Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 2000

Last change on this file since 2000 was 2000, checked in by knoop, 8 years ago
Forced header and separation lines into 80 columns
Property svn:keywords set to `Id`
File size: 77.5 KB

Rev	Line
[1682]	1	!> @file data_output_2d.f90
[2000]	2	!------------------------------------------------------------------------------!
[1036]	3	! This file is part of PALM.
	4	!
[2000]	5	! PALM is free software: you can redistribute it and/or modify it under the
	6	! terms of the GNU General Public License as published by the Free Software
	7	! Foundation, either version 3 of the License, or (at your option) any later
	8	! version.
[1036]	9	!
	10	! PALM is distributed in the hope that it will be useful, but WITHOUT ANY
	11	! WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
	12	! A PARTICULAR PURPOSE. See the GNU General Public License for more details.
	13	!
	14	! You should have received a copy of the GNU General Public License along with
	15	! PALM. If not, see <http://www.gnu.org/licenses/>.
	16	!
[1818]	17	! Copyright 1997-2016 Leibniz Universitaet Hannover
[2000]	18	!------------------------------------------------------------------------------!
[1036]	19	!
[254]	20	! Current revisions:
[1]	21	! -----------------
[2000]	22	! Forced header and separation lines into 80 columns
[1961]	23	!
[1552]	24	! Former revisions:
	25	! -----------------
	26	! $Id: data_output_2d.f90 2000 2016-08-20 18:09:15Z knoop $
	27	!
[1981]	28	! 1980 2016-07-29 15:51:57Z suehring
	29	! Bugfix, in order to steer user-defined output, setting flag found explicitly
	30	! to .F.
	31	!
[1977]	32	! 1976 2016-07-27 13:28:04Z maronga
	33	! Output of radiation quantities is now done directly in the respective module
	34	!
[1973]	35	! 1972 2016-07-26 07:52:02Z maronga
[1976]	36	! Output of land surface quantities is now done directly in the respective
	37	! module
[1973]	38	!
[1961]	39	! 1960 2016-07-12 16:34:24Z suehring
	40	! Scalar surface flux added
	41	! Rename INTEGER variable s into s_ind, as s is already assigned to scalar
	42	!
[1851]	43	! 1849 2016-04-08 11:33:18Z hoffmann
	44	! precipitation_amount, precipitation_rate, prr moved to arrays_3d
	45	!
[1823]	46	! 1822 2016-04-07 07:49:42Z hoffmann
	47	! Output of bulk cloud physics simplified.
	48	!
[1789]	49	! 1788 2016-03-10 11:01:04Z maronga
	50	! Added output of z0q
	51	!
[1784]	52	! 1783 2016-03-06 18:36:17Z raasch
	53	! name change of netcdf routines and module + related changes
	54	!
[1746]	55	! 1745 2016-02-05 13:06:51Z gronemeier
	56	! Bugfix: test if time axis limit exceeds moved to point after call of check_open
	57	!
[1704]	58	! 1703 2015-11-02 12:38:44Z raasch
	59	! bugfix for output of single (*) xy-sections in case of parallel netcdf I/O
	60	!
[1702]	61	! 1701 2015-11-02 07:43:04Z maronga
	62	! Bugfix in output of RRTGM data
	63	!
[1692]	64	! 1691 2015-10-26 16:17:44Z maronga
	65	! Added output of Obukhov length (ol) and radiative heating rates for RRTMG.
	66	! Formatting corrections.
	67	!
[1683]	68	! 1682 2015-10-07 23:56:08Z knoop
	69	! Code annotations made doxygen readable
	70	!
[1586]	71	! 1585 2015-04-30 07:05:52Z maronga
	72	! Added support for RRTMG
	73	!
[1556]	74	! 1555 2015-03-04 17:44:27Z maronga
	75	! Added output of r_a and r_s
	76	!
[1552]	77	! 1551 2015-03-03 14:18:16Z maronga
[1551]	78	! Added suppport for land surface model and radiation model output. In the course
	79	! of this action, the limits for vertical loops have been changed (from nzb and
	80	! nzt+1 to nzb_do and nzt_do, respectively in order to allow soil model output).
	81	! Moreover, a new vertical grid zs was introduced.
[1329]	82	!
[1360]	83	! 1359 2014-04-11 17:15:14Z hoffmann
	84	! New particle structure integrated.
	85	!
[1354]	86	! 1353 2014-04-08 15:21:23Z heinze
	87	! REAL constants provided with KIND-attribute
	88	!
[1329]	89	! 1327 2014-03-21 11:00:16Z raasch
	90	! parts concerning iso2d output removed,
	91	! -netcdf output queries
	92	!
[1321]	93	! 1320 2014-03-20 08:40:49Z raasch
[1320]	94	! ONLY-attribute added to USE-statements,
	95	! kind-parameters added to all INTEGER and REAL declaration statements,
	96	! kinds are defined in new module kinds,
	97	! revision history before 2012 removed,
	98	! comment fields (!:) to be used for variable explanations added to
	99	! all variable declaration statements
[1309]	100	!
[1319]	101	! 1318 2014-03-17 13:35:16Z raasch
	102	! barrier argument removed from cpu_log.
	103	! module interfaces removed
	104	!
[1312]	105	! 1311 2014-03-14 12:13:39Z heinze
	106	! bugfix: close #if defined( __netcdf )
	107	!
[1309]	108	! 1308 2014-03-13 14:58:42Z fricke
[1308]	109	! +local_2d_sections, local_2d_sections_l, ns
	110	! Check, if the limit of the time dimension is exceeded for parallel output
	111	! To increase the performance for parallel output, the following is done:
	112	! - Update of time axis is only done by PE0
	113	! - Cross sections are first stored on a local array and are written
	114	! collectively to the output file by all PEs.
[674]	115	!
[1116]	116	! 1115 2013-03-26 18:16:16Z hoffmann
	117	! ql is calculated by calc_liquid_water_content
	118	!
[1077]	119	! 1076 2012-12-05 08:30:18Z hoffmann
	120	! Bugfix in output of ql
	121	!
[1066]	122	! 1065 2012-11-22 17:42:36Z hoffmann
	123	! Bugfix: Output of cross sections of ql
	124	!
[1054]	125	! 1053 2012-11-13 17:11:03Z hoffmann
	126	! +qr, nr, qc and cross sections
	127	!
[1037]	128	! 1036 2012-10-22 13:43:42Z raasch
	129	! code put under GPL (PALM 3.9)
	130	!
[1035]	131	! 1031 2012-10-19 14:35:30Z raasch
	132	! netCDF4 without parallel file support implemented
	133	!
[1008]	134	! 1007 2012-09-19 14:30:36Z franke
	135	! Bugfix: missing calculation of ql_vp added
	136	!
[979]	137	! 978 2012-08-09 08:28:32Z fricke
	138	! +z0h
	139	!
[1]	140	! Revision 1.1 1997/08/11 06:24:09 raasch
	141	! Initial revision
	142	!
	143	!
	144	! Description:
	145	! ------------
[1682]	146	!> Data output of horizontal cross-sections in netCDF format or binary format
	147	!> compatible to old graphic software iso2d.
	148	!> Attention: The position of the sectional planes is still not always computed
	149	!> --------- correctly. (zu is used always)!
[1]	150	!------------------------------------------------------------------------------!
[1682]	151	SUBROUTINE data_output_2d( mode, av )
	152
[1]	153
[1320]	154	USE arrays_3d, &
[1849]	155	ONLY: dzw, e, nr, ol, p, pt, precipitation_amount, precipitation_rate,&
[1960]	156	prr,q, qc, ql, ql_c, ql_v, ql_vp, qr, qsws, rho, s, sa, shf, &
	157	ssws, tend, ts, u, us, v, vpt, w, z0, z0h, z0q, zu, zw
[1320]	158
[1]	159	USE averaging
[1320]	160
	161	USE cloud_parameters, &
[1849]	162	ONLY: hyrho, l_d_cp, pt_d_t
[1320]	163
	164	USE control_parameters, &
	165	ONLY: cloud_physics, data_output_2d_on_each_pe, data_output_xy, &
	166	data_output_xz, data_output_yz, do2d, &
	167	do2d_xy_last_time, do2d_xy_n, do2d_xy_time_count, &
	168	do2d_xz_last_time, do2d_xz_n, do2d_xz_time_count, &
	169	do2d_yz_last_time, do2d_yz_n, do2d_yz_time_count, &
[1822]	170	ibc_uv_b, io_blocks, io_group, message_string, &
	171	ntdim_2d_xy, ntdim_2d_xz, ntdim_2d_yz, &
	172	psolver, section, simulated_time, simulated_time_chr, &
	173	time_since_reference_point
[1320]	174
	175	USE cpulog, &
	176	ONLY: cpu_log, log_point
	177
	178	USE grid_variables, &
	179	ONLY: dx, dy
	180
	181	USE indices, &
	182	ONLY: nbgp, nx, nxl, nxlg, nxr, nxrg, ny, nyn, nyng, nys, nysg, &
	183	nz, nzb, nzt
	184
	185	USE kinds
[1551]	186
	187	USE land_surface_model_mod, &
[1972]	188	ONLY: land_surface, lsm_data_output_2d, zs
[1551]	189
[1783]	190	#if defined( __netcdf )
	191	USE NETCDF
	192	#endif
[1320]	193
[1783]	194	USE netcdf_interface, &
	195	ONLY: id_set_xy, id_set_xz, id_set_yz, id_var_do2d, id_var_time_xy, &
	196	id_var_time_xz, id_var_time_yz, nc_stat, netcdf_data_format, &
	197	netcdf_handle_error
	198
[1320]	199	USE particle_attributes, &
[1359]	200	ONLY: grid_particles, number_of_particles, particle_advection_start, &
	201	particles, prt_count
[1320]	202
[1]	203	USE pegrid
	204
[1551]	205	USE radiation_model_mod, &
[1976]	206	ONLY: radiation, radiation_data_output_2d
[1551]	207
[1]	208	IMPLICIT NONE
	209
[1682]	210	CHARACTER (LEN=2) :: do2d_mode !<
	211	CHARACTER (LEN=2) :: mode !<
	212	CHARACTER (LEN=4) :: grid !<
	213	CHARACTER (LEN=25) :: section_chr !<
	214	CHARACTER (LEN=50) :: rtext !<
[1320]	215
[1682]	216	INTEGER(iwp) :: av !<
	217	INTEGER(iwp) :: ngp !<
	218	INTEGER(iwp) :: file_id !<
	219	INTEGER(iwp) :: i !<
	220	INTEGER(iwp) :: if !<
	221	INTEGER(iwp) :: is !<
	222	INTEGER(iwp) :: iis !<
	223	INTEGER(iwp) :: j !<
	224	INTEGER(iwp) :: k !<
	225	INTEGER(iwp) :: l !<
	226	INTEGER(iwp) :: layer_xy !<
	227	INTEGER(iwp) :: n !<
[1703]	228	INTEGER(iwp) :: nis !<
[1682]	229	INTEGER(iwp) :: ns !<
	230	INTEGER(iwp) :: nzb_do !< lower limit of the data field (usually nzb)
	231	INTEGER(iwp) :: nzt_do !< upper limit of the data field (usually nzt+1)
	232	INTEGER(iwp) :: psi !<
[1960]	233	INTEGER(iwp) :: s_ind !<
[1682]	234	INTEGER(iwp) :: sender !<
	235	INTEGER(iwp) :: ind(4) !<
[1320]	236
[1682]	237	LOGICAL :: found !<
	238	LOGICAL :: resorted !<
	239	LOGICAL :: two_d !<
[1320]	240
[1682]	241	REAL(wp) :: mean_r !<
	242	REAL(wp) :: s_r2 !<
	243	REAL(wp) :: s_r3 !<
[1320]	244
[1682]	245	REAL(wp), DIMENSION(:), ALLOCATABLE :: level_z !<
	246	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d !<
	247	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d_l !<
	248	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !<
	249	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections !<
	250	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections_l !<
[1359]	251
[1]	252	#if defined( __parallel )
[1682]	253	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: total_2d !<
[1]	254	#endif
[1682]	255	REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !<
[1]	256
	257	NAMELIST /LOCAL/ rtext
	258
	259	!
	260	!-- Immediate return, if no output is requested (no respective sections
	261	!-- found in parameter data_output)
	262	IF ( mode == 'xy' .AND. .NOT. data_output_xy(av) ) RETURN
	263	IF ( mode == 'xz' .AND. .NOT. data_output_xz(av) ) RETURN
	264	IF ( mode == 'yz' .AND. .NOT. data_output_yz(av) ) RETURN
	265
[1308]	266	CALL cpu_log (log_point(3),'data_output_2d','start')
	267
[1]	268	two_d = .FALSE. ! local variable to distinguish between output of pure 2D
	269	! arrays and cross-sections of 3D arrays.
	270
	271	!
	272	!-- Depending on the orientation of the cross-section, the respective output
	273	!-- files have to be opened.
	274	SELECT CASE ( mode )
	275
	276	CASE ( 'xy' )
[1960]	277	s_ind = 1
[667]	278	ALLOCATE( level_z(nzb:nzt+1), local_2d(nxlg:nxrg,nysg:nyng) )
[1]	279
[1308]	280	IF ( netcdf_data_format > 4 ) THEN
	281	ns = 1
[1960]	282	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	283	ns = ns + 1
	284	ENDDO
	285	ns = ns - 1
	286	ALLOCATE( local_2d_sections(nxlg:nxrg,nysg:nyng,1:ns) )
[1353]	287	local_2d_sections = 0.0_wp
[1308]	288	ENDIF
	289
[493]	290	!
[1031]	291	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	292	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	293	CALL check_open( 101+av*10 )
	294	ENDIF
[1]	295	IF ( data_output_2d_on_each_pe ) THEN
	296	CALL check_open( 21 )
	297	ELSE
	298	IF ( myid == 0 ) THEN
	299	#if defined( __parallel )
[667]	300	ALLOCATE( total_2d(-nbgp:nx+nbgp,-nbgp:ny+nbgp) )
[1]	301	#endif
	302	ENDIF
	303	ENDIF
	304
	305	CASE ( 'xz' )
[1960]	306	s_ind = 2
[667]	307	ALLOCATE( local_2d(nxlg:nxrg,nzb:nzt+1) )
[1]	308
[1308]	309	IF ( netcdf_data_format > 4 ) THEN
	310	ns = 1
[1960]	311	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	312	ns = ns + 1
	313	ENDDO
	314	ns = ns - 1
	315	ALLOCATE( local_2d_sections(nxlg:nxrg,1:ns,nzb:nzt+1) )
	316	ALLOCATE( local_2d_sections_l(nxlg:nxrg,1:ns,nzb:nzt+1) )
[1353]	317	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
[1308]	318	ENDIF
	319
[493]	320	!
[1031]	321	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	322	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	323	CALL check_open( 102+av*10 )
	324	ENDIF
[1]	325
	326	IF ( data_output_2d_on_each_pe ) THEN
	327	CALL check_open( 22 )
	328	ELSE
	329	IF ( myid == 0 ) THEN
	330	#if defined( __parallel )
[667]	331	ALLOCATE( total_2d(-nbgp:nx+nbgp,nzb:nzt+1) )
[1]	332	#endif
	333	ENDIF
	334	ENDIF
	335
	336	CASE ( 'yz' )
[1960]	337	s_ind = 3
[667]	338	ALLOCATE( local_2d(nysg:nyng,nzb:nzt+1) )
[1]	339
[1308]	340	IF ( netcdf_data_format > 4 ) THEN
	341	ns = 1
[1960]	342	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	343	ns = ns + 1
	344	ENDDO
	345	ns = ns - 1
	346	ALLOCATE( local_2d_sections(1:ns,nysg:nyng,nzb:nzt+1) )
	347	ALLOCATE( local_2d_sections_l(1:ns,nysg:nyng,nzb:nzt+1) )
[1353]	348	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
[1308]	349	ENDIF
	350
[493]	351	!
[1031]	352	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	353	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	354	CALL check_open( 103+av*10 )
	355	ENDIF
[1]	356
	357	IF ( data_output_2d_on_each_pe ) THEN
	358	CALL check_open( 23 )
	359	ELSE
	360	IF ( myid == 0 ) THEN
	361	#if defined( __parallel )
[667]	362	ALLOCATE( total_2d(-nbgp:ny+nbgp,nzb:nzt+1) )
[1]	363	#endif
	364	ENDIF
	365	ENDIF
	366
	367	CASE DEFAULT
[254]	368	message_string = 'unknown cross-section: ' // TRIM( mode )
	369	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
[1]	370
	371	END SELECT
	372
	373	!
[1745]	374	!-- For parallel netcdf output the time axis must be limited. Return, if this
	375	!-- limit is exceeded. This could be the case, if the simulated time exceeds
	376	!-- the given end time by the length of the given output interval.
	377	IF ( netcdf_data_format > 4 ) THEN
	378	IF ( mode == 'xy' .AND. do2d_xy_time_count(av) + 1 > &
	379	ntdim_2d_xy(av) ) THEN
	380	WRITE ( message_string, * ) 'Output of xy cross-sections is not ', &
	381	'given at t=', simulated_time, '&because the', &
	382	' maximum number of output time levels is exceeded.'
	383	CALL message( 'data_output_2d', 'PA0384', 0, 1, 0, 6, 0 )
	384	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	385	RETURN
	386	ENDIF
	387	IF ( mode == 'xz' .AND. do2d_xz_time_count(av) + 1 > &
	388	ntdim_2d_xz(av) ) THEN
	389	WRITE ( message_string, * ) 'Output of xz cross-sections is not ', &
	390	'given at t=', simulated_time, '&because the', &
	391	' maximum number of output time levels is exceeded.'
	392	CALL message( 'data_output_2d', 'PA0385', 0, 1, 0, 6, 0 )
	393	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	394	RETURN
	395	ENDIF
	396	IF ( mode == 'yz' .AND. do2d_yz_time_count(av) + 1 > &
	397	ntdim_2d_yz(av) ) THEN
	398	WRITE ( message_string, * ) 'Output of yz cross-sections is not ', &
	399	'given at t=', simulated_time, '&because the', &
	400	' maximum number of output time levels is exceeded.'
	401	CALL message( 'data_output_2d', 'PA0386', 0, 1, 0, 6, 0 )
	402	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	403	RETURN
	404	ENDIF
	405	ENDIF
	406
	407	!
[1]	408	!-- Allocate a temporary array for resorting (kji -> ijk).
[667]	409	ALLOCATE( local_pf(nxlg:nxrg,nysg:nyng,nzb:nzt+1) )
[1]	410
	411	!
	412	!-- Loop of all variables to be written.
	413	!-- Output dimensions chosen
	414	if = 1
	415	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
	416	do2d_mode = do2d(av,if)(l-1:l)
	417
	418	DO WHILE ( do2d(av,if)(1:1) /= ' ' )
	419
	420	IF ( do2d_mode == mode ) THEN
[1980]	421	!
	422	!-- Set flag to steer output of radiation, land-surface, or user-defined
	423	!-- quantities
	424	found = .FALSE.
[1551]	425
	426	nzb_do = nzb
	427	nzt_do = nzt+1
[1]	428	!
	429	!-- Store the array chosen on the temporary array.
	430	resorted = .FALSE.
	431	SELECT CASE ( TRIM( do2d(av,if) ) )
	432
	433	CASE ( 'e_xy', 'e_xz', 'e_yz' )
	434	IF ( av == 0 ) THEN
	435	to_be_resorted => e
	436	ELSE
	437	to_be_resorted => e_av
	438	ENDIF
	439	IF ( mode == 'xy' ) level_z = zu
	440
[771]	441	CASE ( 'lpt_xy', 'lpt_xz', 'lpt_yz' )
	442	IF ( av == 0 ) THEN
	443	to_be_resorted => pt
	444	ELSE
	445	to_be_resorted => lpt_av
	446	ENDIF
	447	IF ( mode == 'xy' ) level_z = zu
	448
[1]	449	CASE ( 'lwp*_xy' ) ! 2d-array
	450	IF ( av == 0 ) THEN
[667]	451	DO i = nxlg, nxrg
	452	DO j = nysg, nyng
[1320]	453	local_pf(i,j,nzb+1) = SUM( ql(nzb:nzt,j,i) * &
[1]	454	dzw(1:nzt+1) )
	455	ENDDO
	456	ENDDO
	457	ELSE
[667]	458	DO i = nxlg, nxrg
	459	DO j = nysg, nyng
[1]	460	local_pf(i,j,nzb+1) = lwp_av(j,i)
	461	ENDDO
	462	ENDDO
	463	ENDIF
	464	resorted = .TRUE.
