Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 1977

Last change on this file since 1977 was 1977, checked in by maronga, 8 years ago
last commit documented
Property svn:keywords set to `Id`
File size: 77.2 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	! -----------------
[1961]	21	!
[1977]	22	!
[1552]	23	! Former revisions:
	24	! -----------------
	25	! $Id: data_output_2d.f90 1977 2016-07-27 13:28:18Z maronga $
	26	!
[1977]	27	! 1976 2016-07-27 13:28:04Z maronga
	28	! Output of radiation quantities is now done directly in the respective module
	29	!
[1973]	30	! 1972 2016-07-26 07:52:02Z maronga
[1976]	31	! Output of land surface quantities is now done directly in the respective
	32	! module
[1973]	33	!
[1961]	34	! 1960 2016-07-12 16:34:24Z suehring
	35	! Scalar surface flux added
	36	! Rename INTEGER variable s into s_ind, as s is already assigned to scalar
	37	!
[1851]	38	! 1849 2016-04-08 11:33:18Z hoffmann
	39	! precipitation_amount, precipitation_rate, prr moved to arrays_3d
	40	!
[1823]	41	! 1822 2016-04-07 07:49:42Z hoffmann
	42	! Output of bulk cloud physics simplified.
	43	!
[1789]	44	! 1788 2016-03-10 11:01:04Z maronga
	45	! Added output of z0q
	46	!
[1784]	47	! 1783 2016-03-06 18:36:17Z raasch
	48	! name change of netcdf routines and module + related changes
	49	!
[1746]	50	! 1745 2016-02-05 13:06:51Z gronemeier
	51	! Bugfix: test if time axis limit exceeds moved to point after call of check_open
	52	!
[1704]	53	! 1703 2015-11-02 12:38:44Z raasch
	54	! bugfix for output of single (*) xy-sections in case of parallel netcdf I/O
	55	!
[1702]	56	! 1701 2015-11-02 07:43:04Z maronga
	57	! Bugfix in output of RRTGM data
	58	!
[1692]	59	! 1691 2015-10-26 16:17:44Z maronga
	60	! Added output of Obukhov length (ol) and radiative heating rates for RRTMG.
	61	! Formatting corrections.
	62	!
[1683]	63	! 1682 2015-10-07 23:56:08Z knoop
	64	! Code annotations made doxygen readable
	65	!
[1586]	66	! 1585 2015-04-30 07:05:52Z maronga
	67	! Added support for RRTMG
	68	!
[1556]	69	! 1555 2015-03-04 17:44:27Z maronga
	70	! Added output of r_a and r_s
	71	!
[1552]	72	! 1551 2015-03-03 14:18:16Z maronga
[1551]	73	! Added suppport for land surface model and radiation model output. In the course
	74	! of this action, the limits for vertical loops have been changed (from nzb and
	75	! nzt+1 to nzb_do and nzt_do, respectively in order to allow soil model output).
	76	! Moreover, a new vertical grid zs was introduced.
[1329]	77	!
[1360]	78	! 1359 2014-04-11 17:15:14Z hoffmann
	79	! New particle structure integrated.
	80	!
[1354]	81	! 1353 2014-04-08 15:21:23Z heinze
	82	! REAL constants provided with KIND-attribute
	83	!
[1329]	84	! 1327 2014-03-21 11:00:16Z raasch
	85	! parts concerning iso2d output removed,
	86	! -netcdf output queries
	87	!
[1321]	88	! 1320 2014-03-20 08:40:49Z raasch
[1320]	89	! ONLY-attribute added to USE-statements,
	90	! kind-parameters added to all INTEGER and REAL declaration statements,
	91	! kinds are defined in new module kinds,
	92	! revision history before 2012 removed,
	93	! comment fields (!:) to be used for variable explanations added to
	94	! all variable declaration statements
[1309]	95	!
[1319]	96	! 1318 2014-03-17 13:35:16Z raasch
	97	! barrier argument removed from cpu_log.
	98	! module interfaces removed
	99	!
[1312]	100	! 1311 2014-03-14 12:13:39Z heinze
	101	! bugfix: close #if defined( __netcdf )
	102	!
[1309]	103	! 1308 2014-03-13 14:58:42Z fricke
[1308]	104	! +local_2d_sections, local_2d_sections_l, ns
	105	! Check, if the limit of the time dimension is exceeded for parallel output
	106	! To increase the performance for parallel output, the following is done:
	107	! - Update of time axis is only done by PE0
	108	! - Cross sections are first stored on a local array and are written
	109	! collectively to the output file by all PEs.
[674]	110	!
[1116]	111	! 1115 2013-03-26 18:16:16Z hoffmann
	112	! ql is calculated by calc_liquid_water_content
	113	!
[1077]	114	! 1076 2012-12-05 08:30:18Z hoffmann
	115	! Bugfix in output of ql
	116	!
[1066]	117	! 1065 2012-11-22 17:42:36Z hoffmann
	118	! Bugfix: Output of cross sections of ql
	119	!
[1054]	120	! 1053 2012-11-13 17:11:03Z hoffmann
	121	! +qr, nr, qc and cross sections
	122	!
[1037]	123	! 1036 2012-10-22 13:43:42Z raasch
	124	! code put under GPL (PALM 3.9)
	125	!
[1035]	126	! 1031 2012-10-19 14:35:30Z raasch
	127	! netCDF4 without parallel file support implemented
	128	!
[1008]	129	! 1007 2012-09-19 14:30:36Z franke
	130	! Bugfix: missing calculation of ql_vp added
	131	!
[979]	132	! 978 2012-08-09 08:28:32Z fricke
	133	! +z0h
	134	!
[1]	135	! Revision 1.1 1997/08/11 06:24:09 raasch
	136	! Initial revision
	137	!
	138	!
	139	! Description:
	140	! ------------
[1682]	141	!> Data output of horizontal cross-sections in netCDF format or binary format
	142	!> compatible to old graphic software iso2d.
	143	!> Attention: The position of the sectional planes is still not always computed
	144	!> --------- correctly. (zu is used always)!
[1]	145	!------------------------------------------------------------------------------!
[1682]	146	SUBROUTINE data_output_2d( mode, av )
	147
[1]	148
[1320]	149	USE arrays_3d, &
[1849]	150	ONLY: dzw, e, nr, ol, p, pt, precipitation_amount, precipitation_rate,&
[1960]	151	prr,q, qc, ql, ql_c, ql_v, ql_vp, qr, qsws, rho, s, sa, shf, &
	152	ssws, tend, ts, u, us, v, vpt, w, z0, z0h, z0q, zu, zw
[1320]	153
[1]	154	USE averaging
[1320]	155
	156	USE cloud_parameters, &
[1849]	157	ONLY: hyrho, l_d_cp, pt_d_t
[1320]	158
	159	USE control_parameters, &
	160	ONLY: cloud_physics, data_output_2d_on_each_pe, data_output_xy, &
	161	data_output_xz, data_output_yz, do2d, &
	162	do2d_xy_last_time, do2d_xy_n, do2d_xy_time_count, &
	163	do2d_xz_last_time, do2d_xz_n, do2d_xz_time_count, &
	164	do2d_yz_last_time, do2d_yz_n, do2d_yz_time_count, &
[1822]	165	ibc_uv_b, io_blocks, io_group, message_string, &
	166	ntdim_2d_xy, ntdim_2d_xz, ntdim_2d_yz, &
	167	psolver, section, simulated_time, simulated_time_chr, &
	168	time_since_reference_point
[1320]	169
	170	USE cpulog, &
	171	ONLY: cpu_log, log_point
	172
	173	USE grid_variables, &
	174	ONLY: dx, dy
	175
	176	USE indices, &
	177	ONLY: nbgp, nx, nxl, nxlg, nxr, nxrg, ny, nyn, nyng, nys, nysg, &
	178	nz, nzb, nzt
	179
	180	USE kinds
[1551]	181
	182	USE land_surface_model_mod, &
[1972]	183	ONLY: land_surface, lsm_data_output_2d, zs
[1551]	184
[1783]	185	#if defined( __netcdf )
	186	USE NETCDF
	187	#endif
[1320]	188
[1783]	189	USE netcdf_interface, &
	190	ONLY: id_set_xy, id_set_xz, id_set_yz, id_var_do2d, id_var_time_xy, &
	191	id_var_time_xz, id_var_time_yz, nc_stat, netcdf_data_format, &
	192	netcdf_handle_error
	193
[1320]	194	USE particle_attributes, &
[1359]	195	ONLY: grid_particles, number_of_particles, particle_advection_start, &
	196	particles, prt_count
[1320]	197
[1]	198	USE pegrid
	199
[1551]	200	USE radiation_model_mod, &
[1976]	201	ONLY: radiation, radiation_data_output_2d
[1551]	202
[1]	203	IMPLICIT NONE
	204
[1682]	205	CHARACTER (LEN=2) :: do2d_mode !<
	206	CHARACTER (LEN=2) :: mode !<
	207	CHARACTER (LEN=4) :: grid !<
	208	CHARACTER (LEN=25) :: section_chr !<
	209	CHARACTER (LEN=50) :: rtext !<
[1320]	210
[1682]	211	INTEGER(iwp) :: av !<
	212	INTEGER(iwp) :: ngp !<
	213	INTEGER(iwp) :: file_id !<
	214	INTEGER(iwp) :: i !<
	215	INTEGER(iwp) :: if !<
	216	INTEGER(iwp) :: is !<
	217	INTEGER(iwp) :: iis !<
	218	INTEGER(iwp) :: j !<
	219	INTEGER(iwp) :: k !<
	220	INTEGER(iwp) :: l !<
	221	INTEGER(iwp) :: layer_xy !<
	222	INTEGER(iwp) :: n !<
[1703]	223	INTEGER(iwp) :: nis !<
[1682]	224	INTEGER(iwp) :: ns !<
	225	INTEGER(iwp) :: nzb_do !< lower limit of the data field (usually nzb)
	226	INTEGER(iwp) :: nzt_do !< upper limit of the data field (usually nzt+1)
	227	INTEGER(iwp) :: psi !<
[1960]	228	INTEGER(iwp) :: s_ind !<
[1682]	229	INTEGER(iwp) :: sender !<
	230	INTEGER(iwp) :: ind(4) !<
[1320]	231
[1682]	232	LOGICAL :: found !<
	233	LOGICAL :: resorted !<
	234	LOGICAL :: two_d !<
[1320]	235
[1682]	236	REAL(wp) :: mean_r !<
	237	REAL(wp) :: s_r2 !<
	238	REAL(wp) :: s_r3 !<
[1320]	239
[1682]	240	REAL(wp), DIMENSION(:), ALLOCATABLE :: level_z !<
	241	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d !<
	242	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d_l !<
	243	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !<
	244	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections !<
	245	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections_l !<
[1359]	246
[1]	247	#if defined( __parallel )
[1682]	248	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: total_2d !<
[1]	249	#endif
[1682]	250	REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !<
[1]	251
	252	NAMELIST /LOCAL/ rtext
	253
	254	!
	255	!-- Immediate return, if no output is requested (no respective sections
	256	!-- found in parameter data_output)
	257	IF ( mode == 'xy' .AND. .NOT. data_output_xy(av) ) RETURN
	258	IF ( mode == 'xz' .AND. .NOT. data_output_xz(av) ) RETURN
	259	IF ( mode == 'yz' .AND. .NOT. data_output_yz(av) ) RETURN
	260
[1308]	261	CALL cpu_log (log_point(3),'data_output_2d','start')
	262
[1]	263	two_d = .FALSE. ! local variable to distinguish between output of pure 2D
	264	! arrays and cross-sections of 3D arrays.
	265
	266	!
	267	!-- Depending on the orientation of the cross-section, the respective output
	268	!-- files have to be opened.
	269	SELECT CASE ( mode )
	270
	271	CASE ( 'xy' )
[1960]	272	s_ind = 1
[667]	273	ALLOCATE( level_z(nzb:nzt+1), local_2d(nxlg:nxrg,nysg:nyng) )
[1]	274
[1308]	275	IF ( netcdf_data_format > 4 ) THEN
	276	ns = 1
[1960]	277	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	278	ns = ns + 1
	279	ENDDO
	280	ns = ns - 1
	281	ALLOCATE( local_2d_sections(nxlg:nxrg,nysg:nyng,1:ns) )
[1353]	282	local_2d_sections = 0.0_wp
[1308]	283	ENDIF
	284
[493]	285	!
[1031]	286	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	287	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	288	CALL check_open( 101+av*10 )
	289	ENDIF
[1]	290	IF ( data_output_2d_on_each_pe ) THEN
	291	CALL check_open( 21 )
	292	ELSE
	293	IF ( myid == 0 ) THEN
	294	#if defined( __parallel )
[667]	295	ALLOCATE( total_2d(-nbgp:nx+nbgp,-nbgp:ny+nbgp) )
[1]	296	#endif
	297	ENDIF
	298	ENDIF
	299
	300	CASE ( 'xz' )
[1960]	301	s_ind = 2
[667]	302	ALLOCATE( local_2d(nxlg:nxrg,nzb:nzt+1) )
[1]	303
[1308]	304	IF ( netcdf_data_format > 4 ) THEN
	305	ns = 1
[1960]	306	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	307	ns = ns + 1
	308	ENDDO
	309	ns = ns - 1
	310	ALLOCATE( local_2d_sections(nxlg:nxrg,1:ns,nzb:nzt+1) )
	311	ALLOCATE( local_2d_sections_l(nxlg:nxrg,1:ns,nzb:nzt+1) )
[1353]	312	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
[1308]	313	ENDIF
	314
[493]	315	!
[1031]	316	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	317	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	318	CALL check_open( 102+av*10 )
	319	ENDIF
[1]	320
	321	IF ( data_output_2d_on_each_pe ) THEN
	322	CALL check_open( 22 )
	323	ELSE
	324	IF ( myid == 0 ) THEN
	325	#if defined( __parallel )
[667]	326	ALLOCATE( total_2d(-nbgp:nx+nbgp,nzb:nzt+1) )
[1]	327	#endif
	328	ENDIF
	329	ENDIF
	330
	331	CASE ( 'yz' )
[1960]	332	s_ind = 3
[667]	333	ALLOCATE( local_2d(nysg:nyng,nzb:nzt+1) )
[1]	334
[1308]	335	IF ( netcdf_data_format > 4 ) THEN
	336	ns = 1
[1960]	337	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	338	ns = ns + 1
	339	ENDDO
	340	ns = ns - 1
	341	ALLOCATE( local_2d_sections(1:ns,nysg:nyng,nzb:nzt+1) )
	342	ALLOCATE( local_2d_sections_l(1:ns,nysg:nyng,nzb:nzt+1) )
[1353]	343	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
[1308]	344	ENDIF
	345
[493]	346	!
[1031]	347	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	348	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	349	CALL check_open( 103+av*10 )
	350	ENDIF
[1]	351
	352	IF ( data_output_2d_on_each_pe ) THEN
	353	CALL check_open( 23 )
	354	ELSE
	355	IF ( myid == 0 ) THEN
	356	#if defined( __parallel )
[667]	357	ALLOCATE( total_2d(-nbgp:ny+nbgp,nzb:nzt+1) )
[1]	358	#endif
	359	ENDIF
	360	ENDIF
	361
	362	CASE DEFAULT
[254]	363	message_string = 'unknown cross-section: ' // TRIM( mode )
	364	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
[1]	365
	366	END SELECT
	367
	368	!
[1745]	369	!-- For parallel netcdf output the time axis must be limited. Return, if this
	370	!-- limit is exceeded. This could be the case, if the simulated time exceeds
	371	!-- the given end time by the length of the given output interval.
	372	IF ( netcdf_data_format > 4 ) THEN
	373	IF ( mode == 'xy' .AND. do2d_xy_time_count(av) + 1 > &
	374	ntdim_2d_xy(av) ) THEN
	375	WRITE ( message_string, * ) 'Output of xy cross-sections is not ', &
	376	'given at t=', simulated_time, '&because the', &
	377	' maximum number of output time levels is exceeded.'
	378	CALL message( 'data_output_2d', 'PA0384', 0, 1, 0, 6, 0 )
	379	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	380	RETURN
	381	ENDIF
	382	IF ( mode == 'xz' .AND. do2d_xz_time_count(av) + 1 > &
	383	ntdim_2d_xz(av) ) THEN
	384	WRITE ( message_string, * ) 'Output of xz cross-sections is not ', &
	385	'given at t=', simulated_time, '&because the', &
	386	' maximum number of output time levels is exceeded.'
	387	CALL message( 'data_output_2d', 'PA0385', 0, 1, 0, 6, 0 )
	388	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	389	RETURN
	390	ENDIF
	391	IF ( mode == 'yz' .AND. do2d_yz_time_count(av) + 1 > &
	392	ntdim_2d_yz(av) ) THEN
	393	WRITE ( message_string, * ) 'Output of yz cross-sections is not ', &
	394	'given at t=', simulated_time, '&because the', &
	395	' maximum number of output time levels is exceeded.'
	396	CALL message( 'data_output_2d', 'PA0386', 0, 1, 0, 6, 0 )
	397	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	398	RETURN
	399	ENDIF
	400	ENDIF
	401
	402	!
[1]	403	!-- Allocate a temporary array for resorting (kji -> ijk).
[667]	404	ALLOCATE( local_pf(nxlg:nxrg,nysg:nyng,nzb:nzt+1) )
[1]	405
	406	!
	407	!-- Loop of all variables to be written.
	408	!-- Output dimensions chosen
	409	if = 1
	410	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
	411	do2d_mode = do2d(av,if)(l-1:l)
	412
	413	DO WHILE ( do2d(av,if)(1:1) /= ' ' )
	414
	415	IF ( do2d_mode == mode ) THEN
[1551]	416
	417	nzb_do = nzb
	418	nzt_do = nzt+1
[1]	419	!
	420	!-- Store the array chosen on the temporary array.
	421	resorted = .FALSE.
	422	SELECT CASE ( TRIM( do2d(av,if) ) )
	423
	424	CASE ( 'e_xy', 'e_xz', 'e_yz' )
	425	IF ( av == 0 ) THEN
	426	to_be_resorted => e
	427	ELSE
	428	to_be_resorted => e_av
	429	ENDIF
	430	IF ( mode == 'xy' ) level_z = zu
	431
[771]	432	CASE ( 'lpt_xy', 'lpt_xz', 'lpt_yz' )
	433	IF ( av == 0 ) THEN
	434	to_be_resorted => pt
	435	ELSE
	436	to_be_resorted => lpt_av
	437	ENDIF
	438	IF ( mode == 'xy' ) level_z = zu
	439
[1]	440	CASE ( 'lwp*_xy' ) ! 2d-array
	441	IF ( av == 0 ) THEN
[667]	442	DO i = nxlg, nxrg
	443	DO j = nysg, nyng
[1320]	444	local_pf(i,j,nzb+1) = SUM( ql(nzb:nzt,j,i) * &
[1]	445	dzw(1:nzt+1) )
	446	ENDDO
	447	ENDDO
	448	ELSE
[667]	449	DO i = nxlg, nxrg
	450	DO j = nysg, nyng
[1]	451	local_pf(i,j,nzb+1) = lwp_av(j,i)
	452	ENDDO
	453	ENDDO
	454	ENDIF
	455	resorted = .TRUE.
	456	two_d = .TRUE.
	457	level_z(nzb+1) = zu(nzb+1)
	458
[1053]	459	CASE ( 'nr_xy', 'nr_xz', 'nr_yz' )
	460	IF ( av == 0 ) THEN
	461	to_be_resorted => nr
	462	ELSE
	463	to_be_resorted => nr_av
	464	ENDIF
	465	IF ( mode == 'xy' ) level_z = zu
	466
[1691]	467	CASE ( 'ol*_xy' ) ! 2d-array
	468	IF ( av == 0 ) THEN
	469	DO i = nxlg, nxrg
	470	DO j = nysg, nyng
	471	local_pf(i,j,nzb+1) = ol(j,i)
	472	ENDDO
	473	ENDDO
	474	ELSE
	475	DO i = nxlg, nxrg
	476	DO j = nysg, nyng
	477	local_pf(i,j,nzb+1) = ol_av(j,i)
	478	ENDDO
	479	ENDDO
	480	ENDIF
	481	resorted = .TRUE.
	482	two_d = .TRUE.
	483	level_z(nzb+1) = zu(nzb+1)
	484
[1]	485	CASE ( 'p_xy', 'p_xz', 'p_yz' )
	486	IF ( av == 0 ) THEN
[729]	487	IF ( psolver /= 'sor' ) CALL exchange_horiz( p, nbgp )
[1]	488	to_be_resorted => p
	489	ELSE
[729]	490	IF ( psolver /= 'sor' ) CALL exchange_horiz( p_av, nbgp )
[1]	491	to_be_resorted => p_av
	492	ENDIF
	493	IF ( mode == 'xy' ) level_z = zu
	494
	495	CASE ( 'pc_xy', 'pc_xz', 'pc_yz' ) ! particle concentration
	496	IF ( av == 0 ) THEN
[215]	497	IF ( simulated_time >= particle_advection_start ) THEN
	498	tend = prt_count
[667]	499	CALL exchange_horiz( tend, nbgp )
[215]	500	ELSE
[1353]	501	tend = 0.0_wp
[215]	502	ENDIF
[667]	503	DO i = nxlg, nxrg
	504	DO j = nysg, nyng
[1]	505	DO k = nzb, nzt+1
	506	local_pf(i,j,k) = tend(k,j,i)
	507	ENDDO
	508	ENDDO
	509	ENDDO
	510	resorted = .TRUE.
	511	ELSE
[667]	512	CALL exchange_horiz( pc_av, nbgp )
[1]	513	to_be_resorted => pc_av
	514	ENDIF
	515
[1359]	516	CASE ( 'pr_xy', 'pr_xz', 'pr_yz' ) ! mean particle radius (effective radius)
[1]	517	IF ( av == 0 ) THEN
[215]	518	IF ( simulated_time >= particle_advection_start ) THEN
	519	DO i = nxl, nxr
	520	DO j = nys, nyn
	521	DO k = nzb, nzt+1
[1359]	522	number_of_particles = prt_count(k,j,i)
	523	IF (number_of_particles <= 0) CYCLE
	524	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
	525	s_r2 = 0.0_wp
[1353]	526	s_r3 = 0.0_wp
[1359]	527	DO n = 1, number_of_particles
	528	IF ( particles(n)%particle_mask ) THEN
	529	s_r2 = s_r2 + particles(n)%radius*2 &
	530	particles(n)%weight_factor
	531	s_r3 = s_r3 + particles(n)%radius*3 &
	532	particles(n)%weight_factor
	533	ENDIF
[215]	534	ENDDO
[1359]	535	IF ( s_r2 > 0.0_wp ) THEN
	536	mean_r = s_r3 / s_r2
[215]	537	ELSE
[1353]	538	mean_r = 0.0_wp
[215]	539	ENDIF
	540	tend(k,j,i) = mean_r
[1]	541	ENDDO
	542	ENDDO
	543	ENDDO
[667]	544	CALL exchange_horiz( tend, nbgp )
[215]	545	ELSE
[1353]	546	tend = 0.0_wp
[1359]	547	ENDIF
[667]	548	DO i = nxlg, nxrg
	549	DO j = nysg, nyng
[1]	550	DO k = nzb, nzt+1
	551	local_pf(i,j,k) = tend(k,j,i)
	552	ENDDO
	553	ENDDO
	554	ENDDO
	555	resorted = .TRUE.
	556	ELSE
[667]	557	CALL exchange_horiz( pr_av, nbgp )
[1]	558	to_be_resorted => pr_av
	559	ENDIF
	560
[72]	561	CASE ( 'pra*_xy' ) ! 2d-array / integral quantity => no av
	562	CALL exchange_horiz_2d( precipitation_amount )
[667]	563	DO i = nxlg, nxrg
	564	DO j = nysg, nyng
[72]	565	local_pf(i,j,nzb+1) = precipitation_amount(j,i)
	566	ENDDO
	567	ENDDO
[1353]	568	precipitation_amount = 0.0_wp ! reset for next integ. interval
[72]	569	resorted = .TRUE.
	570	two_d = .TRUE.
	571	level_z(nzb+1) = zu(nzb+1)
	572
	573	CASE ( 'prr*_xy' ) ! 2d-array
[1822]	574	IF ( av == 0 ) THEN
	575	CALL exchange_horiz_2d( prr(nzb+1,:,:) )
	576	DO i = nxlg, nxrg
	577	DO j = nysg, nyng
	578	local_pf(i,j,nzb+1) = prr(nzb+1,j,i) * hyrho(nzb+1)
[1053]	579	ENDDO
[1822]	580	ENDDO
[1053]	581	ELSE
[1822]	582	CALL exchange_horiz_2d( prr_av(nzb+1,:,:) )
	583	DO i = nxlg, nxrg
	584	DO j = nysg, nyng
	585	local_pf(i,j,nzb+1) = prr_av(nzb+1,j,i) * hyrho(nzb+1)
[1053]	586	ENDDO
[1822]	587	ENDDO
[1053]	588	ENDIF
	589	resorted = .TRUE.
	590	two_d = .TRUE.
	591	level_z(nzb+1) = zu(nzb+1)
	592
	593	CASE ( 'prr_xy', 'prr_xz', 'prr_yz' )
[72]	594	IF ( av == 0 ) THEN
[1053]	595	CALL exchange_horiz( prr, nbgp )
[667]	596	DO i = nxlg, nxrg
	597	DO j = nysg, nyng
[1053]	598	DO k = nzb, nzt+1
[1822]	599	local_pf(i,j,k) = prr(k,j,i) * hyrho(nzb+1)
[1053]	600	ENDDO
[72]	601	ENDDO
	602	ENDDO
	603	ELSE
[1053]	604	CALL exchange_horiz( prr_av, nbgp )
[667]	605	DO i = nxlg, nxrg
	606	DO j = nysg, nyng
[1053]	607	DO k = nzb, nzt+1
[1822]	608	local_pf(i,j,k) = prr_av(k,j,i) * hyrho(nzb+1)
[1053]	609	ENDDO
[72]	610	ENDDO
	611	ENDDO
	612	ENDIF
	613	resorted = .TRUE.
[1053]	614	IF ( mode == 'xy' ) level_z = zu
[72]	615
[1]	616	CASE ( 'pt_xy', 'pt_xz', 'pt_yz' )
	617	IF ( av == 0 ) THEN
	618	IF ( .NOT. cloud_physics ) THEN
	619	to_be_resorted => pt
	620	ELSE
[667]	621	DO i = nxlg, nxrg
	622	DO j = nysg, nyng
[1]	623	DO k = nzb, nzt+1
[1320]	624	local_pf(i,j,k) = pt(k,j,i) + l_d_cp * &
	625	pt_d_t(k) * &
[1]	626	ql(k,j,i)
	627	ENDDO
	628	ENDDO
	629	ENDDO
	630	resorted = .TRUE.
	631	ENDIF
	632	ELSE
	633	to_be_resorted => pt_av
	634	ENDIF
	635	IF ( mode == 'xy' ) level_z = zu
	636
	637	CASE ( 'q_xy', 'q_xz', 'q_yz' )
	638	IF ( av == 0 ) THEN
	639	to_be_resorted => q
	640	ELSE
	641	to_be_resorted => q_av
	642	ENDIF
	643	IF ( mode == 'xy' ) level_z = zu
	644
[1053]	645	CASE ( 'qc_xy', 'qc_xz', 'qc_yz' )
[1]	646	IF ( av == 0 ) THEN
[1115]	647	to_be_resorted => qc
[1]	648	ELSE
[1115]	649	to_be_resorted => qc_av
[1]	650	ENDIF
	651	IF ( mode == 'xy' ) level_z = zu
	652
[1053]	653	CASE ( 'ql_xy', 'ql_xz', 'ql_yz' )
	654	IF ( av == 0 ) THEN
[1115]	655	to_be_resorted => ql
[1053]	656	ELSE
[1115]	657	to_be_resorted => ql_av
[1053]	658	ENDIF
	659	IF ( mode == 'xy' ) level_z = zu
	660
[1]	661	CASE ( 'ql_c_xy', 'ql_c_xz', 'ql_c_yz' )
	662	IF ( av == 0 ) THEN
	663	to_be_resorted => ql_c
	664	ELSE
	665	to_be_resorted => ql_c_av
	666	ENDIF
	667	IF ( mode == 'xy' ) level_z = zu
	668
	669	CASE ( 'ql_v_xy', 'ql_v_xz', 'ql_v_yz' )
	670	IF ( av == 0 ) THEN
	671	to_be_resorted => ql_v
	672	ELSE
	673	to_be_resorted => ql_v_av
	674	ENDIF
	675	IF ( mode == 'xy' ) level_z = zu
	676
	677	CASE ( 'ql_vp_xy', 'ql_vp_xz', 'ql_vp_yz' )
	678	IF ( av == 0 ) THEN
[1007]	679	IF ( simulated_time >= particle_advection_start ) THEN
	680	DO i = nxl, nxr
	681	DO j = nys, nyn
	682	DO k = nzb, nzt+1
[1359]	683	number_of_particles = prt_count(k,j,i)
	684	IF (number_of_particles <= 0) CYCLE
	685	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
	686	DO n = 1, number_of_particles
	687	IF ( particles(n)%particle_mask ) THEN
	688	tend(k,j,i) = tend(k,j,i) + &
	689	particles(n)%weight_factor / &
	690	prt_count(k,j,i)
	691	ENDIF
[1007]	692	ENDDO
	693	ENDDO
	694	ENDDO
	695	ENDDO
	696	CALL exchange_horiz( tend, nbgp )
	697	ELSE
[1353]	698	tend = 0.0_wp
[1359]	699	ENDIF
[1007]	700	DO i = nxlg, nxrg
	701	DO j = nysg, nyng
	702	DO k = nzb, nzt+1
	703	local_pf(i,j,k) = tend(k,j,i)
	704	ENDDO
	705	ENDDO
	706	ENDDO
	707	resorted = .TRUE.
