vertical_diffusion.f90 Source File

!> Parametrization of vertical diffusion
!
! Three different processes are modelled by the vertical diffusion scheme:
! 1) Shallow convection - A redistribution of moisture and dry static energy
!    between the lowest two layers where there is conditional instability
! 2) A slow diffusion of moisture in stable conditions
! 3) A fast redistribution of dry static energy where the lapse rate is close to
!    the dry adiabatic limit
module vertical_diffusion
    use types, only : p
    use params

    implicit none

    private
    public get_vertical_diffusion_tend

    ! Constants for vertical diffusion and shallow convection.
    real(p), parameter :: trshc = 6.0  !! Relaxation time (in hours) for shallow convection
    real(p), parameter :: trvdi = 24.0 !! Relaxation time (in hours) for moisture diffusion
    real(p), parameter :: trvds = 6.0  !! Relaxation time (in hours) for super-adiabatic conditions
    real(p), parameter :: redshc = 0.5  !! Reduction factor of shallow convection in areas of deep
    !! convection
    real(p), parameter :: rhgrad = 0.5  !! Maximum gradient of relative humidity (d_RH/d_sigma)
    real(p), parameter :: segrad = 0.1  !! Minimum gradient of dry static energy (d_DSE/d_phi)

contains
    !> Compute tendencies of momentum, energy and moisture due to vertical diffusion
    !  and shallow convection
    subroutine get_vertical_diffusion_tend(se, rh, qa, qsat, phi, icnv, utenvd, vtenvd, &
            ttenvd, qtenvd, fsg, dhs, sigh)
        use physical_constants, only : cp, alhc

        real(p), intent(in) :: se(ix, il, kx)     !! Dry static energy
        real(p), intent(in) :: rh(ix, il, kx)     !! Relative humidity
        real(p), intent(in) :: qa(ix, il, kx)     !! Specific humidity [g/kg]
        real(p), intent(in) :: qsat(ix, il, kx)   !! Saturated specific humidity [g/kg]
        real(p), intent(in) :: phi(ix, il, kx)    !! Geopotential
        integer, intent(in) :: icnv(ix, il)      !! Sigma-level index of deep convection
        real(p), intent(out) :: utenvd(ix, il, kx) !! u-wind tendency
        real(p), intent(out) :: vtenvd(ix, il, kx) !! v-wind tendency
        real(p), intent(out) :: ttenvd(ix, il, kx) !! Temperature tendency
        real(p), intent(out) :: qtenvd(ix, il, kx) !! Specific humidity tendency

        real(p), intent(in) :: dhs(kx)   !! Sigma level thicknesses
        real(p), intent(in) :: fsg(kx)   !! Full sigma levels
        real(p), intent(in) :: sigh(0:kx) !! Half-level sigma

        integer :: nl1, i, j, k, k1
        real(p) :: cshc, cvdi, fshcq, fshcse, fvdiq, fvdise, drh0, fvdiq2, dmse, drh
        real(p) :: fluxse, fluxq, fcnv, se0
        real(p), dimension(kx) :: rsig, rsig1

        ! 1. Initalization

        ! N.B. In this routine, fluxes of dry static energy and humidity
        !      are scaled in such a way that:
        !      d_T/dt = d_F'(SE)/d_sigma,  d_Q/dt = d_F'(Q)/d_sigma

        nl1 = kx - 1
        cshc = dhs(kx) / 3600.0
        cvdi = (sigh(nl1) - sigh(1)) / ((nl1 - 1) * 3600.0)

        fshcq = cshc / trshc
        fshcse = cshc / (trshc * cp)

        fvdiq = cvdi / trvdi
        fvdise = cvdi / (trvds * cp)

        do k = 1, nl1
            rsig(k) = 1.0 / dhs(k)
            rsig1(k) = 1.0 / (1.0 - sigh(k))
        end do
        rsig(kx) = 1.0 / dhs(kx)

        utenvd = 0.0
        vtenvd = 0.0
        ttenvd = 0.0
        qtenvd = 0.0

        ! 2. Shallow convection
        drh0 = rhgrad * (fsg(kx) - fsg(nl1))
        fvdiq2 = fvdiq * sigh(nl1)

        do i = 1, ix
            do j = 1, il
                dmse = se(i, j, kx) - se(i, j, nl1) + alhc * (qa(i, j, kx) - qsat(i, j, nl1))
                drh = rh(i, j, kx) - rh(i, j, nl1)
                fcnv = 1.0

                if (dmse >= 0.0) then
                    if (icnv(i, j) > 0) fcnv = redshc

                    fluxse = fcnv * fshcse * dmse
                    ttenvd(i, j, nl1) = fluxse * rsig(nl1)
                    ttenvd(i, j, kx) = -fluxse * rsig(kx)

                    if (drh >= 0.0) then
                        fluxq = fcnv * fshcq * qsat(i, j, kx) * drh
                        qtenvd(i, j, nl1) = fluxq * rsig(nl1)
                        qtenvd(i, j, kx) = -fluxq * rsig(kx)
                    end if
                else if (drh > drh0) then
                    fluxq = fvdiq2 * qsat(i, j, nl1) * drh
                    qtenvd(i, j, nl1) = fluxq * rsig(nl1)
                    qtenvd(i, j, kx) = -fluxq * rsig(kx)
                end if
            end do
        end do

        ! 3. Vertical diffusion of moisture above the PBL
        do k = 3, kx - 2
            if (sigh(k) > 0.5) then
                drh0 = rhgrad * (fsg(k + 1) - fsg(k))
                fvdiq2 = fvdiq * sigh(k)

                do i = 1, ix
                    do j = 1, il
                        drh = rh(i, j, k + 1) - rh(i, j, k)
                        if (drh >= drh0) then
                            fluxq = fvdiq2 * qsat(i, j, k) * drh
                            qtenvd(i, j, k) = qtenvd(i, j, k) + fluxq * rsig(k)
                            qtenvd(i, j, k + 1) = qtenvd(i, j, k + 1) - fluxq * rsig(k + 1)
                        end if
                    end do
                end do
            end if
        end do

        ! 4. Damping of super-adiabatic lapse rate
        do k = 1, nl1
            do i = 1, ix
                do j = 1, il
                    se0 = se(i, j, k + 1) + segrad * (phi(i, j, k) - phi(i, j, k + 1))

                    if (se(i, j, k) < se0) then
                        fluxse = fvdise * (se0 - se(i, j, k))
                        ttenvd(i, j, k) = ttenvd(i, j, k) + fluxse * rsig(k)
                        do k1 = k + 1, kx
                            ttenvd(i, j, k1) = ttenvd(i, j, k1) - fluxse * rsig1(k)
                        end do
                    end if
                end do
            end do
        end do
    end
end module