	465	two_d = .TRUE.
	466	level_z(nzb+1) = zu(nzb+1)
	467
[1053]	468	CASE ( 'nr_xy', 'nr_xz', 'nr_yz' )
	469	IF ( av == 0 ) THEN
	470	to_be_resorted => nr
	471	ELSE
	472	to_be_resorted => nr_av
	473	ENDIF
	474	IF ( mode == 'xy' ) level_z = zu
	475
[1691]	476	CASE ( 'ol*_xy' ) ! 2d-array
	477	IF ( av == 0 ) THEN
	478	DO i = nxlg, nxrg
	479	DO j = nysg, nyng
	480	local_pf(i,j,nzb+1) = ol(j,i)
	481	ENDDO
	482	ENDDO
	483	ELSE
	484	DO i = nxlg, nxrg
	485	DO j = nysg, nyng
	486	local_pf(i,j,nzb+1) = ol_av(j,i)
	487	ENDDO
	488	ENDDO
	489	ENDIF
	490	resorted = .TRUE.
	491	two_d = .TRUE.
	492	level_z(nzb+1) = zu(nzb+1)
	493
[1]	494	CASE ( 'p_xy', 'p_xz', 'p_yz' )
	495	IF ( av == 0 ) THEN
[729]	496	IF ( psolver /= 'sor' ) CALL exchange_horiz( p, nbgp )
[1]	497	to_be_resorted => p
	498	ELSE
[729]	499	IF ( psolver /= 'sor' ) CALL exchange_horiz( p_av, nbgp )
[1]	500	to_be_resorted => p_av
	501	ENDIF
	502	IF ( mode == 'xy' ) level_z = zu
	503
	504	CASE ( 'pc_xy', 'pc_xz', 'pc_yz' ) ! particle concentration
	505	IF ( av == 0 ) THEN
[215]	506	IF ( simulated_time >= particle_advection_start ) THEN
	507	tend = prt_count
[667]	508	CALL exchange_horiz( tend, nbgp )
[215]	509	ELSE
[1353]	510	tend = 0.0_wp
[215]	511	ENDIF
[667]	512	DO i = nxlg, nxrg
	513	DO j = nysg, nyng
[1]	514	DO k = nzb, nzt+1
	515	local_pf(i,j,k) = tend(k,j,i)
	516	ENDDO
	517	ENDDO
	518	ENDDO
	519	resorted = .TRUE.
	520	ELSE
[667]	521	CALL exchange_horiz( pc_av, nbgp )
[1]	522	to_be_resorted => pc_av
	523	ENDIF
	524
[1359]	525	CASE ( 'pr_xy', 'pr_xz', 'pr_yz' ) ! mean particle radius (effective radius)
[1]	526	IF ( av == 0 ) THEN
[215]	527	IF ( simulated_time >= particle_advection_start ) THEN
	528	DO i = nxl, nxr
	529	DO j = nys, nyn
	530	DO k = nzb, nzt+1
[1359]	531	number_of_particles = prt_count(k,j,i)
	532	IF (number_of_particles <= 0) CYCLE
	533	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
	534	s_r2 = 0.0_wp
[1353]	535	s_r3 = 0.0_wp
[1359]	536	DO n = 1, number_of_particles
	537	IF ( particles(n)%particle_mask ) THEN
	538	s_r2 = s_r2 + particles(n)%radius*2 &
	539	particles(n)%weight_factor
	540	s_r3 = s_r3 + particles(n)%radius*3 &
	541	particles(n)%weight_factor
	542	ENDIF
[215]	543	ENDDO
[1359]	544	IF ( s_r2 > 0.0_wp ) THEN
	545	mean_r = s_r3 / s_r2
[215]	546	ELSE
[1353]	547	mean_r = 0.0_wp
[215]	548	ENDIF
	549	tend(k,j,i) = mean_r
[1]	550	ENDDO
	551	ENDDO
	552	ENDDO
[667]	553	CALL exchange_horiz( tend, nbgp )
[215]	554	ELSE
[1353]	555	tend = 0.0_wp
[1359]	556	ENDIF
[667]	557	DO i = nxlg, nxrg
	558	DO j = nysg, nyng
[1]	559	DO k = nzb, nzt+1
	560	local_pf(i,j,k) = tend(k,j,i)
	561	ENDDO
	562	ENDDO
	563	ENDDO
	564	resorted = .TRUE.
	565	ELSE
[667]	566	CALL exchange_horiz( pr_av, nbgp )
[1]	567	to_be_resorted => pr_av
	568	ENDIF
	569
[72]	570	CASE ( 'pra*_xy' ) ! 2d-array / integral quantity => no av
	571	CALL exchange_horiz_2d( precipitation_amount )
[667]	572	DO i = nxlg, nxrg
	573	DO j = nysg, nyng
[72]	574	local_pf(i,j,nzb+1) = precipitation_amount(j,i)
	575	ENDDO
	576	ENDDO
[1353]	577	precipitation_amount = 0.0_wp ! reset for next integ. interval
[72]	578	resorted = .TRUE.
	579	two_d = .TRUE.
	580	level_z(nzb+1) = zu(nzb+1)
	581
	582	CASE ( 'prr*_xy' ) ! 2d-array
[1822]	583	IF ( av == 0 ) THEN
	584	CALL exchange_horiz_2d( prr(nzb+1,:,:) )
	585	DO i = nxlg, nxrg
	586	DO j = nysg, nyng
	587	local_pf(i,j,nzb+1) = prr(nzb+1,j,i) * hyrho(nzb+1)
[1053]	588	ENDDO
[1822]	589	ENDDO
[1053]	590	ELSE
[1822]	591	CALL exchange_horiz_2d( prr_av(nzb+1,:,:) )
	592	DO i = nxlg, nxrg
	593	DO j = nysg, nyng
	594	local_pf(i,j,nzb+1) = prr_av(nzb+1,j,i) * hyrho(nzb+1)
[1053]	595	ENDDO
[1822]	596	ENDDO
[1053]	597	ENDIF
	598	resorted = .TRUE.
	599	two_d = .TRUE.
	600	level_z(nzb+1) = zu(nzb+1)
	601
	602	CASE ( 'prr_xy', 'prr_xz', 'prr_yz' )
[72]	603	IF ( av == 0 ) THEN
[1053]	604	CALL exchange_horiz( prr, 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(k,j,i) * hyrho(nzb+1)
[1053]	609	ENDDO
[72]	610	ENDDO
	611	ENDDO
	612	ELSE
[1053]	613	CALL exchange_horiz( prr_av, nbgp )
[667]	614	DO i = nxlg, nxrg
	615	DO j = nysg, nyng
[1053]	616	DO k = nzb, nzt+1
[1822]	617	local_pf(i,j,k) = prr_av(k,j,i) * hyrho(nzb+1)
[1053]	618	ENDDO
[72]	619	ENDDO
	620	ENDDO
	621	ENDIF
	622	resorted = .TRUE.
[1053]	623	IF ( mode == 'xy' ) level_z = zu
[72]	624
[1]	625	CASE ( 'pt_xy', 'pt_xz', 'pt_yz' )
	626	IF ( av == 0 ) THEN
	627	IF ( .NOT. cloud_physics ) THEN
	628	to_be_resorted => pt
	629	ELSE
[667]	630	DO i = nxlg, nxrg
	631	DO j = nysg, nyng
[1]	632	DO k = nzb, nzt+1
[1320]	633	local_pf(i,j,k) = pt(k,j,i) + l_d_cp * &
	634	pt_d_t(k) * &
[1]	635	ql(k,j,i)
	636	ENDDO
	637	ENDDO
	638	ENDDO
	639	resorted = .TRUE.
	640	ENDIF
	641	ELSE
	642	to_be_resorted => pt_av
	643	ENDIF
	644	IF ( mode == 'xy' ) level_z = zu
	645
	646	CASE ( 'q_xy', 'q_xz', 'q_yz' )
	647	IF ( av == 0 ) THEN
	648	to_be_resorted => q
	649	ELSE
	650	to_be_resorted => q_av
	651	ENDIF
	652	IF ( mode == 'xy' ) level_z = zu
	653
[1053]	654	CASE ( 'qc_xy', 'qc_xz', 'qc_yz' )
[1]	655	IF ( av == 0 ) THEN
[1115]	656	to_be_resorted => qc
[1]	657	ELSE
[1115]	658	to_be_resorted => qc_av
[1]	659	ENDIF
	660	IF ( mode == 'xy' ) level_z = zu
	661
[1053]	662	CASE ( 'ql_xy', 'ql_xz', 'ql_yz' )
	663	IF ( av == 0 ) THEN
[1115]	664	to_be_resorted => ql
[1053]	665	ELSE
[1115]	666	to_be_resorted => ql_av
[1053]	667	ENDIF
	668	IF ( mode == 'xy' ) level_z = zu
	669
[1]	670	CASE ( 'ql_c_xy', 'ql_c_xz', 'ql_c_yz' )
	671	IF ( av == 0 ) THEN
	672	to_be_resorted => ql_c
	673	ELSE
	674	to_be_resorted => ql_c_av
	675	ENDIF
	676	IF ( mode == 'xy' ) level_z = zu
	677
	678	CASE ( 'ql_v_xy', 'ql_v_xz', 'ql_v_yz' )
	679	IF ( av == 0 ) THEN
	680	to_be_resorted => ql_v
	681	ELSE
	682	to_be_resorted => ql_v_av
	683	ENDIF
	684	IF ( mode == 'xy' ) level_z = zu
	685
	686	CASE ( 'ql_vp_xy', 'ql_vp_xz', 'ql_vp_yz' )
	687	IF ( av == 0 ) THEN
[1007]	688	IF ( simulated_time >= particle_advection_start ) THEN
	689	DO i = nxl, nxr
	690	DO j = nys, nyn
	691	DO k = nzb, nzt+1
[1359]	692	number_of_particles = prt_count(k,j,i)
	693	IF (number_of_particles <= 0) CYCLE
	694	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
	695	DO n = 1, number_of_particles
	696	IF ( particles(n)%particle_mask ) THEN
	697	tend(k,j,i) = tend(k,j,i) + &
	698	particles(n)%weight_factor / &
	699	prt_count(k,j,i)
	700	ENDIF
[1007]	701	ENDDO
	702	ENDDO
	703	ENDDO
	704	ENDDO
	705	CALL exchange_horiz( tend, nbgp )
	706	ELSE
[1353]	707	tend = 0.0_wp
[1359]	708	ENDIF
[1007]	709	DO i = nxlg, nxrg
	710	DO j = nysg, nyng
	711	DO k = nzb, nzt+1
	712	local_pf(i,j,k) = tend(k,j,i)
	713	ENDDO
	714	ENDDO
	715	ENDDO
	716	resorted = .TRUE.
	717	ELSE
	718	CALL exchange_horiz( ql_vp_av, nbgp )
[1]	719	to_be_resorted => ql_vp
	720	ENDIF
	721	IF ( mode == 'xy' ) level_z = zu
	722
[1053]	723	CASE ( 'qr_xy', 'qr_xz', 'qr_yz' )
	724	IF ( av == 0 ) THEN
	725	to_be_resorted => qr
	726	ELSE
	727	to_be_resorted => qr_av
	728	ENDIF
	729	IF ( mode == 'xy' ) level_z = zu
	730
[354]	731	CASE ( 'qsws*_xy' ) ! 2d-array
	732	IF ( av == 0 ) THEN
[667]	733	DO i = nxlg, nxrg
	734	DO j = nysg, nyng
[354]	735	local_pf(i,j,nzb+1) = qsws(j,i)
	736	ENDDO
	737	ENDDO
	738	ELSE
[667]	739	DO i = nxlg, nxrg
	740	DO j = nysg, nyng
[354]	741	local_pf(i,j,nzb+1) = qsws_av(j,i)
	742	ENDDO
	743	ENDDO
	744	ENDIF
	745	resorted = .TRUE.
	746	two_d = .TRUE.
	747	level_z(nzb+1) = zu(nzb+1)
	748
[1]	749	CASE ( 'qv_xy', 'qv_xz', 'qv_yz' )
	750	IF ( av == 0 ) THEN
[667]	751	DO i = nxlg, nxrg
	752	DO j = nysg, nyng
[1]	753	DO k = nzb, nzt+1
	754	local_pf(i,j,k) = q(k,j,i) - ql(k,j,i)
	755	ENDDO
	756	ENDDO
	757	ENDDO
	758	resorted = .TRUE.
	759	ELSE
	760	to_be_resorted => qv_av
	761	ENDIF
	762	IF ( mode == 'xy' ) level_z = zu
	763
[1551]	764
[1585]	765
[96]	766	CASE ( 'rho_xy', 'rho_xz', 'rho_yz' )
	767	IF ( av == 0 ) THEN
	768	to_be_resorted => rho
	769	ELSE
	770	to_be_resorted => rho_av
	771	ENDIF
	772
[1]	773	CASE ( 's_xy', 's_xz', 's_yz' )
	774	IF ( av == 0 ) THEN
[1960]	775	to_be_resorted => s
[1]	776	ELSE
[355]	777	to_be_resorted => s_av
[1]	778	ENDIF
	779
[96]	780	CASE ( 'sa_xy', 'sa_xz', 'sa_yz' )
	781	IF ( av == 0 ) THEN
	782	to_be_resorted => sa
	783	ELSE
	784	to_be_resorted => sa_av
	785	ENDIF
	786
[354]	787	CASE ( 'shf*_xy' ) ! 2d-array
	788	IF ( av == 0 ) THEN
[667]	789	DO i = nxlg, nxrg
	790	DO j = nysg, nyng
[354]	791	local_pf(i,j,nzb+1) = shf(j,i)
	792	ENDDO
	793	ENDDO
	794	ELSE
[667]	795	DO i = nxlg, nxrg
	796	DO j = nysg, nyng
[354]	797	local_pf(i,j,nzb+1) = shf_av(j,i)
	798	ENDDO
	799	ENDDO
	800	ENDIF
	801	resorted = .TRUE.
	802	two_d = .TRUE.
	803	level_z(nzb+1) = zu(nzb+1)
[1960]	804
	805	CASE ( 'ssws*_xy' ) ! 2d-array
	806	IF ( av == 0 ) THEN
	807	DO i = nxlg, nxrg
	808	DO j = nysg, nyng
	809	local_pf(i,j,nzb+1) = ssws(j,i)
	810	ENDDO
	811	ENDDO
	812	ELSE
	813	DO i = nxlg, nxrg
	814	DO j = nysg, nyng
	815	local_pf(i,j,nzb+1) = ssws_av(j,i)
	816	ENDDO
	817	ENDDO
	818	ENDIF
	819	resorted = .TRUE.
	820	two_d = .TRUE.
	821	level_z(nzb+1) = zu(nzb+1)
[1551]	822
[1]	823	CASE ( 't*_xy' ) ! 2d-array
	824	IF ( av == 0 ) THEN
[667]	825	DO i = nxlg, nxrg
	826	DO j = nysg, nyng
[1]	827	local_pf(i,j,nzb+1) = ts(j,i)
	828	ENDDO
	829	ENDDO
	830	ELSE
[667]	831	DO i = nxlg, nxrg
	832	DO j = nysg, nyng
[1]	833	local_pf(i,j,nzb+1) = ts_av(j,i)
	834	ENDDO
	835	ENDDO
	836	ENDIF
	837	resorted = .TRUE.
	838	two_d = .TRUE.
	839	level_z(nzb+1) = zu(nzb+1)
	840
	841	CASE ( 'u_xy', 'u_xz', 'u_yz' )
	842	IF ( av == 0 ) THEN
	843	to_be_resorted => u
	844	ELSE
	845	to_be_resorted => u_av
	846	ENDIF
	847	IF ( mode == 'xy' ) level_z = zu
	848	!
	849	!-- Substitute the values generated by "mirror" boundary condition
	850	!-- at the bottom boundary by the real surface values.
	851	IF ( do2d(av,if) == 'u_xz' .OR. do2d(av,if) == 'u_yz' ) THEN
[1353]	852	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
[1]	853	ENDIF
	854
	855	CASE ( 'u*_xy' ) ! 2d-array
	856	IF ( av == 0 ) THEN
[667]	857	DO i = nxlg, nxrg
	858	DO j = nysg, nyng
[1]	859	local_pf(i,j,nzb+1) = us(j,i)
	860	ENDDO
	861	ENDDO
	862	ELSE
[667]	863	DO i = nxlg, nxrg
	864	DO j = nysg, nyng
[1]	865	local_pf(i,j,nzb+1) = us_av(j,i)
	866	ENDDO
	867	ENDDO
	868	ENDIF
	869	resorted = .TRUE.
	870	two_d = .TRUE.
	871	level_z(nzb+1) = zu(nzb+1)
	872
	873	CASE ( 'v_xy', 'v_xz', 'v_yz' )
	874	IF ( av == 0 ) THEN
	875	to_be_resorted => v
	876	ELSE
	877	to_be_resorted => v_av
	878	ENDIF
	879	IF ( mode == 'xy' ) level_z = zu
	880	!
	881	!-- Substitute the values generated by "mirror" boundary condition
	882	!-- at the bottom boundary by the real surface values.
	883	IF ( do2d(av,if) == 'v_xz' .OR. do2d(av,if) == 'v_yz' ) THEN
[1353]	884	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
[1]	885	ENDIF
	886
	887	CASE ( 'vpt_xy', 'vpt_xz', 'vpt_yz' )
	888	IF ( av == 0 ) THEN
	889	to_be_resorted => vpt
	890	ELSE
	891	to_be_resorted => vpt_av
	892	ENDIF
	893	IF ( mode == 'xy' ) level_z = zu
	894
	895	CASE ( 'w_xy', 'w_xz', 'w_yz' )
	896	IF ( av == 0 ) THEN
	897	to_be_resorted => w
	898	ELSE
	899	to_be_resorted => w_av
	900	ENDIF
	901	IF ( mode == 'xy' ) level_z = zw
	902
[72]	903	CASE ( 'z0*_xy' ) ! 2d-array
	904	IF ( av == 0 ) THEN
[667]	905	DO i = nxlg, nxrg
	906	DO j = nysg, nyng
[72]	907	local_pf(i,j,nzb+1) = z0(j,i)
	908	ENDDO
	909	ENDDO
	910	ELSE
[667]	911	DO i = nxlg, nxrg
	912	DO j = nysg, nyng
[72]	913	local_pf(i,j,nzb+1) = z0_av(j,i)
	914	ENDDO
	915	ENDDO
	916	ENDIF
	917	resorted = .TRUE.
	918	two_d = .TRUE.
	919	level_z(nzb+1) = zu(nzb+1)
	920
[978]	921	CASE ( 'z0h*_xy' ) ! 2d-array
	922	IF ( av == 0 ) THEN
	923	DO i = nxlg, nxrg
	924	DO j = nysg, nyng
	925	local_pf(i,j,nzb+1) = z0h(j,i)
	926	ENDDO
	927	ENDDO
	928	ELSE
	929	DO i = nxlg, nxrg
	930	DO j = nysg, nyng
	931	local_pf(i,j,nzb+1) = z0h_av(j,i)
	932	ENDDO
	933	ENDDO
	934	ENDIF
	935	resorted = .TRUE.
	936	two_d = .TRUE.
	937	level_z(nzb+1) = zu(nzb+1)
	938
[1788]	939	CASE ( 'z0q*_xy' ) ! 2d-array
	940	IF ( av == 0 ) THEN
	941	DO i = nxlg, nxrg
	942	DO j = nysg, nyng
	943	local_pf(i,j,nzb+1) = z0q(j,i)
	944	ENDDO
	945	ENDDO
	946	ELSE
	947	DO i = nxlg, nxrg
	948	DO j = nysg, nyng
	949	local_pf(i,j,nzb+1) = z0q_av(j,i)
	950	ENDDO
	951	ENDDO
	952	ENDIF
	953	resorted = .TRUE.
	954	two_d = .TRUE.
	955	level_z(nzb+1) = zu(nzb+1)
	956
[1]	957	CASE DEFAULT
[1972]	958
[1]	959	!
[1972]	960	!-- Land surface model quantity
	961	IF ( land_surface ) THEN
	962	CALL lsm_data_output_2d( av, do2d(av,if), found, grid, mode,&
	963	local_pf, two_d, nzb_do, nzt_do )
	964	ENDIF
	965
	966	!
