Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 2191

Last change on this file since 2191 was 2191, checked in by raasch, 7 years ago
last commit documented
Property svn:keywords set to `Id`
File size: 77.8 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	!
[2101]	17	! Copyright 1997-2017 Leibniz Universitaet Hannover
[2000]	18	!------------------------------------------------------------------------------!
[1036]	19	!
[254]	20	! Current revisions:
[1]	21	! -----------------
[1961]	22	!
[2191]	23	!
[1552]	24	! Former revisions:
	25	! -----------------
	26	! $Id: data_output_2d.f90 2191 2017-03-21 12:18:56Z raasch $
	27	!
[2191]	28	! 2190 2017-03-21 12:16:43Z raasch
	29	! bugfix for r2031: string rho replaced by rho_ocean
	30	!
[2032]	31	! 2031 2016-10-21 15:11:58Z knoop
	32	! renamed variable rho to rho_ocean and rho_av to rho_ocean_av
	33	!
[2001]	34	! 2000 2016-08-20 18:09:15Z knoop
	35	! Forced header and separation lines into 80 columns
	36	!
[1981]	37	! 1980 2016-07-29 15:51:57Z suehring
	38	! Bugfix, in order to steer user-defined output, setting flag found explicitly
	39	! to .F.
	40	!
[1977]	41	! 1976 2016-07-27 13:28:04Z maronga
	42	! Output of radiation quantities is now done directly in the respective module
	43	!
[1973]	44	! 1972 2016-07-26 07:52:02Z maronga
[1976]	45	! Output of land surface quantities is now done directly in the respective
	46	! module
[1973]	47	!
[1961]	48	! 1960 2016-07-12 16:34:24Z suehring
	49	! Scalar surface flux added
	50	! Rename INTEGER variable s into s_ind, as s is already assigned to scalar
	51	!
[1851]	52	! 1849 2016-04-08 11:33:18Z hoffmann
	53	! precipitation_amount, precipitation_rate, prr moved to arrays_3d
	54	!
[1823]	55	! 1822 2016-04-07 07:49:42Z hoffmann
	56	! Output of bulk cloud physics simplified.
	57	!
[1789]	58	! 1788 2016-03-10 11:01:04Z maronga
	59	! Added output of z0q
	60	!
[1784]	61	! 1783 2016-03-06 18:36:17Z raasch
	62	! name change of netcdf routines and module + related changes
	63	!
[1746]	64	! 1745 2016-02-05 13:06:51Z gronemeier
	65	! Bugfix: test if time axis limit exceeds moved to point after call of check_open
	66	!
[1704]	67	! 1703 2015-11-02 12:38:44Z raasch
	68	! bugfix for output of single (*) xy-sections in case of parallel netcdf I/O
	69	!
[1702]	70	! 1701 2015-11-02 07:43:04Z maronga
	71	! Bugfix in output of RRTGM data
	72	!
[1692]	73	! 1691 2015-10-26 16:17:44Z maronga
	74	! Added output of Obukhov length (ol) and radiative heating rates for RRTMG.
	75	! Formatting corrections.
	76	!
[1683]	77	! 1682 2015-10-07 23:56:08Z knoop
	78	! Code annotations made doxygen readable
	79	!
[1586]	80	! 1585 2015-04-30 07:05:52Z maronga
	81	! Added support for RRTMG
	82	!
[1556]	83	! 1555 2015-03-04 17:44:27Z maronga
	84	! Added output of r_a and r_s
	85	!
[1552]	86	! 1551 2015-03-03 14:18:16Z maronga
[1551]	87	! Added suppport for land surface model and radiation model output. In the course
	88	! of this action, the limits for vertical loops have been changed (from nzb and
	89	! nzt+1 to nzb_do and nzt_do, respectively in order to allow soil model output).
	90	! Moreover, a new vertical grid zs was introduced.
[1329]	91	!
[1360]	92	! 1359 2014-04-11 17:15:14Z hoffmann
	93	! New particle structure integrated.
	94	!
[1354]	95	! 1353 2014-04-08 15:21:23Z heinze
	96	! REAL constants provided with KIND-attribute
	97	!
[1329]	98	! 1327 2014-03-21 11:00:16Z raasch
	99	! parts concerning iso2d output removed,
	100	! -netcdf output queries
	101	!
[1321]	102	! 1320 2014-03-20 08:40:49Z raasch
[1320]	103	! ONLY-attribute added to USE-statements,
	104	! kind-parameters added to all INTEGER and REAL declaration statements,
	105	! kinds are defined in new module kinds,
	106	! revision history before 2012 removed,
	107	! comment fields (!:) to be used for variable explanations added to
	108	! all variable declaration statements
[1309]	109	!
[1319]	110	! 1318 2014-03-17 13:35:16Z raasch
	111	! barrier argument removed from cpu_log.
	112	! module interfaces removed
	113	!
[1312]	114	! 1311 2014-03-14 12:13:39Z heinze
	115	! bugfix: close #if defined( __netcdf )
	116	!
[1309]	117	! 1308 2014-03-13 14:58:42Z fricke
[1308]	118	! +local_2d_sections, local_2d_sections_l, ns
	119	! Check, if the limit of the time dimension is exceeded for parallel output
	120	! To increase the performance for parallel output, the following is done:
	121	! - Update of time axis is only done by PE0
	122	! - Cross sections are first stored on a local array and are written
	123	! collectively to the output file by all PEs.
[674]	124	!
[1116]	125	! 1115 2013-03-26 18:16:16Z hoffmann
	126	! ql is calculated by calc_liquid_water_content
	127	!
[1077]	128	! 1076 2012-12-05 08:30:18Z hoffmann
	129	! Bugfix in output of ql
	130	!
[1066]	131	! 1065 2012-11-22 17:42:36Z hoffmann
	132	! Bugfix: Output of cross sections of ql
	133	!
[1054]	134	! 1053 2012-11-13 17:11:03Z hoffmann
	135	! +qr, nr, qc and cross sections
	136	!
[1037]	137	! 1036 2012-10-22 13:43:42Z raasch
	138	! code put under GPL (PALM 3.9)
	139	!
[1035]	140	! 1031 2012-10-19 14:35:30Z raasch
	141	! netCDF4 without parallel file support implemented
	142	!
[1008]	143	! 1007 2012-09-19 14:30:36Z franke
	144	! Bugfix: missing calculation of ql_vp added
	145	!
[979]	146	! 978 2012-08-09 08:28:32Z fricke
	147	! +z0h
	148	!
[1]	149	! Revision 1.1 1997/08/11 06:24:09 raasch
	150	! Initial revision
	151	!
	152	!
	153	! Description:
	154	! ------------
[1682]	155	!> Data output of horizontal cross-sections in netCDF format or binary format
	156	!> compatible to old graphic software iso2d.
	157	!> Attention: The position of the sectional planes is still not always computed
	158	!> --------- correctly. (zu is used always)!
[1]	159	!------------------------------------------------------------------------------!
[1682]	160	SUBROUTINE data_output_2d( mode, av )
	161
[1]	162
[1320]	163	USE arrays_3d, &
[1849]	164	ONLY: dzw, e, nr, ol, p, pt, precipitation_amount, precipitation_rate,&
[2031]	165	prr,q, qc, ql, ql_c, ql_v, ql_vp, qr, qsws, rho_ocean, s, sa, shf, &
[1960]	166	ssws, tend, ts, u, us, v, vpt, w, z0, z0h, z0q, zu, zw
[1320]	167
[1]	168	USE averaging
[1320]	169
	170	USE cloud_parameters, &
[1849]	171	ONLY: hyrho, l_d_cp, pt_d_t
[1320]	172
	173	USE control_parameters, &
	174	ONLY: cloud_physics, data_output_2d_on_each_pe, data_output_xy, &
	175	data_output_xz, data_output_yz, do2d, &
	176	do2d_xy_last_time, do2d_xy_n, do2d_xy_time_count, &
	177	do2d_xz_last_time, do2d_xz_n, do2d_xz_time_count, &
	178	do2d_yz_last_time, do2d_yz_n, do2d_yz_time_count, &
[1822]	179	ibc_uv_b, io_blocks, io_group, message_string, &
	180	ntdim_2d_xy, ntdim_2d_xz, ntdim_2d_yz, &
	181	psolver, section, simulated_time, simulated_time_chr, &
	182	time_since_reference_point
[1320]	183
	184	USE cpulog, &
	185	ONLY: cpu_log, log_point
	186
	187	USE grid_variables, &
	188	ONLY: dx, dy
	189
	190	USE indices, &
	191	ONLY: nbgp, nx, nxl, nxlg, nxr, nxrg, ny, nyn, nyng, nys, nysg, &
	192	nz, nzb, nzt
	193
	194	USE kinds
[1551]	195
	196	USE land_surface_model_mod, &
[1972]	197	ONLY: land_surface, lsm_data_output_2d, zs
[1551]	198
[1783]	199	#if defined( __netcdf )
	200	USE NETCDF
	201	#endif
[1320]	202
[1783]	203	USE netcdf_interface, &
	204	ONLY: id_set_xy, id_set_xz, id_set_yz, id_var_do2d, id_var_time_xy, &
	205	id_var_time_xz, id_var_time_yz, nc_stat, netcdf_data_format, &
	206	netcdf_handle_error
	207
[1320]	208	USE particle_attributes, &
[1359]	209	ONLY: grid_particles, number_of_particles, particle_advection_start, &
	210	particles, prt_count
[1320]	211
[1]	212	USE pegrid
	213
[1551]	214	USE radiation_model_mod, &
[1976]	215	ONLY: radiation, radiation_data_output_2d
[1551]	216
[1]	217	IMPLICIT NONE
	218
[1682]	219	CHARACTER (LEN=2) :: do2d_mode !<
	220	CHARACTER (LEN=2) :: mode !<
	221	CHARACTER (LEN=4) :: grid !<
	222	CHARACTER (LEN=25) :: section_chr !<
	223	CHARACTER (LEN=50) :: rtext !<
[1320]	224
[1682]	225	INTEGER(iwp) :: av !<
	226	INTEGER(iwp) :: ngp !<
	227	INTEGER(iwp) :: file_id !<
	228	INTEGER(iwp) :: i !<
	229	INTEGER(iwp) :: if !<
	230	INTEGER(iwp) :: is !<
	231	INTEGER(iwp) :: iis !<
	232	INTEGER(iwp) :: j !<
	233	INTEGER(iwp) :: k !<
	234	INTEGER(iwp) :: l !<
	235	INTEGER(iwp) :: layer_xy !<
	236	INTEGER(iwp) :: n !<
[1703]	237	INTEGER(iwp) :: nis !<
[1682]	238	INTEGER(iwp) :: ns !<
	239	INTEGER(iwp) :: nzb_do !< lower limit of the data field (usually nzb)
	240	INTEGER(iwp) :: nzt_do !< upper limit of the data field (usually nzt+1)
	241	INTEGER(iwp) :: psi !<
[1960]	242	INTEGER(iwp) :: s_ind !<
[1682]	243	INTEGER(iwp) :: sender !<
	244	INTEGER(iwp) :: ind(4) !<
[1320]	245
[1682]	246	LOGICAL :: found !<
	247	LOGICAL :: resorted !<
	248	LOGICAL :: two_d !<
[1320]	249
[1682]	250	REAL(wp) :: mean_r !<
	251	REAL(wp) :: s_r2 !<
	252	REAL(wp) :: s_r3 !<
[1320]	253
[1682]	254	REAL(wp), DIMENSION(:), ALLOCATABLE :: level_z !<
	255	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d !<
	256	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d_l !<
	257	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !<
	258	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections !<
	259	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections_l !<
[1359]	260
[1]	261	#if defined( __parallel )
[1682]	262	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: total_2d !<
[1]	263	#endif
[1682]	264	REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !<
[1]	265
	266	NAMELIST /LOCAL/ rtext
	267
	268	!