	708	ELSE
	709	CALL exchange_horiz( ql_vp_av, nbgp )
[1]	710	to_be_resorted => ql_vp
	711	ENDIF
	712	IF ( mode == 'xy' ) level_z = zu
	713
[1053]	714	CASE ( 'qr_xy', 'qr_xz', 'qr_yz' )
	715	IF ( av == 0 ) THEN
	716	to_be_resorted => qr
	717	ELSE
	718	to_be_resorted => qr_av
	719	ENDIF
	720	IF ( mode == 'xy' ) level_z = zu
	721
[354]	722	CASE ( 'qsws*_xy' ) ! 2d-array
	723	IF ( av == 0 ) THEN
[667]	724	DO i = nxlg, nxrg
	725	DO j = nysg, nyng
[354]	726	local_pf(i,j,nzb+1) = qsws(j,i)
	727	ENDDO
	728	ENDDO
	729	ELSE
[667]	730	DO i = nxlg, nxrg
	731	DO j = nysg, nyng
[354]	732	local_pf(i,j,nzb+1) = qsws_av(j,i)
	733	ENDDO
	734	ENDDO
	735	ENDIF
	736	resorted = .TRUE.
	737	two_d = .TRUE.
	738	level_z(nzb+1) = zu(nzb+1)
	739
[1]	740	CASE ( 'qv_xy', 'qv_xz', 'qv_yz' )
	741	IF ( av == 0 ) THEN
[667]	742	DO i = nxlg, nxrg
	743	DO j = nysg, nyng
[1]	744	DO k = nzb, nzt+1
	745	local_pf(i,j,k) = q(k,j,i) - ql(k,j,i)
	746	ENDDO
	747	ENDDO
	748	ENDDO
	749	resorted = .TRUE.
	750	ELSE
	751	to_be_resorted => qv_av
	752	ENDIF
	753	IF ( mode == 'xy' ) level_z = zu
	754
[1551]	755
[1585]	756
[96]	757	CASE ( 'rho_xy', 'rho_xz', 'rho_yz' )
	758	IF ( av == 0 ) THEN
	759	to_be_resorted => rho
	760	ELSE
	761	to_be_resorted => rho_av
	762	ENDIF
	763
[1]	764	CASE ( 's_xy', 's_xz', 's_yz' )
	765	IF ( av == 0 ) THEN
[1960]	766	to_be_resorted => s
[1]	767	ELSE
[355]	768	to_be_resorted => s_av
[1]	769	ENDIF
	770
[96]	771	CASE ( 'sa_xy', 'sa_xz', 'sa_yz' )
	772	IF ( av == 0 ) THEN
	773	to_be_resorted => sa
	774	ELSE
	775	to_be_resorted => sa_av
	776	ENDIF
	777
[354]	778	CASE ( 'shf*_xy' ) ! 2d-array
	779	IF ( av == 0 ) THEN
[667]	780	DO i = nxlg, nxrg
	781	DO j = nysg, nyng
[354]	782	local_pf(i,j,nzb+1) = shf(j,i)
	783	ENDDO
	784	ENDDO
	785	ELSE
[667]	786	DO i = nxlg, nxrg
	787	DO j = nysg, nyng
[354]	788	local_pf(i,j,nzb+1) = shf_av(j,i)
	789	ENDDO
	790	ENDDO
	791	ENDIF
	792	resorted = .TRUE.
	793	two_d = .TRUE.
	794	level_z(nzb+1) = zu(nzb+1)
[1960]	795
	796	CASE ( 'ssws*_xy' ) ! 2d-array
	797	IF ( av == 0 ) THEN
	798	DO i = nxlg, nxrg
	799	DO j = nysg, nyng
	800	local_pf(i,j,nzb+1) = ssws(j,i)
	801	ENDDO
	802	ENDDO
	803	ELSE
	804	DO i = nxlg, nxrg
	805	DO j = nysg, nyng
	806	local_pf(i,j,nzb+1) = ssws_av(j,i)
	807	ENDDO
	808	ENDDO
	809	ENDIF
	810	resorted = .TRUE.
	811	two_d = .TRUE.
	812	level_z(nzb+1) = zu(nzb+1)
[1551]	813
[1]	814	CASE ( 't*_xy' ) ! 2d-array
	815	IF ( av == 0 ) THEN
[667]	816	DO i = nxlg, nxrg
	817	DO j = nysg, nyng
[1]	818	local_pf(i,j,nzb+1) = ts(j,i)
	819	ENDDO
	820	ENDDO
	821	ELSE
[667]	822	DO i = nxlg, nxrg
	823	DO j = nysg, nyng
[1]	824	local_pf(i,j,nzb+1) = ts_av(j,i)
	825	ENDDO
	826	ENDDO
	827	ENDIF
	828	resorted = .TRUE.
	829	two_d = .TRUE.
	830	level_z(nzb+1) = zu(nzb+1)
	831
	832	CASE ( 'u_xy', 'u_xz', 'u_yz' )
	833	IF ( av == 0 ) THEN
	834	to_be_resorted => u
	835	ELSE
	836	to_be_resorted => u_av
	837	ENDIF
	838	IF ( mode == 'xy' ) level_z = zu
	839	!
	840	!-- Substitute the values generated by "mirror" boundary condition
	841	!-- at the bottom boundary by the real surface values.
	842	IF ( do2d(av,if) == 'u_xz' .OR. do2d(av,if) == 'u_yz' ) THEN
[1353]	843	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
[1]	844	ENDIF
	845
	846	CASE ( 'u*_xy' ) ! 2d-array
	847	IF ( av == 0 ) THEN
[667]	848	DO i = nxlg, nxrg
	849	DO j = nysg, nyng
[1]	850	local_pf(i,j,nzb+1) = us(j,i)
	851	ENDDO
	852	ENDDO
	853	ELSE
[667]	854	DO i = nxlg, nxrg
	855	DO j = nysg, nyng
[1]	856	local_pf(i,j,nzb+1) = us_av(j,i)
	857	ENDDO
	858	ENDDO
	859	ENDIF
	860	resorted = .TRUE.
	861	two_d = .TRUE.
	862	level_z(nzb+1) = zu(nzb+1)
	863
	864	CASE ( 'v_xy', 'v_xz', 'v_yz' )
	865	IF ( av == 0 ) THEN
	866	to_be_resorted => v
	867	ELSE
	868	to_be_resorted => v_av
	869	ENDIF
	870	IF ( mode == 'xy' ) level_z = zu
	871	!
	872	!-- Substitute the values generated by "mirror" boundary condition
	873	!-- at the bottom boundary by the real surface values.
	874	IF ( do2d(av,if) == 'v_xz' .OR. do2d(av,if) == 'v_yz' ) THEN
[1353]	875	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
[1]	876	ENDIF
	877
	878	CASE ( 'vpt_xy', 'vpt_xz', 'vpt_yz' )
	879	IF ( av == 0 ) THEN
	880	to_be_resorted => vpt
	881	ELSE
	882	to_be_resorted => vpt_av
	883	ENDIF
	884	IF ( mode == 'xy' ) level_z = zu
	885
	886	CASE ( 'w_xy', 'w_xz', 'w_yz' )
	887	IF ( av == 0 ) THEN
	888	to_be_resorted => w
	889	ELSE
	890	to_be_resorted => w_av
	891	ENDIF
	892	IF ( mode == 'xy' ) level_z = zw
	893
[72]	894	CASE ( 'z0*_xy' ) ! 2d-array
	895	IF ( av == 0 ) THEN
[667]	896	DO i = nxlg, nxrg
	897	DO j = nysg, nyng
[72]	898	local_pf(i,j,nzb+1) = z0(j,i)
	899	ENDDO
	900	ENDDO
	901	ELSE
[667]	902	DO i = nxlg, nxrg
	903	DO j = nysg, nyng
[72]	904	local_pf(i,j,nzb+1) = z0_av(j,i)
	905	ENDDO
	906	ENDDO
	907	ENDIF
	908	resorted = .TRUE.
	909	two_d = .TRUE.
	910	level_z(nzb+1) = zu(nzb+1)
	911
[978]	912	CASE ( 'z0h*_xy' ) ! 2d-array
	913	IF ( av == 0 ) THEN
	914	DO i = nxlg, nxrg
	915	DO j = nysg, nyng
	916	local_pf(i,j,nzb+1) = z0h(j,i)
	917	ENDDO
	918	ENDDO
	919	ELSE
	920	DO i = nxlg, nxrg
	921	DO j = nysg, nyng
	922	local_pf(i,j,nzb+1) = z0h_av(j,i)
	923	ENDDO
	924	ENDDO
	925	ENDIF
	926	resorted = .TRUE.
	927	two_d = .TRUE.
	928	level_z(nzb+1) = zu(nzb+1)
	929
[1788]	930	CASE ( 'z0q*_xy' ) ! 2d-array
	931	IF ( av == 0 ) THEN
	932	DO i = nxlg, nxrg
	933	DO j = nysg, nyng
	934	local_pf(i,j,nzb+1) = z0q(j,i)
	935	ENDDO
	936	ENDDO
	937	ELSE
	938	DO i = nxlg, nxrg
	939	DO j = nysg, nyng
	940	local_pf(i,j,nzb+1) = z0q_av(j,i)
	941	ENDDO
	942	ENDDO
	943	ENDIF
	944	resorted = .TRUE.
	945	two_d = .TRUE.
	946	level_z(nzb+1) = zu(nzb+1)
	947
[1]	948	CASE DEFAULT
[1972]	949
[1]	950	!
[1972]	951	!-- Land surface model quantity
	952	IF ( land_surface ) THEN
	953	CALL lsm_data_output_2d( av, do2d(av,if), found, grid, mode,&
	954	local_pf, two_d, nzb_do, nzt_do )
	955	ENDIF
	956
	957	!
[1976]	958	!-- Radiation quantity
	959	IF ( .NOT. found .AND. radiation ) THEN
	960	CALL radiation_data_output_2d( av, do2d(av,if), found, grid,&
	961	mode, local_pf, two_d )
	962	ENDIF
	963
	964	!
[1]	965	!-- User defined quantity
[1972]	966	IF ( .NOT. found ) THEN
	967	CALL user_data_output_2d( av, do2d(av,if), found, grid, &
	968	local_pf, two_d, nzb_do, nzt_do )
	969	ENDIF
	970
[1]	971	resorted = .TRUE.
