl, maskLocS, iceFld, |
556 |
|
|
O af, myThid ) |
557 |
|
|
ELSEIF (advectionScheme.EQ.ENUM_DST3_FLUX_LIMIT ) THEN |
558 |
|
|
CALL GAD_DST3FL_ADV_Y( bi,bj,k, .FALSE., deltaTadvect, |
559 |
|
|
I vTrans, vCfl, maskLocS, iceFld, |
560 |
|
|
O af, myThid ) |
561 |
|
|
ELSE |
562 |
|
|
STOP |
563 |
|
|
& 'THSICE_ADVECTION: adv. scheme incompatibale with mutli-dim' |
564 |
|
|
ENDIF |
565 |
|
|
|
566 |
|
|
C- Advective flux in Y : done |
567 |
|
|
ENDIF |
568 |
|
|
|
569 |
heimbach |
1.4 |
#ifndef ALLOW_AUTODIFF_TAMC |
570 |
jmc |
1.1 |
IF ( overlapOnly .AND. ipass.EQ.1 ) THEN |
571 |
jmc |
1.6 |
CALL FILL_CS_CORNER_TR_RL( .TRUE., .FALSE., |
572 |
|
|
& iceFld, bi,bj, myThid ) |
573 |
jmc |
1.1 |
IF ( .NOT.useGridVolume ) |
574 |
jmc |
1.6 |
& CALL FILL_CS_CORNER_TR_RL( .TRUE., .FALSE., |
575 |
|
|
& iceVol, bi,bj, myThid ) |
576 |
jmc |
1.1 |
ENDIF |
577 |
heimbach |
1.4 |
#endif |
578 |
jmc |
1.1 |
|
579 |
|
|
C- Update the local seaice field where needed: |
580 |
|
|
|
581 |
|
|
C update in overlap-Only |
582 |
|
|
IF ( overlapOnly ) THEN |
583 |
|
|
jMinUpd = 1-OLy+1 |
584 |
|
|
jMaxUpd = sNy+OLy-1 |
585 |
|
|
C- notes: these 2 lines below have no real effect (because recip_hFac=0 |
586 |
|
|
C in corner region) but safer to keep them. |
587 |
|
|
IF ( S_edge ) jMinUpd = 1 |
588 |
|
|
IF ( N_edge ) jMaxUpd = sNy |
589 |
|
|
|
590 |
heimbach |
1.4 |
#ifdef ALLOW_AUTODIFF_TAMC |
591 |
|
|
CADJ STORE iceFld(:,:) = comlev1_thsice_4, key=ikey_4, byte=isbyte |
592 |
|
|
CADJ STORE iceVol(:,:) = comlev1_thsice_4, key=ikey_4, byte=isbyte |
593 |
|
|
#endif |
594 |
jmc |
1.1 |
IF ( W_edge .AND. useGridVolume ) THEN |
595 |
|
|
DO j=jMinUpd,jMaxUpd |
596 |
|
|
DO i=1-OLx,0 |
597 |
|
|
iceFld(i,j) = iceFld(i,j) |
598 |
|
|
& -deltaTadvect*maskOc(i,j) |
599 |
|
|
& *recip_rA(i,j,bi,bj) |
600 |
|
|
& *( af(i,j+1)-af(i,j) ) |
601 |
|
|
ENDDO |
602 |
|
|
ENDDO |
603 |
|
|
ELSEIF ( W_edge ) THEN |
604 |
|
|
DO j=jMinUpd,jMaxUpd |
605 |
|
|
DO i=1-OLx,0 |
606 |
|
|
tmpVol = iceVol(i,j) |
607 |
|
|
iceVol(i,j) = iceVol(i,j) |
608 |
|
|
& -deltaTadvect*maskOc(i,j) |
609 |
|
|
& *( vTrans(i,j+1)-vTrans(i,j) ) |
610 |
|
|
IF ( iceVol(i,j).GT.iceEps ) |
611 |
|
|
& iceFld(i,j) = ( iceFld(i,j)*tmpVol |
612 |
|
|
& -deltaTadvect*maskOc(i,j) |
613 |
|
|
& *( af(i,j+1)-af(i,j) ) |
614 |
|
|
& )/iceVol(i,j) |
615 |
|
|
ENDDO |
616 |
|
|
ENDDO |
617 |
|
|
ENDIF |
618 |
|
|
IF ( E_edge .AND. useGridVolume ) THEN |
619 |
|
|
DO j=jMinUpd,jMaxUpd |
620 |
|
|
DO i=sNx+1,sNx+OLx |
621 |
|
|
iceFld(i,j) = iceFld(i,j) |
622 |
|
|
& -deltaTadvect*maskOc(i,j) |
623 |
|
|
& *recip_rA(i,j,bi,bj) |
624 |
|
|
& *( af(i,j+1)-af(i,j) ) |
625 |
|
|
ENDDO |
626 |
|
|
ENDDO |
627 |
|
|
ELSEIF ( E_edge ) THEN |
628 |
|
|
DO j=jMinUpd,jMaxUpd |
629 |
|
|
DO i=sNx+1,sNx+OLx |
630 |
|
|
tmpVol = iceVol(i,j) |
631 |
|
|
iceVol(i,j) = iceVol(i,j) |
632 |
|
|
& -deltaTadvect*maskOc(i,j) |
633 |
|
|
& *( vTrans(i,j+1)-vTrans(i,j) ) |
634 |
|
|
IF ( iceVol(i,j).GT.iceEps ) |
635 |
|
|
& iceFld(i,j) = ( iceFld(i,j)*tmpVol |
636 |
|
|
& -deltaTadvect*maskOc(i,j) |
637 |
|
|
& *( af(i,j+1)-af(i,j) ) |
638 |
|
|
& )/iceVol(i,j) |
639 |
|
|
ENDDO |
640 |
|
|
ENDDO |
641 |
|
|
ENDIF |
642 |
|
|
C-- keep advective flux (for diagnostics) |
643 |
|
|
IF ( W_edge ) THEN |
644 |
|
|
DO j=1-OLy+1,sNy+OLy |
645 |
|
|
DO i=1-OLx,0 |
646 |
|
|
afy(i,j) = af(i,j) |
647 |
|
|
ENDDO |
648 |
|
|
ENDDO |
649 |
|
|
ENDIF |
650 |
|
|
IF ( E_edge ) THEN |
651 |
|
|
DO j=1-OLy+1,sNy+OLy |
652 |
|
|
DO i=sNx+1,sNx+OLx |
653 |
|
|
afy(i,j) = af(i,j) |
654 |
|
|
ENDDO |
655 |
|
|
ENDDO |
656 |
|
|
ENDIF |
657 |
|
|
|
658 |
|
|
ELSE |
659 |
|
|
C do not only update the overlap |
660 |
|
|
iMinUpd = 1-OLx |
661 |
|
|
iMaxUpd = sNx+OLx |
662 |
|
|
IF ( interiorOnly .AND. W_edge ) iMinUpd = 1 |
663 |
|
|
IF ( interiorOnly .AND. E_edge ) iMaxUpd = sNx |
664 |
|
|
IF ( useGridVolume ) THEN |
665 |
|
|
DO j=1-OLy+1,sNy+OLy-1 |
666 |
|
|
DO i=iMinUpd,iMaxUpd |
667 |
|
|
iceFld(i,j) = iceFld(i,j) |
668 |
|
|
& -deltaTadvect*maskOc(i,j) |
669 |
|
|
& *recip_rA(i,j,bi,bj) |
670 |
|
|
& *( af(i,j+1)-af(i,j) ) |
671 |
|
|
ENDDO |
672 |
|
|
ENDDO |
673 |
|
|
ELSE |
674 |
|
|
DO j=1-OLy+1,sNy+OLy-1 |
675 |
|
|
DO i=iMinUpd,iMaxUpd |
676 |
|
|
tmpVol = iceVol(i,j) |
677 |
|
|
iceVol(i,j) = iceVol(i,j) |
678 |
|
|
& -deltaTadvect*maskOc(i,j) |
679 |
|
|
& *( vTrans(i,j+1)-vTrans(i,j) ) |
680 |
|
|
IF ( iceVol(i,j).GT.iceEps ) |
681 |
|
|
& iceFld(i,j) = ( iceFld(i,j)*tmpVol |
682 |
|
|
& -deltaTadvect*maskOc(i,j) |
683 |
|
|
& *( af(i,j+1)-af(i,j) ) |
684 |
|
|
& )/iceVol(i,j) |
685 |
|
|
ENDDO |
686 |
|
|
ENDDO |
687 |
|
|
ENDIF |
688 |
|
|
C-- keep advective flux (for diagnostics) |
689 |
|
|
DO j=1-OLy+1,sNy+OLy |
690 |
|
|
DO i=iMinUpd,iMaxUpd |
691 |
|
|
afy(i,j) = af(i,j) |
692 |
|
|
ENDDO |
693 |
|
|
ENDDO |
694 |
|
|
|
695 |
|
|
C end if/else update overlap-Only |
696 |
|
|
ENDIF |
697 |
|
|
|
698 |
|
|
C-- End of Y direction |
699 |
|
|
ENDIF |
700 |
|
|
|
701 |
|
|
C-- End of ipass loop |
702 |
|
|
ENDDO |
703 |
|
|
|
704 |
|
|
C- explicit advection is done ; add some debugging print |
705 |
|
|
IF ( dBug ) THEN |
706 |
|
|
DO j=1-OLy,sNy+OLy |
707 |
|
|
DO i=1-OLx,sNx+OLx |
708 |
|
|
IF ( i.EQ.12 .AND. j.EQ.11 .AND. bi.EQ.3 ) THEN |
709 |
|
|
tmpFac= deltaTadvect*recip_rA(i,j,bi,bj) |
710 |
|
|
WRITE(6,'(A,1P4E14.6)') 'ICE_adv:', |
711 |
|
|
& afx(i,j)*tmpFac,afx(i+1,j)*tmpFac, |
712 |
|
|
& afy(i,j)*tmpFac,afy(i,j+1)*tmpFac |
713 |
|
|
ENDIF |
714 |
|
|
ENDDO |
715 |
|
|
ENDDO |
716 |
|
|
ENDIF |
717 |
|
|
|
718 |
|
|
#ifdef ALLOW_DEBUG |
719 |
|
|
IF ( debugLevel .GE. debLevB |
720 |
|
|
& .AND. tracerIdentity.EQ.GAD_SI_HICE |
721 |
|
|
& .AND. k.LE.3 .AND. myIter.EQ.1+nIter0 |
722 |
|
|
& .AND. nPx.EQ.1 .AND. nPy.EQ.1 |
723 |
|
|
& .AND. useCubedSphereExchange ) THEN |
724 |
|
|
CALL DEBUG_CS_CORNER_UV( ' afx,afy from THSICE_ADVECTION', |
725 |
|
|
& afx,afy, k, standardMessageUnit,bi,bj,myThid ) |
726 |
|
|
ENDIF |
727 |
|
|
#endif /* ALLOW_DEBUG */ |
728 |
|
|
|
729 |
jmc |
1.2 |
#endif /* ALLOW_GENERIC_ADVDIFF */ |
730 |
|
|
|
731 |
jmc |
1.1 |
RETURN |
732 |
|
|
END |