[1] | 1 | SUBROUTINE advec_w_ups |
---|
| 2 | |
---|
| 3 | !------------------------------------------------------------------------------! |
---|
| 4 | ! Actual revisions: |
---|
| 5 | ! ----------------- |
---|
| 6 | ! |
---|
| 7 | ! |
---|
| 8 | ! Former revisions: |
---|
| 9 | ! ----------------- |
---|
| 10 | ! $Log: advec_w_ups.f90,v $ |
---|
| 11 | ! Revision 1.6 2004/04/30 08:05:05 raasch |
---|
| 12 | ! Enlarged transposition arrays introduced |
---|
| 13 | ! |
---|
| 14 | ! Revision 1.5 2003/03/16 09:26:46 raasch |
---|
| 15 | ! Two underscores (_) are placed in front of all define-strings |
---|
| 16 | ! |
---|
| 17 | ! Revision 1.4 2001/03/29 17:37:53 raasch |
---|
| 18 | ! v_ad is now allocated locally, several temporary arrays removed from |
---|
| 19 | ! argument lists of spline_* and long_filter_* |
---|
| 20 | ! Translation of remaining German identifiers (variables, subroutines, etc.) |
---|
| 21 | ! |
---|
| 22 | ! Revision 1.3 2001/01/22 05:23:31 raasch |
---|
| 23 | ! Module test_variables removed |
---|
| 24 | ! |
---|
| 25 | ! Revision 1.2 2000/01/20 09:57:15 letzel |
---|
| 26 | ! All comments translated into English |
---|
| 27 | ! |
---|
| 28 | ! Revision 1.1 1999/02/05 08:52:09 raasch |
---|
| 29 | ! Initial revision |
---|
| 30 | ! |
---|
| 31 | ! |
---|
| 32 | ! Description: |
---|
| 33 | ! ------------ |
---|
| 34 | ! Upstream-Spline advection of the w velocity-component. The advection process |
---|
| 35 | ! is divided into three subsequent steps, one for each of the dimensions. The |
---|
| 36 | ! result is stored as a tendency in array tend. The computation of the cubic |
---|
| 37 | ! splines and the possible execution of the Long-filter require that all grid |
---|
| 38 | ! points of the relevant dimension are available. For model runs on more than |
---|
| 39 | ! one PE therefore both the advected and the advecting quantities are |
---|
| 40 | ! transposed accordingly. |
---|
| 41 | ! |
---|
| 42 | ! Internally used arrays: |
---|
| 43 | ! v_ad = scalar quantity to be advected, initialised = w at the beginning, also |
---|
| 44 | ! being used as temporary storage after each time step |
---|
| 45 | ! d = advecting component (u, v, or w) |
---|
| 46 | !------------------------------------------------------------------------------! |
---|
| 47 | |
---|
| 48 | USE advection |
---|
| 49 | USE arrays_3d |
---|
| 50 | USE cpulog |
---|
| 51 | USE grid_variables |
---|
| 52 | USE indices |
---|
| 53 | USE interfaces |
---|
| 54 | USE control_parameters |
---|
| 55 | |
---|
| 56 | IMPLICIT NONE |
---|
| 57 | |
---|
| 58 | INTEGER :: i,j,k |
---|
| 59 | REAL, DIMENSION(:,:,:), ALLOCATABLE :: v_ad |
---|
| 60 | |
---|
| 61 | |
---|
| 62 | CALL cpu_log( log_point_s(19), 'advec_w_ups', 'start' ) |
---|
| 63 | |
---|
| 64 | #if defined( __parallel ) |
---|
| 65 | |
---|
| 66 | ! |
---|
| 67 | !-- Advection of w in x-direction: |
---|
| 68 | !-- Store w in temporary array v_ad (component to be advected, boundaries |
---|
| 69 | !-- are not used because they disturb the transposition) |
---|
| 70 | ALLOCATE( v_ad(nzb+1:nzta,nys:nyna,nxl:nxra) ) |
---|
| 71 | v_ad = 0.0 |
---|
| 72 | v_ad(nzb+1:nzt,nys:nyn,nxl:nxr) = w(nzb+1:nzt,nys:nyn,nxl:nxr) |
---|
| 73 | |
---|
| 74 | ! |
---|
| 75 | !-- Enlarge the size of tend, used as a working array for the transpositions |
---|
| 76 | IF ( nxra > nxr .OR. nyna > nyn .OR. nza > nz ) THEN |
---|
| 77 | DEALLOCATE( tend ) |
---|
| 78 | ALLOCATE( tend(1:nza,nys:nyna,nxl:nxra) ) |
---|
| 79 | ENDIF |
---|
| 80 | |
---|
| 81 | ! |
---|
| 82 | !-- Transpose the component to be advected: z --> x |
---|
| 83 | CALL transpose_zx( v_ad, tend, v_ad, tend, v_ad ) |
---|
| 84 | |
---|
| 85 | #else |
---|
| 86 | |
---|
| 87 | ! |
---|
| 88 | !-- Advection of w in x-direction: |
---|
| 89 | !-- Store w in temporary array v_ad (component to be advected) |
---|
| 90 | ALLOCATE( v_ad(nzb:nzt+1,nys-1:nyn+1,nxl-1:nxr+1) ) |
---|
| 91 | v_ad(:,:,:) = w(:,:,:) |
---|
| 92 | |
---|
| 93 | #endif |
---|
| 94 | |
---|
| 95 | ! |
---|
| 96 | !-- Advecting component (u) must be averaged out on the w grid |
---|
| 97 | d = 0.0 |
---|
| 98 | DO i = nxl, nxr |
---|
| 99 | DO j = nys, nyn |
---|
| 100 | DO k = nzb+1, nzt |
---|
| 101 | d(k,j,i) = 0.25 * ( u(k,j,i) + u(k,j,i+1) + & |
---|
| 102 | u(k+1,j,i+1) + u(k+1,j,i) ) - u_gtrans |
---|
| 103 | ENDDO |
---|
| 104 | ENDDO |
---|
| 105 | ENDDO |
---|
| 106 | |
---|
| 107 | #if defined( __parallel ) |
---|
| 108 | |
---|
| 109 | ! |
---|
| 110 | !-- Transpose the component to be advected: z --> x |
---|
| 111 | CALL transpose_zx( d, tend, d, tend, d ) |
---|
| 112 | |
---|
| 113 | #endif |
---|
| 114 | |
---|
| 115 | ! |
---|
| 116 | !-- Upstream-Spline advection of w in x-direction. Array tend comes out |
---|
| 117 | !-- as v_ad before the advection step including cyclic boundaries. |
---|
| 118 | !-- It is needed for the long filter. |
---|
| 119 | CALL spline_x( v_ad, d, 'w' ) |
---|
| 120 | |
---|
| 121 | ! |
---|
| 122 | !-- Advection of w in y-direction: |
---|
| 123 | !-- advecting component (v) must be averaged out on the w grid |
---|
| 124 | DO i = nxl, nxr |
---|
| 125 | DO j = nys, nyn |
---|
| 126 | DO k = nzb+1, nzt |
---|
| 127 | d(k,j,i) = 0.25 * ( v(k,j,i) + v(k,j+1,i) + & |
---|
| 128 | v(k+1,j+1,i) + v(k+1,j,i) ) - v_gtrans |
---|
| 129 | ENDDO |
---|
| 130 | ENDDO |
---|
| 131 | ENDDO |
---|
| 132 | |
---|
| 133 | #if defined( __parallel ) |
---|
| 134 | |
---|
| 135 | ! |
---|
| 136 | !-- Transpose the advecting component: z --> y |
---|
| 137 | CALL transpose_zx( d, tend, d, tend, d ) |
---|
| 138 | CALL transpose_xy( d, tend, d, tend, d ) |
---|
| 139 | |
---|
| 140 | ! |
---|
| 141 | !-- Transpose the component to be advected: x --> y |
---|
| 142 | CALL transpose_xy( v_ad, tend, v_ad, tend, v_ad ) |
---|
| 143 | |
---|
| 144 | #endif |
---|
| 145 | |
---|
| 146 | ! |
---|
| 147 | !-- Upstream-Spline advection of w in y-direction |
---|
| 148 | CALL spline_y( v_ad, d, 'w' ) |
---|
| 149 | |
---|
| 150 | ! |
---|
| 151 | !-- Advection of w in z-direction: |
---|
| 152 | !-- advecting component (d) = component to be advected (v_ad) |
---|
| 153 | d(nzb+1:nzt,nys:nyn,nxl:nxr) = w(nzb+1:nzt,nys:nyn,nxl:nxr) |
---|
| 154 | |
---|
| 155 | #if defined( __parallel ) |
---|
| 156 | |
---|
| 157 | ! |
---|
| 158 | !-- Transpose the component to be advected: y --> z (= y --> x + x --> z) |
---|
| 159 | CALL transpose_yx( v_ad, tend, v_ad, tend, v_ad ) |
---|
| 160 | CALL transpose_xz( v_ad, tend, v_ad, tend, v_ad ) |
---|
| 161 | |
---|
| 162 | ! |
---|
| 163 | !-- Resize tend to its normal size |
---|
| 164 | IF ( nxra > nxr .OR. nyna > nyn .OR. nza > nz ) THEN |
---|
| 165 | DEALLOCATE( tend ) |
---|
| 166 | ALLOCATE( tend(nzb:nzt+1,nys-1:nyn+1,nxl-1:nxr+1) ) |
---|
| 167 | ENDIF |
---|
| 168 | |
---|
| 169 | #endif |
---|
| 170 | |
---|
| 171 | ! |
---|
| 172 | !-- Upstream-Spline advection of w in z-direction |
---|
| 173 | CALL spline_z( v_ad, d, dzw, spl_tri_zw, 'w' ) |
---|
| 174 | |
---|
| 175 | ! |
---|
| 176 | !-- Compute the tendency term |
---|
| 177 | DO i = nxl, nxr |
---|
| 178 | DO j = nys, nyn |
---|
| 179 | DO k = nzb+1, nzt |
---|
| 180 | tend(k,j,i) = ( v_ad(k,j,i) - w(k,j,i) ) / dt_3d |
---|
| 181 | ENDDO |
---|
| 182 | ENDDO |
---|
| 183 | ENDDO |
---|
| 184 | |
---|
| 185 | DEALLOCATE( v_ad ) |
---|
| 186 | |
---|
| 187 | CALL cpu_log( log_point_s(19), 'advec_w_ups', 'stop' ) |
---|
| 188 | |
---|
| 189 | END SUBROUTINE advec_w_ups |
---|