[1976]	967	!-- Radiation quantity
	968	IF ( .NOT. found .AND. radiation ) THEN
	969	CALL radiation_data_output_2d( av, do2d(av,if), found, grid,&
	970	mode, local_pf, two_d )
	971	ENDIF
	972
	973	!
[1]	974	!-- User defined quantity
[1972]	975	IF ( .NOT. found ) THEN
	976	CALL user_data_output_2d( av, do2d(av,if), found, grid, &
	977	local_pf, two_d, nzb_do, nzt_do )
	978	ENDIF
	979
[1]	980	resorted = .TRUE.
	981
	982	IF ( grid == 'zu' ) THEN
	983	IF ( mode == 'xy' ) level_z = zu
	984	ELSEIF ( grid == 'zw' ) THEN
	985	IF ( mode == 'xy' ) level_z = zw
[343]	986	ELSEIF ( grid == 'zu1' ) THEN
	987	IF ( mode == 'xy' ) level_z(nzb+1) = zu(nzb+1)
[1551]	988	ELSEIF ( grid == 'zs' ) THEN
	989	IF ( mode == 'xy' ) level_z = zs
[1]	990	ENDIF
	991
	992	IF ( .NOT. found ) THEN
[1320]	993	message_string = 'no output provided for: ' // &
[274]	994	TRIM( do2d(av,if) )
[254]	995	CALL message( 'data_output_2d', 'PA0181', 0, 0, 0, 6, 0 )
[1]	996	ENDIF
	997
	998	END SELECT
	999
	1000	!
	1001	!-- Resort the array to be output, if not done above
	1002	IF ( .NOT. resorted ) THEN
[667]	1003	DO i = nxlg, nxrg
	1004	DO j = nysg, nyng
[1551]	1005	DO k = nzb_do, nzt_do
[1]	1006	local_pf(i,j,k) = to_be_resorted(k,j,i)
	1007	ENDDO
	1008	ENDDO
	1009	ENDDO
	1010	ENDIF
	1011
	1012	!
	1013	!-- Output of the individual cross-sections, depending on the cross-
	1014	!-- section mode chosen.
	1015	is = 1
[1960]	1016	loop1: DO WHILE ( section(is,s_ind) /= -9999 .OR. two_d )
[1]	1017
	1018	SELECT CASE ( mode )
	1019
	1020	CASE ( 'xy' )
	1021	!
	1022	!-- Determine the cross section index
	1023	IF ( two_d ) THEN
	1024	layer_xy = nzb+1
	1025	ELSE
[1960]	1026	layer_xy = section(is,s_ind)
[1]	1027	ENDIF
	1028
	1029	!
[1551]	1030	!-- Exit the loop for layers beyond the data output domain
	1031	!-- (used for soil model)
[1691]	1032	IF ( layer_xy > nzt_do ) THEN
[1551]	1033	EXIT loop1
	1034	ENDIF
	1035
	1036	!
[1308]	1037	!-- Update the netCDF xy cross section time axis.
	1038	!-- In case of parallel output, this is only done by PE0
	1039	!-- to increase the performance.
	1040	IF ( simulated_time /= do2d_xy_last_time(av) ) THEN
	1041	do2d_xy_time_count(av) = do2d_xy_time_count(av) + 1
	1042	do2d_xy_last_time(av) = simulated_time
	1043	IF ( myid == 0 ) THEN
[1327]	1044	IF ( .NOT. data_output_2d_on_each_pe &
	1045	.OR. netcdf_data_format > 4 ) &
[493]	1046	THEN
[1]	1047	#if defined( __netcdf )
	1048	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1049	id_var_time_xy(av), &
[291]	1050	(/ time_since_reference_point /), &
[1]	1051	start = (/ do2d_xy_time_count(av) /), &
	1052	count = (/ 1 /) )
[1783]	1053	CALL netcdf_handle_error( 'data_output_2d', 53 )
[1]	1054	#endif
	1055	ENDIF
	1056	ENDIF
	1057	ENDIF
	1058	!
	1059	!-- If required, carry out averaging along z
[1960]	1060	IF ( section(is,s_ind) == -1 .AND. .NOT. two_d ) THEN
[1]	1061
[1353]	1062	local_2d = 0.0_wp
[1]	1063	!
	1064	!-- Carry out the averaging (all data are on the PE)
[1551]	1065	DO k = nzb_do, nzt_do
[667]	1066	DO j = nysg, nyng
	1067	DO i = nxlg, nxrg
[1]	1068	local_2d(i,j) = local_2d(i,j) + local_pf(i,j,k)
	1069	ENDDO
	1070	ENDDO
	1071	ENDDO
	1072
[1551]	1073	local_2d = local_2d / ( nzt_do - nzb_do + 1.0_wp)
[1]	1074
	1075	ELSE
	1076	!
	1077	!-- Just store the respective section on the local array
	1078	local_2d = local_pf(:,:,layer_xy)
	1079
	1080	ENDIF
	1081
	1082	#if defined( __parallel )
[1327]	1083	IF ( netcdf_data_format > 4 ) THEN
[1]	1084	!
[1031]	1085	!-- Parallel output in netCDF4/HDF5 format.
[493]	1086	IF ( two_d ) THEN
	1087	iis = 1
	1088	ELSE
	1089	iis = is
	1090	ENDIF
	1091
[1]	1092	#if defined( __netcdf )
[1308]	1093	!
	1094	!-- For parallel output, all cross sections are first stored
	1095	!-- here on a local array and will be written to the output
	1096	!-- file afterwards to increase the performance.
	1097	DO i = nxlg, nxrg
	1098	DO j = nysg, nyng
	1099	local_2d_sections(i,j,iis) = local_2d(i,j)
	1100	ENDDO
	1101	ENDDO
[1]	1102	#endif
[493]	1103	ELSE
[1]	1104
[493]	1105	IF ( data_output_2d_on_each_pe ) THEN
[1]	1106	!
[493]	1107	!-- Output of partial arrays on each PE
	1108	#if defined( __netcdf )
[1327]	1109	IF ( myid == 0 ) THEN
[1320]	1110	WRITE ( 21 ) time_since_reference_point, &
[493]	1111	do2d_xy_time_count(av), av
	1112	ENDIF
	1113	#endif
[759]	1114	DO i = 0, io_blocks-1
	1115	IF ( i == io_group ) THEN
[1551]	1116	WRITE ( 21 ) nxlg, nxrg, nysg, nyng, nysg, nyng
[759]	1117	WRITE ( 21 ) local_2d
	1118	ENDIF
	1119	#if defined( __parallel )
	1120	CALL MPI_BARRIER( comm2d, ierr )
	1121	#endif
	1122	ENDDO
[559]	1123
[493]	1124	ELSE
[1]	1125	!
[493]	1126	!-- PE0 receives partial arrays from all processors and
	1127	!-- then outputs them. Here a barrier has to be set,
	1128	!-- because otherwise "-MPI- FATAL: Remote protocol queue
	1129	!-- full" may occur.
	1130	CALL MPI_BARRIER( comm2d, ierr )
	1131
[667]	1132	ngp = ( nxrg-nxlg+1 ) * ( nyng-nysg+1 )
[493]	1133	IF ( myid == 0 ) THEN
[1]	1134	!
[493]	1135	!-- Local array can be relocated directly.
[667]	1136	total_2d(nxlg:nxrg,nysg:nyng) = local_2d
[1]	1137	!
[493]	1138	!-- Receive data from all other PEs.
	1139	DO n = 1, numprocs-1
[1]	1140	!
[493]	1141	!-- Receive index limits first, then array.
	1142	!-- Index limits are received in arbitrary order from
	1143	!-- the PEs.
[1320]	1144	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1145	MPI_ANY_SOURCE, 0, comm2d, &
[493]	1146	status, ierr )
	1147	sender = status(MPI_SOURCE)
	1148	DEALLOCATE( local_2d )
	1149	ALLOCATE( local_2d(ind(1):ind(2),ind(3):ind(4)) )
[1320]	1150	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1151	MPI_REAL, sender, 1, comm2d, &
[493]	1152	status, ierr )
	1153	total_2d(ind(1):ind(2),ind(3):ind(4)) = local_2d
	1154	ENDDO
[1]	1155	!
[493]	1156	!-- Relocate the local array for the next loop increment
	1157	DEALLOCATE( local_2d )
[667]	1158	ALLOCATE( local_2d(nxlg:nxrg,nysg:nyng) )
[1]	1159
	1160	#if defined( __netcdf )
[1327]	1161	IF ( two_d ) THEN
	1162	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1163	id_var_do2d(av,if), &
	1164	total_2d(0:nx+1,0:ny+1), &
	1165	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
	1166	count = (/ nx+2, ny+2, 1, 1 /) )
	1167	ELSE
	1168	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1169	id_var_do2d(av,if), &
	1170	total_2d(0:nx+1,0:ny+1), &
	1171	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
	1172	count = (/ nx+2, ny+2, 1, 1 /) )
[1]	1173	ENDIF
[1783]	1174	CALL netcdf_handle_error( 'data_output_2d', 54 )
[1]	1175	#endif
	1176
[493]	1177	ELSE
[1]	1178	!
[493]	1179	!-- First send the local index limits to PE0
[667]	1180	ind(1) = nxlg; ind(2) = nxrg
	1181	ind(3) = nysg; ind(4) = nyng
[1320]	1182	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1183	comm2d, ierr )
[1]	1184	!
[493]	1185	!-- Send data to PE0
[1320]	1186	CALL MPI_SEND( local_2d(nxlg,nysg), ngp, &
[493]	1187	MPI_REAL, 0, 1, comm2d, ierr )
	1188	ENDIF
	1189	!
	1190	!-- A barrier has to be set, because otherwise some PEs may
	1191	!-- proceed too fast so that PE0 may receive wrong data on
	1192	!-- tag 0
	1193	CALL MPI_BARRIER( comm2d, ierr )
[1]	1194	ENDIF
[493]	1195
[1]	1196	ENDIF
	1197	#else
	1198	#if defined( __netcdf )
[1327]	1199	IF ( two_d ) THEN
	1200	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1201	id_var_do2d(av,if), &
	1202	local_2d(nxl:nxr+1,nys:nyn+1), &
	1203	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
	1204	count = (/ nx+2, ny+2, 1, 1 /) )
	1205	ELSE
	1206	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1207	id_var_do2d(av,if), &
	1208	local_2d(nxl:nxr+1,nys:nyn+1), &
	1209	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
	1210	count = (/ nx+2, ny+2, 1, 1 /) )
[1]	1211	ENDIF
[1783]	1212	CALL netcdf_handle_error( 'data_output_2d', 447 )
[1]	1213	#endif
	1214	#endif
	1215	do2d_xy_n = do2d_xy_n + 1
	1216	!
	1217	!-- For 2D-arrays (e.g. u*) only one cross-section is available.
	1218	!-- Hence exit loop of output levels.
	1219	IF ( two_d ) THEN
[1703]	1220	IF ( netcdf_data_format < 5 ) two_d = .FALSE.
[1]	1221	EXIT loop1
	1222	ENDIF
	1223
	1224	CASE ( 'xz' )
	1225	!
[1308]	1226	!-- Update the netCDF xz cross section time axis.
	1227	!-- In case of parallel output, this is only done by PE0
	1228	!-- to increase the performance.
	1229	IF ( simulated_time /= do2d_xz_last_time(av) ) THEN
	1230	do2d_xz_time_count(av) = do2d_xz_time_count(av) + 1
	1231	do2d_xz_last_time(av) = simulated_time
	1232	IF ( myid == 0 ) THEN
[1327]	1233	IF ( .NOT. data_output_2d_on_each_pe &
	1234	.OR. netcdf_data_format > 4 ) &
[493]	1235	THEN
[1]	1236	#if defined( __netcdf )
	1237	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1238	id_var_time_xz(av), &
[291]	1239	(/ time_since_reference_point /), &
[1]	1240	start = (/ do2d_xz_time_count(av) /), &
	1241	count = (/ 1 /) )
[1783]	1242	CALL netcdf_handle_error( 'data_output_2d', 56 )
[1]	1243	#endif
	1244	ENDIF
	1245	ENDIF
	1246	ENDIF
[667]	1247
[1]	1248	!
	1249	!-- If required, carry out averaging along y
[1960]	1250	IF ( section(is,s_ind) == -1 ) THEN
[1]	1251
[1551]	1252	ALLOCATE( local_2d_l(nxlg:nxrg,nzb_do:nzt_do) )
[1353]	1253	local_2d_l = 0.0_wp
[1551]	1254	ngp = ( nxrg-nxlg + 1 ) * ( nzt_do-nzb_do + 1 )
[1]	1255	!
	1256	!-- First local averaging on the PE
[1551]	1257	DO k = nzb_do, nzt_do
[1]	1258	DO j = nys, nyn
[667]	1259	DO i = nxlg, nxrg
[1320]	1260	local_2d_l(i,k) = local_2d_l(i,k) + &
[1]	1261	local_pf(i,j,k)
	1262	ENDDO
	1263	ENDDO
	1264	ENDDO
	1265	#if defined( __parallel )
	1266	!
	1267	!-- Now do the averaging over all PEs along y
[622]	1268	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1269	CALL MPI_ALLREDUCE( local_2d_l(nxlg,nzb_do), &
	1270	local_2d(nxlg,nzb_do), ngp, MPI_REAL, &
[1]	1271	MPI_SUM, comm1dy, ierr )
	1272	#else
	1273	local_2d = local_2d_l
	1274	#endif
[1353]	1275	local_2d = local_2d / ( ny + 1.0_wp )
[1]	1276
	1277	DEALLOCATE( local_2d_l )
	1278
	1279	ELSE
	1280	!
	1281	!-- Just store the respective section on the local array
	1282	!-- (but only if it is available on this PE!)
[1960]	1283	IF ( section(is,s_ind) >= nys .AND. section(is,s_ind) <= nyn ) &
[1]	1284	THEN
[1960]	1285	local_2d = local_pf(:,section(is,s_ind),nzb_do:nzt_do)
[1]	1286	ENDIF
	1287
	1288	ENDIF
	1289
	1290	#if defined( __parallel )
[1327]	1291	IF ( netcdf_data_format > 4 ) THEN
[1]	1292	!
[1031]	1293	!-- Output in netCDF4/HDF5 format.
[493]	1294	!-- Output only on those PEs where the respective cross
	1295	!-- sections reside. Cross sections averaged along y are
	1296	!-- output on the respective first PE along y (myidy=0).
[1960]	1297	IF ( ( section(is,s_ind) >= nys .AND. &
	1298	section(is,s_ind) <= nyn ) .OR. &
	1299	( section(is,s_ind) == -1 .AND. myidy == 0 ) ) THEN
[1]	1300	#if defined( __netcdf )
[493]	1301	!
[1308]	1302	!-- For parallel output, all cross sections are first
	1303	!-- stored here on a local array and will be written to the
	1304	!-- output file afterwards to increase the performance.
	1305	DO i = nxlg, nxrg
[1551]	1306	DO k = nzb_do, nzt_do
[1308]	1307	local_2d_sections_l(i,is,k) = local_2d(i,k)
	1308	ENDDO
	1309	ENDDO
[1]	1310	#endif
	1311	ENDIF
	1312
	1313	ELSE
	1314
[493]	1315	IF ( data_output_2d_on_each_pe ) THEN
[1]	1316	!
[493]	1317	!-- Output of partial arrays on each PE. If the cross
	1318	!-- section does not reside on the PE, output special
	1319	!-- index values.
	1320	#if defined( __netcdf )
[1327]	1321	IF ( myid == 0 ) THEN
[1320]	1322	WRITE ( 22 ) time_since_reference_point, &
[493]	1323	do2d_xz_time_count(av), av
	1324	ENDIF
	1325	#endif
[759]	1326	DO i = 0, io_blocks-1
	1327	IF ( i == io_group ) THEN
[1960]	1328	IF ( ( section(is,s_ind) >= nys .AND. &
	1329	section(is,s_ind) <= nyn ) .OR. &
	1330	( section(is,s_ind) == -1 .AND. &
[1320]	1331	nys-1 == -1 ) ) &
[759]	1332	THEN
[1551]	1333	WRITE (22) nxlg, nxrg, nzb_do, nzt_do, nzb, nzt+1
[759]	1334	WRITE (22) local_2d
	1335	ELSE
[1551]	1336	WRITE (22) -1, -1, -1, -1, -1, -1
[759]	1337	ENDIF
	1338	ENDIF
	1339	#if defined( __parallel )
	1340	CALL MPI_BARRIER( comm2d, ierr )
	1341	#endif
	1342	ENDDO
[493]	1343
	1344	ELSE
[1]	1345	!
[493]	1346	!-- PE0 receives partial arrays from all processors of the
	1347	!-- respective cross section and outputs them. Here a
	1348	!-- barrier has to be set, because otherwise
	1349	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
	1350	CALL MPI_BARRIER( comm2d, ierr )
	1351
[1551]	1352	ngp = ( nxrg-nxlg + 1 ) * ( nzt_do-nzb_do + 1 )
[493]	1353	IF ( myid == 0 ) THEN
[1]	1354	!
[493]	1355	!-- Local array can be relocated directly.
[1960]	1356	IF ( ( section(is,s_ind) >= nys .AND. &
	1357	section(is,s_ind) <= nyn ) .OR. &
	1358	( section(is,s_ind) == -1 .AND. &
	1359	nys-1 == -1 ) ) THEN
[1551]	1360	total_2d(nxlg:nxrg,nzb_do:nzt_do) = local_2d
[493]	1361	ENDIF
[1]	1362	!
[493]	1363	!-- Receive data from all other PEs.
	1364	DO n = 1, numprocs-1
	1365	!
	1366	!-- Receive index limits first, then array.
	1367	!-- Index limits are received in arbitrary order from
	1368	!-- the PEs.
[1320]	1369	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1370	MPI_ANY_SOURCE, 0, comm2d, &
[1]	1371	status, ierr )
[493]	1372	!
	1373	!-- Not all PEs have data for XZ-cross-section.
	1374	IF ( ind(1) /= -9999 ) THEN
	1375	sender = status(MPI_SOURCE)
	1376	DEALLOCATE( local_2d )
[1320]	1377	ALLOCATE( local_2d(ind(1):ind(2), &
[493]	1378	ind(3):ind(4)) )
	1379	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1380	MPI_REAL, sender, 1, comm2d, &
	1381	status, ierr )
[1320]	1382	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
[493]	1383	local_2d
	1384	ENDIF
	1385	ENDDO
	1386	!
	1387	!-- Relocate the local array for the next loop increment
	1388	DEALLOCATE( local_2d )
[1551]	1389	ALLOCATE( local_2d(nxlg:nxrg,nzb_do:nzt_do) )
[1]	1390
	1391	#if defined( __netcdf )
[1327]	1392	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1393	id_var_do2d(av,if), &
[1551]	1394	total_2d(0:nx+1,nzb_do:nzt_do),&
[1327]	1395	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
[1551]	1396	count = (/ nx+2, 1, nzt_do-nzb_do+1, 1 /) )
[1783]	1397	CALL netcdf_handle_error( 'data_output_2d', 58 )
[1]	1398	#endif
	1399
[493]	1400	ELSE
[1]	1401	!
[493]	1402	!-- If the cross section resides on the PE, send the
	1403	!-- local index limits, otherwise send -9999 to PE0.
[1960]	1404	IF ( ( section(is,s_ind) >= nys .AND. &
	1405	section(is,s_ind) <= nyn ) .OR. &
	1406	( section(is,s_ind) == -1 .AND. nys-1 == -1 ) ) &
[493]	1407	THEN
[667]	1408	ind(1) = nxlg; ind(2) = nxrg
[1551]	1409	ind(3) = nzb_do; ind(4) = nzt_do
[493]	1410	ELSE
	1411	ind(1) = -9999; ind(2) = -9999
	1412	ind(3) = -9999; ind(4) = -9999
	1413	ENDIF
[1320]	1414	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1415	comm2d, ierr )
	1416	!
	1417	!-- If applicable, send data to PE0.
	1418	IF ( ind(1) /= -9999 ) THEN
[1551]	1419	CALL MPI_SEND( local_2d(nxlg,nzb_do), ngp, &
[493]	1420	MPI_REAL, 0, 1, comm2d, ierr )
	1421	ENDIF
[1]	1422	ENDIF
	1423	!
[493]	1424	!-- A barrier has to be set, because otherwise some PEs may
	1425	!-- proceed too fast so that PE0 may receive wrong data on
	1426	!-- tag 0
	1427	CALL MPI_BARRIER( comm2d, ierr )
[1]	1428	ENDIF
[493]	1429
[1]	1430	ENDIF
	1431	#else
	1432	#if defined( __netcdf )
[1327]	1433	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1434	id_var_do2d(av,if), &
[1551]	1435	local_2d(nxl:nxr+1,nzb_do:nzt_do), &
[1327]	1436	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
[1551]	1437	count = (/ nx+2, 1, nzt_do-nzb_do+1, 1 /) )
[1783]	1438	CALL netcdf_handle_error( 'data_output_2d', 451 )
[1]	1439	#endif
	1440	#endif
	1441	do2d_xz_n = do2d_xz_n + 1
	1442
	1443	CASE ( 'yz' )
	1444	!
[1308]	1445	!-- Update the netCDF yz cross section time axis.
	1446	!-- In case of parallel output, this is only done by PE0
	1447	!-- to increase the performance.
	1448	IF ( simulated_time /= do2d_yz_last_time(av) ) THEN
	1449	do2d_yz_time_count(av) = do2d_yz_time_count(av) + 1
	1450	do2d_yz_last_time(av) = simulated_time
	1451	IF ( myid == 0 ) THEN
[1327]	1452	IF ( .NOT. data_output_2d_on_each_pe &
	1453	.OR. netcdf_data_format > 4 ) &
[493]	1454	THEN
[1]	1455	#if defined( __netcdf )
	1456	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1457	id_var_time_yz(av), &
[291]	1458	(/ time_since_reference_point /), &
[1]	1459	start = (/ do2d_yz_time_count(av) /), &
	1460	count = (/ 1 /) )
[1783]	1461	CALL netcdf_handle_error( 'data_output_2d', 59 )
[1]	1462	#endif
	1463	ENDIF
	1464	ENDIF
[1308]	1465	ENDIF
[493]	1466
[1]	1467	!
	1468	!-- If required, carry out averaging along x
[1960]	1469	IF ( section(is,s_ind) == -1 ) THEN
[1]	1470
[1551]	1471	ALLOCATE( local_2d_l(nysg:nyng,nzb_do:nzt_do) )
[1353]	1472	local_2d_l = 0.0_wp
[1551]	1473	ngp = ( nyng-nysg+1 ) * ( nzt_do-nzb_do+1 )
[1]	1474	!
	1475	!-- First local averaging on the PE
[1551]	1476	DO k = nzb_do, nzt_do
[667]	1477	DO j = nysg, nyng
[1]	1478	DO i = nxl, nxr
[1320]	1479	local_2d_l(j,k) = local_2d_l(j,k) + &
[1]	1480	local_pf(i,j,k)
	1481	ENDDO
	1482	ENDDO
	1483	ENDDO
	1484	#if defined( __parallel )
	1485	!
	1486	!-- Now do the averaging over all PEs along x
[622]	1487	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1488	CALL MPI_ALLREDUCE( local_2d_l(nysg,nzb_do), &
	1489	local_2d(nysg,nzb_do), ngp, MPI_REAL, &
[1]	1490	MPI_SUM, comm1dx, ierr )
	1491	#else
	1492	local_2d = local_2d_l
	1493	#endif
[1353]	1494	local_2d = local_2d / ( nx + 1.0_wp )
[1]	1495
	1496	DEALLOCATE( local_2d_l )
	1497
	1498	ELSE
	1499	!
	1500	!-- Just store the respective section on the local array
	1501	!-- (but only if it is available on this PE!)
[1960]	1502	IF ( section(is,s_ind) >= nxl .AND. section(is,s_ind) <= nxr ) &
[1]	1503	THEN
[1960]	1504	local_2d = local_pf(section(is,s_ind),:,nzb_do:nzt_do)
[1]	1505	ENDIF
	1506
	1507	ENDIF
	1508
	1509	#if defined( __parallel )
[1327]	1510	IF ( netcdf_data_format > 4 ) THEN
[1]	1511	!
[1031]	1512	!-- Output in netCDF4/HDF5 format.
[493]	1513	!-- Output only on those PEs where the respective cross
	1514	!-- sections reside. Cross sections averaged along x are
	1515	!-- output on the respective first PE along x (myidx=0).
[1960]	1516	IF ( ( section(is,s_ind) >= nxl .AND. &
	1517	section(is,s_ind) <= nxr ) .OR. &
	1518	( section(is,s_ind) == -1 .AND. myidx == 0 ) ) THEN
[1]	1519	#if defined( __netcdf )
[493]	1520	!
[1308]	1521	!-- For parallel output, all cross sections are first
	1522	!-- stored here on a local array and will be written to the
	1523	!-- output file afterwards to increase the performance.
	1524	DO j = nysg, nyng
[1551]	1525	DO k = nzb_do, nzt_do
[1308]	1526	local_2d_sections_l(is,j,k) = local_2d(j,k)
	1527	ENDDO
	1528	ENDDO
[1]	1529	#endif
	1530	ENDIF
	1531
	1532	ELSE
	1533
[493]	1534	IF ( data_output_2d_on_each_pe ) THEN
[1]	1535	!
[493]	1536	!-- Output of partial arrays on each PE. If the cross
	1537	!-- section does not reside on the PE, output special
	1538	!-- index values.
	1539	#if defined( __netcdf )
[1327]	1540	IF ( myid == 0 ) THEN
[1320]	1541	WRITE ( 23 ) time_since_reference_point, &
[493]	1542	do2d_yz_time_count(av), av
	1543	ENDIF
	1544	#endif
[759]	1545	DO i = 0, io_blocks-1
	1546	IF ( i == io_group ) THEN
[1960]	1547	IF ( ( section(is,s_ind) >= nxl .AND. &
	1548	section(is,s_ind) <= nxr ) .OR. &
	1549	( section(is,s_ind) == -1 .AND. &
[1320]	1550	nxl-1 == -1 ) ) &
[759]	1551	THEN
[1551]	1552	WRITE (23) nysg, nyng, nzb_do, nzt_do, nzb, nzt+1
[759]	1553	WRITE (23) local_2d
	1554	ELSE
[1551]	1555	WRITE (23) -1, -1, -1, -1, -1, -1
[759]	1556	ENDIF
	1557	ENDIF
	1558	#if defined( __parallel )
	1559	CALL MPI_BARRIER( comm2d, ierr )
	1560	#endif
	1561	ENDDO
[493]	1562
	1563	ELSE
[1]	1564	!
[493]	1565	!-- PE0 receives partial arrays from all processors of the
	1566	!-- respective cross section and outputs them. Here a
	1567	!-- barrier has to be set, because otherwise
	1568	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
	1569	CALL MPI_BARRIER( comm2d, ierr )
	1570
[1551]	1571	ngp = ( nyng-nysg+1 ) * ( nzt_do-nzb_do+1 )
[493]	1572	IF ( myid == 0 ) THEN
[1]	1573	!
[493]	1574	!-- Local array can be relocated directly.
[1960]	1575	IF ( ( section(is,s_ind) >= nxl .AND. &
	1576	section(is,s_ind) <= nxr ) .OR. &
	1577	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
[493]	1578	THEN
[1551]	1579	total_2d(nysg:nyng,nzb_do:nzt_do) = local_2d
[493]	1580	ENDIF
[1]	1581	!
[493]	1582	!-- Receive data from all other PEs.
	1583	DO n = 1, numprocs-1
	1584	!
	1585	!-- Receive index limits first, then array.
	1586	!-- Index limits are received in arbitrary order from
	1587	!-- the PEs.
[1320]	1588	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1589	MPI_ANY_SOURCE, 0, comm2d, &
[1]	1590	status, ierr )
[493]	1591	!
	1592	!-- Not all PEs have data for YZ-cross-section.
	1593	IF ( ind(1) /= -9999 ) THEN
	1594	sender = status(MPI_SOURCE)
	1595	DEALLOCATE( local_2d )
[1320]	1596	ALLOCATE( local_2d(ind(1):ind(2), &
[493]	1597	ind(3):ind(4)) )
	1598	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1599	MPI_REAL, sender, 1, comm2d, &
	1600	status, ierr )
[1320]	1601	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
[493]	1602	local_2d
	1603	ENDIF
	1604	ENDDO
	1605	!
	1606	!-- Relocate the local array for the next loop increment
	1607	DEALLOCATE( local_2d )
[1551]	1608	ALLOCATE( local_2d(nysg:nyng,nzb_do:nzt_do) )
[1]	1609
	1610	#if defined( __netcdf )
[1327]	1611	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1612	id_var_do2d(av,if), &
[1551]	1613	total_2d(0:ny+1,nzb_do:nzt_do),&
[1327]	1614	start = (/ is, 1, 1, do2d_yz_time_count(av) /), &
[1551]	1615	count = (/ 1, ny+2, nzt_do-nzb_do+1, 1 /) )
[1783]	1616	CALL netcdf_handle_error( 'data_output_2d', 61 )
[1]	1617	#endif
	1618
[493]	1619	ELSE
[1]	1620	!
[493]	1621	!-- If the cross section resides on the PE, send the
	1622	!-- local index limits, otherwise send -9999 to PE0.
[1960]	1623	IF ( ( section(is,s_ind) >= nxl .AND. &
	1624	section(is,s_ind) <= nxr ) .OR. &
	1625	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
[493]	1626	THEN
[667]	1627	ind(1) = nysg; ind(2) = nyng
[1551]	1628	ind(3) = nzb_do; ind(4) = nzt_do
[493]	1629	ELSE
	1630	ind(1) = -9999; ind(2) = -9999
	1631	ind(3) = -9999; ind(4) = -9999
	1632	ENDIF
[1320]	1633	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1634	comm2d, ierr )
	1635	!
	1636	!-- If applicable, send data to PE0.
	1637	IF ( ind(1) /= -9999 ) THEN
[1551]	1638	CALL MPI_SEND( local_2d(nysg,nzb_do), ngp, &
[493]	1639	MPI_REAL, 0, 1, comm2d, ierr )
	1640	ENDIF
[1]	1641	ENDIF
	1642	!
[493]	1643	!-- A barrier has to be set, because otherwise some PEs may
	1644	!-- proceed too fast so that PE0 may receive wrong data on
	1645	!-- tag 0
	1646	CALL MPI_BARRIER( comm2d, ierr )
[1]	1647	ENDIF
[493]	1648
[1]	1649	ENDIF
	1650	#else
	1651	#if defined( __netcdf )
[1327]	1652	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1653	id_var_do2d(av,if), &
[1551]	1654	local_2d(nys:nyn+1,nzb_do:nzt_do), &
[1327]	1655	start = (/ is, 1, 1, do2d_xz_time_count(av) /), &
[1551]	1656	count = (/ 1, ny+2, nzt_do-nzb_do+1, 1 /) )
[1783]	1657	CALL netcdf_handle_error( 'data_output_2d', 452 )
[1]	1658	#endif
	1659	#endif
	1660	do2d_yz_n = do2d_yz_n + 1
	1661
	1662	END SELECT
	1663
	1664	is = is + 1
	1665	ENDDO loop1
	1666
[1308]	1667	!
	1668	!-- For parallel output, all data were collected before on a local array
	1669	!-- and are written now to the netcdf file. This must be done to increase
	1670	!-- the performance of the parallel output.
	1671	#if defined( __netcdf )
[1327]	1672	IF ( netcdf_data_format > 4 ) THEN
[1308]	1673
	1674	SELECT CASE ( mode )
	1675
	1676	CASE ( 'xy' )
	1677	IF ( two_d ) THEN
[1703]	1678	nis = 1
	1679	two_d = .FALSE.
[1308]	1680	ELSE
[1703]	1681	nis = ns
[1308]	1682	ENDIF
	1683	!
	1684	!-- Do not output redundant ghost point data except for the
	1685	!-- boundaries of the total domain.
	1686	IF ( nxr == nx .AND. nyn /= ny ) THEN
	1687	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1688	id_var_do2d(av,if), &
	1689	local_2d_sections(nxl:nxr+1, &
[1703]	1690	nys:nyn,1:nis), &
[1308]	1691	start = (/ nxl+1, nys+1, 1, &
	1692	do2d_xy_time_count(av) /), &
	1693	count = (/ nxr-nxl+2, &
[1703]	1694	nyn-nys+1, nis, 1 &
[1308]	1695	/) )
	1696	ELSEIF ( nxr /= nx .AND. nyn == ny ) THEN
	1697	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1698	id_var_do2d(av,if), &
	1699	local_2d_sections(nxl:nxr, &
[1703]	1700	nys:nyn+1,1:nis), &
[1308]	1701	start = (/ nxl+1, nys+1, 1, &
	1702	do2d_xy_time_count(av) /), &
	1703	count = (/ nxr-nxl+1, &
[1703]	1704	nyn-nys+2, nis, 1 &
[1308]	1705	/) )
	1706	ELSEIF ( nxr == nx .AND. nyn == ny ) THEN
	1707	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1708	id_var_do2d(av,if), &
	1709	local_2d_sections(nxl:nxr+1, &
[1703]	1710	nys:nyn+1,1:nis), &
[1308]	1711	start = (/ nxl+1, nys+1, 1, &
	1712	do2d_xy_time_count(av) /), &
	1713	count = (/ nxr-nxl+2, &
[1703]	1714	nyn-nys+2, nis, 1 &
[1308]	1715	/) )
	1716	ELSE
	1717	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1718	id_var_do2d(av,if), &
	1719	local_2d_sections(nxl:nxr, &
[1703]	1720	nys:nyn,1:nis), &
[1308]	1721	start = (/ nxl+1, nys+1, 1, &
	1722	do2d_xy_time_count(av) /), &
	1723	count = (/ nxr-nxl+1, &
[1703]	1724	nyn-nys+1, nis, 1 &
[1308]	1725	/) )
	1726	ENDIF
	1727
[1783]	1728	CALL netcdf_handle_error( 'data_output_2d', 55 )
[1308]	1729
	1730	CASE ( 'xz' )
	1731	!
	1732	!-- First, all PEs get the information of all cross-sections.
	1733	!-- Then the data are written to the output file by all PEs
	1734	!-- while NF90_COLLECTIVE is set in subroutine
	1735	!-- define_netcdf_header. Although redundant information are
	1736	!-- written to the output file in that case, the performance
	1737	!-- is significantly better compared to the case where only
	1738	!-- the first row of PEs in x-direction (myidx = 0) is given
	1739	!-- the output while NF90_INDEPENDENT is set.
	1740	IF ( npey /= 1 ) THEN
	1741
	1742	#if defined( __parallel )
	1743	!
	1744	!-- Distribute data over all PEs along y
[1551]	1745	ngp = ( nxrg-nxlg+1 ) * ( nzt_do-nzb_do+1 ) * ns
[1308]	1746	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1747	CALL MPI_ALLREDUCE( local_2d_sections_l(nxlg,1,nzb_do), &
	1748	local_2d_sections(nxlg,1,nzb_do), &
[1308]	1749	ngp, MPI_REAL, MPI_SUM, comm1dy, &
	1750	ierr )
	1751	#else
	1752	local_2d_sections = local_2d_sections_l
	1753	#endif
	1754	ENDIF
	1755	!
	1756	!-- Do not output redundant ghost point data except for the
	1757	!-- boundaries of the total domain.
	1758	IF ( nxr == nx ) THEN
	1759	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1760	id_var_do2d(av,if), &
	1761	local_2d_sections(nxl:nxr+1,1: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+2, ns, nzt_do-nzb_do+1, &
[1308]	1766	1 /) )
	1767	ELSE
	1768	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1769	id_var_do2d(av,if), &
	1770	local_2d_sections(nxl:nxr,1:ns, &
[1551]	1771	nzb_do:nzt_do), &
[1308]	1772	start = (/ nxl+1, 1, 1, &
	1773	do2d_xz_time_count(av) /), &
[1551]	1774	count = (/ nxr-nxl+1, ns, nzt_do-nzb_do+1, &
[1308]	1775	1 /) )
	1776	ENDIF
	1777
[1783]	1778	CALL netcdf_handle_error( 'data_output_2d', 57 )
[1308]	1779
	1780	CASE ( 'yz' )
	1781	!
	1782	!-- First, all PEs get the information of all cross-sections.
	1783	!-- Then the data are written to the output file by all PEs
	1784	!-- while NF90_COLLECTIVE is set in subroutine
	1785	!-- define_netcdf_header. Although redundant information are
	1786	!-- written to the output file in that case, the performance
	1787	!-- is significantly better compared to the case where only
	1788	!-- the first row of PEs in y-direction (myidy = 0) is given
	1789	!-- the output while NF90_INDEPENDENT is set.
	1790	IF ( npex /= 1 ) THEN
	1791
	1792	#if defined( __parallel )
	1793	!
	1794	!-- Distribute data over all PEs along x
	1795	ngp = ( nyng-nysg+1 ) * ( nzt-nzb + 2 ) * ns
	1796	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1797	CALL MPI_ALLREDUCE( local_2d_sections_l(1,nysg,nzb_do), &
	1798	local_2d_sections(1,nysg,nzb_do), &
[1308]	1799	ngp, MPI_REAL, MPI_SUM, comm1dx, &
	1800	ierr )
	1801	#else
	1802	local_2d_sections = local_2d_sections_l
	1803	#endif
	1804	ENDIF
	1805	!
	1806	!-- Do not output redundant ghost point data except for the
	1807	!-- boundaries of the total domain.
	1808	IF ( nyn == ny ) THEN
	1809	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1810	id_var_do2d(av,if), &
	1811	local_2d_sections(1:ns, &
[1551]	1812	nys:nyn+1,nzb_do:nzt_do), &
[1308]	1813	start = (/ 1, nys+1, 1, &
	1814	do2d_yz_time_count(av) /), &
	1815	count = (/ ns, nyn-nys+2, &
[1551]	1816	nzt_do-nzb_do+1, 1 /) )
[1308]	1817	ELSE
	1818	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1819	id_var_do2d(av,if), &
	1820	local_2d_sections(1:ns,nys:nyn, &
[1551]	1821	nzb_do:nzt_do), &
[1308]	1822	start = (/ 1, nys+1, 1, &
	1823	do2d_yz_time_count(av) /), &
	1824	count = (/ ns, nyn-nys+1, &
[1551]	1825	nzt_do-nzb_do+1, 1 /) )
[1308]	1826	ENDIF
	1827
[1783]	1828	CALL netcdf_handle_error( 'data_output_2d', 60 )
[1308]	1829
	1830	CASE DEFAULT
	1831	message_string = 'unknown cross-section: ' // TRIM( mode )
	1832	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
	1833
	1834	END SELECT
	1835
	1836	ENDIF
[1311]	1837	#endif
[1]	1838	ENDIF
	1839
	1840	if = if + 1
	1841	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
	1842	do2d_mode = do2d(av,if)(l-1:l)
	1843
	1844	ENDDO
	1845
	1846	!
	1847	!-- Deallocate temporary arrays.
	1848	IF ( ALLOCATED( level_z ) ) DEALLOCATE( level_z )
[1308]	1849	IF ( netcdf_data_format > 4 ) THEN
	1850	DEALLOCATE( local_pf, local_2d, local_2d_sections )
	1851	IF( mode == 'xz' .OR. mode == 'yz' ) DEALLOCATE( local_2d_sections_l )
	1852	ENDIF
[1]	1853	#if defined( __parallel )
	1854	IF ( .NOT. data_output_2d_on_each_pe .AND. myid == 0 ) THEN
	1855	DEALLOCATE( total_2d )
	1856	ENDIF
	1857	#endif
	1858
	1859	!
	1860	!-- Close plot output file.
[1960]	1861	file_id = 20 + s_ind
[1]	1862
	1863	IF ( data_output_2d_on_each_pe ) THEN
[759]	1864	DO i = 0, io_blocks-1
	1865	IF ( i == io_group ) THEN
	1866	CALL close_file( file_id )
	1867	ENDIF
	1868	#if defined( __parallel )
	1869	CALL MPI_BARRIER( comm2d, ierr )
	1870	#endif
	1871	ENDDO
[1]	1872	ELSE
	1873	IF ( myid == 0 ) CALL close_file( file_id )
	1874	ENDIF
	1875
[1318]	1876	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
[1]	1877
	1878	END SUBROUTINE data_output_2d

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |