Changeset 14072 for NEMO/trunk/src/OCE/TRA/traadv_ubs.F90
- Timestamp:
- 2020-12-04T08:48:38+01:00 (3 years ago)
- File:
-
- 1 edited
-
NEMO/trunk/src/OCE/TRA/traadv_ubs.F90 (modified) (14 diffs)
Legend:
- Unmodified
- Added
- Removed
-
NEMO/trunk/src/OCE/TRA/traadv_ubs.F90
r13982 r14072 10 10 !!---------------------------------------------------------------------- 11 11 !! tra_adv_ubs : update the tracer trend with the horizontal 12 !! advection trends using a third order biaised scheme 12 !! advection trends using a third order biaised scheme 13 13 !!---------------------------------------------------------------------- 14 14 USE oce ! ocean dynamics and active tracers … … 16 16 USE trc_oce ! share passive tracers/Ocean variables 17 17 USE trd_oce ! trends: ocean variables 18 USE traadv_fct ! acces to routine interp_4th_cpt 19 USE trdtra ! trends manager: tracers 18 USE traadv_fct ! acces to routine interp_4th_cpt 19 USE trdtra ! trends manager: tracers 20 20 USE diaptr ! poleward transport diagnostics 21 21 USE diaar5 ! AR5 diagnostics … … 25 25 USE lib_mpp ! massively parallel library 26 26 USE lbclnk ! ocean lateral boundary condition (or mpp link) 27 USE lib_fortran ! Fortran utilities (allows no signed zero when 'key_nosignedzero' defined) 27 USE lib_fortran ! Fortran utilities (allows no signed zero when 'key_nosignedzero' defined) 28 28 29 29 IMPLICIT NONE … … 51 51 !!---------------------------------------------------------------------- 52 52 !! *** ROUTINE tra_adv_ubs *** 53 !! 53 !! 54 54 !! ** Purpose : Compute the now trend due to the advection of tracers 55 55 !! and add it to the general trend of passive tracer equations. … … 60 60 !! For example the i-component of the advective fluxes are given by : 61 61 !! ! e2u e3u un ( mi(Tn) - zltu(i ) ) if un(i) >= 0 62 !! ztu = ! or 62 !! ztu = ! or 63 63 !! ! e2u e3u un ( mi(Tn) - zltu(i+1) ) if un(i) < 0 64 64 !! where zltu is the second derivative of the before temperature field: 65 65 !! zltu = 1/e3t di[ e2u e3u / e1u di[Tb] ] 66 !! This results in a dissipatively dominant (i.e. hyper-diffusive) 67 !! truncation error. The overall performance of the advection scheme 68 !! is similar to that reported in (Farrow and Stevens, 1995). 66 !! This results in a dissipatively dominant (i.e. hyper-diffusive) 67 !! truncation error. The overall performance of the advection scheme 68 !! is similar to that reported in (Farrow and Stevens, 1995). 69 69 !! For stability reasons, the first term of the fluxes which corresponds 70 !! to a second order centered scheme is evaluated using the now velocity 71 !! (centered in time) while the second term which is the diffusive part 72 !! of the scheme, is evaluated using the before velocity (forward in time). 70 !! to a second order centered scheme is evaluated using the now velocity 71 !! (centered in time) while the second term which is the diffusive part 72 !! of the scheme, is evaluated using the before velocity (forward in time). 73 73 !! Note that UBS is not positive. Do not use it on passive tracers. 74 74 !! On the vertical, the advection is evaluated using a FCT scheme, 75 !! as the UBS have been found to be too diffusive. 