Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 1972

Last change on this file since 1972 was 1972, checked in by maronga, 8 years ago
further modularization of land surface model (2D/3D output and restart data)
Property svn:keywords set to `Id`
File size: 80.1 KB

Rev	Line
[1682]	1	!> @file data_output_2d.f90
[1036]	2	!--------------------------------------------------------------------------------!
	3	! This file is part of PALM.
	4	!
	5	! PALM is free software: you can redistribute it and/or modify it under the terms
	6	! of the GNU General Public License as published by the Free Software Foundation,
	7	! either version 3 of the License, or (at your option) any later version.
	8	!
	9	! PALM is distributed in the hope that it will be useful, but WITHOUT ANY
	10	! WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
	11	! A PARTICULAR PURPOSE. See the GNU General Public License for more details.
	12	!
	13	! You should have received a copy of the GNU General Public License along with
	14	! PALM. If not, see <http://www.gnu.org/licenses/>.
	15	!
[1818]	16	! Copyright 1997-2016 Leibniz Universitaet Hannover
[1704]	17	!--------------------------------------------------------------------------------!
[1036]	18	!
[254]	19	! Current revisions:
[1]	20	! -----------------
[1972]	21	! Output of land surface quantities is now done directly in the respective module
[1961]	22	!
[1552]	23	! Former revisions:
	24	! -----------------
	25	! $Id: data_output_2d.f90 1972 2016-07-26 07:52:02Z maronga $
	26	!
[1961]	27	! 1960 2016-07-12 16:34:24Z suehring
	28	! Scalar surface flux added
	29	! Rename INTEGER variable s into s_ind, as s is already assigned to scalar
	30	!
[1851]	31	! 1849 2016-04-08 11:33:18Z hoffmann
	32	! precipitation_amount, precipitation_rate, prr moved to arrays_3d
	33	!
[1823]	34	! 1822 2016-04-07 07:49:42Z hoffmann
	35	! Output of bulk cloud physics simplified.
	36	!
[1789]	37	! 1788 2016-03-10 11:01:04Z maronga
	38	! Added output of z0q
	39	!
[1784]	40	! 1783 2016-03-06 18:36:17Z raasch
	41	! name change of netcdf routines and module + related changes
	42	!
[1746]	43	! 1745 2016-02-05 13:06:51Z gronemeier
	44	! Bugfix: test if time axis limit exceeds moved to point after call of check_open
	45	!
[1704]	46	! 1703 2015-11-02 12:38:44Z raasch
	47	! bugfix for output of single (*) xy-sections in case of parallel netcdf I/O
	48	!
[1702]	49	! 1701 2015-11-02 07:43:04Z maronga
	50	! Bugfix in output of RRTGM data
	51	!
[1692]	52	! 1691 2015-10-26 16:17:44Z maronga
	53	! Added output of Obukhov length (ol) and radiative heating rates for RRTMG.
	54	! Formatting corrections.
	55	!
[1683]	56	! 1682 2015-10-07 23:56:08Z knoop
	57	! Code annotations made doxygen readable
	58	!
[1586]	59	! 1585 2015-04-30 07:05:52Z maronga
	60	! Added support for RRTMG
	61	!
[1556]	62	! 1555 2015-03-04 17:44:27Z maronga
	63	! Added output of r_a and r_s
	64	!
[1552]	65	! 1551 2015-03-03 14:18:16Z maronga
[1551]	66	! Added suppport for land surface model and radiation model output. In the course
	67	! of this action, the limits for vertical loops have been changed (from nzb and
	68	! nzt+1 to nzb_do and nzt_do, respectively in order to allow soil model output).
	69	! Moreover, a new vertical grid zs was introduced.
[1329]	70	!
[1360]	71	! 1359 2014-04-11 17:15:14Z hoffmann
	72	! New particle structure integrated.
	73	!
[1354]	74	! 1353 2014-04-08 15:21:23Z heinze
	75	! REAL constants provided with KIND-attribute
	76	!
[1329]	77	! 1327 2014-03-21 11:00:16Z raasch
	78	! parts concerning iso2d output removed,
	79	! -netcdf output queries
	80	!
[1321]	81	! 1320 2014-03-20 08:40:49Z raasch
[1320]	82	! ONLY-attribute added to USE-statements,
	83	! kind-parameters added to all INTEGER and REAL declaration statements,
	84	! kinds are defined in new module kinds,
	85	! revision history before 2012 removed,
	86	! comment fields (!:) to be used for variable explanations added to
	87	! all variable declaration statements
[1309]	88	!
[1319]	89	! 1318 2014-03-17 13:35:16Z raasch
	90	! barrier argument removed from cpu_log.
	91	! module interfaces removed
	92	!
[1312]	93	! 1311 2014-03-14 12:13:39Z heinze
	94	! bugfix: close #if defined( __netcdf )
	95	!
[1309]	96	! 1308 2014-03-13 14:58:42Z fricke
[1308]	97	! +local_2d_sections, local_2d_sections_l, ns
	98	! Check, if the limit of the time dimension is exceeded for parallel output
	99	! To increase the performance for parallel output, the following is done:
	100	! - Update of time axis is only done by PE0
	101	! - Cross sections are first stored on a local array and are written
	102	! collectively to the output file by all PEs.
[674]	103	!
[1116]	104	! 1115 2013-03-26 18:16:16Z hoffmann
	105	! ql is calculated by calc_liquid_water_content
	106	!
[1077]	107	! 1076 2012-12-05 08:30:18Z hoffmann
	108	! Bugfix in output of ql
	109	!
[1066]	110	! 1065 2012-11-22 17:42:36Z hoffmann
	111	! Bugfix: Output of cross sections of ql
	112	!
[1054]	113	! 1053 2012-11-13 17:11:03Z hoffmann
	114	! +qr, nr, qc and cross sections
	115	!
[1037]	116	! 1036 2012-10-22 13:43:42Z raasch
	117	! code put under GPL (PALM 3.9)
	118	!
[1035]	119	! 1031 2012-10-19 14:35:30Z raasch
	120	! netCDF4 without parallel file support implemented
	121	!
[1008]	122	! 1007 2012-09-19 14:30:36Z franke
	123	! Bugfix: missing calculation of ql_vp added
	124	!
[979]	125	! 978 2012-08-09 08:28:32Z fricke
	126	! +z0h
	127	!
[1]	128	! Revision 1.1 1997/08/11 06:24:09 raasch
	129	! Initial revision
	130	!
	131	!
	132	! Description:
	133	! ------------
[1682]	134	!> Data output of horizontal cross-sections in netCDF format or binary format
	135	!> compatible to old graphic software iso2d.
	136	!> Attention: The position of the sectional planes is still not always computed
	137	!> --------- correctly. (zu is used always)!
[1]	138	!------------------------------------------------------------------------------!
[1682]	139	SUBROUTINE data_output_2d( mode, av )
	140
[1]	141
[1320]	142	USE arrays_3d, &
[1849]	143	ONLY: dzw, e, nr, ol, p, pt, precipitation_amount, precipitation_rate,&
[1960]	144	prr,q, qc, ql, ql_c, ql_v, ql_vp, qr, qsws, rho, s, sa, shf, &
	145	ssws, tend, ts, u, us, v, vpt, w, z0, z0h, z0q, zu, zw
[1320]	146
[1]	147	USE averaging
[1320]	148
	149	USE cloud_parameters, &
[1849]	150	ONLY: hyrho, l_d_cp, pt_d_t
[1320]	151
	152	USE control_parameters, &
	153	ONLY: cloud_physics, data_output_2d_on_each_pe, data_output_xy, &
	154	data_output_xz, data_output_yz, do2d, &
	155	do2d_xy_last_time, do2d_xy_n, do2d_xy_time_count, &
	156	do2d_xz_last_time, do2d_xz_n, do2d_xz_time_count, &
	157	do2d_yz_last_time, do2d_yz_n, do2d_yz_time_count, &
[1822]	158	ibc_uv_b, io_blocks, io_group, message_string, &
	159	ntdim_2d_xy, ntdim_2d_xz, ntdim_2d_yz, &
	160	psolver, section, simulated_time, simulated_time_chr, &
	161	time_since_reference_point
[1320]	162
	163	USE cpulog, &
	164	ONLY: cpu_log, log_point
	165
	166	USE grid_variables, &
	167	ONLY: dx, dy
	168
	169	USE indices, &
	170	ONLY: nbgp, nx, nxl, nxlg, nxr, nxrg, ny, nyn, nyng, nys, nysg, &
	171	nz, nzb, nzt
	172
	173	USE kinds
[1551]	174
	175	USE land_surface_model_mod, &
[1972]	176	ONLY: land_surface, lsm_data_output_2d, zs
[1551]	177
[1783]	178	#if defined( __netcdf )
	179	USE NETCDF
	180	#endif
[1320]	181
[1783]	182	USE netcdf_interface, &
	183	ONLY: id_set_xy, id_set_xz, id_set_yz, id_var_do2d, id_var_time_xy, &
	184	id_var_time_xz, id_var_time_yz, nc_stat, netcdf_data_format, &
	185	netcdf_handle_error
	186
[1320]	187	USE particle_attributes, &
[1359]	188	ONLY: grid_particles, number_of_particles, particle_advection_start, &
	189	particles, prt_count
[1320]	190
[1]	191	USE pegrid
	192
[1551]	193	USE radiation_model_mod, &
[1585]	194	ONLY: rad_net, rad_net_av, rad_sw_in, rad_sw_in_av, rad_sw_out, &
[1691]	195	rad_sw_out_av, rad_sw_cs_hr, rad_sw_cs_hr_av, rad_sw_hr, &
	196	rad_sw_hr_av, rad_lw_in, rad_lw_in_av, rad_lw_out, &
	197	rad_lw_out_av, rad_lw_cs_hr, rad_lw_cs_hr_av, rad_lw_hr, &
	198	rad_lw_hr_av
[1551]	199
[1]	200	IMPLICIT NONE
	201
[1682]	202	CHARACTER (LEN=2) :: do2d_mode !<
	203	CHARACTER (LEN=2) :: mode !<
	204	CHARACTER (LEN=4) :: grid !<
	205	CHARACTER (LEN=25) :: section_chr !<
	206	CHARACTER (LEN=50) :: rtext !<
[1320]	207
[1682]	208	INTEGER(iwp) :: av !<
	209	INTEGER(iwp) :: ngp !<
	210	INTEGER(iwp) :: file_id !<
	211	INTEGER(iwp) :: i !<
	212	INTEGER(iwp) :: if !<
	213	INTEGER(iwp) :: is !<
	214	INTEGER(iwp) :: iis !<
	215	INTEGER(iwp) :: j !<
	216	INTEGER(iwp) :: k !<
	217	INTEGER(iwp) :: l !<
	218	INTEGER(iwp) :: layer_xy !<
	219	INTEGER(iwp) :: n !<
[1703]	220	INTEGER(iwp) :: nis !<
[1682]	221	INTEGER(iwp) :: ns !<
	222	INTEGER(iwp) :: nzb_do !< lower limit of the data field (usually nzb)
	223	INTEGER(iwp) :: nzt_do !< upper limit of the data field (usually nzt+1)
	224	INTEGER(iwp) :: psi !<
[1960]	225	INTEGER(iwp) :: s_ind !<
[1682]	226	INTEGER(iwp) :: sender !<
	227	INTEGER(iwp) :: ind(4) !<
[1320]	228
[1682]	229	LOGICAL :: found !<
	230	LOGICAL :: resorted !<
	231	LOGICAL :: two_d !<
[1320]	232
[1682]	233	REAL(wp) :: mean_r !<
	234	REAL(wp) :: s_r2 !<
	235	REAL(wp) :: s_r3 !<
[1320]	236
[1682]	237	REAL(wp), DIMENSION(:), ALLOCATABLE :: level_z !<
	238	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d !<
	239	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d_l !<
	240	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !<
	241	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections !<
	242	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections_l !<
[1359]	243
[1]	244	#if defined( __parallel )
[1682]	245	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: total_2d !<
[1]	246	#endif
[1682]	247	REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !<
[1]	248
	249	NAMELIST /LOCAL/ rtext
	250
	251	!
	252	!-- Immediate return, if no output is requested (no respective sections
	253	!-- found in parameter data_output)
	254	IF ( mode == 'xy' .AND. .NOT. data_output_xy(av) ) RETURN
	255	IF ( mode == 'xz' .AND. .NOT. data_output_xz(av) ) RETURN
	256	IF ( mode == 'yz' .AND. .NOT. data_output_yz(av) ) RETURN
	257
[1308]	258	CALL cpu_log (log_point(3),'data_output_2d','start')
	259
[1]	260	two_d = .FALSE. ! local variable to distinguish between output of pure 2D
	261	! arrays and cross-sections of 3D arrays.
	262
	263	!
	264	!-- Depending on the orientation of the cross-section, the respective output
	265	!-- files have to be opened.
	266	SELECT CASE ( mode )
	267
	268	CASE ( 'xy' )
[1960]	269	s_ind = 1
[667]	270	ALLOCATE( level_z(nzb:nzt+1), local_2d(nxlg:nxrg,nysg:nyng) )
[1]	271
[1308]	272	IF ( netcdf_data_format > 4 ) THEN
	273	ns = 1
[1960]	274	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	275	ns = ns + 1
	276	ENDDO
	277	ns = ns - 1
	278	ALLOCATE( local_2d_sections(nxlg:nxrg,nysg:nyng,1:ns) )
[1353]	279	local_2d_sections = 0.0_wp
[1308]	280	ENDIF
	281
[493]	282	!
[1031]	283	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	284	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	285	CALL check_open( 101+av*10 )
	286	ENDIF
[1]	287	IF ( data_output_2d_on_each_pe ) THEN
	288	CALL check_open( 21 )
	289	ELSE
	290	IF ( myid == 0 ) THEN
	291	#if defined( __parallel )
[667]	292	ALLOCATE( total_2d(-nbgp:nx+nbgp,-nbgp:ny+nbgp) )
[1]	293	#endif
	294	ENDIF
	295	ENDIF
	296
	297	CASE ( 'xz' )
[1960]	298	s_ind = 2
[667]	299	ALLOCATE( local_2d(nxlg:nxrg,nzb:nzt+1) )
[1]	300
[1308]	301	IF ( netcdf_data_format > 4 ) THEN
	302	ns = 1
[1960]	303	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	304	ns = ns + 1
	305	ENDDO
	306	ns = ns - 1
	307	ALLOCATE( local_2d_sections(nxlg:nxrg,1:ns,nzb:nzt+1) )
	308	ALLOCATE( local_2d_sections_l(nxlg:nxrg,1:ns,nzb:nzt+1) )
[1353]	309	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
[1308]	310	ENDIF
	311
[493]	312	!
[1031]	313	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	314	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	315	CALL check_open( 102+av*10 )
	316	ENDIF
[1]	317
	318	IF ( data_output_2d_on_each_pe ) THEN
	319	CALL check_open( 22 )
	320	ELSE
	321	IF ( myid == 0 ) THEN
	322	#if defined( __parallel )
[667]	323	ALLOCATE( total_2d(-nbgp:nx+nbgp,nzb:nzt+1) )
[1]	324	#endif
	325	ENDIF
	326	ENDIF
	327
	328	CASE ( 'yz' )
[1960]	329	s_ind = 3
[667]	330	ALLOCATE( local_2d(nysg:nyng,nzb:nzt+1) )
[1]	331
[1308]	332	IF ( netcdf_data_format > 4 ) THEN
	333	ns = 1
[1960]	334	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	335	ns = ns + 1
	336	ENDDO
	337	ns = ns - 1
	338	ALLOCATE( local_2d_sections(1:ns,nysg:nyng,nzb:nzt+1) )
	339	ALLOCATE( local_2d_sections_l(1:ns,nysg:nyng,nzb:nzt+1) )
[1353]	340	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
[1308]	341	ENDIF
	342
[493]	343	!
[1031]	344	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	345	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	346	CALL check_open( 103+av*10 )
	347	ENDIF
[1]	348
	349	IF ( data_output_2d_on_each_pe ) THEN
	350	CALL check_open( 23 )
	351	ELSE
	352	IF ( myid == 0 ) THEN
	353	#if defined( __parallel )
[667]	354	ALLOCATE( total_2d(-nbgp:ny+nbgp,nzb:nzt+1) )
[1]	355	#endif
	356	ENDIF
	357	ENDIF
	358
	359	CASE DEFAULT
[254]	360	message_string = 'unknown cross-section: ' // TRIM( mode )
	361	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
[1]	362
	363	END SELECT
	364
	365	!
[1745]	366	!-- For parallel netcdf output the time axis must be limited. Return, if this
	367	!-- limit is exceeded. This could be the case, if the simulated time exceeds
	368	!-- the given end time by the length of the given output interval.
	369	IF ( netcdf_data_format > 4 ) THEN
	370	IF ( mode == 'xy' .AND. do2d_xy_time_count(av) + 1 > &
	371	ntdim_2d_xy(av) ) THEN
	372	WRITE ( message_string, * ) 'Output of xy cross-sections is not ', &
	373	'given at t=', simulated_time, '&because the', &
	374	' maximum number of output time levels is exceeded.'
	375	CALL message( 'data_output_2d', 'PA0384', 0, 1, 0, 6, 0 )
	376	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	377	RETURN
	378	ENDIF
	379	IF ( mode == 'xz' .AND. do2d_xz_time_count(av) + 1 > &
	380	ntdim_2d_xz(av) ) THEN
	381	WRITE ( message_string, * ) 'Output of xz cross-sections is not ', &
	382	'given at t=', simulated_time, '&because the', &
	383	' maximum number of output time levels is exceeded.'
	384	CALL message( 'data_output_2d', 'PA0385', 0, 1, 0, 6, 0 )
	385	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	386	RETURN
	387	ENDIF
	388	IF ( mode == 'yz' .AND. do2d_yz_time_count(av) + 1 > &
	389	ntdim_2d_yz(av) ) THEN
	390	WRITE ( message_string, * ) 'Output of yz cross-sections is not ', &
	391	'given at t=', simulated_time, '&because the', &
	392	' maximum number of output time levels is exceeded.'
	393	CALL message( 'data_output_2d', 'PA0386', 0, 1, 0, 6, 0 )
	394	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	395	RETURN
	396	ENDIF
	397	ENDIF
	398
	399	!
[1]	400	!-- Allocate a temporary array for resorting (kji -> ijk).
[667]	401	ALLOCATE( local_pf(nxlg:nxrg,nysg:nyng,nzb:nzt+1) )
[1]	402
	403	!
	404	!-- Loop of all variables to be written.
	405	!-- Output dimensions chosen
	406	if = 1
	407	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
	408	do2d_mode = do2d(av,if)(l-1:l)
	409
	410	DO WHILE ( do2d(av,if)(1:1) /= ' ' )
	411
	412	IF ( do2d_mode == mode ) THEN
[1551]	413
	414	nzb_do = nzb
	415	nzt_do = nzt+1
[1]	416	!
	417	!-- Store the array chosen on the temporary array.
	418	resorted = .FALSE.
	419	SELECT CASE ( TRIM( do2d(av,if) ) )
	420
	421	CASE ( 'e_xy', 'e_xz', 'e_yz' )
	422	IF ( av == 0 ) THEN
	423	to_be_resorted => e
	424	ELSE
	425	to_be_resorted => e_av
	426	ENDIF
	427	IF ( mode == 'xy' ) level_z = zu
	428
[771]	429	CASE ( 'lpt_xy', 'lpt_xz', 'lpt_yz' )
	430	IF ( av == 0 ) THEN
	431	to_be_resorted => pt
	432	ELSE
	433	to_be_resorted => lpt_av
	434	ENDIF
	435	IF ( mode == 'xy' ) level_z = zu
	436
[1]	437	CASE ( 'lwp*_xy' ) ! 2d-array
	438	IF ( av == 0 ) THEN
[667]	439	DO i = nxlg, nxrg
	440	DO j = nysg, nyng
[1320]	441	local_pf(i,j,nzb+1) = SUM( ql(nzb:nzt,j,i) * &
[1]	442	dzw(1:nzt+1) )
	443	ENDDO
	444	ENDDO
	445	ELSE
[667]	446	DO i = nxlg, nxrg
	447	DO j = nysg, nyng
[1]	448	local_pf(i,j,nzb+1) = lwp_av(j,i)
	449	ENDDO
	450	ENDDO
	451	ENDIF
	452	resorted = .TRUE.
	453	two_d = .TRUE.
	454	level_z(nzb+1) = zu(nzb+1)
	455
[1053]	456	CASE ( 'nr_xy', 'nr_xz', 'nr_yz' )
	457	IF ( av == 0 ) THEN
	458	to_be_resorted => nr
	459	ELSE
	460	to_be_resorted => nr_av
	461	ENDIF
	462	IF ( mode == 'xy' ) level_z = zu
	463
[1691]	464	CASE ( 'ol*_xy' ) ! 2d-array
	465	IF ( av == 0 ) THEN
	466	DO i = nxlg, nxrg
	467	DO j = nysg, nyng
	468	local_pf(i,j,nzb+1) = ol(j,i)
	469	ENDDO
	470	ENDDO
	471	ELSE
	472	DO i = nxlg, nxrg
	473	DO j = nysg, nyng
	474	local_pf(i,j,nzb+1) = ol_av(j,i)
	475	ENDDO
	476	ENDDO
	477	ENDIF
	478	resorted = .TRUE.
	479	two_d = .TRUE.
	480	level_z(nzb+1) = zu(nzb+1)
	481
[1]	482	CASE ( 'p_xy', 'p_xz', 'p_yz' )
	483	IF ( av == 0 ) THEN
[729]	484	IF ( psolver /= 'sor' ) CALL exchange_horiz( p, nbgp )
[1]	485	to_be_resorted => p
	486	ELSE
[729]	487	IF ( psolver /= 'sor' ) CALL exchange_horiz( p_av, nbgp )
[1]	488	to_be_resorted => p_av
	489	ENDIF
	490	IF ( mode == 'xy' ) level_z = zu
	491
	492	CASE ( 'pc_xy', 'pc_xz', 'pc_yz' ) ! particle concentration
	493	IF ( av == 0 ) THEN
[215]	494	IF ( simulated_time >= particle_advection_start ) THEN
	495	tend = prt_count
[667]	496	CALL exchange_horiz( tend, nbgp )
[215]	497	ELSE
[1353]	498	tend = 0.0_wp
[215]	499	ENDIF
[667]	500	DO i = nxlg, nxrg
	501	DO j = nysg, nyng
[1]	502	DO k = nzb, nzt+1
	503	local_pf(i,j,k) = tend(k,j,i)
	504	ENDDO
	505	ENDDO
	506	ENDDO
	507	resorted = .TRUE.
	508	ELSE
[667]	509	CALL exchange_horiz( pc_av, nbgp )
[1]	510	to_be_resorted => pc_av
	511	ENDIF
	512
[1359]	513	CASE ( 'pr_xy', 'pr_xz', 'pr_yz' ) ! mean particle radius (effective radius)
[1]	514	IF ( av == 0 ) THEN
[215]	515	IF ( simulated_time >= particle_advection_start ) THEN
	516	DO i = nxl, nxr
	517	DO j = nys, nyn
	518	DO k = nzb, nzt+1
[1359]	519	number_of_particles = prt_count(k,j,i)
	520	IF (number_of_particles <= 0) CYCLE
	521	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
	522	s_r2 = 0.0_wp
[1353]	523	s_r3 = 0.0_wp
[1359]	524	DO n = 1, number_of_particles
	525	IF ( particles(n)%particle_mask ) THEN
	526	s_r2 = s_r2 + particles(n)%radius*2 &
	527	particles(n)%weight_factor
	528	s_r3 = s_r3 + particles(n)%radius*3 &
	529	particles(n)%weight_factor
	530	ENDIF
[215]	531	ENDDO
[1359]	532	IF ( s_r2 > 0.0_wp ) THEN
	533	mean_r = s_r3 / s_r2
[215]	534	ELSE
[1353]	535	mean_r = 0.0_wp
[215]	536	ENDIF
	537	tend(k,j,i) = mean_r
[1]	538	ENDDO
	539	ENDDO
	540	ENDDO
[667]	541	CALL exchange_horiz( tend, nbgp )
[215]	542	ELSE
[1353]	543	tend = 0.0_wp
[1359]	544	ENDIF
[667]	545	DO i = nxlg, nxrg
	546	DO j = nysg, nyng
[1]	547	DO k = nzb, nzt+1
	548	local_pf(i,j,k) = tend(k,j,i)
	549	ENDDO
	550	ENDDO
	551	ENDDO
	552	resorted = .TRUE.
	553	ELSE
[667]	554	CALL exchange_horiz( pr_av, nbgp )
[1]	555	to_be_resorted => pr_av
	556	ENDIF
	557
[72]	558	CASE ( 'pra*_xy' ) ! 2d-array / integral quantity => no av
	559	CALL exchange_horiz_2d( precipitation_amount )
[667]	560	DO i = nxlg, nxrg
	561	DO j = nysg, nyng
[72]	562	local_pf(i,j,nzb+1) = precipitation_amount(j,i)
	563	ENDDO
	564	ENDDO
[1353]	565	precipitation_amount = 0.0_wp ! reset for next integ. interval
[72]	566	resorted = .TRUE.
	567	two_d = .TRUE.
	568	level_z(nzb+1) = zu(nzb+1)
	569
	570	CASE ( 'prr*_xy' ) ! 2d-array
[1822]	571	IF ( av == 0 ) THEN
	572	CALL exchange_horiz_2d( prr(nzb+1,:,:) )
	573	DO i = nxlg, nxrg
	574	DO j = nysg, nyng
	575	local_pf(i,j,nzb+1) = prr(nzb+1,j,i) * hyrho(nzb+1)
[1053]	576	ENDDO
[1822]	577	ENDDO
[1053]	578	ELSE
[1822]	579	CALL exchange_horiz_2d( prr_av(nzb+1,:,:) )
	580	DO i = nxlg, nxrg
	581	DO j = nysg, nyng
	582	local_pf(i,j,nzb+1) = prr_av(nzb+1,j,i) * hyrho(nzb+1)
[1053]	583	ENDDO
[1822]	584	ENDDO
[1053]	585	ENDIF
	586	resorted = .TRUE.
	587	two_d = .TRUE.
	588	level_z(nzb+1) = zu(nzb+1)
	589
	590	CASE ( 'prr_xy', 'prr_xz', 'prr_yz' )
[72]	591	IF ( av == 0 ) THEN
[1053]	592	CALL exchange_horiz( prr, nbgp )
[667]	593	DO i = nxlg, nxrg
	594	DO j = nysg, nyng
[1053]	595	DO k = nzb, nzt+1
[1822]	596	local_pf(i,j,k) = prr(k,j,i) * hyrho(nzb+1)
[1053]	597	ENDDO
[72]	598	ENDDO
	599	ENDDO
	600	ELSE
[1053]	601	CALL exchange_horiz( prr_av, nbgp )
[667]	602	DO i = nxlg, nxrg
	603	DO j = nysg, nyng
[1053]	604	DO k = nzb, nzt+1
[1822]	605	local_pf(i,j,k) = prr_av(k,j,i) * hyrho(nzb+1)
[1053]	606	ENDDO
[72]	607	ENDDO
	608	ENDDO
	609	ENDIF
	610	resorted = .TRUE.
[1053]	611	IF ( mode == 'xy' ) level_z = zu
[72]	612
[1]	613	CASE ( 'pt_xy', 'pt_xz', 'pt_yz' )
	614	IF ( av == 0 ) THEN
	615	IF ( .NOT. cloud_physics ) THEN
	616	to_be_resorted => pt
	617	ELSE
[667]	618	DO i = nxlg, nxrg
	619	DO j = nysg, nyng
[1]	620	DO k = nzb, nzt+1
[1320]	621	local_pf(i,j,k) = pt(k,j,i) + l_d_cp * &
	622	pt_d_t(k) * &
[1]	623	ql(k,j,i)
	624	ENDDO
	625	ENDDO
	626	ENDDO
	627	resorted = .TRUE.
	628	ENDIF
	629	ELSE
	630	to_be_resorted => pt_av
	631	ENDIF
	632	IF ( mode == 'xy' ) level_z = zu
	633
	634	CASE ( 'q_xy', 'q_xz', 'q_yz' )
	635	IF ( av == 0 ) THEN
	636	to_be_resorted => q
	637	ELSE
	638	to_be_resorted => q_av
	639	ENDIF
	640	IF ( mode == 'xy' ) level_z = zu
	641
[1053]	642	CASE ( 'qc_xy', 'qc_xz', 'qc_yz' )
[1]	643	IF ( av == 0 ) THEN
[1115]	644	to_be_resorted => qc
[1]	645	ELSE
[1115]	646	to_be_resorted => qc_av
[1]	647	ENDIF
	648	IF ( mode == 'xy' ) level_z = zu
	649
[1053]	650	CASE ( 'ql_xy', 'ql_xz', 'ql_yz' )
	651	IF ( av == 0 ) THEN
[1115]	652	to_be_resorted => ql
[1053]	653	ELSE
[1115]	654	to_be_resorted => ql_av
[1053]	655	ENDIF
	656	IF ( mode == 'xy' ) level_z = zu
	657
[1]	658	CASE ( 'ql_c_xy', 'ql_c_xz', 'ql_c_yz' )
	659	IF ( av == 0 ) THEN
	660	to_be_resorted => ql_c
	661	ELSE
	662	to_be_resorted => ql_c_av
	663	ENDIF
	664	IF ( mode == 'xy' ) level_z = zu
	665
	666	CASE ( 'ql_v_xy', 'ql_v_xz', 'ql_v_yz' )
	667	IF ( av == 0 ) THEN
	668	to_be_resorted => ql_v
	669	ELSE
	670	to_be_resorted => ql_v_av
	671	ENDIF
	672	IF ( mode == 'xy' ) level_z = zu
	673
	674	CASE ( 'ql_vp_xy', 'ql_vp_xz', 'ql_vp_yz' )
	675	IF ( av == 0 ) THEN
[1007]	676	IF ( simulated_time >= particle_advection_start ) THEN
	677	DO i = nxl, nxr
	678	DO j = nys, nyn
	679	DO k = nzb, nzt+1
[1359]	680	number_of_particles = prt_count(k,j,i)
	681	IF (number_of_particles <= 0) CYCLE
	682	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
	683	DO n = 1, number_of_particles
	684	IF ( particles(n)%particle_mask ) THEN
	685	tend(k,j,i) = tend(k,j,i) + &
	686	particles(n)%weight_factor / &
	687	prt_count(k,j,i)
	688	ENDIF
[1007]	689	ENDDO
	690	ENDDO
	691	ENDDO
	692	ENDDO
	693	CALL exchange_horiz( tend, nbgp )
	694	ELSE
[1353]	695	tend = 0.0_wp
[1359]	696	ENDIF
[1007]	697	DO i = nxlg, nxrg
	698	DO j = nysg, nyng
	699	DO k = nzb, nzt+1
	700	local_pf(i,j,k) = tend(k,j,i)
	701	ENDDO
	702	ENDDO
	703	ENDDO
	704	resorted = .TRUE.
	705	ELSE
	706	CALL exchange_horiz( ql_vp_av, nbgp )
[1]	707	to_be_resorted => ql_vp
	708	ENDIF
	709	IF ( mode == 'xy' ) level_z = zu
	710
[1053]	711	CASE ( 'qr_xy', 'qr_xz', 'qr_yz' )
	712	IF ( av == 0 ) THEN
	713	to_be_resorted => qr
	714	ELSE
	715	to_be_resorted => qr_av
	716	ENDIF
	717	IF ( mode == 'xy' ) level_z = zu
	718
[354]	719	CASE ( 'qsws*_xy' ) ! 2d-array
	720	IF ( av == 0 ) THEN
[667]	721	DO i = nxlg, nxrg
	722	DO j = nysg, nyng
[354]	723	local_pf(i,j,nzb+1) = qsws(j,i)
	724	ENDDO
	725	ENDDO
	726	ELSE
[667]	727	DO i = nxlg, nxrg
	728	DO j = nysg, nyng
[354]	729	local_pf(i,j,nzb+1) = qsws_av(j,i)
	730	ENDDO
	731	ENDDO
	732	ENDIF
	733	resorted = .TRUE.
	734	two_d = .TRUE.
	735	level_z(nzb+1) = zu(nzb+1)
	736
[1]	737	CASE ( 'qv_xy', 'qv_xz', 'qv_yz' )
	738	IF ( av == 0 ) THEN
[667]	739	DO i = nxlg, nxrg
	740	DO j = nysg, nyng
[1]	741	DO k = nzb, nzt+1
	742	local_pf(i,j,k) = q(k,j,i) - ql(k,j,i)
	743	ENDDO
	744	ENDDO
	745	ENDDO
	746	resorted = .TRUE.
	747	ELSE
	748	to_be_resorted => qv_av
	749	ENDIF
	750	IF ( mode == 'xy' ) level_z = zu
	751
[1551]	752	CASE ( 'rad_net*_xy' ) ! 2d-array
	753	IF ( av == 0 ) THEN
	754	DO i = nxlg, nxrg
	755	DO j = nysg, nyng
	756	local_pf(i,j,nzb+1) = rad_net(j,i)
	757	ENDDO
	758	ENDDO
	759	ELSE
	760	DO i = nxlg, nxrg
	761	DO j = nysg, nyng
	762	local_pf(i,j,nzb+1) = rad_net_av(j,i)
	763	ENDDO
	764	ENDDO
	765	ENDIF
	766	resorted = .TRUE.
	767	two_d = .TRUE.
	768	level_z(nzb+1) = zu(nzb+1)
	769
[1585]	770
	771	CASE ( 'rad_lw_in_xy', 'rad_lw_in_xz', 'rad_lw_in_yz' )
	772	IF ( av == 0 ) THEN
	773	to_be_resorted => rad_lw_in
[1551]	774	ELSE
[1585]	775	to_be_resorted => rad_lw_in_av
[1551]	776	ENDIF
[1701]	777	IF ( mode == 'xy' ) level_z = zu
[1551]	778
[1585]	779	CASE ( 'rad_lw_out_xy', 'rad_lw_out_xz', 'rad_lw_out_yz' )
	780	IF ( av == 0 ) THEN
	781	to_be_resorted => rad_lw_out
	782	ELSE
	783	to_be_resorted => rad_lw_out_av
	784	ENDIF
[1701]	785	IF ( mode == 'xy' ) level_z = zu
[1585]	786
[1691]	787	CASE ( 'rad_lw_cs_hr_xy', 'rad_lw_cs_hr_xz', 'rad_lw_cs_hr_yz' )
	788	IF ( av == 0 ) THEN
	789	to_be_resorted => rad_lw_cs_hr
	790	ELSE
	791	to_be_resorted => rad_lw_cs_hr_av
	792	ENDIF
[1701]	793	IF ( mode == 'xy' ) level_z = zw
[1691]	794
	795	CASE ( 'rad_lw_hr_xy', 'rad_lw_hr_xz', 'rad_lw_hr_yz' )
	796	IF ( av == 0 ) THEN
	797	to_be_resorted => rad_lw_hr
	798	ELSE
	799	to_be_resorted => rad_lw_hr_av
	800	ENDIF
[1701]	801	IF ( mode == 'xy' ) level_z = zw
[1691]	802
[1585]	803	CASE ( 'rad_sw_in_xy', 'rad_sw_in_xz', 'rad_sw_in_yz' )
	804	IF ( av == 0 ) THEN
	805	to_be_resorted => rad_sw_in
	806	ELSE
	807	to_be_resorted => rad_sw_in_av
	808	ENDIF
[1701]	809	IF ( mode == 'xy' ) level_z = zu
[1585]	810
	811	CASE ( 'rad_sw_out_xy', 'rad_sw_out_xz', 'rad_sw_out_yz' )
	812	IF ( av == 0 ) THEN
	813	to_be_resorted => rad_sw_out
	814	ELSE
	815	to_be_resorted => rad_sw_out_av
	816	ENDIF
[1701]	817	IF ( mode == 'xy' ) level_z = zu
[1585]	818
[1691]	819	CASE ( 'rad_sw_cs_hr_xy', 'rad_sw_cs_hr_xz', 'rad_sw_cs_hr_yz' )
	820	IF ( av == 0 ) THEN
	821	to_be_resorted => rad_sw_cs_hr
	822	ELSE
	823	to_be_resorted => rad_sw_cs_hr_av
	824	ENDIF
[1701]	825	IF ( mode == 'xy' ) level_z = zw
[1691]	826
	827	CASE ( 'rad_sw_hr_xy', 'rad_sw_hr_xz', 'rad_sw_hr_yz' )
	828	IF ( av == 0 ) THEN
	829	to_be_resorted => rad_sw_hr
	830	ELSE
	831	to_be_resorted => rad_sw_hr_av
	832	ENDIF
[1701]	833	IF ( mode == 'xy' ) level_z = zw
[1691]	834
[96]	835	CASE ( 'rho_xy', 'rho_xz', 'rho_yz' )
	836	IF ( av == 0 ) THEN
	837	to_be_resorted => rho
	838	ELSE
	839	to_be_resorted => rho_av
	840	ENDIF
	841
[1]	842	CASE ( 's_xy', 's_xz', 's_yz' )
	843	IF ( av == 0 ) THEN
[1960]	844	to_be_resorted => s
[1]	845	ELSE
[355]	846	to_be_resorted => s_av
[1]	847	ENDIF
	848
[96]	849	CASE ( 'sa_xy', 'sa_xz', 'sa_yz' )
	850	IF ( av == 0 ) THEN
	851	to_be_resorted => sa
	852	ELSE
	853	to_be_resorted => sa_av
	854	ENDIF
	855
[354]	856	CASE ( 'shf*_xy' ) ! 2d-array
	857	IF ( av == 0 ) THEN
[667]	858	DO i = nxlg, nxrg
	859	DO j = nysg, nyng
[354]	860	local_pf(i,j,nzb+1) = shf(j,i)
	861	ENDDO
	862	ENDDO
	863	ELSE
[667]	864	DO i = nxlg, nxrg
	865	DO j = nysg, nyng
[354]	866	local_pf(i,j,nzb+1) = shf_av(j,i)
	867	ENDDO
	868	ENDDO
	869	ENDIF
	870	resorted = .TRUE.
	871	two_d = .TRUE.
	872	level_z(nzb+1) = zu(nzb+1)
[1960]	873
	874	CASE ( 'ssws*_xy' ) ! 2d-array
	875	IF ( av == 0 ) THEN
	876	DO i = nxlg, nxrg
	877	DO j = nysg, nyng
	878	local_pf(i,j,nzb+1) = ssws(j,i)
	879	ENDDO
	880	ENDDO
	881	ELSE
	882	DO i = nxlg, nxrg
	883	DO j = nysg, nyng
	884	local_pf(i,j,nzb+1) = ssws_av(j,i)
	885	ENDDO
	886	ENDDO
	887	ENDIF
	888	resorted = .TRUE.
	889	two_d = .TRUE.
	890	level_z(nzb+1) = zu(nzb+1)
[1551]	891
[1]	892	CASE ( 't*_xy' ) ! 2d-array
	893	IF ( av == 0 ) THEN
[667]	894	DO i = nxlg, nxrg
	895	DO j = nysg, nyng
[1]	896	local_pf(i,j,nzb+1) = ts(j,i)
	897	ENDDO
	898	ENDDO
	899	ELSE
[667]	900	DO i = nxlg, nxrg
	901	DO j = nysg, nyng
[1]	902	local_pf(i,j,nzb+1) = ts_av(j,i)
	903	ENDDO
	904	ENDDO
	905	ENDIF
	906	resorted = .TRUE.
	907	two_d = .TRUE.
	908	level_z(nzb+1) = zu(nzb+1)
	909
	910	CASE ( 'u_xy', 'u_xz', 'u_yz' )
	911	IF ( av == 0 ) THEN
	912	to_be_resorted => u
	913	ELSE
	914	to_be_resorted => u_av
	915	ENDIF
	916	IF ( mode == 'xy' ) level_z = zu
	917	!
	918	!-- Substitute the values generated by "mirror" boundary condition
	919	!-- at the bottom boundary by the real surface values.
	920	IF ( do2d(av,if) == 'u_xz' .OR. do2d(av,if) == 'u_yz' ) THEN
[1353]	921	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
[1]	922	ENDIF
	923
	924	CASE ( 'u*_xy' ) ! 2d-array
	925	IF ( av == 0 ) THEN
[667]	926	DO i = nxlg, nxrg
	927	DO j = nysg, nyng
[1]	928	local_pf(i,j,nzb+1) = us(j,i)
	929	ENDDO
	930	ENDDO
	931	ELSE
[667]	932	DO i = nxlg, nxrg
	933	DO j = nysg, nyng
[1]	934	local_pf(i,j,nzb+1) = us_av(j,i)
	935	ENDDO
	936	ENDDO
	937	ENDIF
	938	resorted = .TRUE.
	939	two_d = .TRUE.
	940	level_z(nzb+1) = zu(nzb+1)
	941
	942	CASE ( 'v_xy', 'v_xz', 'v_yz' )
	943	IF ( av == 0 ) THEN
	944	to_be_resorted => v
	945	ELSE
	946	to_be_resorted => v_av
	947	ENDIF
	948	IF ( mode == 'xy' ) level_z = zu
	949	!
	950	!-- Substitute the values generated by "mirror" boundary condition
	951	!-- at the bottom boundary by the real surface values.
	952	IF ( do2d(av,if) == 'v_xz' .OR. do2d(av,if) == 'v_yz' ) THEN
[1353]	953	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
[1]	954	ENDIF
	955
	956	CASE ( 'vpt_xy', 'vpt_xz', 'vpt_yz' )
	957	IF ( av == 0 ) THEN
	958	to_be_resorted => vpt
	959	ELSE
	960	to_be_resorted => vpt_av
	961	ENDIF
	962	IF ( mode == 'xy' ) level_z = zu
	963
	964	CASE ( 'w_xy', 'w_xz', 'w_yz' )
	965	IF ( av == 0 ) THEN
	966	to_be_resorted => w
	967	ELSE
	968	to_be_resorted => w_av
	969	ENDIF
	970	IF ( mode == 'xy' ) level_z = zw
	971
[72]	972	CASE ( 'z0*_xy' ) ! 2d-array
	973	IF ( av == 0 ) THEN
[667]	974	DO i = nxlg, nxrg
	975	DO j = nysg, nyng
[72]	976	local_pf(i,j,nzb+1) = z0(j,i)
	977	ENDDO
	978	ENDDO
	979	ELSE
[667]	980	DO i = nxlg, nxrg
	981	DO j = nysg, nyng
[72]	982	local_pf(i,j,nzb+1) = z0_av(j,i)
	983	ENDDO
	984	ENDDO
	985	ENDIF
	986	resorted = .TRUE.
	987	two_d = .TRUE.
	988	level_z(nzb+1) = zu(nzb+1)
	989
[978]	990	CASE ( 'z0h*_xy' ) ! 2d-array
	991	IF ( av == 0 ) THEN
	992	DO i = nxlg, nxrg
	993	DO j = nysg, nyng
	994	local_pf(i,j,nzb+1) = z0h(j,i)
	995	ENDDO
	996	ENDDO
	997	ELSE
	998	DO i = nxlg, nxrg
	999	DO j = nysg, nyng
	1000	local_pf(i,j,nzb+1) = z0h_av(j,i)
	1001	ENDDO
	1002	ENDDO
	1003	ENDIF
	1004	resorted = .TRUE.
	1005	two_d = .TRUE.
	1006	level_z(nzb+1) = zu(nzb+1)
	1007
[1788]	1008	CASE ( 'z0q*_xy' ) ! 2d-array
	1009	IF ( av == 0 ) THEN
	1010	DO i = nxlg, nxrg
	1011	DO j = nysg, nyng
	1012	local_pf(i,j,nzb+1) = z0q(j,i)
	1013	ENDDO
	1014	ENDDO
	1015	ELSE
	1016	DO i = nxlg, nxrg
	1017	DO j = nysg, nyng
	1018	local_pf(i,j,nzb+1) = z0q_av(j,i)
	1019	ENDDO
	1020	ENDDO
	1021	ENDIF
	1022	resorted = .TRUE.
	1023	two_d = .TRUE.
	1024	level_z(nzb+1) = zu(nzb+1)
	1025
[1]	1026	CASE DEFAULT
[1972]	1027
[1]	1028	!
[1972]	1029	!-- Land surface model quantity
	1030	IF ( land_surface ) THEN
	1031	CALL lsm_data_output_2d( av, do2d(av,if), found, grid, mode,&
	1032	local_pf, two_d, nzb_do, nzt_do )
	1033	ENDIF
	1034
	1035	!
[1]	1036	!-- User defined quantity
[1972]	1037	IF ( .NOT. found ) THEN
	1038	CALL user_data_output_2d( av, do2d(av,if), found, grid, &
	1039	local_pf, two_d, nzb_do, nzt_do )
	1040	ENDIF
	1041
[1]	1042	resorted = .TRUE.
	1043
	1044	IF ( grid == 'zu' ) THEN
	1045	IF ( mode == 'xy' ) level_z = zu
	1046	ELSEIF ( grid == 'zw' ) THEN
	1047	IF ( mode == 'xy' ) level_z = zw
[343]	1048	ELSEIF ( grid == 'zu1' ) THEN
	1049	IF ( mode == 'xy' ) level_z(nzb+1) = zu(nzb+1)
[1551]	1050	ELSEIF ( grid == 'zs' ) THEN
	1051	IF ( mode == 'xy' ) level_z = zs
[1]	1052	ENDIF
	1053
	1054	IF ( .NOT. found ) THEN
[1320]	1055	message_string = 'no output provided for: ' // &
[274]	1056	TRIM( do2d(av,if) )
[254]	1057	CALL message( 'data_output_2d', 'PA0181', 0, 0, 0, 6, 0 )
[1]	1058	ENDIF
	1059
	1060	END SELECT
	1061
	1062	!
	1063	!-- Resort the array to be output, if not done above
	1064	IF ( .NOT. resorted ) THEN
[667]	1065	DO i = nxlg, nxrg
	1066	DO j = nysg, nyng
[1551]	1067	DO k = nzb_do, nzt_do
[1]	1068	local_pf(i,j,k) = to_be_resorted(k,j,i)
	1069	ENDDO
	1070	ENDDO
	1071	ENDDO
	1072	ENDIF
	1073
	1074	!
	1075	!-- Output of the individual cross-sections, depending on the cross-
	1076	!-- section mode chosen.
	1077	is = 1
[1960]	1078	loop1: DO WHILE ( section(is,s_ind) /= -9999 .OR. two_d )
[1]	1079
	1080	SELECT CASE ( mode )
	1081
	1082	CASE ( 'xy' )
	1083	!
	1084	!-- Determine the cross section index
	1085	IF ( two_d ) THEN
	1086	layer_xy = nzb+1
	1087	ELSE
[1960]	1088	layer_xy = section(is,s_ind)
[1]	1089	ENDIF
	1090
	1091	!
[1551]	1092	!-- Exit the loop for layers beyond the data output domain
	1093	!-- (used for soil model)
[1691]	1094	IF ( layer_xy > nzt_do ) THEN
[1551]	1095	EXIT loop1
	1096	ENDIF
	1097
	1098	!
[1308]	1099	!-- Update the netCDF xy cross section time axis.
	1100	!-- In case of parallel output, this is only done by PE0
	1101	!-- to increase the performance.
	1102	IF ( simulated_time /= do2d_xy_last_time(av) ) THEN
	1103	do2d_xy_time_count(av) = do2d_xy_time_count(av) + 1
	1104	do2d_xy_last_time(av) = simulated_time
	1105	IF ( myid == 0 ) THEN
[1327]	1106	IF ( .NOT. data_output_2d_on_each_pe &
	1107	.OR. netcdf_data_format > 4 ) &
[493]	1108	THEN
[1]	1109	#if defined( __netcdf )
	1110	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1111	id_var_time_xy(av), &
[291]	1112	(/ time_since_reference_point /), &
[1]	1113	start = (/ do2d_xy_time_count(av) /), &
	1114	count = (/ 1 /) )
[1783]	1115	CALL netcdf_handle_error( 'data_output_2d', 53 )
[1]	1116	#endif
	1117	ENDIF
	1118	ENDIF
	1119	ENDIF
	1120	!
	1121	!-- If required, carry out averaging along z
[1960]	1122	IF ( section(is,s_ind) == -1 .AND. .NOT. two_d ) THEN
[1]	1123
[1353]	1124	local_2d = 0.0_wp
[1]	1125	!
	1126	!-- Carry out the averaging (all data are on the PE)
[1551]	1127	DO k = nzb_do, nzt_do
[667]	1128	DO j = nysg, nyng
	1129	DO i = nxlg, nxrg
[1]	1130	local_2d(i,j) = local_2d(i,j) + local_pf(i,j,k)
	1131	ENDDO
	1132	ENDDO
	1133	ENDDO
	1134
[1551]	1135	local_2d = local_2d / ( nzt_do - nzb_do + 1.0_wp)
[1]	1136
	1137	ELSE
	1138	!
	1139	!-- Just store the respective section on the local array
	1140	local_2d = local_pf(:,:,layer_xy)
	1141
	1142	ENDIF
	1143
	1144	#if defined( __parallel )
[1327]	1145	IF ( netcdf_data_format > 4 ) THEN
[1]	1146	!
[1031]	1147	!-- Parallel output in netCDF4/HDF5 format.
[493]	1148	IF ( two_d ) THEN
	1149	iis = 1
	1150	ELSE
	1151	iis = is
	1152	ENDIF
	1153
[1]	1154	#if defined( __netcdf )
[1308]	1155	!
	1156	!-- For parallel output, all cross sections are first stored
	1157	!-- here on a local array and will be written to the output
	1158	!-- file afterwards to increase the performance.
	1159	DO i = nxlg, nxrg
	1160	DO j = nysg, nyng
	1161	local_2d_sections(i,j,iis) = local_2d(i,j)
	1162	ENDDO
	1163	ENDDO
[1]	1164	#endif
[493]	1165	ELSE
[1]	1166
[493]	1167	IF ( data_output_2d_on_each_pe ) THEN
[1]	1168	!
[493]	1169	!-- Output of partial arrays on each PE
	1170	#if defined( __netcdf )
[1327]	1171	IF ( myid == 0 ) THEN
[1320]	1172	WRITE ( 21 ) time_since_reference_point, &
[493]	1173	do2d_xy_time_count(av), av
	1174	ENDIF
	1175	#endif
[759]	1176	DO i = 0, io_blocks-1
	1177	IF ( i == io_group ) THEN
[1551]	1178	WRITE ( 21 ) nxlg, nxrg, nysg, nyng, nysg, nyng
[759]	1179	WRITE ( 21 ) local_2d
	1180	ENDIF
	1181	#if defined( __parallel )
	1182	CALL MPI_BARRIER( comm2d, ierr )
	1183	#endif
	1184	ENDDO
[559]	1185
[493]	1186	ELSE
[1]	1187	!
[493]	1188	!-- PE0 receives partial arrays from all processors and
	1189	!-- then outputs them. Here a barrier has to be set,
	1190	!-- because otherwise "-MPI- FATAL: Remote protocol queue
	1191	!-- full" may occur.
	1192	CALL MPI_BARRIER( comm2d, ierr )
	1193
[667]	1194	ngp = ( nxrg-nxlg+1 ) * ( nyng-nysg+1 )
[493]	1195	IF ( myid == 0 ) THEN
[1]	1196	!
[493]	1197	!-- Local array can be relocated directly.
[667]	1198	total_2d(nxlg:nxrg,nysg:nyng) = local_2d
[1]	1199	!
[493]	1200	!-- Receive data from all other PEs.
	1201	DO n = 1, numprocs-1
[1]	1202	!
[493]	1203	!-- Receive index limits first, then array.
	1204	!-- Index limits are received in arbitrary order from
	1205	!-- the PEs.
[1320]	1206	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1207	MPI_ANY_SOURCE, 0, comm2d, &
[493]	1208	status, ierr )
	1209	sender = status(MPI_SOURCE)
	1210	DEALLOCATE( local_2d )
	1211	ALLOCATE( local_2d(ind(1):ind(2),ind(3):ind(4)) )
[1320]	1212	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1213	MPI_REAL, sender, 1, comm2d, &
[493]	1214	status, ierr )
	1215	total_2d(ind(1):ind(2),ind(3):ind(4)) = local_2d
	1216	ENDDO
[1]	1217	!
[493]	1218	!-- Relocate the local array for the next loop increment
	1219	DEALLOCATE( local_2d )
[667]	1220	ALLOCATE( local_2d(nxlg:nxrg,nysg:nyng) )
[1]	1221
	1222	#if defined( __netcdf )
[1327]	1223	IF ( two_d ) THEN
	1224	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1225	id_var_do2d(av,if), &
	1226	total_2d(0:nx+1,0:ny+1), &
	1227	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
	1228	count = (/ nx+2, ny+2, 1, 1 /) )
	1229	ELSE
	1230	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1231	id_var_do2d(av,if), &
	1232	total_2d(0:nx+1,0:ny+1), &
	1233	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
	1234	count = (/ nx+2, ny+2, 1, 1 /) )
[1]	1235	ENDIF
[1783]	1236	CALL netcdf_handle_error( 'data_output_2d', 54 )
[1]	1237	#endif
	1238
[493]	1239	ELSE
[1]	1240	!
[493]	1241	!-- First send the local index limits to PE0
[667]	1242	ind(1) = nxlg; ind(2) = nxrg
	1243	ind(3) = nysg; ind(4) = nyng
[1320]	1244	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1245	comm2d, ierr )
[1]	1246	!
[493]	1247	!-- Send data to PE0
[1320]	1248	CALL MPI_SEND( local_2d(nxlg,nysg), ngp, &
[493]	1249	MPI_REAL, 0, 1, comm2d, ierr )
	1250	ENDIF
	1251	!
	1252	!-- A barrier has to be set, because otherwise some PEs may
	1253	!-- proceed too fast so that PE0 may receive wrong data on
	1254	!-- tag 0
	1255	CALL MPI_BARRIER( comm2d, ierr )
[1]	1256	ENDIF
[493]	1257
[1]	1258	ENDIF
	1259	#else
	1260	#if defined( __netcdf )
[1327]	1261	IF ( two_d ) THEN
	1262	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1263	id_var_do2d(av,if), &
	1264	local_2d(nxl:nxr+1,nys:nyn+1), &
	1265	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
	1266	count = (/ nx+2, ny+2, 1, 1 /) )
	1267	ELSE
	1268	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1269	id_var_do2d(av,if), &
	1270	local_2d(nxl:nxr+1,nys:nyn+1), &
	1271	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
	1272	count = (/ nx+2, ny+2, 1, 1 /) )
[1]	1273	ENDIF
[1783]	1274	CALL netcdf_handle_error( 'data_output_2d', 447 )
[1]	1275	#endif
	1276	#endif
	1277	do2d_xy_n = do2d_xy_n + 1
	1278	!
	1279	!-- For 2D-arrays (e.g. u*) only one cross-section is available.
	1280	!-- Hence exit loop of output levels.
	1281	IF ( two_d ) THEN
[1703]	1282	IF ( netcdf_data_format < 5 ) two_d = .FALSE.
[1]	1283	EXIT loop1
	1284	ENDIF
	1285
	1286	CASE ( 'xz' )
	1287	!
[1308]	1288	!-- Update the netCDF xz cross section time axis.
	1289	!-- In case of parallel output, this is only done by PE0
	1290	!-- to increase the performance.
	1291	IF ( simulated_time /= do2d_xz_last_time(av) ) THEN
	1292	do2d_xz_time_count(av) = do2d_xz_time_count(av) + 1
	1293	do2d_xz_last_time(av) = simulated_time
	1294	IF ( myid == 0 ) THEN
[1327]	1295	IF ( .NOT. data_output_2d_on_each_pe &
	1296	.OR. netcdf_data_format > 4 ) &
[493]	1297	THEN
[1]	1298	#if defined( __netcdf )
	1299	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1300	id_var_time_xz(av), &
[291]	1301	(/ time_since_reference_point /), &
[1]	1302	start = (/ do2d_xz_time_count(av) /), &
	1303	count = (/ 1 /) )
[1783]	1304	CALL netcdf_handle_error( 'data_output_2d', 56 )
[1]	1305	#endif
	1306	ENDIF
	1307	ENDIF
	1308	ENDIF
[667]	1309
[1]	1310	!
	1311	!-- If required, carry out averaging along y
[1960]	1312	IF ( section(is,s_ind) == -1 ) THEN
[1]	1313
[1551]	1314	ALLOCATE( local_2d_l(nxlg:nxrg,nzb_do:nzt_do) )
[1353]	1315	local_2d_l = 0.0_wp
[1551]	1316	ngp = ( nxrg-nxlg + 1 ) * ( nzt_do-nzb_do + 1 )
[1]	1317	!
	1318	!-- First local averaging on the PE
[1551]	1319	DO k = nzb_do, nzt_do
[1]	1320	DO j = nys, nyn
[667]	1321	DO i = nxlg, nxrg
[1320]	1322	local_2d_l(i,k) = local_2d_l(i,k) + &
[1]	1323	local_pf(i,j,k)
	1324	ENDDO
	1325	ENDDO
	1326	ENDDO
	1327	#if defined( __parallel )
	1328	!
	1329	!-- Now do the averaging over all PEs along y
[622]	1330	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1331	CALL MPI_ALLREDUCE( local_2d_l(nxlg,nzb_do), &
	1332	local_2d(nxlg,nzb_do), ngp, MPI_REAL, &
[1]	1333	MPI_SUM, comm1dy, ierr )
	1334	#else
	1335	local_2d = local_2d_l
	1336	#endif
[1353]	1337	local_2d = local_2d / ( ny + 1.0_wp )
[1]	1338
	1339	DEALLOCATE( local_2d_l )
	1340
	1341	ELSE
	1342	!
	1343	!-- Just store the respective section on the local array
	1344	!-- (but only if it is available on this PE!)
[1960]	1345	IF ( section(is,s_ind) >= nys .AND. section(is,s_ind) <= nyn ) &
[1]	1346	THEN
[1960]	1347	local_2d = local_pf(:,section(is,s_ind),nzb_do:nzt_do)
[1]	1348	ENDIF
	1349
	1350	ENDIF
	1351
	1352	#if defined( __parallel )
[1327]	1353	IF ( netcdf_data_format > 4 ) THEN
[1]	1354	!
[1031]	1355	!-- Output in netCDF4/HDF5 format.
[493]	1356	!-- Output only on those PEs where the respective cross
	1357	!-- sections reside. Cross sections averaged along y are
	1358	!-- output on the respective first PE along y (myidy=0).
[1960]	1359	IF ( ( section(is,s_ind) >= nys .AND. &
	1360	section(is,s_ind) <= nyn ) .OR. &
	1361	( section(is,s_ind) == -1 .AND. myidy == 0 ) ) THEN
[1]	1362	#if defined( __netcdf )
[493]	1363	!
[1308]	1364	!-- For parallel output, all cross sections are first
	1365	!-- stored here on a local array and will be written to the
	1366	!-- output file afterwards to increase the performance.
	1367	DO i = nxlg, nxrg
[1551]	1368	DO k = nzb_do, nzt_do
[1308]	1369	local_2d_sections_l(i,is,k) = local_2d(i,k)
	1370	ENDDO
	1371	ENDDO
[1]	1372	#endif
	1373	ENDIF
	1374
	1375	ELSE
	1376
[493]	1377	IF ( data_output_2d_on_each_pe ) THEN
[1]	1378	!
[493]	1379	!-- Output of partial arrays on each PE. If the cross
	1380	!-- section does not reside on the PE, output special
	1381	!-- index values.
	1382	#if defined( __netcdf )
[1327]	1383	IF ( myid == 0 ) THEN
[1320]	1384	WRITE ( 22 ) time_since_reference_point, &
[493]	1385	do2d_xz_time_count(av), av
	1386	ENDIF
	1387	#endif
[759]	1388	DO i = 0, io_blocks-1
	1389	IF ( i == io_group ) THEN
[1960]	1390	IF ( ( section(is,s_ind) >= nys .AND. &
	1391	section(is,s_ind) <= nyn ) .OR. &
	1392	( section(is,s_ind) == -1 .AND. &
[1320]	1393	nys-1 == -1 ) ) &
[759]	1394	THEN
[1551]	1395	WRITE (22) nxlg, nxrg, nzb_do, nzt_do, nzb, nzt+1
[759]	1396	WRITE (22) local_2d
	1397	ELSE
[1551]	1398	WRITE (22) -1, -1, -1, -1, -1, -1
[759]	1399	ENDIF
	1400	ENDIF
	1401	#if defined( __parallel )
	1402	CALL MPI_BARRIER( comm2d, ierr )
	1403	#endif
	1404	ENDDO
[493]	1405
	1406	ELSE
[1]	1407	!
[493]	1408	!-- PE0 receives partial arrays from all processors of the
	1409	!-- respective cross section and outputs them. Here a
	1410	!-- barrier has to be set, because otherwise
	1411	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
	1412	CALL MPI_BARRIER( comm2d, ierr )
	1413
[1551]	1414	ngp = ( nxrg-nxlg + 1 ) * ( nzt_do-nzb_do + 1 )
[493]	1415	IF ( myid == 0 ) THEN
[1]	1416	!
[493]	1417	!-- Local array can be relocated directly.
[1960]	1418	IF ( ( section(is,s_ind) >= nys .AND. &
	1419	section(is,s_ind) <= nyn ) .OR. &
	1420	( section(is,s_ind) == -1 .AND. &
	1421	nys-1 == -1 ) ) THEN
[1551]	1422	total_2d(nxlg:nxrg,nzb_do:nzt_do) = local_2d
[493]	1423	ENDIF
[1]	1424	!
[493]	1425	!-- Receive data from all other PEs.
	1426	DO n = 1, numprocs-1
	1427	!
	1428	!-- Receive index limits first, then array.
	1429	!-- Index limits are received in arbitrary order from
	1430	!-- the PEs.
[1320]	1431	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1432	MPI_ANY_SOURCE, 0, comm2d, &
[1]	1433	status, ierr )
[493]	1434	!
	1435	!-- Not all PEs have data for XZ-cross-section.
	1436	IF ( ind(1) /= -9999 ) THEN
	1437	sender = status(MPI_SOURCE)
	1438	DEALLOCATE( local_2d )
[1320]	1439	ALLOCATE( local_2d(ind(1):ind(2), &
[493]	1440	ind(3):ind(4)) )
	1441	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1442	MPI_REAL, sender, 1, comm2d, &
	1443	status, ierr )
[1320]	1444	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
[493]	1445	local_2d
	1446	ENDIF
	1447	ENDDO
	1448	!
	1449	!-- Relocate the local array for the next loop increment
	1450	DEALLOCATE( local_2d )
[1551]	1451	ALLOCATE( local_2d(nxlg:nxrg,nzb_do:nzt_do) )
[1]	1452
	1453	#if defined( __netcdf )
[1327]	1454	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1455	id_var_do2d(av,if), &
[1551]	1456	total_2d(0:nx+1,nzb_do:nzt_do),&
[1327]	1457	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
[1551]	1458	count = (/ nx+2, 1, nzt_do-nzb_do+1, 1 /) )
[1783]	1459	CALL netcdf_handle_error( 'data_output_2d', 58 )
[1]	1460	#endif
	1461
[493]	1462	ELSE
[1]	1463	!
[493]	1464	!-- If the cross section resides on the PE, send the
	1465	!-- local index limits, otherwise send -9999 to PE0.
[1960]	1466	IF ( ( section(is,s_ind) >= nys .AND. &
	1467	section(is,s_ind) <= nyn ) .OR. &
	1468	( section(is,s_ind) == -1 .AND. nys-1 == -1 ) ) &
[493]	1469	THEN
[667]	1470	ind(1) = nxlg; ind(2) = nxrg
[1551]	1471	ind(3) = nzb_do; ind(4) = nzt_do
[493]	1472	ELSE
	1473	ind(1) = -9999; ind(2) = -9999
	1474	ind(3) = -9999; ind(4) = -9999
	1475	ENDIF
[1320]	1476	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1477	comm2d, ierr )
	1478	!
	1479	!-- If applicable, send data to PE0.
	1480	IF ( ind(1) /= -9999 ) THEN
[1551]	1481	CALL MPI_SEND( local_2d(nxlg,nzb_do), ngp, &
[493]	1482	MPI_REAL, 0, 1, comm2d, ierr )
	1483	ENDIF
[1]	1484	ENDIF
	1485	!
[493]	1486	!-- A barrier has to be set, because otherwise some PEs may
	1487	!-- proceed too fast so that PE0 may receive wrong data on
	1488	!-- tag 0
	1489	CALL MPI_BARRIER( comm2d, ierr )
[1]	1490	ENDIF
[493]	1491
[1]	1492	ENDIF
	1493	#else
	1494	#if defined( __netcdf )
[1327]	1495	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1496	id_var_do2d(av,if), &
[1551]	1497	local_2d(nxl:nxr+1,nzb_do:nzt_do), &
[1327]	1498	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
[1551]	1499	count = (/ nx+2, 1, nzt_do-nzb_do+1, 1 /) )
[1783]	1500	CALL netcdf_handle_error( 'data_output_2d', 451 )
[1]	1501	#endif
	1502	#endif
	1503	do2d_xz_n = do2d_xz_n + 1
	1504
	1505	CASE ( 'yz' )
	1506	!
[1308]	1507	!-- Update the netCDF yz cross section time axis.
	1508	!-- In case of parallel output, this is only done by PE0
	1509	!-- to increase the performance.
	1510	IF ( simulated_time /= do2d_yz_last_time(av) ) THEN
	1511	do2d_yz_time_count(av) = do2d_yz_time_count(av) + 1
	1512	do2d_yz_last_time(av) = simulated_time
	1513	IF ( myid == 0 ) THEN
[1327]	1514	IF ( .NOT. data_output_2d_on_each_pe &
	1515	.OR. netcdf_data_format > 4 ) &
[493]	1516	THEN
[1]	1517	#if defined( __netcdf )
	1518	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1519	id_var_time_yz(av), &
[291]	1520	(/ time_since_reference_point /), &
[1]	1521	start = (/ do2d_yz_time_count(av) /), &
	1522	count = (/ 1 /) )
[1783]	1523	CALL netcdf_handle_error( 'data_output_2d', 59 )
[1]	1524	#endif
	1525	ENDIF
	1526	ENDIF
[1308]	1527	ENDIF
[493]	1528
[1]	1529	!
	1530	!-- If required, carry out averaging along x
[1960]	1531	IF ( section(is,s_ind) == -1 ) THEN
[1]	1532
[1551]	1533	ALLOCATE( local_2d_l(nysg:nyng,nzb_do:nzt_do) )
[1353]	1534	local_2d_l = 0.0_wp
[1551]	1535	ngp = ( nyng-nysg+1 ) * ( nzt_do-nzb_do+1 )
[1]	1536	!
	1537	!-- First local averaging on the PE
[1551]	1538	DO k = nzb_do, nzt_do
[667]	1539	DO j = nysg, nyng
[1]	1540	DO i = nxl, nxr
[1320]	1541	local_2d_l(j,k) = local_2d_l(j,k) + &
[1]	1542	local_pf(i,j,k)
	1543	ENDDO
	1544	ENDDO
	1545	ENDDO
	1546	#if defined( __parallel )
	1547	!
	1548	!-- Now do the averaging over all PEs along x
[622]	1549	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1550	CALL MPI_ALLREDUCE( local_2d_l(nysg,nzb_do), &
	1551	local_2d(nysg,nzb_do), ngp, MPI_REAL, &
[1]	1552	MPI_SUM, comm1dx, ierr )
	1553	#else
	1554	local_2d = local_2d_l
	1555	#endif
[1353]	1556	local_2d = local_2d / ( nx + 1.0_wp )
[1]	1557
	1558	DEALLOCATE( local_2d_l )
	1559
	1560	ELSE
	1561	!
	1562	!-- Just store the respective section on the local array
	1563	!-- (but only if it is available on this PE!)
[1960]	1564	IF ( section(is,s_ind) >= nxl .AND. section(is,s_ind) <= nxr ) &
[1]	1565	THEN
[1960]	1566	local_2d = local_pf(section(is,s_ind),:,nzb_do:nzt_do)
[1]	1567	ENDIF
	1568
	1569	ENDIF
	1570
	1571	#if defined( __parallel )
[1327]	1572	IF ( netcdf_data_format > 4 ) THEN
[1]	1573	!
[1031]	1574	!-- Output in netCDF4/HDF5 format.
[493]	1575	!-- Output only on those PEs where the respective cross
	1576	!-- sections reside. Cross sections averaged along x are
	1577	!-- output on the respective first PE along x (myidx=0).
[1960]	1578	IF ( ( section(is,s_ind) >= nxl .AND. &
	1579	section(is,s_ind) <= nxr ) .OR. &
	1580	( section(is,s_ind) == -1 .AND. myidx == 0 ) ) THEN
[1]	1581	#if defined( __netcdf )
[493]	1582	!
[1308]	1583	!-- For parallel output, all cross sections are first
	1584	!-- stored here on a local array and will be written to the
	1585	!-- output file afterwards to increase the performance.
	1586	DO j = nysg, nyng
[1551]	1587	DO k = nzb_do, nzt_do
[1308]	1588	local_2d_sections_l(is,j,k) = local_2d(j,k)
	1589	ENDDO
	1590	ENDDO
[1]	1591	#endif
	1592	ENDIF
	1593
	1594	ELSE
	1595
[493]	1596	IF ( data_output_2d_on_each_pe ) THEN
[1]	1597	!
[493]	1598	!-- Output of partial arrays on each PE. If the cross
	1599	!-- section does not reside on the PE, output special
	1600	!-- index values.
	1601	#if defined( __netcdf )
[1327]	1602	IF ( myid == 0 ) THEN
[1320]	1603	WRITE ( 23 ) time_since_reference_point, &
[493]	1604	do2d_yz_time_count(av), av
	1605	ENDIF
	1606	#endif
[759]	1607	DO i = 0, io_blocks-1
	1608	IF ( i == io_group ) THEN
[1960]	1609	IF ( ( section(is,s_ind) >= nxl .AND. &
	1610	section(is,s_ind) <= nxr ) .OR. &
	1611	( section(is,s_ind) == -1 .AND. &
[1320]	1612	nxl-1 == -1 ) ) &
[759]	1613	THEN
[1551]	1614	WRITE (23) nysg, nyng, nzb_do, nzt_do, nzb, nzt+1
[759]	1615	WRITE (23) local_2d
	1616	ELSE
[1551]	1617	WRITE (23) -1, -1, -1, -1, -1, -1
[759]	1618	ENDIF
	1619	ENDIF
	1620	#if defined( __parallel )
	1621	CALL MPI_BARRIER( comm2d, ierr )
	1622	#endif
	1623	ENDDO
[493]	1624
	1625	ELSE
[1]	1626	!
[493]	1627	!-- PE0 receives partial arrays from all processors of the
	1628	!-- respective cross section and outputs them. Here a
	1629	!-- barrier has to be set, because otherwise
	1630	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
	1631	CALL MPI_BARRIER( comm2d, ierr )
	1632
[1551]	1633	ngp = ( nyng-nysg+1 ) * ( nzt_do-nzb_do+1 )
[493]	1634	IF ( myid == 0 ) THEN
[1]	1635	!
[493]	1636	!-- Local array can be relocated directly.
[1960]	1637	IF ( ( section(is,s_ind) >= nxl .AND. &
	1638	section(is,s_ind) <= nxr ) .OR. &
	1639	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
[493]	1640	THEN
[1551]	1641	total_2d(nysg:nyng,nzb_do:nzt_do) = local_2d
[493]	1642	ENDIF
[1]	1643	!
[493]	1644	!-- Receive data from all other PEs.
	1645	DO n = 1, numprocs-1
	1646	!
	1647	!-- Receive index limits first, then array.
	1648	!-- Index limits are received in arbitrary order from
	1649	!-- the PEs.
[1320]	1650	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1651	MPI_ANY_SOURCE, 0, comm2d, &
[1]	1652	status, ierr )
[493]	1653	!
	1654	!-- Not all PEs have data for YZ-cross-section.
	1655	IF ( ind(1) /= -9999 ) THEN
	1656	sender = status(MPI_SOURCE)
	1657	DEALLOCATE( local_2d )
[1320]	1658	ALLOCATE( local_2d(ind(1):ind(2), &
[493]	1659	ind(3):ind(4)) )
	1660	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1661	MPI_REAL, sender, 1, comm2d, &
	1662	status, ierr )
[1320]	1663	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
[493]	1664	local_2d
	1665	ENDIF
	1666	ENDDO
	1667	!
	1668	!-- Relocate the local array for the next loop increment
	1669	DEALLOCATE( local_2d )
[1551]	1670	ALLOCATE( local_2d(nysg:nyng,nzb_do:nzt_do) )
[1]	1671
	1672	#if defined( __netcdf )
[1327]	1673	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1674	id_var_do2d(av,if), &
[1551]	1675	total_2d(0:ny+1,nzb_do:nzt_do),&
[1327]	1676	start = (/ is, 1, 1, do2d_yz_time_count(av) /), &
[1551]	1677	count = (/ 1, ny+2, nzt_do-nzb_do+1, 1 /) )
[1783]	1678	CALL netcdf_handle_error( 'data_output_2d', 61 )
[1]	1679	#endif
	1680
[493]	1681	ELSE
[1]	1682	!
[493]	1683	!-- If the cross section resides on the PE, send the
	1684	!-- local index limits, otherwise send -9999 to PE0.
[1960]	1685	IF ( ( section(is,s_ind) >= nxl .AND. &
	1686	section(is,s_ind) <= nxr ) .OR. &
	1687	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
[493]	1688	THEN
[667]	1689	ind(1) = nysg; ind(2) = nyng
[1551]	1690	ind(3) = nzb_do; ind(4) = nzt_do
[493]	1691	ELSE
	1692	ind(1) = -9999; ind(2) = -9999
	1693	ind(3) = -9999; ind(4) = -9999
	1694	ENDIF
[1320]	1695	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1696	comm2d, ierr )
	1697	!
	1698	!-- If applicable, send data to PE0.
	1699	IF ( ind(1) /= -9999 ) THEN
[1551]	1700	CALL MPI_SEND( local_2d(nysg,nzb_do), ngp, &
[493]	1701	MPI_REAL, 0, 1, comm2d, ierr )
	1702	ENDIF
[1]	1703	ENDIF
	1704	!
[493]	1705	!-- A barrier has to be set, because otherwise some PEs may
	1706	!-- proceed too fast so that PE0 may receive wrong data on
	1707	!-- tag 0
	1708	CALL MPI_BARRIER( comm2d, ierr )
[1]	1709	ENDIF
[493]	1710
[1]	1711	ENDIF
	1712	#else
	1713	#if defined( __netcdf )
[1327]	1714	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1715	id_var_do2d(av,if), &
[1551]	1716	local_2d(nys:nyn+1,nzb_do:nzt_do), &
[1327]	1717	start = (/ is, 1, 1, do2d_xz_time_count(av) /), &
[1551]	1718	count = (/ 1, ny+2, nzt_do-nzb_do+1, 1 /) )
[1783]	1719	CALL netcdf_handle_error( 'data_output_2d', 452 )
[1]	1720	#endif
	1721	#endif
	1722	do2d_yz_n = do2d_yz_n + 1
	1723
	1724	END SELECT
	1725
	1726	is = is + 1
	1727	ENDDO loop1
	1728
[1308]	1729	!
	1730	!-- For parallel output, all data were collected before on a local array
	1731	!-- and are written now to the netcdf file. This must be done to increase
	1732	!-- the performance of the parallel output.
	1733	#if defined( __netcdf )
[1327]	1734	IF ( netcdf_data_format > 4 ) THEN
[1308]	1735
	1736	SELECT CASE ( mode )
	1737
	1738	CASE ( 'xy' )
	1739	IF ( two_d ) THEN
[1703]	1740	nis = 1
	1741	two_d = .FALSE.
[1308]	1742	ELSE
[1703]	1743	nis = ns
[1308]	1744	ENDIF
	1745	!
	1746	!-- Do not output redundant ghost point data except for the
	1747	!-- boundaries of the total domain.
	1748	IF ( nxr == nx .AND. nyn /= ny ) THEN
	1749	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1750	id_var_do2d(av,if), &
	1751	local_2d_sections(nxl:nxr+1, &
[1703]	1752	nys:nyn,1:nis), &
[1308]	1753	start = (/ nxl+1, nys+1, 1, &
	1754	do2d_xy_time_count(av) /), &
	1755	count = (/ nxr-nxl+2, &
[1703]	1756	nyn-nys+1, nis, 1 &
[1308]	1757	/) )
	1758	ELSEIF ( nxr /= nx .AND. nyn == ny ) THEN
	1759	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1760	id_var_do2d(av,if), &
	1761	local_2d_sections(nxl:nxr, &
[1703]	1762	nys:nyn+1,1:nis), &
[1308]	1763	start = (/ nxl+1, nys+1, 1, &
	1764	do2d_xy_time_count(av) /), &
	1765	count = (/ nxr-nxl+1, &
[1703]	1766	nyn-nys+2, nis, 1 &
[1308]	1767	/) )
	1768	ELSEIF ( nxr == nx .AND. nyn == ny ) THEN
	1769	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1770	id_var_do2d(av,if), &
	1771	local_2d_sections(nxl:nxr+1, &
[1703]	1772	nys:nyn+1,1:nis), &
[1308]	1773	start = (/ nxl+1, nys+1, 1, &
	1774	do2d_xy_time_count(av) /), &
	1775	count = (/ nxr-nxl+2, &
[1703]	1776	nyn-nys+2, nis, 1 &
[1308]	1777	/) )
	1778	ELSE
	1779	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1780	id_var_do2d(av,if), &
	1781	local_2d_sections(nxl:nxr, &
[1703]	1782	nys:nyn,1:nis), &
[1308]	1783	start = (/ nxl+1, nys+1, 1, &
	1784	do2d_xy_time_count(av) /), &
	1785	count = (/ nxr-nxl+1, &
[1703]	1786	nyn-nys+1, nis, 1 &
[1308]	1787	/) )
	1788	ENDIF
	1789
[1783]	1790	CALL netcdf_handle_error( 'data_output_2d', 55 )
[1308]	1791
	1792	CASE ( 'xz' )
	1793	!
	1794	!-- First, all PEs get the information of all cross-sections.
	1795	!-- Then the data are written to the output file by all PEs
	1796	!-- while NF90_COLLECTIVE is set in subroutine
	1797	!-- define_netcdf_header. Although redundant information are
	1798	!-- written to the output file in that case, the performance
	1799	!-- is significantly better compared to the case where only
	1800	!-- the first row of PEs in x-direction (myidx = 0) is given
	1801	!-- the output while NF90_INDEPENDENT is set.
	1802	IF ( npey /= 1 ) THEN
	1803
	1804	#if defined( __parallel )
	1805	!
	1806	!-- Distribute data over all PEs along y
[1551]	1807	ngp = ( nxrg-nxlg+1 ) * ( nzt_do-nzb_do+1 ) * ns
[1308]	1808	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1809	CALL MPI_ALLREDUCE( local_2d_sections_l(nxlg,1,nzb_do), &
	1810	local_2d_sections(nxlg,1,nzb_do), &
[1308]	1811	ngp, MPI_REAL, MPI_SUM, comm1dy, &
	1812	ierr )
	1813	#else
	1814	local_2d_sections = local_2d_sections_l
	1815	#endif
	1816	ENDIF
	1817	!
	1818	!-- Do not output redundant ghost point data except for the
	1819	!-- boundaries of the total domain.
	1820	IF ( nxr == nx ) THEN
	1821	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1822	id_var_do2d(av,if), &
	1823	local_2d_sections(nxl:nxr+1,1:ns, &
[1551]	1824	nzb_do:nzt_do), &
[1308]	1825	start = (/ nxl+1, 1, 1, &
	1826	do2d_xz_time_count(av) /), &
[1551]	1827	count = (/ nxr-nxl+2, ns, nzt_do-nzb_do+1, &
[1308]	1828	1 /) )
	1829	ELSE
	1830	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1831	id_var_do2d(av,if), &
	1832	local_2d_sections(nxl:nxr,1:ns, &
[1551]	1833	nzb_do:nzt_do), &
[1308]	1834	start = (/ nxl+1, 1, 1, &
	1835	do2d_xz_time_count(av) /), &
[1551]	1836	count = (/ nxr-nxl+1, ns, nzt_do-nzb_do+1, &
[1308]	1837	1 /) )
	1838	ENDIF
	1839
[1783]	1840	CALL netcdf_handle_error( 'data_output_2d', 57 )
[1308]	1841
	1842	CASE ( 'yz' )
	1843	!
	1844	!-- First, all PEs get the information of all cross-sections.
	1845	!-- Then the data are written to the output file by all PEs
	1846	!-- while NF90_COLLECTIVE is set in subroutine
	1847	!-- define_netcdf_header. Although redundant information are
	1848	!-- written to the output file in that case, the performance
	1849	!-- is significantly better compared to the case where only
	1850	!-- the first row of PEs in y-direction (myidy = 0) is given
	1851	!-- the output while NF90_INDEPENDENT is set.
	1852	IF ( npex /= 1 ) THEN
	1853
	1854	#if defined( __parallel )
	1855	!
	1856	!-- Distribute data over all PEs along x
	1857	ngp = ( nyng-nysg+1 ) * ( nzt-nzb + 2 ) * ns
	1858	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1859	CALL MPI_ALLREDUCE( local_2d_sections_l(1,nysg,nzb_do), &
	1860	local_2d_sections(1,nysg,nzb_do), &
[1308]	1861	ngp, MPI_REAL, MPI_SUM, comm1dx, &
	1862	ierr )
	1863	#else
	1864	local_2d_sections = local_2d_sections_l
	1865	#endif
	1866	ENDIF
	1867	!
	1868	!-- Do not output redundant ghost point data except for the
	1869	!-- boundaries of the total domain.
	1870	IF ( nyn == ny ) THEN
	1871	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1872	id_var_do2d(av,if), &
	1873	local_2d_sections(1:ns, &
[1551]	1874	nys:nyn+1,nzb_do:nzt_do), &
[1308]	1875	start = (/ 1, nys+1, 1, &
	1876	do2d_yz_time_count(av) /), &
	1877	count = (/ ns, nyn-nys+2, &
[1551]	1878	nzt_do-nzb_do+1, 1 /) )
[1308]	1879	ELSE
	1880	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1881	id_var_do2d(av,if), &
	1882	local_2d_sections(1:ns,nys:nyn, &
[1551]	1883	nzb_do:nzt_do), &
[1308]	1884	start = (/ 1, nys+1, 1, &
	1885	do2d_yz_time_count(av) /), &
	1886	count = (/ ns, nyn-nys+1, &
[1551]	1887	nzt_do-nzb_do+1, 1 /) )
[1308]	1888	ENDIF
	1889
[1783]	1890	CALL netcdf_handle_error( 'data_output_2d', 60 )
[1308]	1891
	1892	CASE DEFAULT
	1893	message_string = 'unknown cross-section: ' // TRIM( mode )
	1894	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
	1895
	1896	END SELECT
	1897
	1898	ENDIF
[1311]	1899	#endif
[1]	1900	ENDIF
	1901
	1902	if = if + 1
	1903	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
	1904	do2d_mode = do2d(av,if)(l-1:l)
	1905
	1906	ENDDO
	1907
	1908	!
	1909	!-- Deallocate temporary arrays.
	1910	IF ( ALLOCATED( level_z ) ) DEALLOCATE( level_z )
[1308]	1911	IF ( netcdf_data_format > 4 ) THEN
	1912	DEALLOCATE( local_pf, local_2d, local_2d_sections )
	1913	IF( mode == 'xz' .OR. mode == 'yz' ) DEALLOCATE( local_2d_sections_l )
	1914	ENDIF
[1]	1915	#if defined( __parallel )
	1916	IF ( .NOT. data_output_2d_on_each_pe .AND. myid == 0 ) THEN
	1917	DEALLOCATE( total_2d )
	1918	ENDIF
	1919	#endif
	1920
	1921	!
	1922	!-- Close plot output file.
[1960]	1923	file_id = 20 + s_ind
[1]	1924
	1925	IF ( data_output_2d_on_each_pe ) THEN
[759]	1926	DO i = 0, io_blocks-1
	1927	IF ( i == io_group ) THEN
	1928	CALL close_file( file_id )
	1929	ENDIF
	1930	#if defined( __parallel )
	1931	CALL MPI_BARRIER( comm2d, ierr )
	1932	#endif
	1933	ENDDO
[1]	1934	ELSE
	1935	IF ( myid == 0 ) CALL close_file( file_id )
	1936	ENDIF
	1937
[1318]	1938	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
[1]	1939
	1940	END SUBROUTINE data_output_2d

Note: See TracBrowser for help on using the repository browser.

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |