source: LMDZ6/branches/Amaury_dev/libf/phylmd/tend_to_tke.F90 @ 5220

!***************************************************************************************
! tend_to_tke.F90
!*************

! Subroutine that adds a tendency on the TKE created by the 
! fluxes of momentum retrieved from the wind speed tendencies
! of the physics.

! The basic concept is the following:
! the TKE equation writes  de/dt = -u'w' du/dz -v'w' dv/dz +g/theta dtheta/dz +......


! We expect contributions to the term u'w' and v'w' that do not come from the Yamada
! scheme, for instance: gravity waves, drag from high vegetation..... These contributions
! need to be accounted for.
! we explicitely calculate the fluxes, integrating the wind speed 
!                        tendency from the top of the atmospher



! contacts: Frederic Hourdin, Etienne Vignon

! History:
!---------
! - 1st redaction, Etienne, 15/10/2016
! Ajout des 4 sous surfaces pour la tke
! on sort l'ajout des tendances du if sur les deux cas, pour ne pas
! dupliuqer les lignes
! on enleve le pas de temps qui disprait dans les calculs


!**************************************************************************************

SUBROUTINE tend_to_tke(dt, plev, exner, temp, windu, windv, dt_a, du_a, dv_a, pctsrf, tke)

  USE dimphy, ONLY: klon, klev
  USE indice_sol_mod, ONLY: nbsrf
  USE lmdz_yomcst

  IMPLICIT NONE

  ! Declarations
  !==============


  ! Inputs
  !-------
  REAL dt                   ! Time step [s]
  REAL plev(klon, klev + 1)    ! inter-layer pressure [Pa]
  REAL temp(klon, klev)      ! temperature [K], grid-cell average or for a one subsurface
  REAL windu(klon, klev)     ! zonal wind [m/s], grid-cell average or for a one subsurface
  REAL windv(klon, klev)     ! meridonal wind [m/s], grid-cell average or for a one subsurface
  REAL exner(klon, klev)     ! Fonction d'Exner = T/theta
  REAL dt_a(klon, klev)      ! Temperature tendency [K], grid-cell average or for a one subsurface
  REAL du_a(klon, klev)      ! Zonal wind speed tendency [m/s], grid-cell average or for a one subsurface
  REAL dv_a(klon, klev)      ! Meridional wind speed tendency [m/s], grid-cell average or for a one subsurface
  REAL pctsrf(klon, nbsrf + 1)       ! Turbulent Kinetic energy [m2/s2], grid-cell average or for a subsurface

  ! Inputs/Outputs
  !---------------
  REAL tke(klon, klev + 1, nbsrf + 1)       ! Turbulent Kinetic energy [m2/s2], grid-cell average or for a subsurface


  ! Local
  !-------

  INTEGER i, k, isrf                 ! indices
  REAL    masse(klon, klev)          ! mass in the layers [kg/m2]
  REAL    unsmasse(klon, klev + 1)     ! linear mass in the layers [kg/m2]
  REAL    flux_rhotw(klon, klev + 1)   ! flux massique de tempe. pot. rho*u'*theta'
  REAL    flux_rhouw(klon, klev + 1)   ! flux massique de quantit?? de mouvement rho*u'*w' [kg/m/s2]
  REAL    flux_rhovw(klon, klev + 1)   ! flux massique de quantit?? de mouvement rho*v'*w' [kg/m/s2]
  REAL    tendt(klon, klev)        ! new temperature tke tendency [m2/s2/s]
  REAL    tendu(klon, klev)        ! new zonal tke tendency [m2/s2/s]
  REAL    tendv(klon, klev)        ! new meridonal tke tendency [m2/s2/s]




  ! First calculations:
  !=====================

  unsmasse(:, :) = 0.
  DO k = 1, klev
    masse(:, k) = (plev(:, k) - plev(:, k + 1)) / RG
    unsmasse(:, k) = unsmasse(:, k) + 0.5 / masse(:, k)
    unsmasse(:, k + 1) = unsmasse(:, k + 1) + 0.5 / masse(:, k)
  END DO

  tendu(:, :) = 0.0
  tendv(:, :) = 0.0

  ! Method 1: Calculation of fluxes using a downward integration
  !============================================================



  ! Flux calculation

  flux_rhotw(:, klev + 1) = 0.
  flux_rhouw(:, klev + 1) = 0.
  flux_rhovw(:, klev + 1) = 0.

  DO k = klev, 1, -1
    flux_rhotw(:, k) = flux_rhotw(:, k + 1) + masse(:, k) * dt_a(:, k) / exner(:, k)
    flux_rhouw(:, k) = flux_rhouw(:, k + 1) + masse(:, k) * du_a(:, k)
    flux_rhovw(:, k) = flux_rhovw(:, k + 1) + masse(:, k) * dv_a(:, k)
  ENDDO


  ! TKE update:

  DO k = 2, klev
    tendt(:, k) = -flux_rhotw(:, k) * (exner(:, k) - exner(:, k - 1)) * unsmasse(:, k) * RCPD
    tendu(:, k) = -flux_rhouw(:, k) * (windu(:, k) - windu(:, k - 1)) * unsmasse(:, k)
    tendv(:, k) = -flux_rhovw(:, k) * (windv(:, k) - windv(:, k - 1)) * unsmasse(:, k)
  ENDDO
  tendt(:, 1) = -flux_rhotw(:, 1) * (exner(:, 1) - 1.) * unsmasse(:, 1) * RCPD
  tendu(:, 1) = -1. * flux_rhouw(:, 1) * windu(:, 1) * unsmasse(:, 1)
  tendv(:, 1) = -1. * flux_rhovw(:, 1) * windv(:, 1) * unsmasse(:, 1)

  DO isrf = 1, nbsrf
    DO k = 1, klev
      DO i = 1, klon
        IF (pctsrf(i, isrf)>0.) THEN
          tke(i, k, isrf) = tke(i, k, isrf) + tendu(i, k) + tendv(i, k) + tendt(i, k)
          tke(i, k, isrf) = max(tke(i, k, isrf), 1.e-10)
        ENDIF
      ENDDO
    ENDDO
  ENDDO


  !  IF (klon==1) THEN
  !  CALL iophys_ecrit('u',klev,'u','',windu)
  !  CALL iophys_ecrit('v',klev,'v','',windu)
  !  CALL iophys_ecrit('t',klev,'t','',temp)
  !  CALL iophys_ecrit('tke1',klev,'tke1','',tke(:,1:klev,1))
  !  CALL iophys_ecrit('tke2',klev,'tke2','',tke(:,1:klev,2))
  !  CALL iophys_ecrit('tke3',klev,'tke3','',tke(:,1:klev,3))
  !  CALL iophys_ecrit('tke4',klev,'tke4','',tke(:,1:klev,4))
  !  CALL iophys_ecrit('theta',klev,'theta','',temp/exner)
  !  CALL iophys_ecrit('Duv',klev,'Duv','',tendu(:,1:klev)+tendv(:,1:klev))
  !  CALL iophys_ecrit('Dt',klev,'Dt','',tendt(:,1:klev))
  !  ENDIF

END SUBROUTINE tend_to_tke