	269	!-- Immediate return, if no output is requested (no respective sections
	270	!-- found in parameter data_output)
	271	IF ( mode == 'xy' .AND. .NOT. data_output_xy(av) ) RETURN
	272	IF ( mode == 'xz' .AND. .NOT. data_output_xz(av) ) RETURN
	273	IF ( mode == 'yz' .AND. .NOT. data_output_yz(av) ) RETURN
	274
[1308]	275	CALL cpu_log (log_point(3),'data_output_2d','start')
	276
[1]	277	two_d = .FALSE. ! local variable to distinguish between output of pure 2D
	278	! arrays and cross-sections of 3D arrays.
	279
	280	!
	281	!-- Depending on the orientation of the cross-section, the respective output
	282	!-- files have to be opened.
	283	SELECT CASE ( mode )
	284
	285	CASE ( 'xy' )
[1960]	286	s_ind = 1
[667]	287	ALLOCATE( level_z(nzb:nzt+1), local_2d(nxlg:nxrg,nysg:nyng) )
[1]	288
[1308]	289	IF ( netcdf_data_format > 4 ) THEN
	290	ns = 1
[1960]	291	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	292	ns = ns + 1
	293	ENDDO
	294	ns = ns - 1
	295	ALLOCATE( local_2d_sections(nxlg:nxrg,nysg:nyng,1:ns) )
[1353]	296	local_2d_sections = 0.0_wp
[1308]	297	ENDIF
	298
[493]	299	!
[1031]	300	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	301	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	302	CALL check_open( 101+av*10 )
	303	ENDIF
[1]	304	IF ( data_output_2d_on_each_pe ) THEN
	305	CALL check_open( 21 )
	306	ELSE
	307	IF ( myid == 0 ) THEN
	308	#if defined( __parallel )
[667]	309	ALLOCATE( total_2d(-nbgp:nx+nbgp,-nbgp:ny+nbgp) )
[1]	310	#endif
	311	ENDIF
	312	ENDIF
	313
	314	CASE ( 'xz' )
[1960]	315	s_ind = 2
[667]	316	ALLOCATE( local_2d(nxlg:nxrg,nzb:nzt+1) )
[1]	317
[1308]	318	IF ( netcdf_data_format > 4 ) THEN
	319	ns = 1
[1960]	320	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	321	ns = ns + 1
	322	ENDDO
	323	ns = ns - 1
	324	ALLOCATE( local_2d_sections(nxlg:nxrg,1:ns,nzb:nzt+1) )
	325	ALLOCATE( local_2d_sections_l(nxlg:nxrg,1:ns,nzb:nzt+1) )
[1353]	326	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
[1308]	327	ENDIF
	328
[493]	329	!
[1031]	330	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	331	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	332	CALL check_open( 102+av*10 )
	333	ENDIF
[1]	334
	335	IF ( data_output_2d_on_each_pe ) THEN
	336	CALL check_open( 22 )
	337	ELSE
	338	IF ( myid == 0 ) THEN
	339	#if defined( __parallel )
[667]	340	ALLOCATE( total_2d(-nbgp:nx+nbgp,nzb:nzt+1) )
[1]	341	#endif
	342	ENDIF
	343	ENDIF
	344
	345	CASE ( 'yz' )
[1960]	346	s_ind = 3
[667]	347	ALLOCATE( local_2d(nysg:nyng,nzb:nzt+1) )
[1]	348
[1308]	349	IF ( netcdf_data_format > 4 ) THEN
	350	ns = 1
[1960]	351	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	352	ns = ns + 1
	353	ENDDO
	354	ns = ns - 1
	355	ALLOCATE( local_2d_sections(1:ns,nysg:nyng,nzb:nzt+1) )
	356	ALLOCATE( local_2d_sections_l(1:ns,nysg:nyng,nzb:nzt+1) )
[1353]	357	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
[1308]	358	ENDIF
	359
[493]	360	!
[1031]	361	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	362	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	363	CALL check_open( 103+av*10 )
	364	ENDIF
[1]	365
	366	IF ( data_output_2d_on_each_pe ) THEN
	367	CALL check_open( 23 )
	368	ELSE
	369	IF ( myid == 0 ) THEN
	370	#if defined( __parallel )
[667]	371	ALLOCATE( total_2d(-nbgp:ny+nbgp,nzb:nzt+1) )
[1]	372	#endif
	373	ENDIF
	374	ENDIF
	375
	376	CASE DEFAULT
[254]	377	message_string = 'unknown cross-section: ' // TRIM( mode )
	378	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
[1]	379
	380	END SELECT
	381
	382	!
[1745]	383	!-- For parallel netcdf output the time axis must be limited. Return, if this
	384	!-- limit is exceeded. This could be the case, if the simulated time exceeds
	385	!-- the given end time by the length of the given output interval.
	386	IF ( netcdf_data_format > 4 ) THEN
	387	IF ( mode == 'xy' .AND. do2d_xy_time_count(av) + 1 > &
	388	ntdim_2d_xy(av) ) THEN
	389	WRITE ( message_string, * ) 'Output of xy 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', 'PA0384', 0, 1, 0, 6, 0 )
	393	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	394	RETURN
	395	ENDIF
	396	IF ( mode == 'xz' .AND. do2d_xz_time_count(av) + 1 > &
	397	ntdim_2d_xz(av) ) THEN
	398	WRITE ( message_string, * ) 'Output of xz 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', 'PA0385', 0, 1, 0, 6, 0 )
	402	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	403	RETURN
	404	ENDIF
	405	IF ( mode == 'yz' .AND. do2d_yz_time_count(av) + 1 > &
	406	ntdim_2d_yz(av) ) THEN
	407	WRITE ( message_string, * ) 'Output of yz cross-sections is not ', &
	408	'given at t=', simulated_time, '&because the', &
	409	' maximum number of output time levels is exceeded.'
	410	CALL message( 'data_output_2d', 'PA0386', 0, 1, 0, 6, 0 )
	411	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	412	RETURN
	413	ENDIF
	414	ENDIF
	415
	416	!
[1]	417	!-- Allocate a temporary array for resorting (kji -> ijk).
[667]	418	ALLOCATE( local_pf(nxlg:nxrg,nysg:nyng,nzb:nzt+1) )
[1]	419
	420	!
	421	!-- Loop of all variables to be written.
	422	!-- Output dimensions chosen
	423	if = 1
	424	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
	425	do2d_mode = do2d(av,if)(l-1:l)
	426
	427	DO WHILE ( do2d(av,if)(1:1) /= ' ' )
	428
	429	IF ( do2d_mode == mode ) THEN
[1980]	430	!
	431	!-- Set flag to steer output of radiation, land-surface, or user-defined
	432	!-- quantities
	433	found = .FALSE.
[1551]	434
	435	nzb_do = nzb
	436	nzt_do = nzt+1
[1]	437	!
	438	!-- Store the array chosen on the temporary array.
	439	resorted = .FALSE.
	440	SELECT CASE ( TRIM( do2d(av,if) ) )
	441
	442	CASE ( 'e_xy', 'e_xz', 'e_yz' )
	443	IF ( av == 0 ) THEN
	444	to_be_resorted => e
	445	ELSE
	446	to_be_resorted => e_av
	447	ENDIF
	448	IF ( mode == 'xy' ) level_z = zu
	449
[771]	450	CASE ( 'lpt_xy', 'lpt_xz', 'lpt_yz' )
	451	IF ( av == 0 ) THEN
	452	to_be_resorted => pt
	453	ELSE
	454	to_be_resorted => lpt_av
	455	ENDIF
	456	IF ( mode == 'xy' ) level_z = zu
	457
[1]	458	CASE ( 'lwp*_xy' ) ! 2d-array
	459	IF ( av == 0 ) THEN
[667]	460	DO i = nxlg, nxrg
	461	DO j = nysg, nyng
[1320]	462	local_pf(i,j,nzb+1) = SUM( ql(nzb:nzt,j,i) * &
[1]	463	dzw(1:nzt+1) )
	464	ENDDO
	465	ENDDO
	466	ELSE
[667]	467	DO i = nxlg, nxrg
	468	DO j = nysg, nyng
[1]	469	local_pf(i,j,nzb+1) = lwp_av(j,i)
	470	ENDDO
	471	ENDDO
	472	ENDIF
	473	resorted = .TRUE.
	474	two_d = .TRUE.