	972
	973	IF ( grid == 'zu' ) THEN
	974	IF ( mode == 'xy' ) level_z = zu
	975	ELSEIF ( grid == 'zw' ) THEN
	976	IF ( mode == 'xy' ) level_z = zw
[343]	977	ELSEIF ( grid == 'zu1' ) THEN
	978	IF ( mode == 'xy' ) level_z(nzb+1) = zu(nzb+1)
[1551]	979	ELSEIF ( grid == 'zs' ) THEN
	980	IF ( mode == 'xy' ) level_z = zs
[1]	981	ENDIF
	982
	983	IF ( .NOT. found ) THEN
[1320]	984	message_string = 'no output provided for: ' // &
[274]	985	TRIM( do2d(av,if) )
[254]	986	CALL message( 'data_output_2d', 'PA0181', 0, 0, 0, 6, 0 )
[1]	987	ENDIF
	988
	989	END SELECT
	990
	991	!
	992	!-- Resort the array to be output, if not done above
	993	IF ( .NOT. resorted ) THEN
[667]	994	DO i = nxlg, nxrg
	995	DO j = nysg, nyng
[1551]	996	DO k = nzb_do, nzt_do
[1]	997	local_pf(i,j,k) = to_be_resorted(k,j,i)
	998	ENDDO
	999	ENDDO
	1000	ENDDO
	1001	ENDIF
	1002
	1003	!
	1004	!-- Output of the individual cross-sections, depending on the cross-
	1005	!-- section mode chosen.
	1006	is = 1
[1960]	1007	loop1: DO WHILE ( section(is,s_ind) /= -9999 .OR. two_d )
[1]	1008
	1009	SELECT CASE ( mode )
	1010
	1011	CASE ( 'xy' )
	1012	!
	1013	!-- Determine the cross section index
	1014	IF ( two_d ) THEN
	1015	layer_xy = nzb+1
	1016	ELSE
[1960]	1017	layer_xy = section(is,s_ind)
[1]	1018	ENDIF
	1019
	1020	!
[1551]	1021	!-- Exit the loop for layers beyond the data output domain
	1022	!-- (used for soil model)
[1691]	1023	IF ( layer_xy > nzt_do ) THEN
[1551]	1024	EXIT loop1
	1025	ENDIF
	1026
	1027	!
[1308]	1028	!-- Update the netCDF xy cross section time axis.
	1029	!-- In case of parallel output, this is only done by PE0
	1030	!-- to increase the performance.
	1031	IF ( simulated_time /= do2d_xy_last_time(av) ) THEN
	1032	do2d_xy_time_count(av) = do2d_xy_time_count(av) + 1
	1033	do2d_xy_last_time(av) = simulated_time
	1034	IF ( myid == 0 ) THEN
[1327]	1035	IF ( .NOT. data_output_2d_on_each_pe &
	1036	.OR. netcdf_data_format > 4 ) &
[493]	1037	THEN
[1]	1038	#if defined( __netcdf )
	1039	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1040	id_var_time_xy(av), &
[291]	1041	(/ time_since_reference_point /), &
[1]	1042	start = (/ do2d_xy_time_count(av) /), &
	1043	count = (/ 1 /) )
[1783]	1044	CALL netcdf_handle_error( 'data_output_2d', 53 )
[1]	1045	#endif
	1046	ENDIF
	1047	ENDIF
	1048	ENDIF
	1049	!
	1050	!-- If required, carry out averaging along z
[1960]	1051	IF ( section(is,s_ind) == -1 .AND. .NOT. two_d ) THEN
[1]	1052
[1353]	1053	local_2d = 0.0_wp
[1]	1054	!
	1055	!-- Carry out the averaging (all data are on the PE)
[1551]	1056	DO k = nzb_do, nzt_do
[667]	1057	DO j = nysg, nyng
	1058	DO i = nxlg, nxrg
[1]	1059	local_2d(i,j) = local_2d(i,j) + local_pf(i,j,k)
	1060	ENDDO
	1061	ENDDO
	1062	ENDDO
	1063
[1551]	1064	local_2d = local_2d / ( nzt_do - nzb_do + 1.0_wp)
[1]	1065
	1066	ELSE
	1067	!
	1068	!-- Just store the respective section on the local array
	1069	local_2d = local_pf(:,:,layer_xy)
	1070
	1071	ENDIF
	1072
	1073	#if defined( __parallel )
[1327]	1074	IF ( netcdf_data_format > 4 ) THEN
[1]	1075	!
[1031]	1076	!-- Parallel output in netCDF4/HDF5 format.
[493]	1077	IF ( two_d ) THEN
	1078	iis = 1
	1079	ELSE
	1080	iis = is
	1081	ENDIF
	1082
[1]	1083	#if defined( __netcdf )
[1308]	1084	!
	1085	!-- For parallel output, all cross sections are first stored
	1086	!-- here on a local array and will be written to the output
	1087	!-- file afterwards to increase the performance.
	1088	DO i = nxlg, nxrg
	1089	DO j = nysg, nyng
	1090	local_2d_sections(i,j,iis) = local_2d(i,j)
	1091	ENDDO
	1092	ENDDO
[1]	1093	#endif
[493]	1094	ELSE
[1]	1095
[493]	1096	IF ( data_output_2d_on_each_pe ) THEN
[1]	1097	!
[493]	1098	!-- Output of partial arrays on each PE
	1099	#if defined( __netcdf )
[1327]	1100	IF ( myid == 0 ) THEN
[1320]	1101	WRITE ( 21 ) time_since_reference_point, &
[493]	1102	do2d_xy_time_count(av), av
	1103	ENDIF
	1104	#endif
[759]	1105	DO i = 0, io_blocks-1
	1106	IF ( i == io_group ) THEN
[1551]	1107	WRITE ( 21 ) nxlg, nxrg, nysg, nyng, nysg, nyng
[759]	1108	WRITE ( 21 ) local_2d
	1109	ENDIF
	1110	#if defined( __parallel )
	1111	CALL MPI_BARRIER( comm2d, ierr )
	1112	#endif
	1113	ENDDO
[559]	1114
[493]	1115	ELSE
[1]	1116	!
[493]	1117	!-- PE0 receives partial arrays from all processors and
	1118	!-- then outputs them. Here a barrier has to be set,
	1119	!-- because otherwise "-MPI- FATAL: Remote protocol queue
	1120	!-- full" may occur.
	1121	CALL MPI_BARRIER( comm2d, ierr )
	1122
[667]	1123	ngp = ( nxrg-nxlg+1 ) * ( nyng-nysg+1 )
[493]	1124	IF ( myid == 0 ) THEN
[1]	1125	!
[493]	1126	!-- Local array can be relocated directly.
[667]	1127	total_2d(nxlg:nxrg,nysg:nyng) = local_2d
[1]	1128	!
[493]	1129	!-- Receive data from all other PEs.
	1130	DO n = 1, numprocs-1
[1]	1131	!
[493]	1132	!-- Receive index limits first, then array.
	1133	!-- Index limits are received in arbitrary order from
	1134	!-- the PEs.
[1320]	1135	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1136	MPI_ANY_SOURCE, 0, comm2d, &
[493]	1137	status, ierr )
	1138	sender = status(MPI_SOURCE)
	1139	DEALLOCATE( local_2d )
	1140	ALLOCATE( local_2d(ind(1):ind(2),ind(3):ind(4)) )
[1320]	1141	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1142	MPI_REAL, sender, 1, comm2d, &
[493]	1143	status, ierr )
	1144	total_2d(ind(1):ind(2),ind(3):ind(4)) = local_2d
	1145	ENDDO
[1]	1146	!
[493]	1147	!-- Relocate the local array for the next loop increment
	1148	DEALLOCATE( local_2d )
[667]	1149	ALLOCATE( local_2d(nxlg:nxrg,nysg:nyng) )
[1]	1150
	1151	#if defined( __netcdf )
[1327]	1152	IF ( two_d ) THEN
	1153	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1154	id_var_do2d(av,if), &
	1155	total_2d(0:nx+1,0:ny+1), &
	1156	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
	1157	count = (/ nx+2, ny+2, 1, 1 /) )
	1158	ELSE
	1159	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1160	id_var_do2d(av,if), &
	1161	total_2d(0:nx+1,0:ny+1), &
	1162	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
	1163	count = (/ nx+2, ny+2, 1, 1 /) )
[1]	1164	ENDIF
[1783]	1165	CALL netcdf_handle_error( 'data_output_2d', 54 )
[1]	1166	#endif
	1167
[493]	1168	ELSE
[1]	1169	!
[493]	1170	!-- First send the local index limits to PE0
[667]	1171	ind(1) = nxlg; ind(2) = nxrg
	1172	ind(3) = nysg; ind(4) = nyng
[1320]	1173	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1174	comm2d, ierr )
[1]	1175	!
[493]	1176	!-- Send data to PE0
[1320]	1177	CALL MPI_SEND( local_2d(nxlg,nysg), ngp, &
[493]	1178	MPI_REAL, 0, 1, comm2d, ierr )
	1179	ENDIF
	1180	!
	1181	!-- A barrier has to be set, because otherwise some PEs may
	1182	!-- proceed too fast so that PE0 may receive wrong data on
	1183	!-- tag 0
	1184	CALL MPI_BARRIER( comm2d, ierr )
[1]	1185	ENDIF
[493]	1186
[1]	1187	ENDIF
	1188	#else
	1189	#if defined( __netcdf )
[1327]	1190	IF ( two_d ) THEN
	1191	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1192	id_var_do2d(av,if), &
	1193	local_2d(nxl:nxr+1,nys:nyn+1), &
	1194	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
	1195	count = (/ nx+2, ny+2, 1, 1 /) )
	1196	ELSE
	1197	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1198	id_var_do2d(av,if), &
	1199	local_2d(nxl:nxr+1,nys:nyn+1), &
	1200	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
	1201	count = (/ nx+2, ny+2, 1, 1 /) )
[1]	1202	ENDIF
[1783]	1203	CALL netcdf_handle_error( 'data_output_2d', 447 )
[1]	1204	#endif
	1205	#endif
	1206	do2d_xy_n = do2d_xy_n + 1
	1207	!
	1208	!-- For 2D-arrays (e.g. u*) only one cross-section is available.
	1209	!-- Hence exit loop of output levels.
	1210	IF ( two_d ) THEN
[1703]	1211	IF ( netcdf_data_format < 5 ) two_d = .FALSE.
[1]	1212	EXIT loop1
	1213	ENDIF
	1214
	1215	CASE ( 'xz' )
	1216	!
[1308]	1217	!-- Update the netCDF xz cross section time axis.
	1218	!-- In case of parallel output, this is only done by PE0
	1219	!-- to increase the performance.
	1220	IF ( simulated_time /= do2d_xz_last_time(av) ) THEN
	1221	do2d_xz_time_count(av) = do2d_xz_time_count(av) + 1
	1222	do2d_xz_last_time(av) = simulated_time
	1223	IF ( myid == 0 ) THEN
[1327]	1224	IF ( .NOT. data_output_2d_on_each_pe &
	1225	.OR. netcdf_data_format > 4 ) &
[493]	1226	THEN
[1]	1227	#if defined( __netcdf )
	1228	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1229	id_var_time_xz(av), &
[291]	1230	(/ time_since_reference_point /), &
[1]	1231	start = (/ do2d_xz_time_count(av) /), &
	1232	count = (/ 1 /) )
[1783]	1233	CALL netcdf_handle_error( 'data_output_2d', 56 )
[1]	1234	#endif
	1235	ENDIF
	1236	ENDIF
	1237	ENDIF
[667]	1238
[1]	1239	!
	1240	!-- If required, carry out averaging along y
[1960]	1241	IF ( section(is,s_ind) == -1 ) THEN
[1]	1242
[1551]	1243	ALLOCATE( local_2d_l(nxlg:nxrg,nzb_do:nzt_do) )
[1353]	1244	local_2d_l = 0.0_wp
[1551]	1245	ngp = ( nxrg-nxlg + 1 ) * ( nzt_do-nzb_do + 1 )
[1]	1246	!
	1247	!-- First local averaging on the PE
[1551]	1248	DO k = nzb_do, nzt_do
[1]	1249	DO j = nys, nyn
[667]	1250	DO i = nxlg, nxrg
[1320]	1251	local_2d_l(i,k) = local_2d_l(i,k) + &
[1]	1252	local_pf(i,j,k)
	1253	ENDDO
	1254	ENDDO
	1255	ENDDO
	1256	#if defined( __parallel )
	1257	!
	1258	!-- Now do the averaging over all PEs along y
[622]	1259	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1260	CALL MPI_ALLREDUCE( local_2d_l(nxlg,nzb_do), &
	1261	local_2d(nxlg,nzb_do), ngp, MPI_REAL, &
[1]	1262	MPI_SUM, comm1dy, ierr )
	1263	#else
	1264	local_2d = local_2d_l
	1265	#endif
[1353]	1266	local_2d = local_2d / ( ny + 1.0_wp )
[1]	1267
	1268	DEALLOCATE( local_2d_l )
	1269
	1270	ELSE
	1271	!
	1272	!-- Just store the respective section on the local array
	1273	!-- (but only if it is available on this PE!)
[1960]	1274	IF ( section(is,s_ind) >= nys .AND. section(is,s_ind) <= nyn ) &
[1]	1275	THEN
[1960]	1276	local_2d = local_pf(:,section(is,s_ind),nzb_do:nzt_do)
[1]	1277	ENDIF
	1278
	1279	ENDIF
	1280
	1281	#if defined( __parallel )
[1327]	1282	IF ( netcdf_data_format > 4 ) THEN
[1]	1283	!
[1031]	1284	!-- Output in netCDF4/HDF5 format.
[493]	1285	!-- Output only on those PEs where the respective cross
	1286	!-- sections reside. Cross sections averaged along y are
	1287	!-- output on the respective first PE along y (myidy=0).
[1960]	1288	IF ( ( section(is,s_ind) >= nys .AND. &
	1289	section(is,s_ind) <= nyn ) .OR. &
	1290	( section(is,s_ind) == -1 .AND. myidy == 0 ) ) THEN
[1]	1291	#if defined( __netcdf )
[493]	1292	!
[1308]	1293	!-- For parallel output, all cross sections are first
	1294	!-- stored here on a local array and will be written to the
	1295	!-- output file afterwards to increase the performance.
	1296	DO i = nxlg, nxrg
[1551]	1297	DO k = nzb_do, nzt_do
[1308]	1298	local_2d_sections_l(i,is,k) = local_2d(i,k)
	1299	ENDDO
	1300	ENDDO
[1]	1301	#endif
	1302	ENDIF
	1303
	1304	ELSE
	1305
[493]	1306	IF ( data_output_2d_on_each_pe ) THEN
[1]	1307	!
[493]	1308	!-- Output of partial arrays on each PE. If the cross
	1309	!-- section does not reside on the PE, output special
	1310	!-- index values.
	1311	#if defined( __netcdf )
[1327]	1312	IF ( myid == 0 ) THEN
[1320]	1313	WRITE ( 22 ) time_since_reference_point, &
[493]	1314	do2d_xz_time_count(av), av
	1315	ENDIF
	1316	#endif
[759]	1317	DO i = 0, io_blocks-1
	1318	IF ( i == io_group ) THEN
[1960]	1319	IF ( ( section(is,s_ind) >= nys .AND. &
	1320	section(is,s_ind) <= nyn ) .OR. &
	1321	( section(is,s_ind) == -1 .AND. &
[1320]	1322	nys-1 == -1 ) ) &
[759]	1323	THEN
[1551]	1324	WRITE (22) nxlg, nxrg, nzb_do, nzt_do, nzb, nzt+1
[759]	1325	WRITE (22) local_2d
	1326	ELSE
[1551]	1327	WRITE (22) -1, -1, -1, -1, -1, -1
[759]	1328	ENDIF
	1329	ENDIF
	1330	#if defined( __parallel )
	1331	CALL MPI_BARRIER( comm2d, ierr )
	1332	#endif
	1333	ENDDO
[493]	1334
	1335	ELSE
[1]	1336	!
[493]	1337	!-- PE0 receives partial arrays from all processors of the
	1338	!-- respective cross section and outputs them. Here a
	1339	!-- barrier has to be set, because otherwise
	1340	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
	1341	CALL MPI_BARRIER( comm2d, ierr )
	1342
[1551]	1343	ngp = ( nxrg-nxlg + 1 ) * ( nzt_do-nzb_do + 1 )
[493]	1344	IF ( myid == 0 ) THEN
[1]	1345	!
[493]	1346	!-- Local array can be relocated directly.
[1960]	1347	IF ( ( section(is,s_ind) >= nys .AND. &
	1348	section(is,s_ind) <= nyn ) .OR. &
	1349	( section(is,s_ind) == -1 .AND. &
	1350	nys-1 == -1 ) ) THEN
[1551]	1351	total_2d(nxlg:nxrg,nzb_do:nzt_do) = local_2d
[493]	1352	ENDIF
[1]	1353	!
[493]	1354	!-- Receive data from all other PEs.
	1355	DO n = 1, numprocs-1
	1356	!
	1357	!-- Receive index limits first, then array.
	1358	!-- Index limits are received in arbitrary order from
	1359	!-- the PEs.
[1320]	1360	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1361	MPI_ANY_SOURCE, 0, comm2d, &
[1]	1362	status, ierr )
[493]	1363	!
	1364	!-- Not all PEs have data for XZ-cross-section.
	1365	IF ( ind(1) /= -9999 ) THEN
	1366	sender = status(MPI_SOURCE)
	1367	DEALLOCATE( local_2d )
[1320]	1368	ALLOCATE( local_2d(ind(1):ind(2), &
[493]	1369	ind(3):ind(4)) )
	1370	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1371	MPI_REAL, sender, 1, comm2d, &
	1372	status, ierr )
[1320]	1373	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
[493]	1374	local_2d
	1375	ENDIF
	1376	ENDDO
	1377	!
	1378	!-- Relocate the local array for the next loop increment
	1379	DEALLOCATE( local_2d )
[1551]	1380	ALLOCATE( local_2d(nxlg:nxrg,nzb_do:nzt_do) )
[1]	1381
	1382	#if defined( __netcdf )
[1327]	1383	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1384	id_var_do2d(av,if), &
[1551]	1385	total_2d(0:nx+1,nzb_do:nzt_do),&
[1327]	1386	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
[1551]	1387	count = (/ nx+2, 1, nzt_do-nzb_do+1, 1 /) )
[1783]	1388	CALL netcdf_handle_error( 'data_output_2d', 58 )
[1]	1389	#endif
	1390
[493]	1391	ELSE
[1]	1392	!
[493]	1393	!-- If the cross section resides on the PE, send the
	1394	!-- local index limits, otherwise send -9999 to PE0.
[1960]	1395	IF ( ( section(is,s_ind) >= nys .AND. &
	1396	section(is,s_ind) <= nyn ) .OR. &
	1397	( section(is,s_ind) == -1 .AND. nys-1 == -1 ) ) &
[493]	1398	THEN
[667]	1399	ind(1) = nxlg; ind(2) = nxrg
[1551]	1400	ind(3) = nzb_do; ind(4) = nzt_do
[493]	1401	ELSE
	1402	ind(1) = -9999; ind(2) = -9999
	1403	ind(3) = -9999; ind(4) = -9999
	1404	ENDIF
[1320]	1405	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1406	comm2d, ierr )
	1407	!
	1408	!-- If applicable, send data to PE0.
	1409	IF ( ind(1) /= -9999 ) THEN
[1551]	1410	CALL MPI_SEND( local_2d(nxlg,nzb_do), ngp, &
[493]	1411	MPI_REAL, 0, 1, comm2d, ierr )
	1412	ENDIF
[1]	1413	ENDIF
	1414	!