76 !! kn_ubs_v argument controles whether the FCT is based on 77 !! a 2nd order centrered scheme (kn_ubs_v=2) or on a 4th order compact 75 !! as the UBS have been found to be too diffusive. 76 !! kn_ubs_v argument controles whether the FCT is based on 77 !! a 2nd order centrered scheme (kn_ubs_v=2) or on a 4th order compact 78 78 !! scheme (kn_ubs_v=4). 79 79 !! … … 82 82 !! - poleward advective heat and salt transport (ln_diaptr=T) 83 83 !! 84 !! Reference : Shchepetkin, A. F., J. C. McWilliams, 2005, Ocean Modelling, 9, 347-404. 84 !! Reference : Shchepetkin, A. F., J. C. McWilliams, 2005, Ocean Modelling, 9, 347-404. 85 85 !! Farrow, D.E., Stevens, D.P., 1995, J. Phys. Ocean. 25, 1731�1741. 86 86 !!---------------------------------------------------------------------- … … 125 125 DO jn = 1, kjpt ! tracer loop 126 126 ! ! =========== 127 ! 127 ! 128 128 DO jk = 1, jpkm1 !== horizontal laplacian of before tracer ==! 129 129 DO_2D( nn_hls, nn_hls-1, nn_hls, nn_hls-1 ) ! First derivative (masked gradient) … … 138 138 zltv(ji,jj,jk) = ( ztv(ji,jj,jk) - ztv(ji,jj-1,jk) ) * zcoef 139 139 END_2D 140 ! 141 END DO 140 ! 141 END DO 142 142 IF (nn_hls.EQ.1) CALL lbc_lnk_multi( 'traadv_ubs', zltu, 'T', 1.0_wp, zltv, 'T', 1.0_wp ) ! Lateral boundary cond. (unchanged sgn) 143 ! 143 ! 144 144 DO_3D( 1, 0, 1, 0, 1, jpkm1 ) !== Horizontal advective fluxes ==! (UBS) 145 145 zfp_ui = pU(ji,jj,jk) + ABS( pU(ji,jj,jk) ) ! upstream transport (x2) … … 166 166 & * r1_e1e2t(ji,jj) / e3t(ji,jj,jk,Kmm) 167 167 END_2D 168 ! 168 ! 169 169 END DO 170 170 ! … … 177 177 CALL trd_tra( kt, Kmm, Krhs, cdtype, jn, jptra_yad, ztv, pV, pt(:,:,:,jn,Kmm) ) 178 178 END IF 179 ! 179 ! 180 180 ! ! "Poleward" heat and salt transports (contribution of upstream fluxes) 181 181 IF( l_ptr ) CALL dia_ptr_hst( jn, 'adv', ztv(:,:,:) ) … … 188 188 SELECT CASE( kn_ubs_v ) ! select the vertical advection scheme 189 189 ! 190 CASE( 2 ) ! 2nd order FCT 191 ! 190 CASE( 2 ) ! 2nd order FCT 191 ! 192 192 IF( l_trd ) THEN 193 193 DO_3D( 0, 0, 0, 0, 1, jpkm1 ) … … 205 205 IF( ln_isfcav ) THEN ! top of the ice-shelf cavities and at the ocean surface 206 206 DO_2D( 1, 1, 1, 1 ) 207 ztw(ji,jj, mikt(ji,jj) ) = pW(ji,jj,mikt(ji,jj)) * pt(ji,jj,mikt(ji,jj),jn,Kbb) ! linear free surface 207 ztw(ji,jj, mikt(ji,jj) ) = pW(ji,jj,mikt(ji,jj)) * pt(ji,jj,mikt(ji,jj),jn,Kbb) ! linear free surface 208 208 END_2D 209 209 ELSE ! no cavities: only at the ocean surface … … 217 217 ztak = - ( ztw(ji,jj,jk) - ztw(ji,jj,jk+1) ) & 218 218 & * r1_e1e2t(ji,jj) / e3t(ji,jj,jk,Kmm) 219 pt(ji,jj,jk,jn,Krhs) = pt(ji,jj,jk,jn,Krhs) + ztak 219 pt(ji,jj,jk,jn,Krhs) = pt(ji,jj,jk,jn,Krhs) + ztak 220 220 zti(ji,jj,jk) = ( pt(ji,jj,jk,jn,Kbb) + p2dt * ( ztak + zltu(ji,jj,jk) ) ) * tmask(ji,jj,jk) 221 221 END_3D … … 266 266 !!--------------------------------------------------------------------- 267 267 !! *** ROUTINE nonosc_z *** 268 !! 269 !! ** Purpose : compute monotonic tracer fluxes from the upstream 270 !! scheme and the before field by a nonoscillatory algorithm 268 !! 269 !! ** Purpose : compute monotonic tracer fluxes from the upstream 270 !! scheme and the before field by a nonoscillatory algorithm 271 271 !! 272 272 !! ** Method : ... ???
Note: See TracChangeset
for help on using the changeset viewer.