	475	level_z(nzb+1) = zu(nzb+1)
	476
[1053]	477	CASE ( 'nr_xy', 'nr_xz', 'nr_yz' )
	478	IF ( av == 0 ) THEN
	479	to_be_resorted => nr
	480	ELSE
	481	to_be_resorted => nr_av
	482	ENDIF
	483	IF ( mode == 'xy' ) level_z = zu
	484
[1691]	485	CASE ( 'ol*_xy' ) ! 2d-array
	486	IF ( av == 0 ) THEN
	487	DO i = nxlg, nxrg
	488	DO j = nysg, nyng
	489	local_pf(i,j,nzb+1) = ol(j,i)
	490	ENDDO
	491	ENDDO
	492	ELSE
	493	DO i = nxlg, nxrg
	494	DO j = nysg, nyng
	495	local_pf(i,j,nzb+1) = ol_av(j,i)
	496	ENDDO
	497	ENDDO
	498	ENDIF
	499	resorted = .TRUE.
	500	two_d = .TRUE.
	501	level_z(nzb+1) = zu(nzb+1)
	502
[1]	503	CASE ( 'p_xy', 'p_xz', 'p_yz' )
	504	IF ( av == 0 ) THEN
[729]	505	IF ( psolver /= 'sor' ) CALL exchange_horiz( p, nbgp )
[1]	506	to_be_resorted => p
	507	ELSE
[729]	508	IF ( psolver /= 'sor' ) CALL exchange_horiz( p_av, nbgp )
[1]	509	to_be_resorted => p_av
	510	ENDIF
	511	IF ( mode == 'xy' ) level_z = zu
	512
	513	CASE ( 'pc_xy', 'pc_xz', 'pc_yz' ) ! particle concentration
	514	IF ( av == 0 ) THEN
[215]	515	IF ( simulated_time >= particle_advection_start ) THEN
	516	tend = prt_count
[667]	517	CALL exchange_horiz( tend, nbgp )
[215]	518	ELSE
[1353]	519	tend = 0.0_wp
[215]	520	ENDIF
[667]	521	DO i = nxlg, nxrg
	522	DO j = nysg, nyng
[1]	523	DO k = nzb, nzt+1
	524	local_pf(i,j,k) = tend(k,j,i)
	525	ENDDO
	526	ENDDO
	527	ENDDO
	528	resorted = .TRUE.
	529	ELSE
[667]	530	CALL exchange_horiz( pc_av, nbgp )
[1]	531	to_be_resorted => pc_av
	532	ENDIF
	533
[1359]	534	CASE ( 'pr_xy', 'pr_xz', 'pr_yz' ) ! mean particle radius (effective radius)
[1]	535	IF ( av == 0 ) THEN
[215]	536	IF ( simulated_time >= particle_advection_start ) THEN
	537	DO i = nxl, nxr
	538	DO j = nys, nyn
	539	DO k = nzb, nzt+1
[1359]	540	number_of_particles = prt_count(k,j,i)
	541	IF (number_of_particles <= 0) CYCLE
	542	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
	543	s_r2 = 0.0_wp
[1353]	544	s_r3 = 0.0_wp
[1359]	545	DO n = 1, number_of_particles
	546	IF ( particles(n)%particle_mask ) THEN
	547	s_r2 = s_r2 + particles(n)%radius*2 &
	548	particles(n)%weight_factor
	549	s_r3 = s_r3 + particles(n)%radius*3 &
	550	particles(n)%weight_factor
	551	ENDIF
[215]	552	ENDDO
[1359]	553	IF ( s_r2 > 0.0_wp ) THEN
	554	mean_r = s_r3 / s_r2
[215]	555	ELSE
[1353]	556	mean_r = 0.0_wp
[215]	557	ENDIF
	558	tend(k,j,i) = mean_r
[1]	559	ENDDO
	560	ENDDO
	561	ENDDO
[667]	562	CALL exchange_horiz( tend, nbgp )
[215]	563	ELSE
[1353]	564	tend = 0.0_wp
[1359]	565	ENDIF
[667]	566	DO i = nxlg, nxrg
	567	DO j = nysg, nyng
[1]	568	DO k = nzb, nzt+1
	569	local_pf(i,j,k) = tend(k,j,i)
	570	ENDDO
	571	ENDDO
	572	ENDDO
	573	resorted = .TRUE.
	574	ELSE
[667]	575	CALL exchange_horiz( pr_av, nbgp )
[1]	576	to_be_resorted => pr_av
	577	ENDIF
	578
[72]	579	CASE ( 'pra*_xy' ) ! 2d-array / integral quantity => no av
	580	CALL exchange_horiz_2d( precipitation_amount )
[667]	581	DO i = nxlg, nxrg
	582	DO j = nysg, nyng
[72]	583	local_pf(i,j,nzb+1) = precipitation_amount(j,i)
	584	ENDDO
	585	ENDDO
[1353]	586	precipitation_amount = 0.0_wp ! reset for next integ. interval
[72]	587	resorted = .TRUE.
	588	two_d = .TRUE.
	589	level_z(nzb+1) = zu(nzb+1)
	590
	591	CASE ( 'prr*_xy' ) ! 2d-array
[1822]	592	IF ( av == 0 ) THEN
	593	CALL exchange_horiz_2d( prr(nzb+1,:,:) )
	594	DO i = nxlg, nxrg
	595	DO j = nysg, nyng
	596	local_pf(i,j,nzb+1) = prr(nzb+1,j,i) * hyrho(nzb+1)
[1053]	597	ENDDO
[1822]	598	ENDDO
[1053]	599	ELSE
[1822]	600	CALL exchange_horiz_2d( prr_av(nzb+1,:,:) )
	601	DO i = nxlg, nxrg
	602	DO j = nysg, nyng
	603	local_pf(i,j,nzb+1) = prr_av(nzb+1,j,i) * hyrho(nzb+1)
[1053]	604	ENDDO
[1822]	605	ENDDO
[1053]	606	ENDIF
	607	resorted = .TRUE.
	608	two_d = .TRUE.
	609	level_z(nzb+1) = zu(nzb+1)
	610
	611	CASE ( 'prr_xy', 'prr_xz', 'prr_yz' )
[72]	612	IF ( av == 0 ) THEN
[1053]	613	CALL exchange_horiz( prr, 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(k,j,i) * hyrho(nzb+1)
[1053]	618	ENDDO
[72]	619	ENDDO
	620	ENDDO
	621	ELSE
[1053]	622	CALL exchange_horiz( prr_av, nbgp )
[667]	623	DO i = nxlg, nxrg
	624	DO j = nysg, nyng
[1053]	625	DO k = nzb, nzt+1
[1822]	626	local_pf(i,j,k) = prr_av(k,j,i) * hyrho(nzb+1)
[1053]	627	ENDDO
[72]	628	ENDDO
	629	ENDDO
	630	ENDIF
	631	resorted = .TRUE.
[1053]	632	IF ( mode == 'xy' ) level_z = zu
[72]	633
[1]	634	CASE ( 'pt_xy', 'pt_xz', 'pt_yz' )
	635	IF ( av == 0 ) THEN
	636	IF ( .NOT. cloud_physics ) THEN
	637	to_be_resorted => pt
	638	ELSE
[667]	639	DO i = nxlg, nxrg
	640	DO j = nysg, nyng
[1]	641	DO k = nzb, nzt+1
[1320]	642	local_pf(i,j,k) = pt(k,j,i) + l_d_cp * &
	643	pt_d_t(k) * &
[1]	644	ql(k,j,i)
	645	ENDDO
	646	ENDDO
	647	ENDDO
	648	resorted = .TRUE.
	649	ENDIF
	650	ELSE
	651	to_be_resorted => pt_av
	652	ENDIF
	653	IF ( mode == 'xy' ) level_z = zu
	654
	655	CASE ( 'q_xy', 'q_xz', 'q_yz' )
	656	IF ( av == 0 ) THEN
	657	to_be_resorted => q
	658	ELSE
	659	to_be_resorted => q_av
	660	ENDIF
	661	IF ( mode == 'xy' ) level_z = zu
	662
[1053]	663	CASE ( 'qc_xy', 'qc_xz', 'qc_yz' )
[1]	664	IF ( av == 0 ) THEN
[1115]	665	to_be_resorted => qc
[1]	666	ELSE
[1115]	667	to_be_resorted => qc_av
[1]	668	ENDIF
	669	IF ( mode == 'xy' ) level_z = zu
	670
[1053]	671	CASE ( 'ql_xy', 'ql_xz', 'ql_yz' )
	672	IF ( av == 0 ) THEN
[1115]	673	to_be_resorted => ql
[1053]	674	ELSE
[1115]	675	to_be_resorted => ql_av
[1053]	676	ENDIF
	677	IF ( mode == 'xy' ) level_z = zu
	678
[1]	679	CASE ( 'ql_c_xy', 'ql_c_xz', 'ql_c_yz' )
	680	IF ( av == 0 ) THEN
	681	to_be_resorted => ql_c
	682	ELSE
	683	to_be_resorted => ql_c_av
	684	ENDIF
	685	IF ( mode == 'xy' ) level_z = zu
	686
	687	CASE ( 'ql_v_xy', 'ql_v_xz', 'ql_v_yz' )
	688	IF ( av == 0 ) THEN
	689	to_be_resorted => ql_v
	690	ELSE
	691	to_be_resorted => ql_v_av
	692	ENDIF
	693	IF ( mode == 'xy' ) level_z = zu
	694
	695	CASE ( 'ql_vp_xy', 'ql_vp_xz', 'ql_vp_yz' )
	696	IF ( av == 0 ) THEN
[1007]	697	IF ( simulated_time >= particle_advection_start ) THEN
	698	DO i = nxl, nxr
	699	DO j = nys, nyn
	700	DO k = nzb, nzt+1
[1359]	701	number_of_particles = prt_count(k,j,i)
	702	IF (number_of_particles <= 0) CYCLE
	703	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
	704	DO n = 1, number_of_particles
	705	IF ( particles(n)%particle_mask ) THEN
	706	tend(k,j,i) = tend(k,j,i) + &
	707	particles(n)%weight_factor / &
	708	prt_count(k,j,i)
	709	ENDIF
[1007]	710	ENDDO
	711	ENDDO
	712	ENDDO
	713	ENDDO
	714	CALL exchange_horiz( tend, nbgp )
	715	ELSE
[1353]	716	tend = 0.0_wp
[1359]	717	ENDIF
[1007]	718	DO i = nxlg, nxrg
	719	DO j = nysg, nyng
	720	DO k = nzb, nzt+1
	721	local_pf(i,j,k) = tend(k,j,i)
	722	ENDDO
	723	ENDDO
	724	ENDDO
	725	resorted = .TRUE.
	726	ELSE
	727	CALL exchange_horiz( ql_vp_av, nbgp )
[1]	728	to_be_resorted => ql_vp
	729	ENDIF
	730	IF ( mode == 'xy' ) level_z = zu
	731
[1053]	732	CASE ( 'qr_xy', 'qr_xz', 'qr_yz' )
	733	IF ( av == 0 ) THEN
	734	to_be_resorted => qr
	735	ELSE
	736	to_be_resorted => qr_av
	737	ENDIF
	738	IF ( mode == 'xy' ) level_z = zu
	739
[354]	740	CASE ( 'qsws*_xy' ) ! 2d-array
	741	IF ( av == 0 ) THEN
[667]	742	DO i = nxlg, nxrg
	743	DO j = nysg, nyng
[354]	744	local_pf(i,j,nzb+1) = qsws(j,i)
	745	ENDDO
	746	ENDDO
	747	ELSE
[667]	748	DO i = nxlg, nxrg
	749	DO j = nysg, nyng
[354]	750	local_pf(i,j,nzb+1) = qsws_av(j,i)
	751	ENDDO
	752	ENDDO
	753	ENDIF
	754	resorted = .TRUE.
	755	two_d = .TRUE.
	756	level_z(nzb+1) = zu(nzb+1)
	757
[1]	758	CASE ( 'qv_xy', 'qv_xz', 'qv_yz' )
	759	IF ( av == 0 ) THEN
[667]	760	DO i = nxlg, nxrg
	761	DO j = nysg, nyng
[1]	762	DO k = nzb, nzt+1
	763	local_pf(i,j,k) = q(k,j,i) - ql(k,j,i)
	764	ENDDO
	765	ENDDO
	766	ENDDO
	767	resorted = .TRUE.
	768	ELSE
	769	to_be_resorted => qv_av
	770	ENDIF
	771	IF ( mode == 'xy' ) level_z = zu
	772
[1551]	773
[1585]	774
[2190]	775	CASE ( 'rho_ocean_xy', 'rho_ocean_xz', 'rho_ocean_yz' )
[96]	776	IF ( av == 0 ) THEN
[2031]	777	to_be_resorted => rho_ocean
[96]	778	ELSE
[2031]	779	to_be_resorted => rho_ocean_av
[96]	780	ENDIF
	781
[1]	782	CASE ( 's_xy', 's_xz', 's_yz' )
	783	IF ( av == 0 ) THEN
[1960]	784	to_be_resorted => s
[1]	785	ELSE
[355]	786	to_be_resorted => s_av
[1]	787	ENDIF
	788
[96]	789	CASE ( 'sa_xy', 'sa_xz', 'sa_yz' )
	790	IF ( av == 0 ) THEN
	791	to_be_resorted => sa
	792	ELSE
	793	to_be_resorted => sa_av
	794	ENDIF
	795
[354]	796	CASE ( 'shf*_xy' ) ! 2d-array
	797	IF ( av == 0 ) THEN
[667]	798	DO i = nxlg, nxrg
	799	DO j = nysg, nyng
[354]	800	local_pf(i,j,nzb+1) = shf(j,i)
	801	ENDDO
	802	ENDDO
	803	ELSE
[667]	804	DO i = nxlg, nxrg
	805	DO j = nysg, nyng
[354]	806	local_pf(i,j,nzb+1) = shf_av(j,i)
	807	ENDDO
	808	ENDDO
	809	ENDIF
	810	resorted = .TRUE.
	811	two_d = .TRUE.
	812	level_z(nzb+1) = zu(nzb+1)
[1960]	813
	814	CASE ( 'ssws*_xy' ) ! 2d-array
	815	IF ( av == 0 ) THEN
	816	DO i = nxlg, nxrg
	817	DO j = nysg, nyng
	818	local_pf(i,j,nzb+1) = ssws(j,i)
	819	ENDDO
	820	ENDDO
	821	ELSE
	822	DO i = nxlg, nxrg
	823	DO j = nysg, nyng
	824	local_pf(i,j,nzb+1) = ssws_av(j,i)
	825	ENDDO
	826	ENDDO
	827	ENDIF
	828	resorted = .TRUE.
	829	two_d = .TRUE.
	830	level_z(nzb+1) = zu(nzb+1)
[1551]	831
[1]	832	CASE ( 't*_xy' ) ! 2d-array
	833	IF ( av == 0 ) THEN
[667]	834	DO i = nxlg, nxrg
	835	DO j = nysg, nyng
[1]	836	local_pf(i,j,nzb+1) = ts(j,i)
	837	ENDDO
	838	ENDDO
	839	ELSE
[667]	840	DO i = nxlg, nxrg
	841	DO j = nysg, nyng
[1]	842	local_pf(i,j,nzb+1) = ts_av(j,i)
	843	ENDDO
	844	ENDDO
	845	ENDIF
	846	resorted = .TRUE.
	847	two_d = .TRUE.
	848	level_z(nzb+1) = zu(nzb+1)
	849
	850	CASE ( 'u_xy', 'u_xz', 'u_yz' )
	851	IF ( av == 0 ) THEN
	852	to_be_resorted => u
	853	ELSE
	854	to_be_resorted => u_av
	855	ENDIF
	856	IF ( mode == 'xy' ) level_z = zu
	857	!
	858	!-- Substitute the values generated by "mirror" boundary condition
	859	!-- at the bottom boundary by the real surface values.
	860	IF ( do2d(av,if) == 'u_xz' .OR. do2d(av,if) == 'u_yz' ) THEN
[1353]	861	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
[1]	862	ENDIF
	863
	864	CASE ( 'u*_xy' ) ! 2d-array
	865	IF ( av == 0 ) THEN
[667]	866	DO i = nxlg, nxrg
	867	DO j = nysg, nyng
[1]	868	local_pf(i,j,nzb+1) = us(j,i)
	869	ENDDO
	870	ENDDO
	871	ELSE
[667]	872	DO i = nxlg, nxrg
	873	DO j = nysg, nyng
[1]	874	local_pf(i,j,nzb+1) = us_av(j,i)
	875	ENDDO
	876	ENDDO
	877	ENDIF
	878	resorted = .TRUE.
	879	two_d = .TRUE.
	880	level_z(nzb+1) = zu(nzb+1)
	881
	882	CASE ( 'v_xy', 'v_xz', 'v_yz' )
	883	IF ( av == 0 ) THEN
	884	to_be_resorted => v
	885	ELSE
	886	to_be_resorted => v_av
	887	ENDIF
	888	IF ( mode == 'xy' ) level_z = zu
	889	!
	890	!-- Substitute the values generated by "mirror" boundary condition
	891	!-- at the bottom boundary by the real surface values.
	892	IF ( do2d(av,if) == 'v_xz' .OR. do2d(av,if) == 'v_yz' ) THEN
[1353]	893	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
[1]	894	ENDIF
	895
	896	CASE ( 'vpt_xy', 'vpt_xz', 'vpt_yz' )
	897	IF ( av == 0 ) THEN
	898	to_be_resorted => vpt
	899	ELSE
	900	to_be_resorted => vpt_av
	901	ENDIF
	902	IF ( mode == 'xy' ) level_z = zu
	903
	904	CASE ( 'w_xy', 'w_xz', 'w_yz' )
	905	IF ( av == 0 ) THEN
	906	to_be_resorted => w
	907	ELSE
	908	to_be_resorted => w_av
	909	ENDIF
	910	IF ( mode == 'xy' ) level_z = zw
	911
[72]	912	CASE ( 'z0*_xy' ) ! 2d-array
	913	IF ( av == 0 ) THEN
[667]	914	DO i = nxlg, nxrg
	915	DO j = nysg, nyng
[72]	916	local_pf(i,j,nzb+1) = z0(j,i)
	917	ENDDO
	918	ENDDO
	919	ELSE
[667]	920	DO i = nxlg, nxrg
	921	DO j = nysg, nyng
[72]	922	local_pf(i,j,nzb+1) = z0_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
[978]	930	CASE ( 'z0h*_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) = z0h(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) = z0h_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
[1788]	948	CASE ( 'z0q*_xy' ) ! 2d-array
	949	IF ( av == 0 ) THEN
	950	DO i = nxlg, nxrg
	951	DO j = nysg, nyng
	952	local_pf(i,j,nzb+1) = z0q(j,i)
	953	ENDDO
	954	ENDDO
	955	ELSE
	956	DO i = nxlg, nxrg
	957	DO j = nysg, nyng
	958	local_pf(i,j,nzb+1) = z0q_av(j,i)
	959	ENDDO
	960	ENDDO
	961	ENDIF
	962	resorted = .TRUE.
	963	two_d = .TRUE.
	964	level_z(nzb+1) = zu(nzb+1)
	965
[1]	966	CASE DEFAULT
[1972]	967
[1]	968	!
[1972]	969	!-- Land surface model quantity
	970	IF ( land_surface ) THEN
	971	CALL lsm_data_output_2d( av, do2d(av,if), found, grid, mode,&
	972	local_pf, two_d, nzb_do, nzt_do )
	973	ENDIF
	974
	975	!
[1976]	976	!-- Radiation quantity
	977	IF ( .NOT. found .AND. radiation ) THEN
	978	CALL radiation_data_output_2d( av, do2d(av,if), found, grid,&
	979	mode, local_pf, two_d )
	980	ENDIF
	981
	982	!
[1]	983	!-- User defined quantity
[1972]	984	IF ( .NOT. found ) THEN
	985	CALL user_data_output_2d( av, do2d(av,if), found, grid, &
	986	local_pf, two_d, nzb_do, nzt_do )
	987	ENDIF
	988
[1]	989	resorted = .TRUE.
