| 1 |
C $Header: /u/gcmpack/models/MITgcmUV/model/src/calc_isoslopes.F,v 1.15 2000/06/09 02:45:04 heimbach Exp $ |
| 2 |
#include "CPP_OPTIONS.h" |
| 3 |
|
| 4 |
CStartOfInterface |
| 5 |
SUBROUTINE CALC_ISOSLOPES( |
| 6 |
I bi, bj, iMin, iMax, jMin, jMax, K, |
| 7 |
I rhoKm1, rhoKp1, rhotmp, |
| 8 |
O K13, K23, K33, KapGM, |
| 9 |
I myThid ) |
| 10 |
C /==========================================================\ |
| 11 |
C | SUBROUTINE CALC_ISOSLOPES | |
| 12 |
C | o Calculate isentropic surface slopes for GM scheme. | |
| 13 |
C |==========================================================| |
| 14 |
C | The small-angle approixmation of the full-slope tensor | |
| 15 |
C | is used here. A maximum slope is also imposed to prevent | |
| 16 |
C | unphysical values in homogenised regions. | |
| 17 |
C | This routine works on a given XY slab of fluid. | |
| 18 |
C \==========================================================/ |
| 19 |
IMPLICIT NONE |
| 20 |
|
| 21 |
C == Global variables == |
| 22 |
#include "SIZE.h" |
| 23 |
#include "GRID.h" |
| 24 |
#include "DYNVARS.h" |
| 25 |
#include "EEPARAMS.h" |
| 26 |
#include "PARAMS.h" |
| 27 |
|
| 28 |
C == Routine arguments == |
| 29 |
C |
| 30 |
_RL rhoKm1(1-Olx:sNx+Olx,1-Oly:sNy+Oly) |
| 31 |
_RL rhoKp1(1-Olx:sNx+Olx,1-Oly:sNy+Oly) |
| 32 |
_RL rhotmp(1-Olx:sNx+Olx,1-Oly:sNy+Oly) |
| 33 |
_RL K13(1-Olx:sNx+Olx,1-Oly:sNy+Oly,Nr) |
| 34 |
_RL K23(1-Olx:sNx+Olx,1-Oly:sNy+Oly,Nr) |
| 35 |
_RL K33(1-Olx:sNx+Olx,1-Oly:sNy+Oly,Nr) |
| 36 |
_RL KapGM(1-Olx:sNx+Olx,1-Oly:sNy+Oly) |
| 37 |
INTEGER bi,bj,iMin,iMax,jMin,jMax,K |
| 38 |
INTEGER myThid |
| 39 |
CEndOfInterface |
| 40 |
|
| 41 |
C == Local variables == |
| 42 |
INTEGER i,j |
| 43 |
_RL dSigmaDx,dSigmaDy,dSigmaDr,SlopeX,SlopeY,dRdSigma,N2 |
| 44 |
_RL gradSmod,stratLimit,rMaxSlope |
| 45 |
_RL Small_Number |
| 46 |
PARAMETER(Small_Number=1. _d -12) |
| 47 |
_RS deltaH,zero_rs |
| 48 |
PARAMETER(zero_rs=0.) |
| 49 |
|
| 50 |
#ifdef ALLOW_AUTODIFF_TAMC |
| 51 |
DO i=1-OLx,sNx+OLx |
| 52 |
K13 (i, 1-OLy,k) = 0.0 |
| 53 |
K23 (i, 1-OLy,k) = 0.0 |
| 54 |
K33 (i, 1-OLy,k) = 0.0 |
| 55 |
K13 (i,sNy+OLy,k) = 0.0 |
| 56 |
K23 (i,sNy+OLy,k) = 0.0 |
| 57 |
K33 (i,sNy+OLy,k) = 0.0 |
| 58 |
ENDDO |
| 59 |
DO j=1-OLy,sNy+OLy |
| 60 |
K13 ( 1-OLx,j,k) = 0.0 |
| 61 |
K23 ( 1-OLx,j,k) = 0.0 |
| 62 |
K33 ( 1-OLx,j,k) = 0.0 |
| 63 |
K13 (sNx+OLx,j,k) = 0.0 |
| 64 |
K23 (sNx+OLx,j,k) = 0.0 |
| 65 |
K33 (sNx+OLx,j,k) = 0.0 |
| 66 |
ENDDO |
| 67 |
#endif |
| 68 |
|
| 69 |
C Calculate terms for mean Richardson number |
| 70 |
C which is used in the "variable K" parameterisaton. |
| 71 |
C Distance between interface above layer and the integration depth (GMdepth) |
| 72 |
deltaH=abs(GMdepth)-abs(rF(k)) |
| 73 |
C If positive we limit this to the layer thickness |
| 74 |
deltaH=min(deltaH,drF(k)) |
| 75 |
C If negative then we are below the integration level |
| 76 |
deltaH=max(deltaH,zero_rs) |
| 77 |
C Now we convert deltaH to a non-dimensional fraction |
| 78 |
deltaH=deltaH/GMdepth |
| 79 |
|
| 80 |
rMaxSlope=0. |
| 81 |
if (GMmaxSlope.ne.0.) rMaxSlope=1. _d 0 / GMmaxSlope |
| 82 |
|
| 83 |
#ifdef ALLOW_AUTODIFF_TAMC |
| 84 |
!HPF$ INDEPENDENT |
| 85 |
#endif |
| 86 |
DO j=1-Oly+1,sNy+Oly-1 |
| 87 |
#ifdef ALLOW_AUTODIFF_TAMC |
| 88 |
!HPF$ INDEPENDENT |
| 89 |
#endif |
| 90 |
DO i=1-Olx+1,sNx+Olx-1 |
| 91 |
|
| 92 |
C Horizontal gradient of Sigma at rVel points |
| 93 |
dSigmaDx=0.25* |
| 94 |
& (_maskW(i+1,j, k ,bi,bj) |
| 95 |
& *_recip_dxC(i+1,j,bi,bj) |
| 96 |
& *(rhoKp1(i+1,j)-rhoKp1( i ,j)) |
| 97 |
& +_maskW( i ,j, k ,bi,bj) |
| 98 |
& *_recip_dxC(i,j,bi,bj) |
| 99 |
& *(rhoKp1( i ,j)-rhoKp1(i-1,j)) |
| 100 |
& +_maskW(i+1,j,k-1,bi,bj) |
| 101 |
& *_recip_dxC(i+1,j,bi,bj) |
| 102 |
& *(rhoKm1(i+1,j)-rhoKm1( i ,j)) |
| 103 |
& +_maskW( i ,j,k-1,bi,bj) |
| 104 |
& *_recip_dxC(i,j,bi,bj) |
| 105 |
& *(rhoKm1( i ,j)-rhoKm1(i-1,j)) ) |
| 106 |
dSigmaDy=0.25* |
| 107 |
& (_maskS(i,j+1, k ,bi,bj) |
| 108 |
& *_recip_dyC(i,j+1,bi,bj) |
| 109 |
& *(rhoKp1(i,j+1)-rhoKp1(i, j )) |
| 110 |
& +_maskS(i, j, k ,bi,bj) |
| 111 |
& *_recip_dyC(i,j,bi,bj) |
| 112 |
& *(rhoKp1(i, j )-rhoKp1(i,j-1)) |
| 113 |
& +_maskS(i,j+1,k-1,bi,bj) |
| 114 |
& *_recip_dyC(i,j+1,bi,bj) |
| 115 |
& *(rhoKm1(i,j+1)-rhoKm1(i, j )) |
| 116 |
& +_maskS(i, j ,k-1,bi,bj) |
| 117 |
& *_recip_dyC(i,j,bi,bj) |
| 118 |
& *(rhoKm1(i, j )-rhoKm1(i,j-1)) ) |
| 119 |
|
| 120 |
C Calculate magnitude of horizontal gradient |
| 121 |
#ifdef ALLOW_AUTODIFF_TAMC |
| 122 |
if (dSigmaDx*dSigmaDx+dSigmaDy*dSigmaDy .eq. 0) then |
| 123 |
gradSmod=0.0 |
| 124 |
else |
| 125 |
gradSmod=sqrt(dSigmaDx*dSigmaDx+dSigmaDy*dSigmaDy) |
| 126 |
endif |
| 127 |
#else |
| 128 |
gradSmod=sqrt(dSigmaDx*dSigmaDx+dSigmaDy*dSigmaDy) |
| 129 |
#endif |
| 130 |
|
| 131 |
C Limit Stratification |
| 132 |
stratLimit=-Small_Number-gradSmod*rMaxSlope |
| 133 |
dSigmaDr=recip_drC(k)*rkFac*(rhotmp(i,j)-rhoKp1(i,j)) |
| 134 |
C Altern -> if (dSigmaDz.gt.stratLimit) dSigmaDz=stratLimit |
| 135 |
dSigmaDr=min(dSigmaDr,stratLimit) |
| 136 |
N2=(-Gravity*recip_Rhonil)*dSigmaDr |
| 137 |
dRdSigma=1. _d 0 / dSigmaDr |
| 138 |
|
| 139 |
C Iso-neutral slopes |
| 140 |
SlopeX=-dSigmaDx*dRdSigma |
| 141 |
SlopeY=-dSigmaDy*dRdSigma |
| 142 |
if (hFacC(i,j,k,bi,bj).eq.0.) then |
| 143 |
SlopeX=0. |
| 144 |
SlopeY=0. |
| 145 |
endif |
| 146 |
|
| 147 |
C Components of Redi/GM tensor |
| 148 |
K13(i,j,k)=2. _d 0 * SlopeX |
| 149 |
K23(i,j,k)=2. _d 0 * SlopeY |
| 150 |
K33(i,j,k)=SlopeX*SlopeX+SlopeY*SlopeY |
| 151 |
|
| 152 |
C Depth average of M^2/N^2 * N |
| 153 |
KapGM(i,j)=KapGM(i,j)+deltaH* |
| 154 |
& GMalpha*GMlength*GMlength* sqrt(K33(i,j,k)*N2) |
| 155 |
|
| 156 |
C Limit range that KapGM can take |
| 157 |
KapGM(i,j)=min(KapGM(i,j),GMmaxval) |
| 158 |
|
| 159 |
ENDDO |
| 160 |
ENDDO |
| 161 |
|
| 162 |
RETURN |
| 163 |
END |
Parent Directory
| 
Revision Log
| 
Revision Graph