!----------------------------------------------------------------------- ! ! Purpose: ! Transfrom variables from spherical harmonic coefficients ! to grid point values during second gaussian latitude scan (scan2) ! ! Method: ! Assemble northern and southern hemisphere grid values from the ! symmetric and antisymmetric fourier coefficients. ! 1. Determine the fourier coefficients for the northern or southern ! hemisphere latitude. ! 2. Transform to gridpoint values ! 3. Clean up ! ! Author: ! Original version: J. Rosinski ! Standardized: J. Rosinski, June 1992 ! Reviewed: B. Boville, J. Hack, August 1992 ! Reviewed: B. Boville, April 1996 ! Modified: P. Worley, September 2002 ! !----------------------------------------------------------------------- ! subroutine spegrd_bft (lat ,nlon_fft, & grdps ,grzs ,grds ,gruhs ,grvhs , & grths ,grpss ,grus ,grvs ,grts , & grpls ,grpms ,grdpa ,grza ,grda , & gruha ,grvha ,grtha ,grpsa ,grua , & grva ,grta ,grpla ,grpma ,fftbuf ) !----------------------------------------------------------------------- ! ! Purpose: ! Preparation for transform of variables from spherical harmonic ! coefficients to grid point values during second gaussian latitude scan ! (scan2) ! ! Original version: J. Rosinski ! Standardized: J. Rosinski, June 1992 ! Reviewed: B. Boville, J. Hack, August 1992 ! Reviewed: B. Boville, April 1996 ! Modified: P. Worley, September 2002 ! !----------------------------------------------------------------------- ! ! $Id$ ! $Author$ ! use shr_kind_mod, only: r8 => shr_kind_r8 use pmgrid, only: plat, plev, plevp use spmd_utils, only: iam use comspe, only: maxm, numm !----------------------------------------------------------------------- implicit none !--------------------------------------------------------------------- ! ! Arguments ! integer, intent(in) :: lat ! latitude index integer, intent(in) :: nlon_fft ! first dimension of FFT work array ! ! Symmetric fourier coefficient arrays for all variables transformed ! from spherical harmonics (see grcalc) ! real(r8), intent(in) :: grdps(2*maxm) ! sum(n) of K(4)*(n(n+1)/a**2)**2*2dt*lnps(n,m)*P(n,m) real(r8), intent(in) :: grzs(2*maxm,plev) ! sum(n) of z(n,m)*P(n,m) real(r8), intent(in) :: grds(2*maxm,plev) ! sum(n) of d(n,m)*P(n,m) real(r8), intent(in) :: gruhs(2*maxm,plev) ! sum(n) of K(2i)*z(n,m)*H(n,m)*a/(n(n+1)) real(r8), intent(in) :: grvhs(2*maxm,plev) ! sum(n) of K(2i)*d(n,m)*H(n,m)*a/(n(n+1)) real(r8), intent(in) :: grths(2*maxm,plev) ! sum(n) of K(2i)*t(n,m)*P(n,m) real(r8), intent(in) :: grpss(2*maxm) ! sum(n) of lnps(n,m)*P(n,m) real(r8), intent(in) :: grus(2*maxm,plev) ! sum(n) of z(n,m)*H(n,m)*a/(n(n+1)) real(r8), intent(in) :: grvs(2*maxm,plev) ! sum(n) of d(n,m)*H(n,m)*a/(n(n+1)) real(r8), intent(in) :: grts(2*maxm,plev) ! sum(n) of t(n,m)*P(n,m) real(r8), intent(in) :: grpls(2*maxm) ! sum(n) of lnps(n,m)*P(n,m)*m/a real(r8), intent(in) :: grpms(2*maxm) ! sum(n) of lnps(n,m)*H(n,m) ! ! Antisymmetric fourier coefficient arrays for all variables transformed ! from spherical harmonics (see grcalc) ! real(r8), intent(in) :: grdpa(2*maxm) ! sum(n) of K(4)*(n(n+1)/a**2)**2*2dt*lnps(n,m)*P(n,m) real(r8), intent(in) :: grza(2*maxm,plev) ! sum(n) of z(n,m)*P(n,m) real(r8), intent(in) :: grda(2*maxm,plev) ! sum(n) of d(n,m)*P(n,m) real(r8), intent(in) :: gruha(2*maxm,plev) ! sum(n)K(2i)*z(n,m)*H(n,m)*a/(n(n+1)) real(r8), intent(in) :: grvha(2*maxm,plev) ! sum(n)K(2i)*d(n,m)*H(n,m)*a/(n(n+1)) real(r8), intent(in) :: grtha(2*maxm,plev) ! sum(n) of K(2i)*t(n,m)*P(n,m) real(r8), intent(in) :: grpsa(2*maxm) ! sum(n) of lnps(n,m)*P(n,m) real(r8), intent(in) :: grua(2*maxm,plev) ! sum(n) of z(n,m)*H(n,m)*a/(n(n+1)) real(r8), intent(in) :: grva(2*maxm,plev) ! sum(n) of d(n,m)*H(n,m)*a/(n(n+1)) real(r8), intent(in) :: grta(2*maxm,plev) ! sum(n) of t(n,m)*P(n,m) real(r8), intent(in) :: grpla(2*maxm) ! sum(n) of lnps(n,m)*P(n,m)*m/a real(r8), intent(in) :: grpma(2*maxm) ! sum(n) of lnps(n,m)*H(n,m) real(r8), intent(out) :: fftbuf(nlon_fft,8,plevp) ! buffer used for in-place FFTs ! !---------------------------Local workspace----------------------------- ! integer i,k ! longitude, level indices integer rmlength ! twice number of local wavenumbers integer, parameter :: vortdex = 1 ! indices into fftbuf integer, parameter :: divdex = 2 integer, parameter :: duhdex = 3 integer, parameter :: dvhdex = 4 integer, parameter :: dthdex = 5 integer, parameter :: u3dex = 6 integer, parameter :: v3dex = 7 integer, parameter :: t3dex = 8 integer, parameter :: dpsdex = 1 integer, parameter :: psdex = 2 integer, parameter :: dpsldex = 3 integer, parameter :: dpsmdex = 4 ! !----------------------------------------------------------------------- ! ! Assemble northern and southern hemisphere grid values from the ! symmetric and antisymmetric fourier coefficients: pre-FFT ! !DIR$ NOSTREAM rmlength = 2*numm(iam) if (lat > plat/2) then ! Northern hemisphere !DIR$ PREFERVECTOR do k=1,plev !cdir loopchg do i=1,rmlength fftbuf(i,vortdex,k) = grzs(i,k) + grza(i,k) fftbuf(i,divdex,k) = grds(i,k) + grda(i,k) fftbuf(i,duhdex,k) = gruhs(i,k) + gruha(i,k) fftbuf(i,dvhdex,k) = grvhs(i,k) + grvha(i,k) fftbuf(i,dthdex,k) = grths(i,k) + grtha(i,k) fftbuf(i,u3dex,k) = grus(i,k) + grua(i,k) fftbuf(i,v3dex,k) = grvs(i,k) + grva(i,k) fftbuf(i,t3dex,k) = grts(i,k) + grta(i,k) end do end do ! !cdir altcode=(loopcnt) do i=1,rmlength fftbuf(i,dpsdex,plevp) = grdps(i) + grdpa(i) fftbuf(i,psdex,plevp) = grpss(i) + grpsa(i) fftbuf(i,dpsldex,plevp) = grpls(i) + grpla(i) fftbuf(i,dpsmdex,plevp) = grpms(i) + grpma(i) end do else ! Southern hemisphere !DIR$ PREFERVECTOR do k=1,plev !cdir loopchg do i=1,rmlength fftbuf(i,vortdex,k) = grzs(i,k) - grza(i,k) fftbuf(i,divdex,k) = grds(i,k) - grda(i,k) fftbuf(i,duhdex,k) = gruhs(i,k) - gruha(i,k) fftbuf(i,dvhdex,k) = grvhs(i,k) - grvha(i,k) fftbuf(i,dthdex,k) = grths(i,k) - grtha(i,k) fftbuf(i,u3dex,k) = grus(i,k) - grua(i,k) fftbuf(i,v3dex,k) = grvs(i,k) - grva(i,k) fftbuf(i,t3dex,k) = grts(i,k) - grta(i,k) end do end do !cdir altcode=(loopcnt) do i=1,rmlength fftbuf(i,dpsdex,plevp) = grdps(i) - grdpa(i) fftbuf(i,psdex,plevp) = grpss(i) - grpsa(i) fftbuf(i,dpsldex,plevp) = grpls(i) - grpla(i) fftbuf(i,dpsmdex,plevp) = grpms(i) - grpma(i) end do end if return end subroutine spegrd_bft subroutine spegrd_ift (nlon_fft_in, nlon_fft_out, fftbuf_in, fftbuf_out) !----------------------------------------------------------------------- ! ! Purpose: ! Inverse Fourier transform of variables from spherical harmonic ! coefficients to grid point values during second gaussian latitude scan ! (scan2) ! ! Author: P. Worley, September 2002 ! !----------------------------------------------------------------------- ! ! $Id$ ! $Author$ ! use shr_kind_mod, only: r8 => shr_kind_r8 use pmgrid, only: plon, plat, plevp, beglat, endlat, plev use rgrid use comspe, only: maxm #if ( defined SPMD ) use mpishorthand #endif use eul_control_mod, only : trig, ifax, pcray use perf_mod !----------------------------------------------------------------------- implicit none !----------------------------------------------------------------------- !--------------------------------------------------------------------- ! ! Arguments ! ! ! Input arguments ! integer, intent(in) :: nlon_fft_in ! first dimension of first FFT work array integer, intent(in) :: nlon_fft_out ! first dimension of second FFT work array #if (defined SPMD) real(r8), intent(in) :: fftbuf_in(nlon_fft_in,8,plevp,plat) ! buffer containing fields dcomposed over wavenumbers #else real(r8), intent(in) :: fftbuf_in(1,1,1,1) ! buffer unused #endif ! ! Input/Output arguments ! real(r8), intent(inout) :: fftbuf_out(nlon_fft_out,8,plevp,beglat:endlat) ! buffer used for in-place FFTs ! !---------------------------Local workspace----------------------------- ! #if ( ! defined USEFFTLIB ) real(r8) work((plon+1)*8*plevp) #else real(r8) work((plon+1)*pcray) ! workspace needed by fft991 #endif integer lat ! latitude index integer isign ! +1 => transform spectral to grid integer ntr ! number of transforms to perform integer inc ! distance between transform elements integer begtrm ! (real) location of first truncated wavenumber integer k, ifld ! level and field indices ! !----------------------------------------------------------------------- ! ! #if ( defined SPMD ) ! ! reorder Fourier coefficients ! call t_barrierf ('sync_realloc4b', mpicom) call t_startf('realloc4b') call realloc4b(nlon_fft_in, nlon_fft_out, fftbuf_in, fftbuf_out) call t_stopf('realloc4b') #endif ! ! Zero elements corresponding to truncated wavenumbers, then ! transform from fourier coefficients to gridpoint values. ! ps,vort,div,duh,dvh,dth,dpsl,dpsm,dps, ! u,v,t (SLT) [If you want to do spectral transport, do q as well] ! begtrm = 2*pmmax+1 inc = 1 isign = +1 #ifdef OUTER_OMP !$OMP PARALLEL DO PRIVATE (LAT, NTR, K, IFLD, WORK) #endif do lat=beglat,endlat ntr = 8 !$OMP PARALLEL DO PRIVATE (K, WORK) do k=1,plev fftbuf_out(begtrm:nlon_fft_out,:,k,lat) = 0.0_r8 call fft991 (fftbuf_out(1,1,k,lat), work, trig(1,lat), ifax(1,lat), inc, & nlon_fft_out, nlon(lat), ntr, isign) enddo ntr = 1 !$OMP PARALLEL DO PRIVATE (IFLD, WORK) do ifld=1,4 fftbuf_out(begtrm:nlon_fft_out,ifld,plevp,lat) = 0.0_r8 call fft991 (fftbuf_out(1,ifld,plevp,lat), work, trig(1,lat), ifax(1,lat), inc, & nlon_fft_out, nlon(lat), ntr, isign) enddo enddo ! return end subroutine spegrd_ift subroutine spegrd_aft (ztodt ,lat ,nlon ,nlon_fft, & cwava ,qfcst , & etamid ,ps ,u3 ,v3 ,t3 , & qminus ,vort ,div ,hw2al ,hw2bl , & hw3al ,hw3bl ,hwxal ,hwxbl ,q3m1 , & dps ,dpsl ,dpsm ,t3m2 ,engy2alat, & engy2blat,difftalat, difftblat,phis,fftbuf ) !----------------------------------------------------------------------- ! ! Purpose: ! Completion of transformation of variables from spherical harmonic ! coefficients to grid point values during second gaussian latitude scan ! (scan2) ! ! Method: ! ! Original version: J. Rosinski ! Standardized: J. Rosinski, June 1992 ! Reviewed: B. Boville, J. Hack, August 1992 ! Reviewed: B. Boville, April 1996 ! Modified: P. Worley, September 2002 ! !----------------------------------------------------------------------- ! ! $Id$ ! $Author$ ! use shr_kind_mod, only: r8 => shr_kind_r8 use pmgrid, only: plon, plat, plev, plevp use pspect use commap use cam_history, only: outfld use physconst, only: rga use constituents, only: pcnst use eul_control_mod use hycoef, only: nprlev !----------------------------------------------------------------------- implicit none !----------------------------------------------------------------------- ! ! Arguments ! integer, intent(in) :: lat ! latitude index integer, intent(in) :: nlon ! number of longitudes integer, intent(in) :: nlon_fft ! first dimension of FFT work arrays real(r8), intent(in) :: ztodt ! twice the timestep unles nstep=0 real(r8), intent(in) :: cwava ! normalization factor (1/g*plon) real(r8), intent(in) :: qfcst(plon,plev,pcnst) real(r8), intent(in) :: qminus(plon,plev,pcnst) real(r8), intent(in) :: etamid(plev) ! vertical coords at midpoints real(r8), intent(inout) :: ps(plon) real(r8), intent(inout) :: u3(plon,plev) real(r8), intent(inout) :: v3(plon,plev) real(r8), intent(inout) :: t3(plon,plev) real(r8), intent(inout) :: vort(plon,plev) real(r8), intent(inout) :: div(plon,plev) real(r8), intent(inout) :: q3m1(plon,plev,pcnst) real(r8), intent(out) :: hw2al(pcnst) ! - real(r8), intent(out) :: hw2bl(pcnst) ! | lat contributions to components real(r8), intent(out) :: hw3al(pcnst) ! | of slt global mass integrals real(r8), intent(out) :: hw3bl(pcnst) ! - real(r8), intent(out) :: hwxal(pcnst,4) real(r8), intent(out) :: hwxbl(pcnst,4) real(r8), intent(out) :: dps(plon) real(r8), intent(out) :: dpsl(plon) real(r8), intent(out) :: dpsm(plon) real(r8), intent(in) :: t3m2(plon,plev) ! temperature real(r8), intent(out) :: engy2alat real(r8), intent(out) :: engy2blat real(r8), intent(out) :: difftalat real(r8), intent(out) :: difftblat real(r8), intent(in) :: phis(plon) real(r8), intent(in) :: fftbuf(nlon_fft,8,plevp) ! buffer used for in-place FFTs ! !---------------------------Local workspace----------------------------- ! real(r8) :: duh(plon,plev) ! real(r8) :: dvh(plon,plev) ! real(r8) :: dth(plon,plev) ! real(r8) pmid(plon,plev) ! pressure at model levels real(r8) pint(plon,plevp) ! pressure at model interfaces real(r8) pdel(plon,plev) ! pdel(k) = pint(k+1) - pint(k) real(r8) pdelb(plon,plev) ! pressure diff bet intfcs (press defined using the "B" part ! of the hybrid grid only) real(r8) hcwavaw ! 0.5*cwava*w(lat) real(r8) sum ! real(r8) rcoslat ! 1./cosine(latitude) real(r8) dotproda ! dot product real(r8) dotprodb ! dot product integer i,k,m ! longitude, level, constituent indices integer klev ! top level where hybrid coordinates apply integer, parameter :: vortdex = 1 ! indices into fftbuf integer, parameter :: divdex = 2 integer, parameter :: duhdex = 3 integer, parameter :: dvhdex = 4 integer, parameter :: dthdex = 5 integer, parameter :: u3dex = 6 integer, parameter :: v3dex = 7 integer, parameter :: t3dex = 8 integer, parameter :: dpsdex = 1 integer, parameter :: psdex = 2 integer, parameter :: dpsldex = 3 integer, parameter :: dpsmdex = 4 ! !----------------------------------------------------------------------- ! ! Copy 3D fields out of FFT buffer, removing cosine(latitude) from momentum variables ! rcoslat = 1._r8/cos(clat(lat)) !$OMP PARALLEL DO PRIVATE (K, I) do k=1,plev do i=1,nlon vort(i,k) = fftbuf(i,vortdex,k) div(i,k) = fftbuf(i,divdex,k) duh(i,k) = fftbuf(i,duhdex,k)*rcoslat dvh(i,k) = fftbuf(i,dvhdex,k)*rcoslat dth(i,k) = fftbuf(i,dthdex,k) u3(i,k) = fftbuf(i,u3dex,k)*rcoslat v3(i,k) = fftbuf(i,v3dex,k)*rcoslat t3(i,k) = fftbuf(i,t3dex,k) end do end do ! ! Copy 2D fields out of FFT buffer, converting ! log(ps) to ps. ! !$OMP PARALLEL DO PRIVATE (I) do i=1,nlon dps(i) = fftbuf(i,dpsdex,plevp) dpsl(i) = fftbuf(i,dpsldex,plevp) dpsm(i) = fftbuf(i,dpsmdex,plevp) ps(i) = exp(fftbuf(i,psdex,plevp)) end do ! ! Diagnose pressure arrays needed by DIFCOR ! call plevs0 (nlon, plon, plev, ps, pint, pmid, pdel) call pdelb0 (ps, pdelb, nlon) ! ! Accumulate mass integrals ! sum = 0._r8 do i=1,nlon sum = sum + ps(i) end do tmass(lat) = w(lat)*rga*sum/nlon ! ! Finish horizontal diffusion: add pressure surface correction term to t and ! q diffusions; add kinetic energy dissipation to internal energy (temperature) ! klev = max(kmnhd4,nprlev) call difcor (klev, ztodt, dps, u3, v3, & q3m1(1,1,1), pdel, pint, t3, dth, & duh, dvh, nlon) ! ! Calculate SLT moisture, constituent, energy, and temperature integrals ! hcwavaw = 0.5_r8*cwava*w(lat) engy2alat = 0._r8 engy2blat = 0._r8 difftalat = 0._r8 difftblat = 0._r8 !$OMP PARALLEL DO PRIVATE (M, K, DOTPRODA, DOTPRODB, I) do m=1,pcnst hw2al(m) = 0._r8 hw2bl(m) = 0._r8 hw3al(m) = 0._r8 hw3bl(m) = 0._r8 hwxal(m,1) = 0._r8 hwxal(m,2) = 0._r8 hwxal(m,3) = 0._r8 hwxal(m,4) = 0._r8 hwxbl(m,1) = 0._r8 hwxbl(m,2) = 0._r8 hwxbl(m,3) = 0._r8 hwxbl(m,4) = 0._r8 do k=1,plev dotproda = 0._r8 dotprodb = 0._r8 do i=1,nlon dotproda = dotproda + qfcst(i,k,m)*pdela(i,k) dotprodb = dotprodb + qfcst(i,k,m)*pdelb(i,k) end do hw2al(m) = hw2al(m) + hcwavaw*dotproda hw2bl(m) = hw2bl(m) + hcwavaw*dotprodb end do end do ! using do loop and select to enable functional parallelism with OpenMP !$OMP PARALLEL DO PRIVATE (I) do i=1,6 select case (i) case (1) call engy_te (cwava ,w(lat) ,t3 ,u3 ,v3 ,phis ,pdela, engy2alat ,nlon) case (2) call engy_te (cwava ,w(lat) ,t3 ,u3 ,v3 ,phis ,pdelb, engy2blat ,nlon) case (3) call engy_tdif(cwava ,w(lat) ,t3 ,t3m2 ,pdela, difftalat ,nlon) case (4) call engy_tdif(cwava ,w(lat) ,t3 ,t3m2 ,pdelb, difftblat ,nlon) case (5) call qmassd (cwava, etamid, w(lat), qminus, qfcst, & pdela, hw3al, nlon) case (6) call qmassd (cwava, etamid, w(lat), qminus, qfcst, & pdelb, hw3bl, nlon) end select end do if (pcnst.gt.1) then call xqmass (cwava, etamid, w(lat), qminus, qfcst, & qminus, qfcst, pdela, pdelb, hwxal, & hwxbl, nlon) end if call outfld ('DTH ',dth ,plon ,lat ) return end subroutine spegrd_aft