	990
	991	IF ( grid == 'zu' ) THEN
	992	IF ( mode == 'xy' ) level_z = zu
	993	ELSEIF ( grid == 'zw' ) THEN
	994	IF ( mode == 'xy' ) level_z = zw
[343]	995	ELSEIF ( grid == 'zu1' ) THEN
	996	IF ( mode == 'xy' ) level_z(nzb+1) = zu(nzb+1)
[1551]	997	ELSEIF ( grid == 'zs' ) THEN
	998	IF ( mode == 'xy' ) level_z = zs
[1]	999	ENDIF
	1000
	1001	IF ( .NOT. found ) THEN
[1320]	1002	message_string = 'no output provided for: ' // &
[274]	1003	TRIM( do2d(av,if) )
[254]	1004	CALL message( 'data_output_2d', 'PA0181', 0, 0, 0, 6, 0 )
[1]	1005	ENDIF
	1006
	1007	END SELECT
	1008
	1009	!
	1010	!-- Resort the array to be output, if not done above
	1011	IF ( .NOT. resorted ) THEN
[667]	1012	DO i = nxlg, nxrg
	1013	DO j = nysg, nyng
[1551]	1014	DO k = nzb_do, nzt_do
[1]	1015	local_pf(i,j,k) = to_be_resorted(k,j,i)
	1016	ENDDO
	1017	ENDDO
	1018	ENDDO
	1019	ENDIF
	1020
	1021	!
	1022	!-- Output of the individual cross-sections, depending on the cross-
	1023	!-- section mode chosen.
	1024	is = 1
[1960]	1025	loop1: DO WHILE ( section(is,s_ind) /= -9999 .OR. two_d )
[1]	1026
	1027	SELECT CASE ( mode )
	1028
	1029	CASE ( 'xy' )
	1030	!
	1031	!-- Determine the cross section index
	1032	IF ( two_d ) THEN
	1033	layer_xy = nzb+1
	1034	ELSE
[1960]	1035	layer_xy = section(is,s_ind)
[1]	1036	ENDIF
	1037
	1038	!
[1551]	1039	!-- Exit the loop for layers beyond the data output domain
	1040	!-- (used for soil model)
[1691]	1041	IF ( layer_xy > nzt_do ) THEN
[1551]	1042	EXIT loop1
	1043	ENDIF
	1044
	1045	!
[1308]	1046	!-- Update the netCDF xy cross section time axis.
	1047	!-- In case of parallel output, this is only done by PE0
	1048	!-- to increase the performance.
	1049	IF ( simulated_time /= do2d_xy_last_time(av) ) THEN
	1050	do2d_xy_time_count(av) = do2d_xy_time_count(av) + 1
	1051	do2d_xy_last_time(av) = simulated_time
	1052	IF ( myid == 0 ) THEN
[1327]	1053	IF ( .NOT. data_output_2d_on_each_pe &
	1054	.OR. netcdf_data_format > 4 ) &
[493]	1055	THEN
[1]	1056	#if defined( __netcdf )
	1057	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1058	id_var_time_xy(av), &
[291]	1059	(/ time_since_reference_point /), &
[1]	1060	start = (/ do2d_xy_time_count(av) /), &
	1061	count = (/ 1 /) )
[1783]	1062	CALL netcdf_handle_error( 'data_output_2d', 53 )
[1]	1063	#endif
	1064	ENDIF
	1065	ENDIF
	1066	ENDIF
	1067	!
	1068	!-- If required, carry out averaging along z
[1960]	1069	IF ( section(is,s_ind) == -1 .AND. .NOT. two_d ) THEN
[1]	1070
[1353]	1071	local_2d = 0.0_wp
[1]	1072	!
	1073	!-- Carry out the averaging (all data are on the PE)
[1551]	1074	DO k = nzb_do, nzt_do
[667]	1075	DO j = nysg, nyng
	1076	DO i = nxlg, nxrg
[1]	1077	local_2d(i,j) = local_2d(i,j) + local_pf(i,j,k)
	1078	ENDDO
	1079	ENDDO
	1080	ENDDO
	1081
[1551]	1082	local_2d = local_2d / ( nzt_do - nzb_do + 1.0_wp)
[1]	1083
	1084	ELSE
	1085	!
	1086	!-- Just store the respective section on the local array
	1087	local_2d = local_pf(:,:,layer_xy)
	1088
	1089	ENDIF
	1090
	1091	#if defined( __parallel )
[1327]	1092	IF ( netcdf_data_format > 4 ) THEN
[1]	1093	!
[1031]	1094	!-- Parallel output in netCDF4/HDF5 format.
[493]	1095	IF ( two_d ) THEN
	1096	iis = 1
	1097	ELSE
	1098	iis = is
	1099	ENDIF
	1100
[1]	1101	#if defined( __netcdf )
[1308]	1102	!
	1103	!-- For parallel output, all cross sections are first stored
	1104	!-- here on a local array and will be written to the output
	1105	!-- file afterwards to increase the performance.
	1106	DO i = nxlg, nxrg
	1107	DO j = nysg, nyng
	1108	local_2d_sections(i,j,iis) = local_2d(i,j)
	1109	ENDDO
	1110	ENDDO
[1]	1111	#endif
[493]	1112	ELSE
[1]	1113
[493]	1114	IF ( data_output_2d_on_each_pe ) THEN
[1]	1115	!
[493]	1116	!-- Output of partial arrays on each PE
	1117	#if defined( __netcdf )
[1327]	1118	IF ( myid == 0 ) THEN
[1320]	1119	WRITE ( 21 ) time_since_reference_point, &
[493]	1120	do2d_xy_time_count(av), av
	1121	ENDIF
	1122	#endif
[759]	1123	DO i = 0, io_blocks-1
	1124	IF ( i == io_group ) THEN
[1551]	1125	WRITE ( 21 ) nxlg, nxrg, nysg, nyng, nysg, nyng
[759]	1126	WRITE ( 21 ) local_2d
	1127	ENDIF
	1128	#if defined( __parallel )
	1129	CALL MPI_BARRIER( comm2d, ierr )
	1130	#endif
	1131	ENDDO
[559]	1132
[493]	1133	ELSE
[1]	1134	!
[493]	1135	!-- PE0 receives partial arrays from all processors and
	1136	!-- then outputs them. Here a barrier has to be set,
	1137	!-- because otherwise "-MPI- FATAL: Remote protocol queue
	1138	!-- full" may occur.
	1139	CALL MPI_BARRIER( comm2d, ierr )
	1140
[667]	1141	ngp = ( nxrg-nxlg+1 ) * ( nyng-nysg+1 )
[493]	1142	IF ( myid == 0 ) THEN
[1]	1143	!
[493]	1144	!-- Local array can be relocated directly.
[667]	1145	total_2d(nxlg:nxrg,nysg:nyng) = local_2d
[1]	1146	!
[493]	1147	!-- Receive data from all other PEs.
	1148	DO n = 1, numprocs-1
[1]	1149	!
[493]	1150	!-- Receive index limits first, then array.
	1151	!-- Index limits are received in arbitrary order from
	1152	!-- the PEs.
[1320]	1153	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1154	MPI_ANY_SOURCE, 0, comm2d, &
[493]	1155	status, ierr )
	1156	sender = status(MPI_SOURCE)
	1157	DEALLOCATE( local_2d )
	1158	ALLOCATE( local_2d(ind(1):ind(2),ind(3):ind(4)) )
[1320]	1159	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1160	MPI_REAL, sender, 1, comm2d, &
[493]	1161	status, ierr )
	1162	total_2d(ind(1):ind(2),ind(3):ind(4)) = local_2d
	1163	ENDDO
[1]	1164	!
[493]	1165	!-- Relocate the local array for the next loop increment
	1166	DEALLOCATE( local_2d )
[667]	1167	ALLOCATE( local_2d(nxlg:nxrg,nysg:nyng) )
[1]	1168
	1169	#if defined( __netcdf )
[1327]	1170	IF ( two_d ) THEN
	1171	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1172	id_var_do2d(av,if), &
	1173	total_2d(0:nx+1,0:ny+1), &
	1174	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
	1175	count = (/ nx+2, ny+2, 1, 1 /) )
	1176	ELSE
	1177	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1178	id_var_do2d(av,if), &
	1179	total_2d(0:nx+1,0:ny+1), &
	1180	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
	1181	count = (/ nx+2, ny+2, 1, 1 /) )
[1]	1182	ENDIF
[1783]	1183	CALL netcdf_handle_error( 'data_output_2d', 54 )
[1]	1184	#endif
	1185
[493]	1186	ELSE
[1]	1187	!
[493]	1188	!-- First send the local index limits to PE0
[667]	1189	ind(1) = nxlg; ind(2) = nxrg
	1190	ind(3) = nysg; ind(4) = nyng
[1320]	1191	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1192	comm2d, ierr )
[1]	1193	!
[493]	1194	!-- Send data to PE0
[1320]	1195	CALL MPI_SEND( local_2d(nxlg,nysg), ngp, &
[493]	1196	MPI_REAL, 0, 1, comm2d, ierr )
	1197	ENDIF
	1198	!
	1199	!-- A barrier has to be set, because otherwise some PEs may
	1200	!-- proceed too fast so that PE0 may receive wrong data on
	1201	!-- tag 0
	1202	CALL MPI_BARRIER( comm2d, ierr )
[1]	1203	ENDIF
[493]	1204
[1]	1205	ENDIF
	1206	#else
	1207	#if defined( __netcdf )
[1327]	1208	IF ( two_d ) THEN
	1209	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1210	id_var_do2d(av,if), &
	1211	local_2d(nxl:nxr+1,nys:nyn+1), &
	1212	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
	1213	count = (/ nx+2, ny+2, 1, 1 /) )
	1214	ELSE
	1215	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1216	id_var_do2d(av,if), &
	1217	local_2d(nxl:nxr+1,nys:nyn+1), &
	1218	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
	1219	count = (/ nx+2, ny+2, 1, 1 /) )
[1]	1220	ENDIF
[1783]	1221	CALL netcdf_handle_error( 'data_output_2d', 447 )
[1]	1222	#endif
	1223	#endif
	1224	do2d_xy_n = do2d_xy_n + 1
	1225	!
	1226	!-- For 2D-arrays (e.g. u*) only one cross-section is available.
	1227	!-- Hence exit loop of output levels.
	1228	IF ( two_d ) THEN
[1703]	1229	IF ( netcdf_data_format < 5 ) two_d = .FALSE.
[1]	1230	EXIT loop1
	1231	ENDIF
	1232
	1233	CASE ( 'xz' )
	1234	!
[1308]	1235	!-- Update the netCDF xz cross section time axis.
	1236	!-- In case of parallel output, this is only done by PE0
	1237	!-- to increase the performance.
	1238	IF ( simulated_time /= do2d_xz_last_time(av) ) THEN
	1239	do2d_xz_time_count(av) = do2d_xz_time_count(av) + 1
	1240	do2d_xz_last_time(av) = simulated_time
	1241	IF ( myid == 0 ) THEN
[1327]	1242	IF ( .NOT. data_output_2d_on_each_pe &
	1243	.OR. netcdf_data_format > 4 ) &
[493]	1244	THEN
[1]	1245	#if defined( __netcdf )
	1246	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1247	id_var_time_xz(av), &
[291]	1248	(/ time_since_reference_point /), &
[1]	1249	start = (/ do2d_xz_time_count(av) /), &
	1250	count = (/ 1 /) )
[1783]	1251	CALL netcdf_handle_error( 'data_output_2d', 56 )
[1]	1252	#endif
	1253	ENDIF
	1254	ENDIF
	1255	ENDIF
[667]	1256
[1]	1257	!
	1258	!-- If required, carry out averaging along y
[1960]	1259	IF ( section(is,s_ind) == -1 ) THEN
[1]	1260
[1551]	1261	ALLOCATE( local_2d_l(nxlg:nxrg,nzb_do:nzt_do) )
[1353]	1262	local_2d_l = 0.0_wp
[1551]	1263	ngp = ( nxrg-nxlg + 1 ) * ( nzt_do-nzb_do + 1 )
[1]	1264	!
	1265	!-- First local averaging on the PE
[1551]	1266	DO k = nzb_do, nzt_do
[1]	1267	DO j = nys, nyn
[667]	1268	DO i = nxlg, nxrg
[1320]	1269	local_2d_l(i,k) = local_2d_l(i,k) + &
[1]	1270	local_pf(i,j,k)
	1271	ENDDO
	1272	ENDDO
	1273	ENDDO
	1274	#if defined( __parallel )
	1275	!
	1276	!-- Now do the averaging over all PEs along y
[622]	1277	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1278	CALL MPI_ALLREDUCE( local_2d_l(nxlg,nzb_do), &
	1279	local_2d(nxlg,nzb_do), ngp, MPI_REAL, &
[1]	1280	MPI_SUM, comm1dy, ierr )
	1281	#else
	1282	local_2d = local_2d_l
	1283	#endif
[1353]	1284	local_2d = local_2d / ( ny + 1.0_wp )
[1]	1285
	1286	DEALLOCATE( local_2d_l )
	1287
	1288	ELSE
	1289	!
	1290	!-- Just store the respective section on the local array
	1291	!-- (but only if it is available on this PE!)
[1960]	1292	IF ( section(is,s_ind) >= nys .AND. section(is,s_ind) <= nyn ) &
[1]	1293	THEN
[1960]	1294	local_2d = local_pf(:,section(is,s_ind),nzb_do:nzt_do)
[1]	1295	ENDIF
	1296
	1297	ENDIF
	1298
	1299	#if defined( __parallel )
[1327]	1300	IF ( netcdf_data_format > 4 ) THEN
[1]	1301	!
[1031]	1302	!-- Output in netCDF4/HDF5 format.
[493]	1303	!-- Output only on those PEs where the respective cross
	1304	!-- sections reside. Cross sections averaged along y are
	1305	!-- output on the respective first PE along y (myidy=0).
[1960]	1306	IF ( ( section(is,s_ind) >= nys .AND. &
	1307	section(is,s_ind) <= nyn ) .OR. &
	1308	( section(is,s_ind) == -1 .AND. myidy == 0 ) ) THEN
[1]	1309	#if defined( __netcdf )
[493]	1310	!
[1308]	1311	!-- For parallel output, all cross sections are first
	1312	!-- stored here on a local array and will be written to the
	1313	!-- output file afterwards to increase the performance.
	1314	DO i = nxlg, nxrg
[1551]	1315	DO k = nzb_do, nzt_do
[1308]	1316	local_2d_sections_l(i,is,k) = local_2d(i,k)
	1317	ENDDO
	1318	ENDDO
[1]	1319	#endif
	1320	ENDIF
	1321
	1322	ELSE
	1323
[493]	1324	IF ( data_output_2d_on_each_pe ) THEN
[1]	1325	!
[493]	1326	!-- Output of partial arrays on each PE. If the cross
	1327	!-- section does not reside on the PE, output special
	1328	!-- index values.
	1329	#if defined( __netcdf )
[1327]	1330	IF ( myid == 0 ) THEN
[1320]	1331	WRITE ( 22 ) time_since_reference_point, &
[493]	1332	do2d_xz_time_count(av), av
	1333	ENDIF
	1334	#endif
[759]	1335	DO i = 0, io_blocks-1
	1336	IF ( i == io_group ) THEN
[1960]	1337	IF ( ( section(is,s_ind) >= nys .AND. &
	1338	section(is,s_ind) <= nyn ) .OR. &
	1339	( section(is,s_ind) == -1 .AND. &
[1320]	1340	nys-1 == -1 ) ) &
[759]	1341	THEN
[1551]	1342	WRITE (22) nxlg, nxrg, nzb_do, nzt_do, nzb, nzt+1
[759]	1343	WRITE (22) local_2d
	1344	ELSE
[1551]	1345	WRITE (22) -1, -1, -1, -1, -1, -1
[759]	1346	ENDIF
	1347	ENDIF
	1348	#if defined( __parallel )
	1349	CALL MPI_BARRIER( comm2d, ierr )
	1350	#endif
	1351	ENDDO
[493]	1352
	1353	ELSE
[1]	1354	!
[493]	1355	!-- PE0 receives partial arrays from all processors of the
	1356	!-- respective cross section and outputs them. Here a
	1357	!-- barrier has to be set, because otherwise
	1358	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
	1359	CALL MPI_BARRIER( comm2d, ierr )
	1360
[1551]	1361	ngp = ( nxrg-nxlg + 1 ) * ( nzt_do-nzb_do + 1 )
[493]	1362	IF ( myid == 0 ) THEN
[1]	1363	!
[493]	1364	!-- Local array can be relocated directly.
[1960]	1365	IF ( ( section(is,s_ind) >= nys .AND. &
	1366	section(is,s_ind) <= nyn ) .OR. &
	1367	( section(is,s_ind) == -1 .AND. &
	1368	nys-1 == -1 ) ) THEN
[1551]	1369	total_2d(nxlg:nxrg,nzb_do:nzt_do) = local_2d
[493]	1370	ENDIF
[1]	1371	!
[493]	1372	!-- Receive data from all other PEs.
	1373	DO n = 1, numprocs-1
	1374	!
	1375	!-- Receive index limits first, then array.
	1376	!-- Index limits are received in arbitrary order from
	1377	!-- the PEs.
[1320]	1378	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1379	MPI_ANY_SOURCE, 0, comm2d, &
[1]	1380	status, ierr )
[493]	1381	!
	1382	!-- Not all PEs have data for XZ-cross-section.
	1383	IF ( ind(1) /= -9999 ) THEN
	1384	sender = status(MPI_SOURCE)
	1385	DEALLOCATE( local_2d )
[1320]	1386	ALLOCATE( local_2d(ind(1):ind(2), &
[493]	1387	ind(3):ind(4)) )
	1388	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1389	MPI_REAL, sender, 1, comm2d, &
	1390	status, ierr )
[1320]	1391	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
[493]	1392	local_2d
	1393	ENDIF
	1394	ENDDO
	1395	!
	1396	!-- Relocate the local array for the next loop increment
	1397	DEALLOCATE( local_2d )
[1551]	1398	ALLOCATE( local_2d(nxlg:nxrg,nzb_do:nzt_do) )
[1]	1399
	1400	#if defined( __netcdf )
[1327]	1401	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1402	id_var_do2d(av,if), &
[1551]	1403	total_2d(0:nx+1,nzb_do:nzt_do),&
[1327]	1404	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
[1551]	1405	count = (/ nx+2, 1, nzt_do-nzb_do+1, 1 /) )
[1783]	1406	CALL netcdf_handle_error( 'data_output_2d', 58 )
[1]	1407	#endif
	1408
[493]	1409	ELSE
[1]	1410	!
[493]	1411	!-- If the cross section resides on the PE, send the
	1412	!-- local index limits, otherwise send -9999 to PE0.
[1960]	1413	IF ( ( section(is,s_ind) >= nys .AND. &
	1414	section(is,s_ind) <= nyn ) .OR. &
	1415	( section(is,s_ind) == -1 .AND. nys-1 == -1 ) ) &
[493]	1416	THEN
[667]	1417	ind(1) = nxlg; ind(2) = nxrg
[1551]	1418	ind(3) = nzb_do; ind(4) = nzt_do
[493]	1419	ELSE
	1420	ind(1) = -9999; ind(2) = -9999
	1421	ind(3) = -9999; ind(4) = -9999
	1422	ENDIF
[1320]	1423	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1424	comm2d, ierr )
	1425	!
	1426	!-- If applicable, send data to PE0.
	1427	IF ( ind(1) /= -9999 ) THEN
[1551]	1428	CALL MPI_SEND( local_2d(nxlg,nzb_do), ngp, &
[493]	1429	MPI_REAL, 0, 1, comm2d, ierr )
	1430	ENDIF
[1]	1431	ENDIF
	1432	!
[493]	1433	!-- A barrier has to be set, because otherwise some PEs may
	1434	!-- proceed too fast so that PE0 may receive wrong data on
	1435	!-- tag 0
	1436	CALL MPI_BARRIER( comm2d, ierr )
[1]	1437	ENDIF
[493]	1438
[1]	1439	ENDIF
	1440	#else
	1441	#if defined( __netcdf )
[1327]	1442	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1443	id_var_do2d(av,if), &
[1551]	1444	local_2d(nxl:nxr+1,nzb_do:nzt_do), &
[1327]	1445	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
[1551]	1446	count = (/ nx+2, 1, nzt_do-nzb_do+1, 1 /) )
[1783]	1447	CALL netcdf_handle_error( 'data_output_2d', 451 )
[1]	1448	#endif
	1449	#endif
	1450	do2d_xz_n = do2d_xz_n + 1
	1451
	1452	CASE ( 'yz' )
	1453	!
[1308]	1454	!-- Update the netCDF yz cross section time axis.
	1455	!-- In case of parallel output, this is only done by PE0
	1456	!-- to increase the performance.
	1457	IF ( simulated_time /= do2d_yz_last_time(av) ) THEN
	1458	do2d_yz_time_count(av) = do2d_yz_time_count(av) + 1
	1459	do2d_yz_last_time(av) = simulated_time
	1460	IF ( myid == 0 ) THEN
[1327]	1461	IF ( .NOT. data_output_2d_on_each_pe &
	1462	.OR. netcdf_data_format > 4 ) &
[493]	1463	THEN
[1]	1464	#if defined( __netcdf )
	1465	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1466	id_var_time_yz(av), &
[291]	1467	(/ time_since_reference_point /), &
[1]	1468	start = (/ do2d_yz_time_count(av) /), &
	1469	count = (/ 1 /) )
[1783]	1470	CALL netcdf_handle_error( 'data_output_2d', 59 )
[1]	1471	#endif
	1472	ENDIF
	1473	ENDIF
[1308]	1474	ENDIF
[493]	1475
[1]	1476	!
	1477	!-- If required, carry out averaging along x
[1960]	1478	IF ( section(is,s_ind) == -1 ) THEN
[1]	1479
[1551]	1480	ALLOCATE( local_2d_l(nysg:nyng,nzb_do:nzt_do) )
[1353]	1481	local_2d_l = 0.0_wp
[1551]	1482	ngp = ( nyng-nysg+1 ) * ( nzt_do-nzb_do+1 )
[1]	1483	!
	1484	!-- First local averaging on the PE
[1551]	1485	DO k = nzb_do, nzt_do
[667]	1486	DO j = nysg, nyng
[1]	1487	DO i = nxl, nxr
[1320]	1488	local_2d_l(j,k) = local_2d_l(j,k) + &
[1]	1489	local_pf(i,j,k)
	1490	ENDDO
	1491	ENDDO
	1492	ENDDO
	1493	#if defined( __parallel )
	1494	!
	1495	!-- Now do the averaging over all PEs along x
[622]	1496	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1497	CALL MPI_ALLREDUCE( local_2d_l(nysg,nzb_do), &
	1498	local_2d(nysg,nzb_do), ngp, MPI_REAL, &
[1]	1499	MPI_SUM, comm1dx, ierr )
	1500	#else
	1501	local_2d = local_2d_l
	1502	#endif
[1353]	1503	local_2d = local_2d / ( nx + 1.0_wp )
[1]	1504
	1505	DEALLOCATE( local_2d_l )
	1506
	1507	ELSE
	1508	!
	1509	!-- Just store the respective section on the local array
	1510	!-- (but only if it is available on this PE!)
[1960]	1511	IF ( section(is,s_ind) >= nxl .AND. section(is,s_ind) <= nxr ) &
[1]	1512	THEN
[1960]	1513	local_2d = local_pf(section(is,s_ind),:,nzb_do:nzt_do)
[1]	1514	ENDIF
	1515
	1516	ENDIF
	1517
	1518	#if defined( __parallel )
[1327]	1519	IF ( netcdf_data_format > 4 ) THEN
[1]	1520	!
[1031]	1521	!-- Output in netCDF4/HDF5 format.
[493]	1522	!-- Output only on those PEs where the respective cross
	1523	!-- sections reside. Cross sections averaged along x are
	1524	!-- output on the respective first PE along x (myidx=0).
[1960]	1525	IF ( ( section(is,s_ind) >= nxl .AND. &
	1526	section(is,s_ind) <= nxr ) .OR. &
	1527	( section(is,s_ind) == -1 .AND. myidx == 0 ) ) THEN
[1]	1528	#if defined( __netcdf )
[493]	1529	!
[1308]	1530	!-- For parallel output, all cross sections are first
	1531	!-- stored here on a local array and will be written to the
	1532	!-- output file afterwards to increase the performance.
	1533	DO j = nysg, nyng
[1551]	1534	DO k = nzb_do, nzt_do
[1308]	1535	local_2d_sections_l(is,j,k) = local_2d(j,k)
	1536	ENDDO
	1537	ENDDO
[1]	1538	#endif
	1539	ENDIF
	1540
	1541	ELSE
	1542
[493]	1543	IF ( data_output_2d_on_each_pe ) THEN
[1]	1544	!
[493]	1545	!-- Output of partial arrays on each PE. If the cross
	1546	!-- section does not reside on the PE, output special
	1547	!-- index values.
	1548	#if defined( __netcdf )
[1327]	1549	IF ( myid == 0 ) THEN
[1320]	1550	WRITE ( 23 ) time_since_reference_point, &
[493]	1551	do2d_yz_time_count(av), av
	1552	ENDIF
	1553	#endif
[759]	1554	DO i = 0, io_blocks-1
	1555	IF ( i == io_group ) THEN
[1960]	1556	IF ( ( section(is,s_ind) >= nxl .AND. &
	1557	section(is,s_ind) <= nxr ) .OR. &
	1558	( section(is,s_ind) == -1 .AND. &
[1320]	1559	nxl-1 == -1 ) ) &
[759]	1560	THEN
[1551]	1561	WRITE (23) nysg, nyng, nzb_do, nzt_do, nzb, nzt+1
[759]	1562	WRITE (23) local_2d
	1563	ELSE
[1551]	1564	WRITE (23) -1, -1, -1, -1, -1, -1
[759]	1565	ENDIF
	1566	ENDIF
	1567	#if defined( __parallel )
	1568	CALL MPI_BARRIER( comm2d, ierr )
	1569	#endif
	1570	ENDDO
[493]	1571
	1572	ELSE
[1]	1573	!
[493]	1574	!-- PE0 receives partial arrays from all processors of the
	1575	!-- respective cross section and outputs them. Here a
	1576	!-- barrier has to be set, because otherwise
	1577	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
	1578	CALL MPI_BARRIER( comm2d, ierr )
	1579
[1551]	1580	ngp = ( nyng-nysg+1 ) * ( nzt_do-nzb_do+1 )
[493]	1581	IF ( myid == 0 ) THEN
[1]	1582	!
[493]	1583	!-- Local array can be relocated directly.
[1960]	1584	IF ( ( section(is,s_ind) >= nxl .AND. &
	1585	section(is,s_ind) <= nxr ) .OR. &
	1586	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
[493]	1587	THEN
[1551]	1588	total_2d(nysg:nyng,nzb_do:nzt_do) = local_2d
[493]	1589	ENDIF
[1]	1590	!
[493]	1591	!-- Receive data from all other PEs.
	1592	DO n = 1, numprocs-1
	1593	!
	1594	!-- Receive index limits first, then array.
	1595	!-- Index limits are received in arbitrary order from
	1596	!-- the PEs.
[1320]	1597	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1598	MPI_ANY_SOURCE, 0, comm2d, &
[1]	1599	status, ierr )
[493]	1600	!
	1601	!-- Not all PEs have data for YZ-cross-section.
	1602	IF ( ind(1) /= -9999 ) THEN
	1603	sender = status(MPI_SOURCE)
	1604	DEALLOCATE( local_2d )
[1320]	1605	ALLOCATE( local_2d(ind(1):ind(2), &
[493]	1606	ind(3):ind(4)) )
	1607	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1608	MPI_REAL, sender, 1, comm2d, &
	1609	status, ierr )
[1320]	1610	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
[493]	1611	local_2d
	1612	ENDIF
	1613	ENDDO
	1614	!
	1615	!-- Relocate the local array for the next loop increment
	1616	DEALLOCATE( local_2d )
[1551]	1617	ALLOCATE( local_2d(nysg:nyng,nzb_do:nzt_do) )
[1]	1618
	1619	#if defined( __netcdf )
[1327]	1620	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1621	id_var_do2d(av,if), &
[1551]	1622	total_2d(0:ny+1,nzb_do:nzt_do),&
[1327]	1623	start = (/ is, 1, 1, do2d_yz_time_count(av) /), &
[1551]	1624	count = (/ 1, ny+2, nzt_do-nzb_do+1, 1 /) )
[1783]	1625	CALL netcdf_handle_error( 'data_output_2d', 61 )
[1]	1626	#endif
	1627
[493]	1628	ELSE
[1]	1629	!
[493]	1630	!-- If the cross section resides on the PE, send the
	1631	!-- local index limits, otherwise send -9999 to PE0.
[1960]	1632	IF ( ( section(is,s_ind) >= nxl .AND. &
	1633	section(is,s_ind) <= nxr ) .OR. &
	1634	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
[493]	1635	THEN
[667]	1636	ind(1) = nysg; ind(2) = nyng
[1551]	1637	ind(3) = nzb_do; ind(4) = nzt_do
[493]	1638	ELSE
	1639	ind(1) = -9999; ind(2) = -9999
	1640	ind(3) = -9999; ind(4) = -9999
	1641	ENDIF
[1320]	1642	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1643	comm2d, ierr )
	1644	!
	1645	!-- If applicable, send data to PE0.
	1646	IF ( ind(1) /= -9999 ) THEN
[1551]	1647	CALL MPI_SEND( local_2d(nysg,nzb_do), ngp, &
[493]	1648	MPI_REAL, 0, 1, comm2d, ierr )
	1649	ENDIF
[1]	1650	ENDIF
	1651	!
[493]	1652	!-- A barrier has to be set, because otherwise some PEs may
	1653	!-- proceed too fast so that PE0 may receive wrong data on
	1654	!-- tag 0
	1655	CALL MPI_BARRIER( comm2d, ierr )
[1]	1656	ENDIF
[493]	1657
[1]	1658	ENDIF
	1659	#else
	1660	#if defined( __netcdf )
[1327]	1661	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1662	id_var_do2d(av,if), &
[1551]	1663	local_2d(nys:nyn+1,nzb_do:nzt_do), &
[1327]	1664	start = (/ is, 1, 1, do2d_xz_time_count(av) /), &
[1551]	1665	count = (/ 1, ny+2, nzt_do-nzb_do+1, 1 /) )
[1783]	1666	CALL netcdf_handle_error( 'data_output_2d', 452 )
[1]	1667	#endif
	1668	#endif
	1669	do2d_yz_n = do2d_yz_n + 1
	1670
	1671	END SELECT
	1672
	1673	is = is + 1
	1674	ENDDO loop1
	1675
[1308]	1676	!
	1677	!-- For parallel output, all data were collected before on a local array
	1678	!-- and are written now to the netcdf file. This must be done to increase
	1679	!-- the performance of the parallel output.
	1680	#if defined( __netcdf )
[1327]	1681	IF ( netcdf_data_format > 4 ) THEN
[1308]	1682
	1683	SELECT CASE ( mode )
	1684
	1685	CASE ( 'xy' )
	1686	IF ( two_d ) THEN
[1703]	1687	nis = 1
	1688	two_d = .FALSE.
[1308]	1689	ELSE
[1703]	1690	nis = ns
[1308]	1691	ENDIF
	1692	!
	1693	!-- Do not output redundant ghost point data except for the
	1694	!-- boundaries of the total domain.
	1695	IF ( nxr == nx .AND. nyn /= ny ) THEN
	1696	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1697	id_var_do2d(av,if), &
	1698	local_2d_sections(nxl:nxr+1, &
[1703]	1699	nys:nyn,1:nis), &
[1308]	1700	start = (/ nxl+1, nys+1, 1, &
	1701	do2d_xy_time_count(av) /), &
	1702	count = (/ nxr-nxl+2, &
[1703]	1703	nyn-nys+1, nis, 1 &
[1308]	1704	/) )
	1705	ELSEIF ( nxr /= nx .AND. nyn == ny ) THEN
	1706	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1707	id_var_do2d(av,if), &
	1708	local_2d_sections(nxl:nxr, &
[1703]	1709	nys:nyn+1,1:nis), &
[1308]	1710	start = (/ nxl+1, nys+1, 1, &
	1711	do2d_xy_time_count(av) /), &
	1712	count = (/ nxr-nxl+1, &
[1703]	1713	nyn-nys+2, nis, 1 &
[1308]	1714	/) )
	1715	ELSEIF ( nxr == nx .AND. nyn == ny ) THEN
	1716	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1717	id_var_do2d(av,if), &
	1718	local_2d_sections(nxl:nxr+1, &
[1703]	1719	nys:nyn+1,1:nis), &
[1308]	1720	start = (/ nxl+1, nys+1, 1, &
	1721	do2d_xy_time_count(av) /), &
	1722	count = (/ nxr-nxl+2, &
[1703]	1723	nyn-nys+2, nis, 1 &
[1308]	1724	/) )
	1725	ELSE
	1726	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1727	id_var_do2d(av,if), &
	1728	local_2d_sections(nxl:nxr, &
[1703]	1729	nys:nyn,1:nis), &
[1308]	1730	start = (/ nxl+1, nys+1, 1, &
	1731	do2d_xy_time_count(av) /), &
	1732	count = (/ nxr-nxl+1, &
[1703]	1733	nyn-nys+1, nis, 1 &
[1308]	1734	/) )
	1735	ENDIF
	1736
[1783]	1737	CALL netcdf_handle_error( 'data_output_2d', 55 )
[1308]	1738
	1739	CASE ( 'xz' )
	1740	!
	1741	!-- First, all PEs get the information of all cross-sections.
	1742	!-- Then the data are written to the output file by all PEs
	1743	!-- while NF90_COLLECTIVE is set in subroutine
	1744	!-- define_netcdf_header. Although redundant information are
	1745	!-- written to the output file in that case, the performance
	1746	!-- is significantly better compared to the case where only
	1747	!-- the first row of PEs in x-direction (myidx = 0) is given
	1748	!-- the output while NF90_INDEPENDENT is set.
	1749	IF ( npey /= 1 ) THEN
	1750
	1751	#if defined( __parallel )
	1752	!
	1753	!-- Distribute data over all PEs along y
[1551]	1754	ngp = ( nxrg-nxlg+1 ) * ( nzt_do-nzb_do+1 ) * ns
[1308]	1755	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1756	CALL MPI_ALLREDUCE( local_2d_sections_l(nxlg,1,nzb_do), &
	1757	local_2d_sections(nxlg,1,nzb_do), &
[1308]	1758	ngp, MPI_REAL, MPI_SUM, comm1dy, &
	1759	ierr )
	1760	#else
	1761	local_2d_sections = local_2d_sections_l
	1762	#endif
	1763	ENDIF
	1764	!
	1765	!-- Do not output redundant ghost point data except for the
	1766	!-- boundaries of the total domain.
	1767	IF ( nxr == nx ) THEN
	1768	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1769	id_var_do2d(av,if), &
	1770	local_2d_sections(nxl:nxr+1,1: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+2, ns, nzt_do-nzb_do+1, &
[1308]	1775	1 /) )
	1776	ELSE
	1777	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1778	id_var_do2d(av,if), &
	1779	local_2d_sections(nxl:nxr,1:ns, &
[1551]	1780	nzb_do:nzt_do), &
[1308]	1781	start = (/ nxl+1, 1, 1, &
	1782	do2d_xz_time_count(av) /), &
[1551]	1783	count = (/ nxr-nxl+1, ns, nzt_do-nzb_do+1, &
[1308]	1784	1 /) )
	1785	ENDIF
	1786
[1783]	1787	CALL netcdf_handle_error( 'data_output_2d', 57 )
[1308]	1788
	1789	CASE ( 'yz' )
	1790	!
	1791	!-- First, all PEs get the information of all cross-sections.
	1792	!-- Then the data are written to the output file by all PEs
	1793	!-- while NF90_COLLECTIVE is set in subroutine
	1794	!-- define_netcdf_header. Although redundant information are
	1795	!-- written to the output file in that case, the performance
	1796	!-- is significantly better compared to the case where only
	1797	!-- the first row of PEs in y-direction (myidy = 0) is given
	1798	!-- the output while NF90_INDEPENDENT is set.
	1799	IF ( npex /= 1 ) THEN
	1800
	1801	#if defined( __parallel )
	1802	!
	1803	!-- Distribute data over all PEs along x
	1804	ngp = ( nyng-nysg+1 ) * ( nzt-nzb + 2 ) * ns
	1805	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1806	CALL MPI_ALLREDUCE( local_2d_sections_l(1,nysg,nzb_do), &
	1807	local_2d_sections(1,nysg,nzb_do), &
[1308]	1808	ngp, MPI_REAL, MPI_SUM, comm1dx, &
	1809	ierr )
	1810	#else
	1811	local_2d_sections = local_2d_sections_l
	1812	#endif
	1813	ENDIF
	1814	!
	1815	!-- Do not output redundant ghost point data except for the
	1816	!-- boundaries of the total domain.
	1817	IF ( nyn == ny ) THEN
	1818	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1819	id_var_do2d(av,if), &
	1820	local_2d_sections(1:ns, &
[1551]	1821	nys:nyn+1,nzb_do:nzt_do), &
[1308]	1822	start = (/ 1, nys+1, 1, &
	1823	do2d_yz_time_count(av) /), &
	1824	count = (/ ns, nyn-nys+2, &
[1551]	1825	nzt_do-nzb_do+1, 1 /) )
[1308]	1826	ELSE
	1827	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1828	id_var_do2d(av,if), &
	1829	local_2d_sections(1:ns,nys:nyn, &
[1551]	1830	nzb_do:nzt_do), &
[1308]	1831	start = (/ 1, nys+1, 1, &
	1832	do2d_yz_time_count(av) /), &
	1833	count = (/ ns, nyn-nys+1, &
[1551]	1834	nzt_do-nzb_do+1, 1 /) )
[1308]	1835	ENDIF
	1836
[1783]	1837	CALL netcdf_handle_error( 'data_output_2d', 60 )
[1308]	1838
	1839	CASE DEFAULT
	1840	message_string = 'unknown cross-section: ' // TRIM( mode )
	1841	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
	1842
	1843	END SELECT
	1844
	1845	ENDIF
[1311]	1846	#endif
[1]	1847	ENDIF
	1848
	1849	if = if + 1
	1850	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
	1851	do2d_mode = do2d(av,if)(l-1:l)
	1852
	1853	ENDDO
	1854
	1855	!
	1856	!-- Deallocate temporary arrays.
	1857	IF ( ALLOCATED( level_z ) ) DEALLOCATE( level_z )
[1308]	1858	IF ( netcdf_data_format > 4 ) THEN
	1859	DEALLOCATE( local_pf, local_2d, local_2d_sections )
	1860	IF( mode == 'xz' .OR. mode == 'yz' ) DEALLOCATE( local_2d_sections_l )
	1861	ENDIF
[1]	1862	#if defined( __parallel )
	1863	IF ( .NOT. data_output_2d_on_each_pe .AND. myid == 0 ) THEN
	1864	DEALLOCATE( total_2d )
	1865	ENDIF
	1866	#endif
	1867
	1868	!
	1869	!-- Close plot output file.
[1960]	1870	file_id = 20 + s_ind
[1]	1871
	1872	IF ( data_output_2d_on_each_pe ) THEN
[759]	1873	DO i = 0, io_blocks-1
	1874	IF ( i == io_group ) THEN
	1875	CALL close_file( file_id )
	1876	ENDIF
	1877	#if defined( __parallel )
	1878	CALL MPI_BARRIER( comm2d, ierr )
	1879	#endif
	1880	ENDDO
[1]	1881	ELSE
	1882	IF ( myid == 0 ) CALL close_file( file_id )
	1883	ENDIF
	1884
[1318]	1885	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
[1]	1886
	1887	END SUBROUTINE data_output_2d

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |