| 1 |
jscott |
1.1 |
|
| 2 |
|
|
#include "ctrparam.h" |
| 3 |
|
|
|
| 4 |
|
|
! ============================================================ |
| 5 |
|
|
! |
| 6 |
|
|
! CHEMEDDY.F: Subroutine for calculating zonal-average eddy |
| 7 |
|
|
! diffusion of MIT Global Chemistry Model |
| 8 |
|
|
! |
| 9 |
|
|
! ------------------------------------------------------------ |
| 10 |
|
|
! |
| 11 |
|
|
! Author: Chien Wang |
| 12 |
|
|
! MIT Joint Program on Science and Policy |
| 13 |
|
|
! of Global Change |
| 14 |
|
|
! |
| 15 |
|
|
! ---------------------------------------------------------- |
| 16 |
|
|
! |
| 17 |
|
|
! Revision History: |
| 18 |
|
|
! |
| 19 |
|
|
! When Who What |
| 20 |
|
|
! ---- ---------- ------- |
| 21 |
|
|
! 013096 Chien Wang rev. |
| 22 |
|
|
! 080100 Chien Wang repack based on CliChem3 & add cpp |
| 23 |
|
|
! 051804 Chien Wang rev. for 46x11 |
| 24 |
|
|
! |
| 25 |
|
|
! ========================================================== |
| 26 |
|
|
|
| 27 |
|
|
Subroutine chemeddy(ifdiff,x00,x11,dta) |
| 28 |
|
|
|
| 29 |
|
|
#include "chem_para" |
| 30 |
|
|
#include "chem_com" |
| 31 |
|
|
#include "BD2G04.COM" |
| 32 |
|
|
|
| 33 |
|
|
dimension x00 (nlon,nlat,nlev) |
| 34 |
|
|
dimension x11 (nlon,nlat,nlev) |
| 35 |
|
|
|
| 36 |
|
|
dimension vc (nlat,nlev) |
| 37 |
|
|
dimension beta5(nlat) |
| 38 |
|
|
|
| 39 |
|
|
dimension dcdy(nlat,nlev) |
| 40 |
|
|
dimension dcdz(nlat,nlev) |
| 41 |
|
|
dimension dcdc(nlat,nlev) |
| 42 |
|
|
|
| 43 |
|
|
! ---------------------------------------------------------- |
| 44 |
|
|
|
| 45 |
|
|
#if ( defined CPL_CHEM ) |
| 46 |
|
|
|
| 47 |
|
|
c------------------------------------------------------- |
| 48 |
|
|
c Definitions of parameters: |
| 49 |
|
|
c |
| 50 |
|
|
|
| 51 |
|
|
istart=1 |
| 52 |
|
|
iend =nlon |
| 53 |
|
|
|
| 54 |
|
|
beta5(1)=0.0 |
| 55 |
|
|
do j=2,nlat1 |
| 56 |
|
|
beta5(j)=0.573*sqrt(beta2(j))/(1.-0.427*beta2(j)**0.302) |
| 57 |
|
|
end do |
| 58 |
|
|
beta5(nlat)=0.0 |
| 59 |
|
|
|
| 60 |
|
|
c===== |
| 61 |
|
|
c Calculate dcdy and dcdz: |
| 62 |
|
|
c |
| 63 |
|
|
do 5 i=istart,iend |
| 64 |
|
|
do 5 j=2,nlat |
| 65 |
|
|
do 5 k=1,nlev |
| 66 |
|
|
dcdy(j,k)=(x11(i,j,k)-x11(i,j-1,k))/dyv(j) |
| 67 |
|
|
5 continue |
| 68 |
|
|
|
| 69 |
|
|
do 6 i=istart,iend |
| 70 |
|
|
do 6 j=1,nlat |
| 71 |
|
|
do 6 k=1,nlev1 |
| 72 |
|
|
dcdz(j,k)=-(x11(i,j,k+1)-x11(i,j,k))*deltap(j,k) |
| 73 |
|
|
6 continue |
| 74 |
|
|
|
| 75 |
|
|
do 61 i=istart,iend |
| 76 |
|
|
do 61 j=1,nlat |
| 77 |
|
|
do 62 k=2,nlev |
| 78 |
|
|
dcdz(j,k)=dcdz(j,k-1)*dp2dz(j,k-1) |
| 79 |
|
|
62 continue |
| 80 |
|
|
dcdz(j,1)=dcdz(j,2) |
| 81 |
|
|
61 continue |
| 82 |
|
|
|
| 83 |
|
|
do 7 j=2,nlat1 |
| 84 |
|
|
do 7 k=2,nlev1 |
| 85 |
|
|
dcdc(j,k)=dcdz(j,k)*4.0 |
| 86 |
|
|
& /(dcdy(j,k) +dcdy(j+1,k) |
| 87 |
|
|
& +dcdy(j,k+1)+dcdy(j+1,k+1)+1.e-20) |
| 88 |
|
|
7 continue |
| 89 |
|
|
|
| 90 |
|
|
do 8 j=2,nlat1 |
| 91 |
|
|
alamor =beta5(j)/beta2(j) |
| 92 |
|
|
alamor2 =alamor /beta2(j) |
| 93 |
|
|
oneoalam1=1./(1.+beta5(j)) |
| 94 |
|
|
do 8 k=2,nlev1 |
| 95 |
|
|
dcdc(j,k)=oneoalam1*beta1(j)*beta3(j,k) |
| 96 |
|
|
& *(1.0+alamor |
| 97 |
|
|
& +beta3(j,k)*0.25*beta1(j)*dcdc(j,k) |
| 98 |
|
|
& *(1.+alamor2)) |
| 99 |
|
|
8 continue |
| 100 |
|
|
|
| 101 |
|
|
c===== |
| 102 |
|
|
c Calculate meridional eddy diffusion: |
| 103 |
|
|
c |
| 104 |
|
|
do 10 k=1,nlev |
| 105 |
|
|
paver = 0.5*(p00(1,1)+p00(1,2)) |
| 106 |
|
|
fluxl =-fkt(2,k) |
| 107 |
|
|
& /dyv(2)*dcdy(2,k)*dta |
| 108 |
|
|
& * paver |
| 109 |
|
|
fluxl=max(-0.5*x00(1,2,k),min(0.5*x00(1,1, k),fluxl)) |
| 110 |
|
|
vc(2,k)=fluxl/(paver+1.e-20) |
| 111 |
|
|
do 11 j=2,nlat1 |
| 112 |
|
|
paver = 0.5*(p00(1,j)+p00(1,j+1)) |
| 113 |
|
|
fluxr =-fkt(j+1,k) |
| 114 |
|
|
& /dyv(j+1)*dcdy(j+1,k)*dta |
| 115 |
|
|
& * paver |
| 116 |
|
|
fluxr=max(-0.5*x00(1,j+1,k),min(0.5*x00(1,j,k),fluxr)) |
| 117 |
|
|
vc (j+1,k)=fluxr/(paver+1.e-20) |
| 118 |
|
|
x00(1,j,k)=x00(1,j,k)-(fluxr-fluxl) |
| 119 |
|
|
fluxl=fluxr |
| 120 |
|
|
11 continue |
| 121 |
|
|
10 continue |
| 122 |
|
|
|
| 123 |
|
|
c===== |
| 124 |
|
|
c Calculate vertical eddy diffusion: |
| 125 |
|
|
c |
| 126 |
|
|
c 112696 changed also in eddypa.f for beta4 |
| 127 |
|
|
c |
| 128 |
|
|
do 12 j=2,nlat1 |
| 129 |
|
|
fluxb=0.0 |
| 130 |
|
|
do 14 k=1,n_tropopause ! ktrop = 7 for both 9 and 11 layer model |
| 131 |
|
|
fluxt=0.25*(vc(j,k)+vc(j,k+1)+vc(j+1,k)+vc(j+1,k+1)) |
| 132 |
|
|
& *dcdc(j,k+1) |
| 133 |
|
|
& *beta4(j,k+1) |
| 134 |
|
|
& *p00(1,j) |
| 135 |
|
|
|
| 136 |
|
|
c fluxt=max(-0.5*x00(1,j,k) *dsig(k), |
| 137 |
|
|
c & min( 0.5*x00(1,j,k+1)*dsig(k+1),fluxt)) |
| 138 |
|
|
c if(fluxt*dcdz(j,k+1).lt.0.0) fluxt=0.0 |
| 139 |
|
|
c x00(1,j,k)=x00(1,j,k)+(fluxt-fluxb)/dsig(k) |
| 140 |
|
|
|
| 141 |
|
|
fluxt=max(-0.5*x00(1,j,k+1)*dsig(k+1), |
| 142 |
|
|
& min( 0.5*x00(1,j,k) *dsig(k),fluxt)) |
| 143 |
|
|
if(fluxt*dcdz(j,k+1).gt.0.0) fluxt=0.0 |
| 144 |
|
|
x00(1,j,k)=x00(1,j,k)-(fluxt-fluxb)/dsig(k) |
| 145 |
|
|
fluxb=fluxt |
| 146 |
|
|
14 continue |
| 147 |
|
|
12 continue |
| 148 |
|
|
|
| 149 |
|
|
c write(6,*)"FKT = " |
| 150 |
|
|
c write(6,*)fkt |
| 151 |
|
|
c write(6,*)"VC = " |
| 152 |
|
|
c write(6,*)vc |
| 153 |
|
|
c write(6,*)"DCDY = " |
| 154 |
|
|
c write(6,*)dcdy |
| 155 |
|
|
c |
| 156 |
|
|
c 040895 test: |
| 157 |
|
|
c |
| 158 |
|
|
|
| 159 |
|
|
1996 continue |
| 160 |
|
|
|
| 161 |
|
|
c ====== |
| 162 |
|
|
c 013096 |
| 163 |
|
|
c Apply horizontal diffussion to some tracers |
| 164 |
|
|
c to reduce initialization errors in the |
| 165 |
|
|
c global distribution: |
| 166 |
|
|
c |
| 167 |
|
|
if(ifdiff.ne.0) call chemdiff(ifdiff,x00,x11,dta) |
| 168 |
|
|
|
| 169 |
|
|
call chemcheck(x00) |
| 170 |
|
|
|
| 171 |
|
|
#endif |
| 172 |
|
|
|
| 173 |
|
|
return |
| 174 |
|
|
end |
| 175 |
|
|
|
Parent Directory
| 
Revision Log
| 
Revision Graph