[493]	1415	!-- A barrier has to be set, because otherwise some PEs may
	1416	!-- proceed too fast so that PE0 may receive wrong data on
	1417	!-- tag 0
	1418	CALL MPI_BARRIER( comm2d, ierr )
[1]	1419	ENDIF
[493]	1420
[1]	1421	ENDIF
	1422	#else
	1423	#if defined( __netcdf )
[1327]	1424	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1425	id_var_do2d(av,if), &
[1551]	1426	local_2d(nxl:nxr+1,nzb_do:nzt_do), &
[1327]	1427	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
[1551]	1428	count = (/ nx+2, 1, nzt_do-nzb_do+1, 1 /) )
[1783]	1429	CALL netcdf_handle_error( 'data_output_2d', 451 )
[1]	1430	#endif
	1431	#endif
	1432	do2d_xz_n = do2d_xz_n + 1
	1433
	1434	CASE ( 'yz' )
	1435	!
[1308]	1436	!-- Update the netCDF yz cross section time axis.
	1437	!-- In case of parallel output, this is only done by PE0
	1438	!-- to increase the performance.
	1439	IF ( simulated_time /= do2d_yz_last_time(av) ) THEN
	1440	do2d_yz_time_count(av) = do2d_yz_time_count(av) + 1
	1441	do2d_yz_last_time(av) = simulated_time
	1442	IF ( myid == 0 ) THEN
[1327]	1443	IF ( .NOT. data_output_2d_on_each_pe &
	1444	.OR. netcdf_data_format > 4 ) &
[493]	1445	THEN
[1]	1446	#if defined( __netcdf )
	1447	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1448	id_var_time_yz(av), &
[291]	1449	(/ time_since_reference_point /), &
[1]	1450	start = (/ do2d_yz_time_count(av) /), &
	1451	count = (/ 1 /) )
[1783]	1452	CALL netcdf_handle_error( 'data_output_2d', 59 )
[1]	1453	#endif
	1454	ENDIF
	1455	ENDIF
[1308]	1456	ENDIF
[493]	1457
[1]	1458	!
	1459	!-- If required, carry out averaging along x
[1960]	1460	IF ( section(is,s_ind) == -1 ) THEN
[1]	1461
[1551]	1462	ALLOCATE( local_2d_l(nysg:nyng,nzb_do:nzt_do) )
[1353]	1463	local_2d_l = 0.0_wp
[1551]	1464	ngp = ( nyng-nysg+1 ) * ( nzt_do-nzb_do+1 )
[1]	1465	!
	1466	!-- First local averaging on the PE
[1551]	1467	DO k = nzb_do, nzt_do
[667]	1468	DO j = nysg, nyng
[1]	1469	DO i = nxl, nxr
[1320]	1470	local_2d_l(j,k) = local_2d_l(j,k) + &
[1]	1471	local_pf(i,j,k)
	1472	ENDDO
	1473	ENDDO
	1474	ENDDO
	1475	#if defined( __parallel )
	1476	!
	1477	!-- Now do the averaging over all PEs along x
[622]	1478	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1479	CALL MPI_ALLREDUCE( local_2d_l(nysg,nzb_do), &
	1480	local_2d(nysg,nzb_do), ngp, MPI_REAL, &
[1]	1481	MPI_SUM, comm1dx, ierr )
	1482	#else
	1483	local_2d = local_2d_l
	1484	#endif
[1353]	1485	local_2d = local_2d / ( nx + 1.0_wp )
[1]	1486
	1487	DEALLOCATE( local_2d_l )
	1488
	1489	ELSE
	1490	!
	1491	!-- Just store the respective section on the local array
	1492	!-- (but only if it is available on this PE!)
[1960]	1493	IF ( section(is,s_ind) >= nxl .AND. section(is,s_ind) <= nxr ) &
[1]	1494	THEN
[1960]	1495	local_2d = local_pf(section(is,s_ind),:,nzb_do:nzt_do)
[1]	1496	ENDIF
	1497
	1498	ENDIF
	1499
	1500	#if defined( __parallel )
[1327]	1501	IF ( netcdf_data_format > 4 ) THEN
[1]	1502	!
[1031]	1503	!-- Output in netCDF4/HDF5 format.
[493]	1504	!-- Output only on those PEs where the respective cross
	1505	!-- sections reside. Cross sections averaged along x are
	1506	!-- output on the respective first PE along x (myidx=0).
[1960]	1507	IF ( ( section(is,s_ind) >= nxl .AND. &
	1508	section(is,s_ind) <= nxr ) .OR. &
	1509	( section(is,s_ind) == -1 .AND. myidx == 0 ) ) THEN
[1]	1510	#if defined( __netcdf )
[493]	1511	!
[1308]	1512	!-- For parallel output, all cross sections are first
	1513	!-- stored here on a local array and will be written to the
	1514	!-- output file afterwards to increase the performance.
	1515	DO j = nysg, nyng
[1551]	1516	DO k = nzb_do, nzt_do
[1308]	1517	local_2d_sections_l(is,j,k) = local_2d(j,k)
	1518	ENDDO
	1519	ENDDO
[1]	1520	#endif
	1521	ENDIF
	1522
	1523	ELSE
	1524
[493]	1525	IF ( data_output_2d_on_each_pe ) THEN
[1]	1526	!
[493]	1527	!-- Output of partial arrays on each PE. If the cross
	1528	!-- section does not reside on the PE, output special
	1529	!-- index values.
	1530	#if defined( __netcdf )
[1327]	1531	IF ( myid == 0 ) THEN
[1320]	1532	WRITE ( 23 ) time_since_reference_point, &
[493]	1533	do2d_yz_time_count(av), av
	1534	ENDIF
	1535	#endif
[759]	1536	DO i = 0, io_blocks-1
	1537	IF ( i == io_group ) THEN
[1960]	1538	IF ( ( section(is,s_ind) >= nxl .AND. &
	1539	section(is,s_ind) <= nxr ) .OR. &
	1540	( section(is,s_ind) == -1 .AND. &
[1320]	1541	nxl-1 == -1 ) ) &
[759]	1542	THEN
[1551]	1543	WRITE (23) nysg, nyng, nzb_do, nzt_do, nzb, nzt+1
[759]	1544	WRITE (23) local_2d
	1545	ELSE
[1551]	1546	WRITE (23) -1, -1, -1, -1, -1, -1
[759]	1547	ENDIF
	1548	ENDIF
	1549	#if defined( __parallel )
	1550	CALL MPI_BARRIER( comm2d, ierr )
	1551	#endif
	1552	ENDDO
[493]	1553
	1554	ELSE
[1]	1555	!
[493]	1556	!-- PE0 receives partial arrays from all processors of the
	1557	!-- respective cross section and outputs them. Here a
	1558	!-- barrier has to be set, because otherwise
	1559	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
	1560	CALL MPI_BARRIER( comm2d, ierr )
	1561
[1551]	1562	ngp = ( nyng-nysg+1 ) * ( nzt_do-nzb_do+1 )
[493]	1563	IF ( myid == 0 ) THEN
[1]	1564	!
[493]	1565	!-- Local array can be relocated directly.
[1960]	1566	IF ( ( section(is,s_ind) >= nxl .AND. &
	1567	section(is,s_ind) <= nxr ) .OR. &
	1568	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
[493]	1569	THEN
[1551]	1570	total_2d(nysg:nyng,nzb_do:nzt_do) = local_2d
[493]	1571	ENDIF
[1]	1572	!
[493]	1573	!-- Receive data from all other PEs.
	1574	DO n = 1, numprocs-1
	1575	!
	1576	!-- Receive index limits first, then array.
	1577	!-- Index limits are received in arbitrary order from
	1578	!-- the PEs.
[1320]	1579	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1580	MPI_ANY_SOURCE, 0, comm2d, &
[1]	1581	status, ierr )
[493]	1582	!
	1583	!-- Not all PEs have data for YZ-cross-section.
	1584	IF ( ind(1) /= -9999 ) THEN
	1585	sender = status(MPI_SOURCE)
	1586	DEALLOCATE( local_2d )
[1320]	1587	ALLOCATE( local_2d(ind(1):ind(2), &
[493]	1588	ind(3):ind(4)) )
	1589	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1590	MPI_REAL, sender, 1, comm2d, &
	1591	status, ierr )
[1320]	1592	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
[493]	1593	local_2d
	1594	ENDIF
	1595	ENDDO
	1596	!
	1597	!-- Relocate the local array for the next loop increment
	1598	DEALLOCATE( local_2d )
[1551]	1599	ALLOCATE( local_2d(nysg:nyng,nzb_do:nzt_do) )
[1]	1600
	1601	#if defined( __netcdf )
[1327]	1602	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1603	id_var_do2d(av,if), &
[1551]	1604	total_2d(0:ny+1,nzb_do:nzt_do),&
[1327]	1605	start = (/ is, 1, 1, do2d_yz_time_count(av) /), &
[1551]	1606	count = (/ 1, ny+2, nzt_do-nzb_do+1, 1 /) )
[1783]	1607	CALL netcdf_handle_error( 'data_output_2d', 61 )
[1]	1608	#endif
	1609
[493]	1610	ELSE
[1]	1611	!
[493]	1612	!-- If the cross section resides on the PE, send the
	1613	!-- local index limits, otherwise send -9999 to PE0.
[1960]	1614	IF ( ( section(is,s_ind) >= nxl .AND. &
	1615	section(is,s_ind) <= nxr ) .OR. &
	1616	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
[493]	1617	THEN
[667]	1618	ind(1) = nysg; ind(2) = nyng
[1551]	1619	ind(3) = nzb_do; ind(4) = nzt_do
[493]	1620	ELSE
	1621	ind(1) = -9999; ind(2) = -9999
	1622	ind(3) = -9999; ind(4) = -9999
	1623	ENDIF
[1320]	1624	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1625	comm2d, ierr )
	1626	!
	1627	!-- If applicable, send data to PE0.
	1628	IF ( ind(1) /= -9999 ) THEN
[1551]	1629	CALL MPI_SEND( local_2d(nysg,nzb_do), ngp, &
[493]	1630	MPI_REAL, 0, 1, comm2d, ierr )
	1631	ENDIF
[1]	1632	ENDIF
	1633	!
[493]	1634	!-- A barrier has to be set, because otherwise some PEs may
	1635	!-- proceed too fast so that PE0 may receive wrong data on
	1636	!-- tag 0
	1637	CALL MPI_BARRIER( comm2d, ierr )
[1]	1638	ENDIF
[493]	1639
[1]	1640	ENDIF
	1641	#else
	1642	#if defined( __netcdf )
[1327]	1643	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1644	id_var_do2d(av,if), &
[1551]	1645	local_2d(nys:nyn+1,nzb_do:nzt_do), &
[1327]	1646	start = (/ is, 1, 1, do2d_xz_time_count(av) /), &
[1551]	1647	count = (/ 1, ny+2, nzt_do-nzb_do+1, 1 /) )
[1783]	1648	CALL netcdf_handle_error( 'data_output_2d', 452 )
[1]	1649	#endif
	1650	#endif
	1651	do2d_yz_n = do2d_yz_n + 1
	1652
	1653	END SELECT
	1654
	1655	is = is + 1
	1656	ENDDO loop1
	1657
[1308]	1658	!
	1659	!-- For parallel output, all data were collected before on a local array
	1660	!-- and are written now to the netcdf file. This must be done to increase
	1661	!-- the performance of the parallel output.
	1662	#if defined( __netcdf )
[1327]	1663	IF ( netcdf_data_format > 4 ) THEN
[1308]	1664
	1665	SELECT CASE ( mode )
	1666
	1667	CASE ( 'xy' )
	1668	IF ( two_d ) THEN
[1703]	1669	nis = 1
	1670	two_d = .FALSE.
[1308]	1671	ELSE
[1703]	1672	nis = ns
[1308]	1673	ENDIF
	1674	!
	1675	!-- Do not output redundant ghost point data except for the
	1676	!-- boundaries of the total domain.
	1677	IF ( nxr == nx .AND. nyn /= ny ) THEN
	1678	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1679	id_var_do2d(av,if), &
	1680	local_2d_sections(nxl:nxr+1, &
[1703]	1681	nys:nyn,1:nis), &
[1308]	1682	start = (/ nxl+1, nys+1, 1, &
	1683	do2d_xy_time_count(av) /), &
	1684	count = (/ nxr-nxl+2, &
[1703]	1685	nyn-nys+1, nis, 1 &
[1308]	1686	/) )
	1687	ELSEIF ( nxr /= nx .AND. nyn == ny ) THEN
	1688	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1689	id_var_do2d(av,if), &
	1690	local_2d_sections(nxl:nxr, &
[1703]	1691	nys:nyn+1,1:nis), &
[1308]	1692	start = (/ nxl+1, nys+1, 1, &
	1693	do2d_xy_time_count(av) /), &
	1694	count = (/ nxr-nxl+1, &
[1703]	1695	nyn-nys+2, nis, 1 &
[1308]	1696	/) )
	1697	ELSEIF ( nxr == nx .AND. nyn == ny ) THEN
	1698	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1699	id_var_do2d(av,if), &
	1700	local_2d_sections(nxl:nxr+1, &
[1703]	1701	nys:nyn+1,1:nis), &
[1308]	1702	start = (/ nxl+1, nys+1, 1, &
	1703	do2d_xy_time_count(av) /), &
	1704	count = (/ nxr-nxl+2, &
[1703]	1705	nyn-nys+2, nis, 1 &
[1308]	1706	/) )
	1707	ELSE
	1708	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1709	id_var_do2d(av,if), &
	1710	local_2d_sections(nxl:nxr, &
[1703]	1711	nys:nyn,1:nis), &
[1308]	1712	start = (/ nxl+1, nys+1, 1, &
	1713	do2d_xy_time_count(av) /), &
	1714	count = (/ nxr-nxl+1, &
[1703]	1715	nyn-nys+1, nis, 1 &
[1308]	1716	/) )
	1717	ENDIF
	1718
[1783]	1719	CALL netcdf_handle_error( 'data_output_2d', 55 )
[1308]	1720
	1721	CASE ( 'xz' )
	1722	!
	1723	!-- First, all PEs get the information of all cross-sections.
	1724	!-- Then the data are written to the output file by all PEs
	1725	!-- while NF90_COLLECTIVE is set in subroutine
	1726	!-- define_netcdf_header. Although redundant information are
	1727	!-- written to the output file in that case, the performance
	1728	!-- is significantly better compared to the case where only
	1729	!-- the first row of PEs in x-direction (myidx = 0) is given
	1730	!-- the output while NF90_INDEPENDENT is set.
	1731	IF ( npey /= 1 ) THEN
	1732
	1733	#if defined( __parallel )
	1734	!
	1735	!-- Distribute data over all PEs along y
[1551]	1736	ngp = ( nxrg-nxlg+1 ) * ( nzt_do-nzb_do+1 ) * ns
[1308]	1737	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1738	CALL MPI_ALLREDUCE( local_2d_sections_l(nxlg,1,nzb_do), &
	1739	local_2d_sections(nxlg,1,nzb_do), &
[1308]	1740	ngp, MPI_REAL, MPI_SUM, comm1dy, &
	1741	ierr )
	1742	#else
	1743	local_2d_sections = local_2d_sections_l
	1744	#endif
	1745	ENDIF
	1746	!
	1747	!-- Do not output redundant ghost point data except for the
	1748	!-- boundaries of the total domain.
	1749	IF ( nxr == nx ) THEN
	1750	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1751	id_var_do2d(av,if), &
	1752	local_2d_sections(nxl:nxr+1,1:ns, &
[1551]	1753	nzb_do:nzt_do), &
[1308]	1754	start = (/ nxl+1, 1, 1, &
	1755	do2d_xz_time_count(av) /), &
[1551]	1756	count = (/ nxr-nxl+2, ns, nzt_do-nzb_do+1, &
[1308]	1757	1 /) )
	1758	ELSE
	1759	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1760	id_var_do2d(av,if), &
	1761	local_2d_sections(nxl:nxr,1:ns, &
[1551]	1762	nzb_do:nzt_do), &
[1308]	1763	start = (/ nxl+1, 1, 1, &
	1764	do2d_xz_time_count(av) /), &
[1551]	1765	count = (/ nxr-nxl+1, ns, nzt_do-nzb_do+1, &
[1308]	1766	1 /) )
	1767	ENDIF
	1768
[1783]	1769	CALL netcdf_handle_error( 'data_output_2d', 57 )
[1308]	1770
	1771	CASE ( 'yz' )
	1772	!
	1773	!-- First, all PEs get the information of all cross-sections.
	1774	!-- Then the data are written to the output file by all PEs
	1775	!-- while NF90_COLLECTIVE is set in subroutine
	1776	!-- define_netcdf_header. Although redundant information are
	1777	!-- written to the output file in that case, the performance
	1778	!-- is significantly better compared to the case where only
	1779	!-- the first row of PEs in y-direction (myidy = 0) is given
	1780	!-- the output while NF90_INDEPENDENT is set.
	1781	IF ( npex /= 1 ) THEN
	1782
	1783	#if defined( __parallel )
	1784	!
	1785	!-- Distribute data over all PEs along x
	1786	ngp = ( nyng-nysg+1 ) * ( nzt-nzb + 2 ) * ns
	1787	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1788	CALL MPI_ALLREDUCE( local_2d_sections_l(1,nysg,nzb_do), &
	1789	local_2d_sections(1,nysg,nzb_do), &
[1308]	1790	ngp, MPI_REAL, MPI_SUM, comm1dx, &
	1791	ierr )
	1792	#else
	1793	local_2d_sections = local_2d_sections_l
	1794	#endif
	1795	ENDIF
	1796	!
	1797	!-- Do not output redundant ghost point data except for the
	1798	!-- boundaries of the total domain.
	1799	IF ( nyn == ny ) THEN
	1800	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1801	id_var_do2d(av,if), &
	1802	local_2d_sections(1:ns, &
[1551]	1803	nys:nyn+1,nzb_do:nzt_do), &
[1308]	1804	start = (/ 1, nys+1, 1, &
	1805	do2d_yz_time_count(av) /), &
	1806	count = (/ ns, nyn-nys+2, &
[1551]	1807	nzt_do-nzb_do+1, 1 /) )
[1308]	1808	ELSE
	1809	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1810	id_var_do2d(av,if), &
	1811	local_2d_sections(1:ns,nys:nyn, &
[1551]	1812	nzb_do:nzt_do), &
[1308]	1813	start = (/ 1, nys+1, 1, &
	1814	do2d_yz_time_count(av) /), &
	1815	count = (/ ns, nyn-nys+1, &
[1551]	1816	nzt_do-nzb_do+1, 1 /) )
[1308]	1817	ENDIF
	1818
[1783]	1819	CALL netcdf_handle_error( 'data_output_2d', 60 )
[1308]	1820
	1821	CASE DEFAULT
	1822	message_string = 'unknown cross-section: ' // TRIM( mode )
	1823	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
	1824
	1825	END SELECT
	1826
	1827	ENDIF
[1311]	1828	#endif
[1]	1829	ENDIF
	1830
	1831	if = if + 1
	1832	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
	1833	do2d_mode = do2d(av,if)(l-1:l)
	1834
	1835	ENDDO
	1836
	1837	!
	1838	!-- Deallocate temporary arrays.
	1839	IF ( ALLOCATED( level_z ) ) DEALLOCATE( level_z )
[1308]	1840	IF ( netcdf_data_format > 4 ) THEN
	1841	DEALLOCATE( local_pf, local_2d, local_2d_sections )
	1842	IF( mode == 'xz' .OR. mode == 'yz' ) DEALLOCATE( local_2d_sections_l )
	1843	ENDIF
[1]	1844	#if defined( __parallel )
	1845	IF ( .NOT. data_output_2d_on_each_pe .AND. myid == 0 ) THEN
	1846	DEALLOCATE( total_2d )
	1847	ENDIF
	1848	#endif
	1849
	1850	!
	1851	!-- Close plot output file.
[1960]	1852	file_id = 20 + s_ind
[1]	1853
	1854	IF ( data_output_2d_on_each_pe ) THEN
[759]	1855	DO i = 0, io_blocks-1
	1856	IF ( i == io_group ) THEN
	1857	CALL close_file( file_id )
	1858	ENDIF
	1859	#if defined( __parallel )
	1860	CALL MPI_BARRIER( comm2d, ierr )
	1861	#endif
	1862	ENDDO
[1]	1863	ELSE
	1864	IF ( myid == 0 ) CALL close_file( file_id )
	1865	ENDIF
	1866
[1318]	1867	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
[1]	1868
	1869	END SUBROUTINE data_output_2d

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |