!> @file data_output_mask.f90 !--------------------------------------------------------------------------------! ! This file is part of PALM. ! ! PALM is free software: you can redistribute it and/or modify it under the terms ! of the GNU General Public License as published by the Free Software Foundation, ! either version 3 of the License, or (at your option) any later version. ! ! PALM is distributed in the hope that it will be useful, but WITHOUT ANY ! WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR ! A PARTICULAR PURPOSE. See the GNU General Public License for more details. ! ! You should have received a copy of the GNU General Public License along with ! PALM. If not, see . ! ! Copyright 1997-2015 Leibniz Universitaet Hannover !--------------------------------------------------------------------------------! ! ! Current revisions: ! ----------------- ! ! ! Former revisions: ! ----------------- ! $Id: data_output_mask.f90 1784 2016-03-06 19:14:40Z knoop $ ! ! 2016-03-06 18:36:17Z raasch ! name change of netcdf routines and module + related changes, ! switch back of netcdf data format moved from time integration routine to here ! ! 1691 2015-10-26 16:17:44Z maronga ! Added output of radiative heating rates for RRTMG ! ! 1682 2015-10-07 23:56:08Z knoop ! Code annotations made doxygen readable ! ! 1585 2015-04-30 07:05:52Z maronga ! Added support for RRTMG ! ! 1438 2014-07-22 14:14:06Z heinze ! +nr, qc, qr ! ! 1359 2014-04-11 17:15:14Z hoffmann ! New particle structure integrated. ! ! 1353 2014-04-08 15:21:23Z heinze ! REAL constants provided with KIND-attribute ! ! 1327 2014-03-21 11:00:16Z raasch ! ! ! 1320 2014-03-20 08:40:49Z raasch ! ONLY-attribute added to USE-statements, ! kind-parameters added to all INTEGER and REAL declaration statements, ! kinds are defined in new module kinds, ! revision history before 2012 removed, ! comment fields (!:) to be used for variable explanations added to ! all variable declaration statements ! ! 1318 2014-03-17 13:35:16Z raasch ! barrier argument removed from cpu_log, ! module interfaces removed ! ! 1092 2013-02-02 11:24:22Z raasch ! unused variables removed ! ! 1036 2012-10-22 13:43:42Z raasch ! code put under GPL (PALM 3.9) ! ! 1031 2012-10-19 14:35:30Z raasch ! netCDF4 without parallel file support implemented ! ! 1007 2012-09-19 14:30:36Z franke ! Bugfix: calculation of pr must depend on the particle weighting factor, ! missing calculation of ql_vp added ! ! 410 2009-12-04 17:05:40Z letzel ! Initial version ! ! Description: ! ------------ !> Masked data output in netCDF format for current mask (current value of mid). !------------------------------------------------------------------------------! SUBROUTINE data_output_mask( av ) #if defined( __netcdf ) USE arrays_3d, & ONLY: e, nr, p, pt, q, qc, ql, ql_c, ql_v, qr, rho, sa, tend, u, & v, vpt, w USE averaging, & ONLY: e_av, lpt_av, nr_av, p_av, pc_av, pr_av, pt_av, q_av, qc_av, & ql_av, ql_c_av, ql_v_av, ql_vp_av, qv_av, qr_av, rho_av, s_av, & sa_av, u_av, v_av, vpt_av, w_av USE cloud_parameters, & ONLY: l_d_cp, pt_d_t USE control_parameters, & ONLY: cloud_physics, domask, domask_no, domask_time_count, mask_i, & mask_j, mask_k, mask_size, mask_size_l, mask_start_l, & max_masks, message_string, mid, nz_do3d, simulated_time USE cpulog, & ONLY: cpu_log, log_point USE indices, & ONLY: nbgp, nxl, nxr, nyn, nys, nzb, nzt USE kinds USE NETCDF USE netcdf_interface, & ONLY: id_set_mask, id_var_domask, id_var_time_mask, nc_stat, & netcdf_data_format, netcdf_handle_error USE particle_attributes, & ONLY: grid_particles, number_of_particles, particles, & particle_advection_start, prt_count USE pegrid USE radiation_model_mod, & ONLY: rad_lw_in, rad_lw_in_av, rad_lw_out, rad_lw_out_av, & rad_lw_cs_hr, rad_lw_cs_hr_av, rad_lw_hr, rad_lw_hr_av, & rad_sw_in, rad_sw_in_av, rad_sw_out, rad_sw_out_av, & rad_sw_cs_hr, rad_sw_cs_hr_av, rad_sw_hr, rad_sw_hr_av IMPLICIT NONE INTEGER(iwp) :: av !< INTEGER(iwp) :: ngp !< INTEGER(iwp) :: i !< INTEGER(iwp) :: if !< INTEGER(iwp) :: j !< INTEGER(iwp) :: k !< INTEGER(iwp) :: n !< INTEGER(iwp) :: netcdf_data_format_save !< INTEGER(iwp) :: psi !< INTEGER(iwp) :: sender !< INTEGER(iwp) :: ind(6) !< LOGICAL :: found !< LOGICAL :: resorted !< REAL(wp) :: mean_r !< REAL(wp) :: s_r2 !< REAL(wp) :: s_r3 !< REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !< #if defined( __parallel ) REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: total_pf !< #endif REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !< ! !-- Return, if nothing to output IF ( domask_no(mid,av) == 0 ) RETURN CALL cpu_log (log_point(49),'data_output_mask','start') ! !-- Parallel netcdf output is not tested so far for masked data, hence !-- netcdf_data_format is switched back to non-paralell output. netcdf_data_format_save = netcdf_data_format IF ( netcdf_data_format == 5 ) netcdf_data_format = 3 IF ( netcdf_data_format == 6 ) netcdf_data_format = 4 ! !-- Open output file. IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN CALL check_open( 200+mid+av*max_masks ) ENDIF ! !-- Allocate total and local output arrays. #if defined( __parallel ) IF ( myid == 0 ) THEN ALLOCATE( total_pf(mask_size(mid,1),mask_size(mid,2),mask_size(mid,3)) ) ENDIF #endif ALLOCATE( local_pf(mask_size_l(mid,1),mask_size_l(mid,2), & mask_size_l(mid,3)) ) ! !-- Update the netCDF time axis. domask_time_count(mid,av) = domask_time_count(mid,av) + 1 IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN nc_stat = NF90_PUT_VAR( id_set_mask(mid,av), id_var_time_mask(mid,av), & (/ simulated_time /), & start = (/ domask_time_count(mid,av) /), & count = (/ 1 /) ) CALL netcdf_handle_error( 'data_output_mask', 460 ) ENDIF ! !-- Loop over all variables to be written. if = 1 DO WHILE ( domask(mid,av,if)(1:1) /= ' ' ) ! !-- Reallocate local_pf on PE 0 since its shape changes during MPI exchange IF ( netcdf_data_format < 5 .AND. myid == 0 .AND. if > 1 ) THEN DEALLOCATE( local_pf ) ALLOCATE( local_pf(mask_size_l(mid,1),mask_size_l(mid,2), & mask_size_l(mid,3)) ) ENDIF ! !-- Store the variable chosen. resorted = .FALSE. SELECT CASE ( TRIM( domask(mid,av,if) ) ) CASE ( 'e' ) IF ( av == 0 ) THEN to_be_resorted => e ELSE to_be_resorted => e_av ENDIF CASE ( 'lpt' ) IF ( av == 0 ) THEN to_be_resorted => pt ELSE to_be_resorted => lpt_av ENDIF CASE ( 'nr' ) IF ( av == 0 ) THEN to_be_resorted => nr ELSE to_be_resorted => nr_av ENDIF CASE ( 'p' ) IF ( av == 0 ) THEN to_be_resorted => p ELSE to_be_resorted => p_av ENDIF CASE ( 'pc' ) ! particle concentration (requires ghostpoint exchange) IF ( av == 0 ) THEN tend = prt_count CALL exchange_horiz( tend, nbgp ) DO i = 1, mask_size_l(mid,1) DO j = 1, mask_size_l(mid,2) DO k = 1, mask_size_l(mid,3) local_pf(i,j,k) = tend(mask_k(mid,k), & mask_j(mid,j),mask_i(mid,i)) ENDDO ENDDO ENDDO resorted = .TRUE. ELSE CALL exchange_horiz( pc_av, nbgp ) to_be_resorted => pc_av ENDIF CASE ( 'pr' ) ! mean particle radius (effective radius) IF ( av == 0 ) THEN IF ( simulated_time >= particle_advection_start ) THEN DO i = nxl, nxr DO j = nys, nyn DO k = nzb, nz_do3d number_of_particles = prt_count(k,j,i) IF (number_of_particles <= 0) CYCLE particles => grid_particles(k,j,i)%particles(1:number_of_particles) s_r2 = 0.0_wp s_r3 = 0.0_wp DO n = 1, number_of_particles IF ( particles(n)%particle_mask ) THEN s_r2 = s_r2 + grid_particles(k,j,i)%particles(n)%radius**2 * & grid_particles(k,j,i)%particles(n)%weight_factor s_r3 = s_r3 + grid_particles(k,j,i)%particles(n)%radius**3 * & grid_particles(k,j,i)%particles(n)%weight_factor ENDIF ENDDO IF ( s_r2 > 0.0_wp ) THEN mean_r = s_r3 / s_r2 ELSE mean_r = 0.0_wp ENDIF tend(k,j,i) = mean_r ENDDO ENDDO ENDDO CALL exchange_horiz( tend, nbgp ) ELSE tend = 0.0_wp ENDIF DO i = 1, mask_size_l(mid,1) DO j = 1, mask_size_l(mid,2) DO k = 1, mask_size_l(mid,3) local_pf(i,j,k) = tend(mask_k(mid,k), & mask_j(mid,j),mask_i(mid,i)) ENDDO ENDDO ENDDO resorted = .TRUE. ELSE CALL exchange_horiz( pr_av, nbgp ) to_be_resorted => pr_av ENDIF CASE ( 'pt' ) IF ( av == 0 ) THEN IF ( .NOT. cloud_physics ) THEN to_be_resorted => pt ELSE DO i = 1, mask_size_l(mid,1) DO j = 1, mask_size_l(mid,2) DO k = 1, mask_size_l(mid,3) local_pf(i,j,k) = & pt(mask_k(mid,k),mask_j(mid,j),mask_i(mid,i)) & + l_d_cp * pt_d_t(mask_k(mid,k)) * & ql(mask_k(mid,k),mask_j(mid,j),mask_i(mid,i)) ENDDO ENDDO ENDDO resorted = .TRUE. ENDIF ELSE to_be_resorted => pt_av ENDIF CASE ( 'q' ) IF ( av == 0 ) THEN to_be_resorted => q ELSE to_be_resorted => q_av ENDIF CASE ( 'qc' ) IF ( av == 0 ) THEN to_be_resorted => qc ELSE to_be_resorted => qc_av ENDIF CASE ( 'ql' ) IF ( av == 0 ) THEN to_be_resorted => ql ELSE to_be_resorted => ql_av ENDIF CASE ( 'ql_c' ) IF ( av == 0 ) THEN to_be_resorted => ql_c ELSE to_be_resorted => ql_c_av ENDIF CASE ( 'ql_v' ) IF ( av == 0 ) THEN to_be_resorted => ql_v ELSE to_be_resorted => ql_v_av ENDIF CASE ( 'ql_vp' ) IF ( av == 0 ) THEN IF ( simulated_time >= particle_advection_start ) THEN DO i = nxl, nxr DO j = nys, nyn DO k = nzb, nz_do3d number_of_particles = prt_count(k,j,i) IF (number_of_particles <= 0) CYCLE particles => grid_particles(k,j,i)%particles(1:number_of_particles) DO n = 1, number_of_particles IF ( particles(n)%particle_mask ) THEN tend(k,j,i) = tend(k,j,i) + & particles(n)%weight_factor / & prt_count(k,j,i) ENDIF ENDDO ENDDO ENDDO ENDDO CALL exchange_horiz( tend, nbgp ) ELSE tend = 0.0_wp ENDIF DO i = 1, mask_size_l(mid,1) DO j = 1, mask_size_l(mid,2) DO k = 1, mask_size_l(mid,3) local_pf(i,j,k) = tend(mask_k(mid,k), & mask_j(mid,j),mask_i(mid,i)) ENDDO ENDDO ENDDO resorted = .TRUE. ELSE CALL exchange_horiz( ql_vp_av, nbgp ) to_be_resorted => ql_vp_av ENDIF CASE ( 'qv' ) IF ( av == 0 ) THEN DO i = 1, mask_size_l(mid,1) DO j = 1, mask_size_l(mid,2) DO k = 1, mask_size_l(mid,3) local_pf(i,j,k) = & q(mask_k(mid,k),mask_j(mid,j),mask_i(mid,i)) - & ql(mask_k(mid,k),mask_j(mid,j),mask_i(mid,i)) ENDDO ENDDO ENDDO resorted = .TRUE. ELSE to_be_resorted => qv_av ENDIF CASE ( 'qr' ) IF ( av == 0 ) THEN to_be_resorted => qr ELSE to_be_resorted => qr_av ENDIF CASE ( 'rad_lw_in' ) IF ( av == 0 ) THEN to_be_resorted => rad_lw_in ELSE to_be_resorted => rad_lw_in_av ENDIF CASE ( 'rad_lw_out' ) IF ( av == 0 ) THEN to_be_resorted => rad_lw_out ELSE to_be_resorted => rad_lw_out_av ENDIF CASE ( 'rad_lw_cs_hr' ) IF ( av == 0 ) THEN to_be_resorted => rad_lw_cs_hr ELSE to_be_resorted => rad_lw_cs_hr_av ENDIF CASE ( 'rad_lw_hr' ) IF ( av == 0 ) THEN to_be_resorted => rad_lw_hr ELSE to_be_resorted => rad_lw_hr_av ENDIF CASE ( 'rad_sw_in' ) IF ( av == 0 ) THEN to_be_resorted => rad_sw_in ELSE to_be_resorted => rad_sw_in_av ENDIF CASE ( 'rad_sw_out' ) IF ( av == 0 ) THEN to_be_resorted => rad_sw_out ELSE to_be_resorted => rad_sw_out_av ENDIF CASE ( 'rad_sw_cs_hr' ) IF ( av == 0 ) THEN to_be_resorted => rad_sw_cs_hr ELSE to_be_resorted => rad_sw_cs_hr_av ENDIF CASE ( 'rad_sw_hr' ) IF ( av == 0 ) THEN to_be_resorted => rad_sw_hr ELSE to_be_resorted => rad_sw_hr_av ENDIF CASE ( 'rho' ) IF ( av == 0 ) THEN to_be_resorted => rho ELSE to_be_resorted => rho_av ENDIF CASE ( 's' ) IF ( av == 0 ) THEN to_be_resorted => q ELSE to_be_resorted => s_av ENDIF CASE ( 'sa' ) IF ( av == 0 ) THEN to_be_resorted => sa ELSE to_be_resorted => sa_av ENDIF CASE ( 'u' ) IF ( av == 0 ) THEN to_be_resorted => u ELSE to_be_resorted => u_av ENDIF CASE ( 'v' ) IF ( av == 0 ) THEN to_be_resorted => v ELSE to_be_resorted => v_av ENDIF CASE ( 'vpt' ) IF ( av == 0 ) THEN to_be_resorted => vpt ELSE to_be_resorted => vpt_av ENDIF CASE ( 'w' ) IF ( av == 0 ) THEN to_be_resorted => w ELSE to_be_resorted => w_av ENDIF CASE DEFAULT ! !-- User defined quantity CALL user_data_output_mask(av, domask(mid,av,if), found, local_pf ) resorted = .TRUE. IF ( .NOT. found ) THEN WRITE ( message_string, * ) 'no output available for: ', & TRIM( domask(mid,av,if) ) CALL message( 'data_output_mask', 'PA0327', 0, 0, 0, 6, 0 ) ENDIF END SELECT ! !-- Resort the array to be output, if not done above IF ( .NOT. resorted ) THEN DO i = 1, mask_size_l(mid,1) DO j = 1, mask_size_l(mid,2) DO k = 1, mask_size_l(mid,3) local_pf(i,j,k) = to_be_resorted(mask_k(mid,k), & mask_j(mid,j),mask_i(mid,i)) ENDDO ENDDO ENDDO ENDIF ! !-- I/O block. I/O methods are implemented !-- (1) for parallel execution !-- a. with netCDF 4 parallel I/O-enabled library !-- b. with netCDF 3 library !-- (2) for serial execution. !-- The choice of method depends on the correct setting of preprocessor !-- directives __parallel and __netcdf4_parallel as well as on the parameter !-- netcdf_data_format. #if defined( __parallel ) #if defined( __netcdf4_parallel ) IF ( netcdf_data_format > 4 ) THEN ! !-- (1) a. Parallel I/O using netCDF 4 (not yet tested) nc_stat = NF90_PUT_VAR( id_set_mask(mid,av), & id_var_domask(mid,av,if), & local_pf, & start = (/ mask_start_l(mid,1), mask_start_l(mid,2), & mask_start_l(mid,3), domask_time_count(mid,av) /), & count = (/ mask_size_l(mid,1), mask_size_l(mid,2), & mask_size_l(mid,3), 1 /) ) CALL netcdf_handle_error( 'data_output_mask', 461 ) ELSE #endif ! !-- (1) b. Conventional I/O only through PE0 !-- PE0 receives partial arrays from all processors of the respective mask !-- and outputs them. Here a barrier has to be set, because otherwise !-- "-MPI- FATAL: Remote protocol queue full" may occur. CALL MPI_BARRIER( comm2d, ierr ) ngp = mask_size_l(mid,1) * mask_size_l(mid,2) * mask_size_l(mid,3) IF ( myid == 0 ) THEN ! !-- Local array can be relocated directly. total_pf( & mask_start_l(mid,1):mask_start_l(mid,1)+mask_size_l(mid,1)-1, & mask_start_l(mid,2):mask_start_l(mid,2)+mask_size_l(mid,2)-1, & mask_start_l(mid,3):mask_start_l(mid,3)+mask_size_l(mid,3)-1 ) & = local_pf ! !-- Receive data from all other PEs. DO n = 1, numprocs-1 ! !-- Receive index limits first, then array. !-- Index limits are received in arbitrary order from the PEs. CALL MPI_RECV( ind(1), 6, MPI_INTEGER, MPI_ANY_SOURCE, 0, & comm2d, status, ierr ) ! !-- Not all PEs have data for the mask IF ( ind(1) /= -9999 ) THEN ngp = ( ind(2)-ind(1)+1 ) * (ind(4)-ind(3)+1 ) * & ( ind(6)-ind(5)+1 ) sender = status(MPI_SOURCE) DEALLOCATE( local_pf ) ALLOCATE(local_pf(ind(1):ind(2),ind(3):ind(4),ind(5):ind(6))) CALL MPI_RECV( local_pf(ind(1),ind(3),ind(5)), ngp, & MPI_REAL, sender, 1, comm2d, status, ierr ) total_pf(ind(1):ind(2),ind(3):ind(4),ind(5):ind(6)) & = local_pf ENDIF ENDDO nc_stat = NF90_PUT_VAR( id_set_mask(mid,av), & id_var_domask(mid,av,if), total_pf, & start = (/ 1, 1, 1, domask_time_count(mid,av) /), & count = (/ mask_size(mid,1), mask_size(mid,2), & mask_size(mid,3), 1 /) ) CALL netcdf_handle_error( 'data_output_mask', 462 ) ELSE ! !-- If at least part of the mask resides on the PE, send the index !-- limits for the target array, otherwise send -9999 to PE0. IF ( mask_size_l(mid,1) > 0 .AND. mask_size_l(mid,2) > 0 .AND. & mask_size_l(mid,3) > 0 ) & THEN ind(1) = mask_start_l(mid,1) ind(2) = mask_start_l(mid,1) + mask_size_l(mid,1) - 1 ind(3) = mask_start_l(mid,2) ind(4) = mask_start_l(mid,2) + mask_size_l(mid,2) - 1 ind(5) = mask_start_l(mid,3) ind(6) = mask_start_l(mid,3) + mask_size_l(mid,3) - 1 ELSE ind(1) = -9999; ind(2) = -9999 ind(3) = -9999; ind(4) = -9999 ind(5) = -9999; ind(6) = -9999 ENDIF CALL MPI_SEND( ind(1), 6, MPI_INTEGER, 0, 0, comm2d, ierr ) ! !-- If applicable, send data to PE0. IF ( ind(1) /= -9999 ) THEN CALL MPI_SEND( local_pf(1,1,1), ngp, MPI_REAL, 0, 1, comm2d, & ierr ) ENDIF ENDIF ! !-- A barrier has to be set, because otherwise some PEs may proceed too !-- fast so that PE0 may receive wrong data on tag 0. CALL MPI_BARRIER( comm2d, ierr ) #if defined( __netcdf4_parallel ) ENDIF #endif #else ! !-- (2) For serial execution of PALM, the single processor (PE0) holds all !-- data and writes them directly to file. nc_stat = NF90_PUT_VAR( id_set_mask(mid,av), & id_var_domask(mid,av,if), & local_pf, & start = (/ 1, 1, 1, domask_time_count(mid,av) /), & count = (/ mask_size_l(mid,1), mask_size_l(mid,2), & mask_size_l(mid,3), 1 /) ) CALL netcdf_handle_error( 'data_output_mask', 463 ) #endif if = if + 1 ENDDO ! !-- Deallocate temporary arrays. DEALLOCATE( local_pf ) #if defined( __parallel ) IF ( myid == 0 ) THEN DEALLOCATE( total_pf ) ENDIF #endif ! !-- Switch back to original format given by user (see beginning of this routine) netcdf_data_format = netcdf_data_format_save CALL cpu_log( log_point(49), 'data_output_mask', 'stop' ) #endif END SUBROUTINE data_output_mask