Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 2076

Last change on this file since 2076 was 2032, checked in by knoop, 8 years ago
last commit documented
Property svn:keywords set to `Id`
File size: 77.7 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	! -----------------
[1961]	22	!
[2032]	23	!
[1552]	24	! Former revisions:
	25	! -----------------
	26	! $Id: data_output_2d.f90 2032 2016-10-21 15:13:51Z raasch $
	27	!
[2032]	28	! 2031 2016-10-21 15:11:58Z knoop
	29	! renamed variable rho to rho_ocean and rho_av to rho_ocean_av
	30	!
[2001]	31	! 2000 2016-08-20 18:09:15Z knoop
	32	! Forced header and separation lines into 80 columns
	33	!
[1981]	34	! 1980 2016-07-29 15:51:57Z suehring
	35	! Bugfix, in order to steer user-defined output, setting flag found explicitly
	36	! to .F.
	37	!
[1977]	38	! 1976 2016-07-27 13:28:04Z maronga
	39	! Output of radiation quantities is now done directly in the respective module
	40	!
[1973]	41	! 1972 2016-07-26 07:52:02Z maronga
[1976]	42	! Output of land surface quantities is now done directly in the respective
	43	! module
[1973]	44	!
[1961]	45	! 1960 2016-07-12 16:34:24Z suehring
	46	! Scalar surface flux added
	47	! Rename INTEGER variable s into s_ind, as s is already assigned to scalar
	48	!
[1851]	49	! 1849 2016-04-08 11:33:18Z hoffmann
	50	! precipitation_amount, precipitation_rate, prr moved to arrays_3d
	51	!
[1823]	52	! 1822 2016-04-07 07:49:42Z hoffmann
	53	! Output of bulk cloud physics simplified.
	54	!
[1789]	55	! 1788 2016-03-10 11:01:04Z maronga
	56	! Added output of z0q
	57	!
[1784]	58	! 1783 2016-03-06 18:36:17Z raasch
	59	! name change of netcdf routines and module + related changes
	60	!
[1746]	61	! 1745 2016-02-05 13:06:51Z gronemeier
	62	! Bugfix: test if time axis limit exceeds moved to point after call of check_open
	63	!
[1704]	64	! 1703 2015-11-02 12:38:44Z raasch
	65	! bugfix for output of single (*) xy-sections in case of parallel netcdf I/O
	66	!
[1702]	67	! 1701 2015-11-02 07:43:04Z maronga
	68	! Bugfix in output of RRTGM data
	69	!
[1692]	70	! 1691 2015-10-26 16:17:44Z maronga
	71	! Added output of Obukhov length (ol) and radiative heating rates for RRTMG.
	72	! Formatting corrections.
	73	!
[1683]	74	! 1682 2015-10-07 23:56:08Z knoop
	75	! Code annotations made doxygen readable
	76	!
[1586]	77	! 1585 2015-04-30 07:05:52Z maronga
	78	! Added support for RRTMG
	79	!
[1556]	80	! 1555 2015-03-04 17:44:27Z maronga
	81	! Added output of r_a and r_s
	82	!
[1552]	83	! 1551 2015-03-03 14:18:16Z maronga
[1551]	84	! Added suppport for land surface model and radiation model output. In the course
	85	! of this action, the limits for vertical loops have been changed (from nzb and
	86	! nzt+1 to nzb_do and nzt_do, respectively in order to allow soil model output).
	87	! Moreover, a new vertical grid zs was introduced.
[1329]	88	!
[1360]	89	! 1359 2014-04-11 17:15:14Z hoffmann
	90	! New particle structure integrated.
	91	!
[1354]	92	! 1353 2014-04-08 15:21:23Z heinze
	93	! REAL constants provided with KIND-attribute
	94	!
[1329]	95	! 1327 2014-03-21 11:00:16Z raasch
	96	! parts concerning iso2d output removed,
	97	! -netcdf output queries
	98	!
[1321]	99	! 1320 2014-03-20 08:40:49Z raasch
[1320]	100	! ONLY-attribute added to USE-statements,
	101	! kind-parameters added to all INTEGER and REAL declaration statements,
	102	! kinds are defined in new module kinds,
	103	! revision history before 2012 removed,
	104	! comment fields (!:) to be used for variable explanations added to
	105	! all variable declaration statements
[1309]	106	!
[1319]	107	! 1318 2014-03-17 13:35:16Z raasch
	108	! barrier argument removed from cpu_log.
	109	! module interfaces removed
	110	!
[1312]	111	! 1311 2014-03-14 12:13:39Z heinze
	112	! bugfix: close #if defined( __netcdf )
	113	!
[1309]	114	! 1308 2014-03-13 14:58:42Z fricke
[1308]	115	! +local_2d_sections, local_2d_sections_l, ns
	116	! Check, if the limit of the time dimension is exceeded for parallel output
	117	! To increase the performance for parallel output, the following is done:
	118	! - Update of time axis is only done by PE0
	119	! - Cross sections are first stored on a local array and are written
	120	! collectively to the output file by all PEs.
[674]	121	!
[1116]	122	! 1115 2013-03-26 18:16:16Z hoffmann
	123	! ql is calculated by calc_liquid_water_content
	124	!
[1077]	125	! 1076 2012-12-05 08:30:18Z hoffmann
	126	! Bugfix in output of ql
	127	!
[1066]	128	! 1065 2012-11-22 17:42:36Z hoffmann
	129	! Bugfix: Output of cross sections of ql
	130	!
[1054]	131	! 1053 2012-11-13 17:11:03Z hoffmann
	132	! +qr, nr, qc and cross sections
	133	!
[1037]	134	! 1036 2012-10-22 13:43:42Z raasch
	135	! code put under GPL (PALM 3.9)
	136	!
[1035]	137	! 1031 2012-10-19 14:35:30Z raasch
	138	! netCDF4 without parallel file support implemented
	139	!
[1008]	140	! 1007 2012-09-19 14:30:36Z franke
	141	! Bugfix: missing calculation of ql_vp added
	142	!
[979]	143	! 978 2012-08-09 08:28:32Z fricke
	144	! +z0h
	145	!
[1]	146	! Revision 1.1 1997/08/11 06:24:09 raasch
	147	! Initial revision
	148	!
	149	!
	150	! Description:
	151	! ------------
[1682]	152	!> Data output of horizontal cross-sections in netCDF format or binary format
	153	!> compatible to old graphic software iso2d.
	154	!> Attention: The position of the sectional planes is still not always computed
	155	!> --------- correctly. (zu is used always)!
[1]	156	!------------------------------------------------------------------------------!
[1682]	157	SUBROUTINE data_output_2d( mode, av )
	158
[1]	159
[1320]	160	USE arrays_3d, &
[1849]	161	ONLY: dzw, e, nr, ol, p, pt, precipitation_amount, precipitation_rate,&
[2031]	162	prr,q, qc, ql, ql_c, ql_v, ql_vp, qr, qsws, rho_ocean, s, sa, shf, &
[1960]	163	ssws, tend, ts, u, us, v, vpt, w, z0, z0h, z0q, zu, zw
[1320]	164
[1]	165	USE averaging
[1320]	166
	167	USE cloud_parameters, &
[1849]	168	ONLY: hyrho, l_d_cp, pt_d_t
[1320]	169
	170	USE control_parameters, &
	171	ONLY: cloud_physics, data_output_2d_on_each_pe, data_output_xy, &
	172	data_output_xz, data_output_yz, do2d, &
	173	do2d_xy_last_time, do2d_xy_n, do2d_xy_time_count, &
	174	do2d_xz_last_time, do2d_xz_n, do2d_xz_time_count, &
	175	do2d_yz_last_time, do2d_yz_n, do2d_yz_time_count, &
[1822]	176	ibc_uv_b, io_blocks, io_group, message_string, &
	177	ntdim_2d_xy, ntdim_2d_xz, ntdim_2d_yz, &
	178	psolver, section, simulated_time, simulated_time_chr, &
	179	time_since_reference_point
[1320]	180
	181	USE cpulog, &
	182	ONLY: cpu_log, log_point
	183
	184	USE grid_variables, &
	185	ONLY: dx, dy
	186
	187	USE indices, &
	188	ONLY: nbgp, nx, nxl, nxlg, nxr, nxrg, ny, nyn, nyng, nys, nysg, &
	189	nz, nzb, nzt
	190
	191	USE kinds
[1551]	192
	193	USE land_surface_model_mod, &
[1972]	194	ONLY: land_surface, lsm_data_output_2d, zs
[1551]	195
[1783]	196	#if defined( __netcdf )
	197	USE NETCDF
	198	#endif
[1320]	199
[1783]	200	USE netcdf_interface, &
	201	ONLY: id_set_xy, id_set_xz, id_set_yz, id_var_do2d, id_var_time_xy, &
	202	id_var_time_xz, id_var_time_yz, nc_stat, netcdf_data_format, &
	203	netcdf_handle_error
	204
[1320]	205	USE particle_attributes, &
[1359]	206	ONLY: grid_particles, number_of_particles, particle_advection_start, &
	207	particles, prt_count
[1320]	208
[1]	209	USE pegrid
	210
[1551]	211	USE radiation_model_mod, &
[1976]	212	ONLY: radiation, radiation_data_output_2d
[1551]	213
[1]	214	IMPLICIT NONE
	215
[1682]	216	CHARACTER (LEN=2) :: do2d_mode !<
	217	CHARACTER (LEN=2) :: mode !<
	218	CHARACTER (LEN=4) :: grid !<
	219	CHARACTER (LEN=25) :: section_chr !<
	220	CHARACTER (LEN=50) :: rtext !<
[1320]	221
[1682]	222	INTEGER(iwp) :: av !<
	223	INTEGER(iwp) :: ngp !<
	224	INTEGER(iwp) :: file_id !<
	225	INTEGER(iwp) :: i !<
	226	INTEGER(iwp) :: if !<
	227	INTEGER(iwp) :: is !<
	228	INTEGER(iwp) :: iis !<
	229	INTEGER(iwp) :: j !<
	230	INTEGER(iwp) :: k !<
	231	INTEGER(iwp) :: l !<
	232	INTEGER(iwp) :: layer_xy !<
	233	INTEGER(iwp) :: n !<
[1703]	234	INTEGER(iwp) :: nis !<
[1682]	235	INTEGER(iwp) :: ns !<
	236	INTEGER(iwp) :: nzb_do !< lower limit of the data field (usually nzb)
	237	INTEGER(iwp) :: nzt_do !< upper limit of the data field (usually nzt+1)
	238	INTEGER(iwp) :: psi !<
[1960]	239	INTEGER(iwp) :: s_ind !<
[1682]	240	INTEGER(iwp) :: sender !<
	241	INTEGER(iwp) :: ind(4) !<
[1320]	242
[1682]	243	LOGICAL :: found !<
	244	LOGICAL :: resorted !<
	245	LOGICAL :: two_d !<
[1320]	246
[1682]	247	REAL(wp) :: mean_r !<
	248	REAL(wp) :: s_r2 !<
	249	REAL(wp) :: s_r3 !<
[1320]	250
[1682]	251	REAL(wp), DIMENSION(:), ALLOCATABLE :: level_z !<
	252	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d !<
	253	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d_l !<
	254	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !<
	255	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections !<
	256	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections_l !<
[1359]	257
[1]	258	#if defined( __parallel )
[1682]	259	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: total_2d !<
[1]	260	#endif
[1682]	261	REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !<
[1]	262
	263	NAMELIST /LOCAL/ rtext
	264
	265	!
	266	!-- Immediate return, if no output is requested (no respective sections
	267	!-- found in parameter data_output)
	268	IF ( mode == 'xy' .AND. .NOT. data_output_xy(av) ) RETURN
	269	IF ( mode == 'xz' .AND. .NOT. data_output_xz(av) ) RETURN
	270	IF ( mode == 'yz' .AND. .NOT. data_output_yz(av) ) RETURN
	271
[1308]	272	CALL cpu_log (log_point(3),'data_output_2d','start')
	273
[1]	274	two_d = .FALSE. ! local variable to distinguish between output of pure 2D
	275	! arrays and cross-sections of 3D arrays.
	276
	277	!
	278	!-- Depending on the orientation of the cross-section, the respective output
	279	!-- files have to be opened.
	280	SELECT CASE ( mode )
	281
	282	CASE ( 'xy' )
[1960]	283	s_ind = 1
[667]	284	ALLOCATE( level_z(nzb:nzt+1), local_2d(nxlg:nxrg,nysg:nyng) )
[1]	285
[1308]	286	IF ( netcdf_data_format > 4 ) THEN
	287	ns = 1
[1960]	288	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	289	ns = ns + 1
	290	ENDDO
	291	ns = ns - 1
	292	ALLOCATE( local_2d_sections(nxlg:nxrg,nysg:nyng,1:ns) )
[1353]	293	local_2d_sections = 0.0_wp
[1308]	294	ENDIF
	295
[493]	296	!
[1031]	297	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	298	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	299	CALL check_open( 101+av*10 )
	300	ENDIF
[1]	301	IF ( data_output_2d_on_each_pe ) THEN
	302	CALL check_open( 21 )
	303	ELSE
	304	IF ( myid == 0 ) THEN
	305	#if defined( __parallel )
[667]	306	ALLOCATE( total_2d(-nbgp:nx+nbgp,-nbgp:ny+nbgp) )
[1]	307	#endif
	308	ENDIF
	309	ENDIF
	310
	311	CASE ( 'xz' )
[1960]	312	s_ind = 2
[667]	313	ALLOCATE( local_2d(nxlg:nxrg,nzb:nzt+1) )
[1]	314
[1308]	315	IF ( netcdf_data_format > 4 ) THEN
	316	ns = 1
[1960]	317	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	318	ns = ns + 1
	319	ENDDO
	320	ns = ns - 1
	321	ALLOCATE( local_2d_sections(nxlg:nxrg,1:ns,nzb:nzt+1) )
	322	ALLOCATE( local_2d_sections_l(nxlg:nxrg,1:ns,nzb:nzt+1) )
[1353]	323	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
[1308]	324	ENDIF
	325
[493]	326	!
[1031]	327	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	328	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	329	CALL check_open( 102+av*10 )
	330	ENDIF
[1]	331
	332	IF ( data_output_2d_on_each_pe ) THEN
	333	CALL check_open( 22 )
	334	ELSE
	335	IF ( myid == 0 ) THEN
	336	#if defined( __parallel )
[667]	337	ALLOCATE( total_2d(-nbgp:nx+nbgp,nzb:nzt+1) )
[1]	338	#endif
	339	ENDIF
	340	ENDIF
	341
	342	CASE ( 'yz' )
[1960]	343	s_ind = 3
[667]	344	ALLOCATE( local_2d(nysg:nyng,nzb:nzt+1) )
[1]	345
[1308]	346	IF ( netcdf_data_format > 4 ) THEN
	347	ns = 1
[1960]	348	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	349	ns = ns + 1
	350	ENDDO
	351	ns = ns - 1
	352	ALLOCATE( local_2d_sections(1:ns,nysg:nyng,nzb:nzt+1) )
	353	ALLOCATE( local_2d_sections_l(1:ns,nysg:nyng,nzb:nzt+1) )
[1353]	354	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
[1308]	355	ENDIF
	356
[493]	357	!
[1031]	358	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	359	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	360	CALL check_open( 103+av*10 )
	361	ENDIF
[1]	362
	363	IF ( data_output_2d_on_each_pe ) THEN
	364	CALL check_open( 23 )
	365	ELSE
	366	IF ( myid == 0 ) THEN
	367	#if defined( __parallel )
[667]	368	ALLOCATE( total_2d(-nbgp:ny+nbgp,nzb:nzt+1) )
[1]	369	#endif
	370	ENDIF
	371	ENDIF
	372
	373	CASE DEFAULT
[254]	374	message_string = 'unknown cross-section: ' // TRIM( mode )
	375	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
[1]	376
	377	END SELECT
	378
	379	!
[1745]	380	!-- For parallel netcdf output the time axis must be limited. Return, if this
	381	!-- limit is exceeded. This could be the case, if the simulated time exceeds
	382	!-- the given end time by the length of the given output interval.
	383	IF ( netcdf_data_format > 4 ) THEN
	384	IF ( mode == 'xy' .AND. do2d_xy_time_count(av) + 1 > &
	385	ntdim_2d_xy(av) ) THEN
	386	WRITE ( message_string, * ) 'Output of xy cross-sections is not ', &
	387	'given at t=', simulated_time, '&because the', &
	388	' maximum number of output time levels is exceeded.'
	389	CALL message( 'data_output_2d', 'PA0384', 0, 1, 0, 6, 0 )
	390	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	391	RETURN
	392	ENDIF
	393	IF ( mode == 'xz' .AND. do2d_xz_time_count(av) + 1 > &
	394	ntdim_2d_xz(av) ) THEN
	395	WRITE ( message_string, * ) 'Output of xz cross-sections is not ', &
	396	'given at t=', simulated_time, '&because the', &
	397	' maximum number of output time levels is exceeded.'
	398	CALL message( 'data_output_2d', 'PA0385', 0, 1, 0, 6, 0 )
	399	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	400	RETURN
	401	ENDIF
	402	IF ( mode == 'yz' .AND. do2d_yz_time_count(av) + 1 > &
	403	ntdim_2d_yz(av) ) THEN
	404	WRITE ( message_string, * ) 'Output of yz cross-sections is not ', &
	405	'given at t=', simulated_time, '&because the', &
	406	' maximum number of output time levels is exceeded.'
	407	CALL message( 'data_output_2d', 'PA0386', 0, 1, 0, 6, 0 )
	408	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	409	RETURN
	410	ENDIF
	411	ENDIF
	412
	413	!
[1]	414	!-- Allocate a temporary array for resorting (kji -> ijk).
[667]	415	ALLOCATE( local_pf(nxlg:nxrg,nysg:nyng,nzb:nzt+1) )
[1]	416
	417	!
	418	!-- Loop of all variables to be written.
	419	!-- Output dimensions chosen
	420	if = 1
	421	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
	422	do2d_mode = do2d(av,if)(l-1:l)
	423
	424	DO WHILE ( do2d(av,if)(1:1) /= ' ' )
	425
	426	IF ( do2d_mode == mode ) THEN
[1980]	427	!
	428	!-- Set flag to steer output of radiation, land-surface, or user-defined
	429	!-- quantities
	430	found = .FALSE.
[1551]	431
	432	nzb_do = nzb
	433	nzt_do = nzt+1
[1]	434	!
	435	!-- Store the array chosen on the temporary array.
	436	resorted = .FALSE.
	437	SELECT CASE ( TRIM( do2d(av,if) ) )
	438
	439	CASE ( 'e_xy', 'e_xz', 'e_yz' )
	440	IF ( av == 0 ) THEN
	441	to_be_resorted => e
	442	ELSE
	443	to_be_resorted => e_av
	444	ENDIF
	445	IF ( mode == 'xy' ) level_z = zu
	446
[771]	447	CASE ( 'lpt_xy', 'lpt_xz', 'lpt_yz' )
	448	IF ( av == 0 ) THEN
	449	to_be_resorted => pt
	450	ELSE
	451	to_be_resorted => lpt_av
	452	ENDIF
	453	IF ( mode == 'xy' ) level_z = zu
	454
[1]	455	CASE ( 'lwp*_xy' ) ! 2d-array
	456	IF ( av == 0 ) THEN
[667]	457	DO i = nxlg, nxrg
	458	DO j = nysg, nyng
[1320]	459	local_pf(i,j,nzb+1) = SUM( ql(nzb:nzt,j,i) * &
[1]	460	dzw(1:nzt+1) )
	461	ENDDO
	462	ENDDO
	463	ELSE
[667]	464	DO i = nxlg, nxrg
	465	DO j = nysg, nyng
[1]	466	local_pf(i,j,nzb+1) = lwp_av(j,i)
	467	ENDDO
	468	ENDDO
	469	ENDIF
	470	resorted = .TRUE.
	471	two_d = .TRUE.
	472	level_z(nzb+1) = zu(nzb+1)
	473
[1053]	474	CASE ( 'nr_xy', 'nr_xz', 'nr_yz' )
	475	IF ( av == 0 ) THEN
	476	to_be_resorted => nr
	477	ELSE
	478	to_be_resorted => nr_av
	479	ENDIF
	480	IF ( mode == 'xy' ) level_z = zu
	481
[1691]	482	CASE ( 'ol*_xy' ) ! 2d-array
	483	IF ( av == 0 ) THEN
	484	DO i = nxlg, nxrg
	485	DO j = nysg, nyng
	486	local_pf(i,j,nzb+1) = ol(j,i)
	487	ENDDO
	488	ENDDO
	489	ELSE
	490	DO i = nxlg, nxrg
	491	DO j = nysg, nyng
	492	local_pf(i,j,nzb+1) = ol_av(j,i)
	493	ENDDO
	494	ENDDO
	495	ENDIF
	496	resorted = .TRUE.
	497	two_d = .TRUE.
	498	level_z(nzb+1) = zu(nzb+1)
	499
[1]	500	CASE ( 'p_xy', 'p_xz', 'p_yz' )
	501	IF ( av == 0 ) THEN
[729]	502	IF ( psolver /= 'sor' ) CALL exchange_horiz( p, nbgp )
[1]	503	to_be_resorted => p
	504	ELSE
[729]	505	IF ( psolver /= 'sor' ) CALL exchange_horiz( p_av, nbgp )
[1]	506	to_be_resorted => p_av
	507	ENDIF
	508	IF ( mode == 'xy' ) level_z = zu
	509
	510	CASE ( 'pc_xy', 'pc_xz', 'pc_yz' ) ! particle concentration
	511	IF ( av == 0 ) THEN
[215]	512	IF ( simulated_time >= particle_advection_start ) THEN
	513	tend = prt_count
[667]	514	CALL exchange_horiz( tend, nbgp )
[215]	515	ELSE
[1353]	516	tend = 0.0_wp
[215]	517	ENDIF
[667]	518	DO i = nxlg, nxrg
	519	DO j = nysg, nyng
[1]	520	DO k = nzb, nzt+1
	521	local_pf(i,j,k) = tend(k,j,i)
	522	ENDDO
	523	ENDDO
	524	ENDDO
	525	resorted = .TRUE.
	526	ELSE
[667]	527	CALL exchange_horiz( pc_av, nbgp )
[1]	528	to_be_resorted => pc_av
	529	ENDIF
	530
[1359]	531	CASE ( 'pr_xy', 'pr_xz', 'pr_yz' ) ! mean particle radius (effective radius)
[1]	532	IF ( av == 0 ) THEN
[215]	533	IF ( simulated_time >= particle_advection_start ) THEN
	534	DO i = nxl, nxr
	535	DO j = nys, nyn
	536	DO k = nzb, nzt+1
[1359]	537	number_of_particles = prt_count(k,j,i)
	538	IF (number_of_particles <= 0) CYCLE
	539	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
	540	s_r2 = 0.0_wp
[1353]	541	s_r3 = 0.0_wp
[1359]	542	DO n = 1, number_of_particles
	543	IF ( particles(n)%particle_mask ) THEN
	544	s_r2 = s_r2 + particles(n)%radius*2 &
	545	particles(n)%weight_factor
	546	s_r3 = s_r3 + particles(n)%radius*3 &
	547	particles(n)%weight_factor
	548	ENDIF
[215]	549	ENDDO
[1359]	550	IF ( s_r2 > 0.0_wp ) THEN
	551	mean_r = s_r3 / s_r2
[215]	552	ELSE
[1353]	553	mean_r = 0.0_wp
[215]	554	ENDIF
	555	tend(k,j,i) = mean_r
[1]	556	ENDDO
	557	ENDDO
	558	ENDDO
[667]	559	CALL exchange_horiz( tend, nbgp )
[215]	560	ELSE
[1353]	561	tend = 0.0_wp
[1359]	562	ENDIF
[667]	563	DO i = nxlg, nxrg
	564	DO j = nysg, nyng
[1]	565	DO k = nzb, nzt+1
	566	local_pf(i,j,k) = tend(k,j,i)
	567	ENDDO
	568	ENDDO
	569	ENDDO
	570	resorted = .TRUE.
	571	ELSE
[667]	572	CALL exchange_horiz( pr_av, nbgp )
[1]	573	to_be_resorted => pr_av
	574	ENDIF
	575
[72]	576	CASE ( 'pra*_xy' ) ! 2d-array / integral quantity => no av
	577	CALL exchange_horiz_2d( precipitation_amount )
[667]	578	DO i = nxlg, nxrg
	579	DO j = nysg, nyng
[72]	580	local_pf(i,j,nzb+1) = precipitation_amount(j,i)
	581	ENDDO
	582	ENDDO
[1353]	583	precipitation_amount = 0.0_wp ! reset for next integ. interval
[72]	584	resorted = .TRUE.
	585	two_d = .TRUE.
	586	level_z(nzb+1) = zu(nzb+1)
	587
	588	CASE ( 'prr*_xy' ) ! 2d-array
[1822]	589	IF ( av == 0 ) THEN
	590	CALL exchange_horiz_2d( prr(nzb+1,:,:) )
	591	DO i = nxlg, nxrg
	592	DO j = nysg, nyng
	593	local_pf(i,j,nzb+1) = prr(nzb+1,j,i) * hyrho(nzb+1)
[1053]	594	ENDDO
[1822]	595	ENDDO
[1053]	596	ELSE
[1822]	597	CALL exchange_horiz_2d( prr_av(nzb+1,:,:) )
	598	DO i = nxlg, nxrg
	599	DO j = nysg, nyng
	600	local_pf(i,j,nzb+1) = prr_av(nzb+1,j,i) * hyrho(nzb+1)
[1053]	601	ENDDO
[1822]	602	ENDDO
[1053]	603	ENDIF
	604	resorted = .TRUE.
	605	two_d = .TRUE.
	606	level_z(nzb+1) = zu(nzb+1)
	607
	608	CASE ( 'prr_xy', 'prr_xz', 'prr_yz' )
[72]	609	IF ( av == 0 ) THEN
[1053]	610	CALL exchange_horiz( prr, nbgp )
[667]	611	DO i = nxlg, nxrg
	612	DO j = nysg, nyng
[1053]	613	DO k = nzb, nzt+1
[1822]	614	local_pf(i,j,k) = prr(k,j,i) * hyrho(nzb+1)
[1053]	615	ENDDO
[72]	616	ENDDO
	617	ENDDO
	618	ELSE
[1053]	619	CALL exchange_horiz( prr_av, nbgp )
[667]	620	DO i = nxlg, nxrg
	621	DO j = nysg, nyng
[1053]	622	DO k = nzb, nzt+1
[1822]	623	local_pf(i,j,k) = prr_av(k,j,i) * hyrho(nzb+1)
[1053]	624	ENDDO
[72]	625	ENDDO
	626	ENDDO
	627	ENDIF
	628	resorted = .TRUE.
[1053]	629	IF ( mode == 'xy' ) level_z = zu
[72]	630
[1]	631	CASE ( 'pt_xy', 'pt_xz', 'pt_yz' )
	632	IF ( av == 0 ) THEN
	633	IF ( .NOT. cloud_physics ) THEN
	634	to_be_resorted => pt
	635	ELSE
[667]	636	DO i = nxlg, nxrg
	637	DO j = nysg, nyng
[1]	638	DO k = nzb, nzt+1
[1320]	639	local_pf(i,j,k) = pt(k,j,i) + l_d_cp * &
	640	pt_d_t(k) * &
[1]	641	ql(k,j,i)
	642	ENDDO
	643	ENDDO
	644	ENDDO
	645	resorted = .TRUE.
	646	ENDIF
	647	ELSE
	648	to_be_resorted => pt_av
	649	ENDIF
	650	IF ( mode == 'xy' ) level_z = zu
	651
	652	CASE ( 'q_xy', 'q_xz', 'q_yz' )
	653	IF ( av == 0 ) THEN
	654	to_be_resorted => q
	655	ELSE
	656	to_be_resorted => q_av
	657	ENDIF
	658	IF ( mode == 'xy' ) level_z = zu
	659
[1053]	660	CASE ( 'qc_xy', 'qc_xz', 'qc_yz' )
[1]	661	IF ( av == 0 ) THEN
[1115]	662	to_be_resorted => qc
[1]	663	ELSE
[1115]	664	to_be_resorted => qc_av
[1]	665	ENDIF
	666	IF ( mode == 'xy' ) level_z = zu
	667
[1053]	668	CASE ( 'ql_xy', 'ql_xz', 'ql_yz' )
	669	IF ( av == 0 ) THEN
[1115]	670	to_be_resorted => ql
[1053]	671	ELSE
[1115]	672	to_be_resorted => ql_av
[1053]	673	ENDIF
	674	IF ( mode == 'xy' ) level_z = zu
	675
[1]	676	CASE ( 'ql_c_xy', 'ql_c_xz', 'ql_c_yz' )
	677	IF ( av == 0 ) THEN
	678	to_be_resorted => ql_c
	679	ELSE
	680	to_be_resorted => ql_c_av
	681	ENDIF
	682	IF ( mode == 'xy' ) level_z = zu
	683
	684	CASE ( 'ql_v_xy', 'ql_v_xz', 'ql_v_yz' )
	685	IF ( av == 0 ) THEN
	686	to_be_resorted => ql_v
	687	ELSE
	688	to_be_resorted => ql_v_av
	689	ENDIF
	690	IF ( mode == 'xy' ) level_z = zu
	691
	692	CASE ( 'ql_vp_xy', 'ql_vp_xz', 'ql_vp_yz' )
	693	IF ( av == 0 ) THEN
[1007]	694	IF ( simulated_time >= particle_advection_start ) THEN
	695	DO i = nxl, nxr
	696	DO j = nys, nyn
	697	DO k = nzb, nzt+1
[1359]	698	number_of_particles = prt_count(k,j,i)
	699	IF (number_of_particles <= 0) CYCLE
	700	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
	701	DO n = 1, number_of_particles
	702	IF ( particles(n)%particle_mask ) THEN
	703	tend(k,j,i) = tend(k,j,i) + &
	704	particles(n)%weight_factor / &
	705	prt_count(k,j,i)
	706	ENDIF
[1007]	707	ENDDO
	708	ENDDO
	709	ENDDO
	710	ENDDO
	711	CALL exchange_horiz( tend, nbgp )
	712	ELSE
[1353]	713	tend = 0.0_wp
[1359]	714	ENDIF
[1007]	715	DO i = nxlg, nxrg
	716	DO j = nysg, nyng
	717	DO k = nzb, nzt+1
	718	local_pf(i,j,k) = tend(k,j,i)
	719	ENDDO
	720	ENDDO
	721	ENDDO
	722	resorted = .TRUE.
	723	ELSE
	724	CALL exchange_horiz( ql_vp_av, nbgp )
[1]	725	to_be_resorted => ql_vp
	726	ENDIF
	727	IF ( mode == 'xy' ) level_z = zu
	728
[1053]	729	CASE ( 'qr_xy', 'qr_xz', 'qr_yz' )
	730	IF ( av == 0 ) THEN
	731	to_be_resorted => qr
	732	ELSE
	733	to_be_resorted => qr_av
	734	ENDIF
	735	IF ( mode == 'xy' ) level_z = zu
	736
[354]	737	CASE ( 'qsws*_xy' ) ! 2d-array
	738	IF ( av == 0 ) THEN
[667]	739	DO i = nxlg, nxrg
	740	DO j = nysg, nyng
[354]	741	local_pf(i,j,nzb+1) = qsws(j,i)
	742	ENDDO
	743	ENDDO
	744	ELSE
[667]	745	DO i = nxlg, nxrg
	746	DO j = nysg, nyng
[354]	747	local_pf(i,j,nzb+1) = qsws_av(j,i)
	748	ENDDO
	749	ENDDO
	750	ENDIF
	751	resorted = .TRUE.
	752	two_d = .TRUE.
	753	level_z(nzb+1) = zu(nzb+1)
	754
[1]	755	CASE ( 'qv_xy', 'qv_xz', 'qv_yz' )
	756	IF ( av == 0 ) THEN
[667]	757	DO i = nxlg, nxrg
	758	DO j = nysg, nyng
[1]	759	DO k = nzb, nzt+1
	760	local_pf(i,j,k) = q(k,j,i) - ql(k,j,i)
	761	ENDDO
	762	ENDDO
	763	ENDDO
	764	resorted = .TRUE.
	765	ELSE
	766	to_be_resorted => qv_av
	767	ENDIF
	768	IF ( mode == 'xy' ) level_z = zu
	769
[1551]	770
[1585]	771
[96]	772	CASE ( 'rho_xy', 'rho_xz', 'rho_yz' )
	773	IF ( av == 0 ) THEN
[2031]	774	to_be_resorted => rho_ocean
[96]	775	ELSE
[2031]	776	to_be_resorted => rho_ocean_av
[96]	777	ENDIF
	778
[1]	779	CASE ( 's_xy', 's_xz', 's_yz' )
	780	IF ( av == 0 ) THEN
[1960]	781	to_be_resorted => s
[1]	782	ELSE
[355]	783	to_be_resorted => s_av
[1]	784	ENDIF
	785
[96]	786	CASE ( 'sa_xy', 'sa_xz', 'sa_yz' )
	787	IF ( av == 0 ) THEN
	788	to_be_resorted => sa
	789	ELSE
	790	to_be_resorted => sa_av
	791	ENDIF
	792
[354]	793	CASE ( 'shf*_xy' ) ! 2d-array
	794	IF ( av == 0 ) THEN
[667]	795	DO i = nxlg, nxrg
	796	DO j = nysg, nyng
[354]	797	local_pf(i,j,nzb+1) = shf(j,i)
	798	ENDDO
	799	ENDDO
	800	ELSE
[667]	801	DO i = nxlg, nxrg
	802	DO j = nysg, nyng
[354]	803	local_pf(i,j,nzb+1) = shf_av(j,i)
	804	ENDDO
	805	ENDDO
	806	ENDIF
	807	resorted = .TRUE.
	808	two_d = .TRUE.
	809	level_z(nzb+1) = zu(nzb+1)
[1960]	810
	811	CASE ( 'ssws*_xy' ) ! 2d-array
	812	IF ( av == 0 ) THEN
	813	DO i = nxlg, nxrg
	814	DO j = nysg, nyng
	815	local_pf(i,j,nzb+1) = ssws(j,i)
	816	ENDDO
	817	ENDDO
	818	ELSE
	819	DO i = nxlg, nxrg
	820	DO j = nysg, nyng
	821	local_pf(i,j,nzb+1) = ssws_av(j,i)
	822	ENDDO
	823	ENDDO
	824	ENDIF
	825	resorted = .TRUE.
	826	two_d = .TRUE.
	827	level_z(nzb+1) = zu(nzb+1)
[1551]	828
[1]	829	CASE ( 't*_xy' ) ! 2d-array
	830	IF ( av == 0 ) THEN
[667]	831	DO i = nxlg, nxrg
	832	DO j = nysg, nyng
[1]	833	local_pf(i,j,nzb+1) = ts(j,i)
	834	ENDDO
	835	ENDDO
	836	ELSE
[667]	837	DO i = nxlg, nxrg
	838	DO j = nysg, nyng
[1]	839	local_pf(i,j,nzb+1) = ts_av(j,i)
	840	ENDDO
	841	ENDDO
	842	ENDIF
	843	resorted = .TRUE.
	844	two_d = .TRUE.
	845	level_z(nzb+1) = zu(nzb+1)
	846
	847	CASE ( 'u_xy', 'u_xz', 'u_yz' )
	848	IF ( av == 0 ) THEN
	849	to_be_resorted => u
	850	ELSE
	851	to_be_resorted => u_av
	852	ENDIF
	853	IF ( mode == 'xy' ) level_z = zu
	854	!
	855	!-- Substitute the values generated by "mirror" boundary condition
	856	!-- at the bottom boundary by the real surface values.
	857	IF ( do2d(av,if) == 'u_xz' .OR. do2d(av,if) == 'u_yz' ) THEN
[1353]	858	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
[1]	859	ENDIF
	860
	861	CASE ( 'u*_xy' ) ! 2d-array
	862	IF ( av == 0 ) THEN
[667]	863	DO i = nxlg, nxrg
	864	DO j = nysg, nyng
[1]	865	local_pf(i,j,nzb+1) = us(j,i)
	866	ENDDO
	867	ENDDO
	868	ELSE
[667]	869	DO i = nxlg, nxrg
	870	DO j = nysg, nyng
[1]	871	local_pf(i,j,nzb+1) = us_av(j,i)
	872	ENDDO
	873	ENDDO
	874	ENDIF
	875	resorted = .TRUE.
	876	two_d = .TRUE.
	877	level_z(nzb+1) = zu(nzb+1)
	878
	879	CASE ( 'v_xy', 'v_xz', 'v_yz' )
	880	IF ( av == 0 ) THEN
	881	to_be_resorted => v
	882	ELSE
	883	to_be_resorted => v_av
	884	ENDIF
	885	IF ( mode == 'xy' ) level_z = zu
	886	!
	887	!-- Substitute the values generated by "mirror" boundary condition
	888	!-- at the bottom boundary by the real surface values.
	889	IF ( do2d(av,if) == 'v_xz' .OR. do2d(av,if) == 'v_yz' ) THEN
[1353]	890	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
[1]	891	ENDIF
	892
	893	CASE ( 'vpt_xy', 'vpt_xz', 'vpt_yz' )
	894	IF ( av == 0 ) THEN
	895	to_be_resorted => vpt
	896	ELSE
	897	to_be_resorted => vpt_av
	898	ENDIF
	899	IF ( mode == 'xy' ) level_z = zu
	900
	901	CASE ( 'w_xy', 'w_xz', 'w_yz' )
	902	IF ( av == 0 ) THEN
	903	to_be_resorted => w
	904	ELSE
	905	to_be_resorted => w_av
	906	ENDIF
	907	IF ( mode == 'xy' ) level_z = zw
	908
[72]	909	CASE ( 'z0*_xy' ) ! 2d-array
	910	IF ( av == 0 ) THEN
[667]	911	DO i = nxlg, nxrg
	912	DO j = nysg, nyng
[72]	913	local_pf(i,j,nzb+1) = z0(j,i)
	914	ENDDO
	915	ENDDO
	916	ELSE
[667]	917	DO i = nxlg, nxrg
	918	DO j = nysg, nyng
[72]	919	local_pf(i,j,nzb+1) = z0_av(j,i)
	920	ENDDO
	921	ENDDO
	922	ENDIF
	923	resorted = .TRUE.
	924	two_d = .TRUE.
	925	level_z(nzb+1) = zu(nzb+1)
	926
[978]	927	CASE ( 'z0h*_xy' ) ! 2d-array
	928	IF ( av == 0 ) THEN
	929	DO i = nxlg, nxrg
	930	DO j = nysg, nyng
	931	local_pf(i,j,nzb+1) = z0h(j,i)
	932	ENDDO
	933	ENDDO
	934	ELSE
	935	DO i = nxlg, nxrg
	936	DO j = nysg, nyng
	937	local_pf(i,j,nzb+1) = z0h_av(j,i)
	938	ENDDO
	939	ENDDO
	940	ENDIF
	941	resorted = .TRUE.
	942	two_d = .TRUE.
	943	level_z(nzb+1) = zu(nzb+1)
	944
[1788]	945	CASE ( 'z0q*_xy' ) ! 2d-array
	946	IF ( av == 0 ) THEN
	947	DO i = nxlg, nxrg
	948	DO j = nysg, nyng
	949	local_pf(i,j,nzb+1) = z0q(j,i)
	950	ENDDO
	951	ENDDO
	952	ELSE
	953	DO i = nxlg, nxrg
	954	DO j = nysg, nyng
	955	local_pf(i,j,nzb+1) = z0q_av(j,i)
	956	ENDDO
	957	ENDDO
	958	ENDIF
	959	resorted = .TRUE.
	960	two_d = .TRUE.
	961	level_z(nzb+1) = zu(nzb+1)
	962
[1]	963	CASE DEFAULT
[1972]	964
[1]	965	!
[1972]	966	!-- Land surface model quantity
	967	IF ( land_surface ) THEN
	968	CALL lsm_data_output_2d( av, do2d(av,if), found, grid, mode,&
	969	local_pf, two_d, nzb_do, nzt_do )
	970	ENDIF
	971
	972	!
[1976]	973	!-- Radiation quantity
	974	IF ( .NOT. found .AND. radiation ) THEN
	975	CALL radiation_data_output_2d( av, do2d(av,if), found, grid,&
	976	mode, local_pf, two_d )
	977	ENDIF
	978
	979	!
[1]	980	!-- User defined quantity
[1972]	981	IF ( .NOT. found ) THEN
	982	CALL user_data_output_2d( av, do2d(av,if), found, grid, &
	983	local_pf, two_d, nzb_do, nzt_do )
	984	ENDIF
	985
[1]	986	resorted = .TRUE.
	987
	988	IF ( grid == 'zu' ) THEN
	989	IF ( mode == 'xy' ) level_z = zu
	990	ELSEIF ( grid == 'zw' ) THEN
	991	IF ( mode == 'xy' ) level_z = zw
[343]	992	ELSEIF ( grid == 'zu1' ) THEN
	993	IF ( mode == 'xy' ) level_z(nzb+1) = zu(nzb+1)
[1551]	994	ELSEIF ( grid == 'zs' ) THEN
	995	IF ( mode == 'xy' ) level_z = zs
[1]	996	ENDIF
	997
	998	IF ( .NOT. found ) THEN
[1320]	999	message_string = 'no output provided for: ' // &
[274]	1000	TRIM( do2d(av,if) )
[254]	1001	CALL message( 'data_output_2d', 'PA0181', 0, 0, 0, 6, 0 )
[1]	1002	ENDIF
	1003
	1004	END SELECT
	1005
	1006	!
	1007	!-- Resort the array to be output, if not done above
	1008	IF ( .NOT. resorted ) THEN
[667]	1009	DO i = nxlg, nxrg
	1010	DO j = nysg, nyng
[1551]	1011	DO k = nzb_do, nzt_do
[1]	1012	local_pf(i,j,k) = to_be_resorted(k,j,i)
	1013	ENDDO
	1014	ENDDO
	1015	ENDDO
	1016	ENDIF
	1017
	1018	!
	1019	!-- Output of the individual cross-sections, depending on the cross-
	1020	!-- section mode chosen.
	1021	is = 1
[1960]	1022	loop1: DO WHILE ( section(is,s_ind) /= -9999 .OR. two_d )
[1]	1023
	1024	SELECT CASE ( mode )
	1025
	1026	CASE ( 'xy' )
	1027	!
	1028	!-- Determine the cross section index
	1029	IF ( two_d ) THEN
	1030	layer_xy = nzb+1
	1031	ELSE
[1960]	1032	layer_xy = section(is,s_ind)
[1]	1033	ENDIF
	1034
	1035	!
[1551]	1036	!-- Exit the loop for layers beyond the data output domain
	1037	!-- (used for soil model)
[1691]	1038	IF ( layer_xy > nzt_do ) THEN
[1551]	1039	EXIT loop1
	1040	ENDIF
	1041
	1042	!
[1308]	1043	!-- Update the netCDF xy cross section time axis.
	1044	!-- In case of parallel output, this is only done by PE0
	1045	!-- to increase the performance.
	1046	IF ( simulated_time /= do2d_xy_last_time(av) ) THEN
	1047	do2d_xy_time_count(av) = do2d_xy_time_count(av) + 1
	1048	do2d_xy_last_time(av) = simulated_time
	1049	IF ( myid == 0 ) THEN
[1327]	1050	IF ( .NOT. data_output_2d_on_each_pe &
	1051	.OR. netcdf_data_format > 4 ) &
[493]	1052	THEN
[1]	1053	#if defined( __netcdf )
	1054	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1055	id_var_time_xy(av), &
[291]	1056	(/ time_since_reference_point /), &
[1]	1057	start = (/ do2d_xy_time_count(av) /), &
	1058	count = (/ 1 /) )
[1783]	1059	CALL netcdf_handle_error( 'data_output_2d', 53 )
[1]	1060	#endif
	1061	ENDIF
	1062	ENDIF
	1063	ENDIF
	1064	!
	1065	!-- If required, carry out averaging along z
[1960]	1066	IF ( section(is,s_ind) == -1 .AND. .NOT. two_d ) THEN
[1]	1067
[1353]	1068	local_2d = 0.0_wp
[1]	1069	!
	1070	!-- Carry out the averaging (all data are on the PE)
[1551]	1071	DO k = nzb_do, nzt_do
[667]	1072	DO j = nysg, nyng
	1073	DO i = nxlg, nxrg
[1]	1074	local_2d(i,j) = local_2d(i,j) + local_pf(i,j,k)
	1075	ENDDO
	1076	ENDDO
	1077	ENDDO
	1078
[1551]	1079	local_2d = local_2d / ( nzt_do - nzb_do + 1.0_wp)
[1]	1080
	1081	ELSE
	1082	!
	1083	!-- Just store the respective section on the local array
	1084	local_2d = local_pf(:,:,layer_xy)
	1085
	1086	ENDIF
	1087
	1088	#if defined( __parallel )
[1327]	1089	IF ( netcdf_data_format > 4 ) THEN
[1]	1090	!
[1031]	1091	!-- Parallel output in netCDF4/HDF5 format.
[493]	1092	IF ( two_d ) THEN
	1093	iis = 1
	1094	ELSE
	1095	iis = is
	1096	ENDIF
	1097
[1]	1098	#if defined( __netcdf )
[1308]	1099	!
	1100	!-- For parallel output, all cross sections are first stored
	1101	!-- here on a local array and will be written to the output
	1102	!-- file afterwards to increase the performance.
	1103	DO i = nxlg, nxrg
	1104	DO j = nysg, nyng
	1105	local_2d_sections(i,j,iis) = local_2d(i,j)
	1106	ENDDO
	1107	ENDDO
[1]	1108	#endif
[493]	1109	ELSE
[1]	1110
[493]	1111	IF ( data_output_2d_on_each_pe ) THEN
[1]	1112	!
[493]	1113	!-- Output of partial arrays on each PE
	1114	#if defined( __netcdf )
[1327]	1115	IF ( myid == 0 ) THEN
[1320]	1116	WRITE ( 21 ) time_since_reference_point, &
[493]	1117	do2d_xy_time_count(av), av
	1118	ENDIF
	1119	#endif
[759]	1120	DO i = 0, io_blocks-1
	1121	IF ( i == io_group ) THEN
[1551]	1122	WRITE ( 21 ) nxlg, nxrg, nysg, nyng, nysg, nyng
[759]	1123	WRITE ( 21 ) local_2d
	1124	ENDIF
	1125	#if defined( __parallel )
	1126	CALL MPI_BARRIER( comm2d, ierr )
	1127	#endif
	1128	ENDDO
[559]	1129
[493]	1130	ELSE
[1]	1131	!
[493]	1132	!-- PE0 receives partial arrays from all processors and
	1133	!-- then outputs them. Here a barrier has to be set,
	1134	!-- because otherwise "-MPI- FATAL: Remote protocol queue
	1135	!-- full" may occur.
	1136	CALL MPI_BARRIER( comm2d, ierr )
	1137
[667]	1138	ngp = ( nxrg-nxlg+1 ) * ( nyng-nysg+1 )
[493]	1139	IF ( myid == 0 ) THEN
[1]	1140	!
[493]	1141	!-- Local array can be relocated directly.
[667]	1142	total_2d(nxlg:nxrg,nysg:nyng) = local_2d
[1]	1143	!
[493]	1144	!-- Receive data from all other PEs.
	1145	DO n = 1, numprocs-1
[1]	1146	!
[493]	1147	!-- Receive index limits first, then array.
	1148	!-- Index limits are received in arbitrary order from
	1149	!-- the PEs.
[1320]	1150	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1151	MPI_ANY_SOURCE, 0, comm2d, &
[493]	1152	status, ierr )
	1153	sender = status(MPI_SOURCE)
	1154	DEALLOCATE( local_2d )
	1155	ALLOCATE( local_2d(ind(1):ind(2),ind(3):ind(4)) )
[1320]	1156	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1157	MPI_REAL, sender, 1, comm2d, &
[493]	1158	status, ierr )
	1159	total_2d(ind(1):ind(2),ind(3):ind(4)) = local_2d
	1160	ENDDO
[1]	1161	!
[493]	1162	!-- Relocate the local array for the next loop increment
	1163	DEALLOCATE( local_2d )
[667]	1164	ALLOCATE( local_2d(nxlg:nxrg,nysg:nyng) )
[1]	1165
	1166	#if defined( __netcdf )
[1327]	1167	IF ( two_d ) THEN
	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, 1, do2d_xy_time_count(av) /), &
	1172	count = (/ nx+2, ny+2, 1, 1 /) )
	1173	ELSE
	1174	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1175	id_var_do2d(av,if), &
	1176	total_2d(0:nx+1,0:ny+1), &
	1177	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
	1178	count = (/ nx+2, ny+2, 1, 1 /) )
[1]	1179	ENDIF
[1783]	1180	CALL netcdf_handle_error( 'data_output_2d', 54 )
[1]	1181	#endif
	1182
[493]	1183	ELSE
[1]	1184	!
[493]	1185	!-- First send the local index limits to PE0
[667]	1186	ind(1) = nxlg; ind(2) = nxrg
	1187	ind(3) = nysg; ind(4) = nyng
[1320]	1188	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1189	comm2d, ierr )
[1]	1190	!
[493]	1191	!-- Send data to PE0
[1320]	1192	CALL MPI_SEND( local_2d(nxlg,nysg), ngp, &
[493]	1193	MPI_REAL, 0, 1, comm2d, ierr )
	1194	ENDIF
	1195	!
	1196	!-- A barrier has to be set, because otherwise some PEs may
	1197	!-- proceed too fast so that PE0 may receive wrong data on
	1198	!-- tag 0
	1199	CALL MPI_BARRIER( comm2d, ierr )
[1]	1200	ENDIF
[493]	1201
[1]	1202	ENDIF
	1203	#else
	1204	#if defined( __netcdf )
[1327]	1205	IF ( two_d ) THEN
	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, 1, do2d_xy_time_count(av) /), &
	1210	count = (/ nx+2, ny+2, 1, 1 /) )
	1211	ELSE
	1212	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1213	id_var_do2d(av,if), &
	1214	local_2d(nxl:nxr+1,nys:nyn+1), &
	1215	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
	1216	count = (/ nx+2, ny+2, 1, 1 /) )
[1]	1217	ENDIF
[1783]	1218	CALL netcdf_handle_error( 'data_output_2d', 447 )
[1]	1219	#endif
	1220	#endif
	1221	do2d_xy_n = do2d_xy_n + 1
	1222	!
	1223	!-- For 2D-arrays (e.g. u*) only one cross-section is available.
	1224	!-- Hence exit loop of output levels.
	1225	IF ( two_d ) THEN
[1703]	1226	IF ( netcdf_data_format < 5 ) two_d = .FALSE.
[1]	1227	EXIT loop1
	1228	ENDIF
	1229
	1230	CASE ( 'xz' )
	1231	!
[1308]	1232	!-- Update the netCDF xz cross section time axis.
	1233	!-- In case of parallel output, this is only done by PE0
	1234	!-- to increase the performance.
	1235	IF ( simulated_time /= do2d_xz_last_time(av) ) THEN
	1236	do2d_xz_time_count(av) = do2d_xz_time_count(av) + 1
	1237	do2d_xz_last_time(av) = simulated_time
	1238	IF ( myid == 0 ) THEN
[1327]	1239	IF ( .NOT. data_output_2d_on_each_pe &
	1240	.OR. netcdf_data_format > 4 ) &
[493]	1241	THEN
[1]	1242	#if defined( __netcdf )
	1243	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1244	id_var_time_xz(av), &
[291]	1245	(/ time_since_reference_point /), &
[1]	1246	start = (/ do2d_xz_time_count(av) /), &
	1247	count = (/ 1 /) )
[1783]	1248	CALL netcdf_handle_error( 'data_output_2d', 56 )
[1]	1249	#endif
	1250	ENDIF
	1251	ENDIF
	1252	ENDIF
[667]	1253
[1]	1254	!
	1255	!-- If required, carry out averaging along y
[1960]	1256	IF ( section(is,s_ind) == -1 ) THEN
[1]	1257
[1551]	1258	ALLOCATE( local_2d_l(nxlg:nxrg,nzb_do:nzt_do) )
[1353]	1259	local_2d_l = 0.0_wp
[1551]	1260	ngp = ( nxrg-nxlg + 1 ) * ( nzt_do-nzb_do + 1 )
[1]	1261	!
	1262	!-- First local averaging on the PE
[1551]	1263	DO k = nzb_do, nzt_do
[1]	1264	DO j = nys, nyn
[667]	1265	DO i = nxlg, nxrg
[1320]	1266	local_2d_l(i,k) = local_2d_l(i,k) + &
[1]	1267	local_pf(i,j,k)
	1268	ENDDO
	1269	ENDDO
	1270	ENDDO
	1271	#if defined( __parallel )
	1272	!
	1273	!-- Now do the averaging over all PEs along y
[622]	1274	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1275	CALL MPI_ALLREDUCE( local_2d_l(nxlg,nzb_do), &
	1276	local_2d(nxlg,nzb_do), ngp, MPI_REAL, &
[1]	1277	MPI_SUM, comm1dy, ierr )
	1278	#else
	1279	local_2d = local_2d_l
	1280	#endif
[1353]	1281	local_2d = local_2d / ( ny + 1.0_wp )
[1]	1282
	1283	DEALLOCATE( local_2d_l )
	1284
	1285	ELSE
	1286	!
	1287	!-- Just store the respective section on the local array
	1288	!-- (but only if it is available on this PE!)
[1960]	1289	IF ( section(is,s_ind) >= nys .AND. section(is,s_ind) <= nyn ) &
[1]	1290	THEN
[1960]	1291	local_2d = local_pf(:,section(is,s_ind),nzb_do:nzt_do)
[1]	1292	ENDIF
	1293
	1294	ENDIF
	1295
	1296	#if defined( __parallel )
[1327]	1297	IF ( netcdf_data_format > 4 ) THEN
[1]	1298	!
[1031]	1299	!-- Output in netCDF4/HDF5 format.
[493]	1300	!-- Output only on those PEs where the respective cross
	1301	!-- sections reside. Cross sections averaged along y are
	1302	!-- output on the respective first PE along y (myidy=0).
[1960]	1303	IF ( ( section(is,s_ind) >= nys .AND. &
	1304	section(is,s_ind) <= nyn ) .OR. &
	1305	( section(is,s_ind) == -1 .AND. myidy == 0 ) ) THEN
[1]	1306	#if defined( __netcdf )
[493]	1307	!
[1308]	1308	!-- For parallel output, all cross sections are first
	1309	!-- stored here on a local array and will be written to the
	1310	!-- output file afterwards to increase the performance.
	1311	DO i = nxlg, nxrg
[1551]	1312	DO k = nzb_do, nzt_do
[1308]	1313	local_2d_sections_l(i,is,k) = local_2d(i,k)
	1314	ENDDO
	1315	ENDDO
[1]	1316	#endif
	1317	ENDIF
	1318
	1319	ELSE
	1320
[493]	1321	IF ( data_output_2d_on_each_pe ) THEN
[1]	1322	!
[493]	1323	!-- Output of partial arrays on each PE. If the cross
	1324	!-- section does not reside on the PE, output special
	1325	!-- index values.
	1326	#if defined( __netcdf )
[1327]	1327	IF ( myid == 0 ) THEN
[1320]	1328	WRITE ( 22 ) time_since_reference_point, &
[493]	1329	do2d_xz_time_count(av), av
	1330	ENDIF
	1331	#endif
[759]	1332	DO i = 0, io_blocks-1
	1333	IF ( i == io_group ) THEN
[1960]	1334	IF ( ( section(is,s_ind) >= nys .AND. &
	1335	section(is,s_ind) <= nyn ) .OR. &
	1336	( section(is,s_ind) == -1 .AND. &
[1320]	1337	nys-1 == -1 ) ) &
[759]	1338	THEN
[1551]	1339	WRITE (22) nxlg, nxrg, nzb_do, nzt_do, nzb, nzt+1
[759]	1340	WRITE (22) local_2d
	1341	ELSE
[1551]	1342	WRITE (22) -1, -1, -1, -1, -1, -1
[759]	1343	ENDIF
	1344	ENDIF
	1345	#if defined( __parallel )
	1346	CALL MPI_BARRIER( comm2d, ierr )
	1347	#endif
	1348	ENDDO
[493]	1349
	1350	ELSE
[1]	1351	!
[493]	1352	!-- PE0 receives partial arrays from all processors of the
	1353	!-- respective cross section and outputs them. Here a
	1354	!-- barrier has to be set, because otherwise
	1355	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
	1356	CALL MPI_BARRIER( comm2d, ierr )
	1357
[1551]	1358	ngp = ( nxrg-nxlg + 1 ) * ( nzt_do-nzb_do + 1 )
[493]	1359	IF ( myid == 0 ) THEN
[1]	1360	!
[493]	1361	!-- Local array can be relocated directly.
[1960]	1362	IF ( ( section(is,s_ind) >= nys .AND. &
	1363	section(is,s_ind) <= nyn ) .OR. &
	1364	( section(is,s_ind) == -1 .AND. &
	1365	nys-1 == -1 ) ) THEN
[1551]	1366	total_2d(nxlg:nxrg,nzb_do:nzt_do) = local_2d
[493]	1367	ENDIF
[1]	1368	!
[493]	1369	!-- Receive data from all other PEs.
	1370	DO n = 1, numprocs-1
	1371	!
	1372	!-- Receive index limits first, then array.
	1373	!-- Index limits are received in arbitrary order from
	1374	!-- the PEs.
[1320]	1375	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1376	MPI_ANY_SOURCE, 0, comm2d, &
[1]	1377	status, ierr )
[493]	1378	!
	1379	!-- Not all PEs have data for XZ-cross-section.
	1380	IF ( ind(1) /= -9999 ) THEN
	1381	sender = status(MPI_SOURCE)
	1382	DEALLOCATE( local_2d )
[1320]	1383	ALLOCATE( local_2d(ind(1):ind(2), &
[493]	1384	ind(3):ind(4)) )
	1385	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1386	MPI_REAL, sender, 1, comm2d, &
	1387	status, ierr )
[1320]	1388	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
[493]	1389	local_2d
	1390	ENDIF
	1391	ENDDO
	1392	!
	1393	!-- Relocate the local array for the next loop increment
	1394	DEALLOCATE( local_2d )
[1551]	1395	ALLOCATE( local_2d(nxlg:nxrg,nzb_do:nzt_do) )
[1]	1396
	1397	#if defined( __netcdf )
[1327]	1398	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1399	id_var_do2d(av,if), &
[1551]	1400	total_2d(0:nx+1,nzb_do:nzt_do),&
[1327]	1401	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
[1551]	1402	count = (/ nx+2, 1, nzt_do-nzb_do+1, 1 /) )
[1783]	1403	CALL netcdf_handle_error( 'data_output_2d', 58 )
[1]	1404	#endif
	1405
[493]	1406	ELSE
[1]	1407	!
[493]	1408	!-- If the cross section resides on the PE, send the
	1409	!-- local index limits, otherwise send -9999 to PE0.
[1960]	1410	IF ( ( section(is,s_ind) >= nys .AND. &
	1411	section(is,s_ind) <= nyn ) .OR. &
	1412	( section(is,s_ind) == -1 .AND. nys-1 == -1 ) ) &
[493]	1413	THEN
[667]	1414	ind(1) = nxlg; ind(2) = nxrg
[1551]	1415	ind(3) = nzb_do; ind(4) = nzt_do
[493]	1416	ELSE
	1417	ind(1) = -9999; ind(2) = -9999
	1418	ind(3) = -9999; ind(4) = -9999
	1419	ENDIF
[1320]	1420	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1421	comm2d, ierr )
	1422	!
	1423	!-- If applicable, send data to PE0.
	1424	IF ( ind(1) /= -9999 ) THEN
[1551]	1425	CALL MPI_SEND( local_2d(nxlg,nzb_do), ngp, &
[493]	1426	MPI_REAL, 0, 1, comm2d, ierr )
	1427	ENDIF
[1]	1428	ENDIF
	1429	!
[493]	1430	!-- A barrier has to be set, because otherwise some PEs may
	1431	!-- proceed too fast so that PE0 may receive wrong data on
	1432	!-- tag 0
	1433	CALL MPI_BARRIER( comm2d, ierr )
[1]	1434	ENDIF
[493]	1435
[1]	1436	ENDIF
	1437	#else
	1438	#if defined( __netcdf )
[1327]	1439	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1440	id_var_do2d(av,if), &
[1551]	1441	local_2d(nxl:nxr+1,nzb_do:nzt_do), &
[1327]	1442	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
[1551]	1443	count = (/ nx+2, 1, nzt_do-nzb_do+1, 1 /) )
[1783]	1444	CALL netcdf_handle_error( 'data_output_2d', 451 )
[1]	1445	#endif
	1446	#endif
	1447	do2d_xz_n = do2d_xz_n + 1
	1448
	1449	CASE ( 'yz' )
	1450	!
[1308]	1451	!-- Update the netCDF yz cross section time axis.
	1452	!-- In case of parallel output, this is only done by PE0
	1453	!-- to increase the performance.
	1454	IF ( simulated_time /= do2d_yz_last_time(av) ) THEN
	1455	do2d_yz_time_count(av) = do2d_yz_time_count(av) + 1
	1456	do2d_yz_last_time(av) = simulated_time
	1457	IF ( myid == 0 ) THEN
[1327]	1458	IF ( .NOT. data_output_2d_on_each_pe &
	1459	.OR. netcdf_data_format > 4 ) &
[493]	1460	THEN
[1]	1461	#if defined( __netcdf )
	1462	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1463	id_var_time_yz(av), &
[291]	1464	(/ time_since_reference_point /), &
[1]	1465	start = (/ do2d_yz_time_count(av) /), &
	1466	count = (/ 1 /) )
[1783]	1467	CALL netcdf_handle_error( 'data_output_2d', 59 )
[1]	1468	#endif
	1469	ENDIF
	1470	ENDIF
[1308]	1471	ENDIF
[493]	1472
[1]	1473	!
	1474	!-- If required, carry out averaging along x
[1960]	1475	IF ( section(is,s_ind) == -1 ) THEN
[1]	1476
[1551]	1477	ALLOCATE( local_2d_l(nysg:nyng,nzb_do:nzt_do) )
[1353]	1478	local_2d_l = 0.0_wp
[1551]	1479	ngp = ( nyng-nysg+1 ) * ( nzt_do-nzb_do+1 )
[1]	1480	!
	1481	!-- First local averaging on the PE
[1551]	1482	DO k = nzb_do, nzt_do
[667]	1483	DO j = nysg, nyng
[1]	1484	DO i = nxl, nxr
[1320]	1485	local_2d_l(j,k) = local_2d_l(j,k) + &
[1]	1486	local_pf(i,j,k)
	1487	ENDDO
	1488	ENDDO
	1489	ENDDO
	1490	#if defined( __parallel )
	1491	!
	1492	!-- Now do the averaging over all PEs along x
[622]	1493	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1494	CALL MPI_ALLREDUCE( local_2d_l(nysg,nzb_do), &
	1495	local_2d(nysg,nzb_do), ngp, MPI_REAL, &
[1]	1496	MPI_SUM, comm1dx, ierr )
	1497	#else
	1498	local_2d = local_2d_l
	1499	#endif
[1353]	1500	local_2d = local_2d / ( nx + 1.0_wp )
[1]	1501
	1502	DEALLOCATE( local_2d_l )
	1503
	1504	ELSE
	1505	!
	1506	!-- Just store the respective section on the local array
	1507	!-- (but only if it is available on this PE!)
[1960]	1508	IF ( section(is,s_ind) >= nxl .AND. section(is,s_ind) <= nxr ) &
[1]	1509	THEN
[1960]	1510	local_2d = local_pf(section(is,s_ind),:,nzb_do:nzt_do)
[1]	1511	ENDIF
	1512
	1513	ENDIF
	1514
	1515	#if defined( __parallel )
[1327]	1516	IF ( netcdf_data_format > 4 ) THEN
[1]	1517	!
[1031]	1518	!-- Output in netCDF4/HDF5 format.
[493]	1519	!-- Output only on those PEs where the respective cross
	1520	!-- sections reside. Cross sections averaged along x are
	1521	!-- output on the respective first PE along x (myidx=0).
[1960]	1522	IF ( ( section(is,s_ind) >= nxl .AND. &
	1523	section(is,s_ind) <= nxr ) .OR. &
	1524	( section(is,s_ind) == -1 .AND. myidx == 0 ) ) THEN
[1]	1525	#if defined( __netcdf )
[493]	1526	!
[1308]	1527	!-- For parallel output, all cross sections are first
	1528	!-- stored here on a local array and will be written to the
	1529	!-- output file afterwards to increase the performance.
	1530	DO j = nysg, nyng
[1551]	1531	DO k = nzb_do, nzt_do
[1308]	1532	local_2d_sections_l(is,j,k) = local_2d(j,k)
	1533	ENDDO
	1534	ENDDO
[1]	1535	#endif
	1536	ENDIF
	1537
	1538	ELSE
	1539
[493]	1540	IF ( data_output_2d_on_each_pe ) THEN
[1]	1541	!
[493]	1542	!-- Output of partial arrays on each PE. If the cross
	1543	!-- section does not reside on the PE, output special
	1544	!-- index values.
	1545	#if defined( __netcdf )
[1327]	1546	IF ( myid == 0 ) THEN
[1320]	1547	WRITE ( 23 ) time_since_reference_point, &
[493]	1548	do2d_yz_time_count(av), av
	1549	ENDIF
	1550	#endif
[759]	1551	DO i = 0, io_blocks-1
	1552	IF ( i == io_group ) THEN
[1960]	1553	IF ( ( section(is,s_ind) >= nxl .AND. &
	1554	section(is,s_ind) <= nxr ) .OR. &
	1555	( section(is,s_ind) == -1 .AND. &
[1320]	1556	nxl-1 == -1 ) ) &
[759]	1557	THEN
[1551]	1558	WRITE (23) nysg, nyng, nzb_do, nzt_do, nzb, nzt+1
[759]	1559	WRITE (23) local_2d
	1560	ELSE
[1551]	1561	WRITE (23) -1, -1, -1, -1, -1, -1
[759]	1562	ENDIF
	1563	ENDIF
	1564	#if defined( __parallel )
	1565	CALL MPI_BARRIER( comm2d, ierr )
	1566	#endif
	1567	ENDDO
[493]	1568
	1569	ELSE
[1]	1570	!
[493]	1571	!-- PE0 receives partial arrays from all processors of the
	1572	!-- respective cross section and outputs them. Here a
	1573	!-- barrier has to be set, because otherwise
	1574	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
	1575	CALL MPI_BARRIER( comm2d, ierr )
	1576
[1551]	1577	ngp = ( nyng-nysg+1 ) * ( nzt_do-nzb_do+1 )
[493]	1578	IF ( myid == 0 ) THEN
[1]	1579	!
[493]	1580	!-- Local array can be relocated directly.
[1960]	1581	IF ( ( section(is,s_ind) >= nxl .AND. &
	1582	section(is,s_ind) <= nxr ) .OR. &
	1583	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
[493]	1584	THEN
[1551]	1585	total_2d(nysg:nyng,nzb_do:nzt_do) = local_2d
[493]	1586	ENDIF
[1]	1587	!
[493]	1588	!-- Receive data from all other PEs.
	1589	DO n = 1, numprocs-1
	1590	!
	1591	!-- Receive index limits first, then array.
	1592	!-- Index limits are received in arbitrary order from
	1593	!-- the PEs.
[1320]	1594	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1595	MPI_ANY_SOURCE, 0, comm2d, &
[1]	1596	status, ierr )
[493]	1597	!
	1598	!-- Not all PEs have data for YZ-cross-section.
	1599	IF ( ind(1) /= -9999 ) THEN
	1600	sender = status(MPI_SOURCE)
	1601	DEALLOCATE( local_2d )
[1320]	1602	ALLOCATE( local_2d(ind(1):ind(2), &
[493]	1603	ind(3):ind(4)) )
	1604	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1605	MPI_REAL, sender, 1, comm2d, &
	1606	status, ierr )
[1320]	1607	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
[493]	1608	local_2d
	1609	ENDIF
	1610	ENDDO
	1611	!
	1612	!-- Relocate the local array for the next loop increment
	1613	DEALLOCATE( local_2d )
[1551]	1614	ALLOCATE( local_2d(nysg:nyng,nzb_do:nzt_do) )
[1]	1615
	1616	#if defined( __netcdf )
[1327]	1617	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1618	id_var_do2d(av,if), &
[1551]	1619	total_2d(0:ny+1,nzb_do:nzt_do),&
[1327]	1620	start = (/ is, 1, 1, do2d_yz_time_count(av) /), &
[1551]	1621	count = (/ 1, ny+2, nzt_do-nzb_do+1, 1 /) )
[1783]	1622	CALL netcdf_handle_error( 'data_output_2d', 61 )
[1]	1623	#endif
	1624
[493]	1625	ELSE
[1]	1626	!
[493]	1627	!-- If the cross section resides on the PE, send the
	1628	!-- local index limits, otherwise send -9999 to PE0.
[1960]	1629	IF ( ( section(is,s_ind) >= nxl .AND. &
	1630	section(is,s_ind) <= nxr ) .OR. &
	1631	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
[493]	1632	THEN
[667]	1633	ind(1) = nysg; ind(2) = nyng
[1551]	1634	ind(3) = nzb_do; ind(4) = nzt_do
[493]	1635	ELSE
	1636	ind(1) = -9999; ind(2) = -9999
	1637	ind(3) = -9999; ind(4) = -9999
	1638	ENDIF
[1320]	1639	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1640	comm2d, ierr )
	1641	!
	1642	!-- If applicable, send data to PE0.
	1643	IF ( ind(1) /= -9999 ) THEN
[1551]	1644	CALL MPI_SEND( local_2d(nysg,nzb_do), ngp, &
[493]	1645	MPI_REAL, 0, 1, comm2d, ierr )
	1646	ENDIF
[1]	1647	ENDIF
	1648	!
[493]	1649	!-- A barrier has to be set, because otherwise some PEs may
	1650	!-- proceed too fast so that PE0 may receive wrong data on
	1651	!-- tag 0
	1652	CALL MPI_BARRIER( comm2d, ierr )
[1]	1653	ENDIF
[493]	1654
[1]	1655	ENDIF
	1656	#else
	1657	#if defined( __netcdf )
[1327]	1658	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1659	id_var_do2d(av,if), &
[1551]	1660	local_2d(nys:nyn+1,nzb_do:nzt_do), &
[1327]	1661	start = (/ is, 1, 1, do2d_xz_time_count(av) /), &
[1551]	1662	count = (/ 1, ny+2, nzt_do-nzb_do+1, 1 /) )
[1783]	1663	CALL netcdf_handle_error( 'data_output_2d', 452 )
[1]	1664	#endif
	1665	#endif
	1666	do2d_yz_n = do2d_yz_n + 1
	1667
	1668	END SELECT
	1669
	1670	is = is + 1
	1671	ENDDO loop1
	1672
[1308]	1673	!
	1674	!-- For parallel output, all data were collected before on a local array
	1675	!-- and are written now to the netcdf file. This must be done to increase
	1676	!-- the performance of the parallel output.
	1677	#if defined( __netcdf )
[1327]	1678	IF ( netcdf_data_format > 4 ) THEN
[1308]	1679
	1680	SELECT CASE ( mode )
	1681
	1682	CASE ( 'xy' )
	1683	IF ( two_d ) THEN
[1703]	1684	nis = 1
	1685	two_d = .FALSE.
[1308]	1686	ELSE
[1703]	1687	nis = ns
[1308]	1688	ENDIF
	1689	!
	1690	!-- Do not output redundant ghost point data except for the
	1691	!-- boundaries of the total domain.
	1692	IF ( nxr == nx .AND. nyn /= ny ) THEN
	1693	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1694	id_var_do2d(av,if), &
	1695	local_2d_sections(nxl:nxr+1, &
[1703]	1696	nys:nyn,1:nis), &
[1308]	1697	start = (/ nxl+1, nys+1, 1, &
	1698	do2d_xy_time_count(av) /), &
	1699	count = (/ nxr-nxl+2, &
[1703]	1700	nyn-nys+1, nis, 1 &
[1308]	1701	/) )
	1702	ELSEIF ( nxr /= nx .AND. nyn == ny ) THEN
	1703	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1704	id_var_do2d(av,if), &
	1705	local_2d_sections(nxl:nxr, &
[1703]	1706	nys:nyn+1,1:nis), &
[1308]	1707	start = (/ nxl+1, nys+1, 1, &
	1708	do2d_xy_time_count(av) /), &
	1709	count = (/ nxr-nxl+1, &
[1703]	1710	nyn-nys+2, nis, 1 &
[1308]	1711	/) )
	1712	ELSEIF ( nxr == nx .AND. nyn == ny ) THEN
	1713	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1714	id_var_do2d(av,if), &
	1715	local_2d_sections(nxl:nxr+1, &
[1703]	1716	nys:nyn+1,1:nis), &
[1308]	1717	start = (/ nxl+1, nys+1, 1, &
	1718	do2d_xy_time_count(av) /), &
	1719	count = (/ nxr-nxl+2, &
[1703]	1720	nyn-nys+2, nis, 1 &
[1308]	1721	/) )
	1722	ELSE
	1723	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1724	id_var_do2d(av,if), &
	1725	local_2d_sections(nxl:nxr, &
[1703]	1726	nys:nyn,1:nis), &
[1308]	1727	start = (/ nxl+1, nys+1, 1, &
	1728	do2d_xy_time_count(av) /), &
	1729	count = (/ nxr-nxl+1, &
[1703]	1730	nyn-nys+1, nis, 1 &
[1308]	1731	/) )
	1732	ENDIF
	1733
[1783]	1734	CALL netcdf_handle_error( 'data_output_2d', 55 )
[1308]	1735
	1736	CASE ( 'xz' )
	1737	!
	1738	!-- First, all PEs get the information of all cross-sections.
	1739	!-- Then the data are written to the output file by all PEs
	1740	!-- while NF90_COLLECTIVE is set in subroutine
	1741	!-- define_netcdf_header. Although redundant information are
	1742	!-- written to the output file in that case, the performance
	1743	!-- is significantly better compared to the case where only
	1744	!-- the first row of PEs in x-direction (myidx = 0) is given
	1745	!-- the output while NF90_INDEPENDENT is set.
	1746	IF ( npey /= 1 ) THEN
	1747
	1748	#if defined( __parallel )
	1749	!
	1750	!-- Distribute data over all PEs along y
[1551]	1751	ngp = ( nxrg-nxlg+1 ) * ( nzt_do-nzb_do+1 ) * ns
[1308]	1752	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1753	CALL MPI_ALLREDUCE( local_2d_sections_l(nxlg,1,nzb_do), &
	1754	local_2d_sections(nxlg,1,nzb_do), &
[1308]	1755	ngp, MPI_REAL, MPI_SUM, comm1dy, &
	1756	ierr )
	1757	#else
	1758	local_2d_sections = local_2d_sections_l
	1759	#endif
	1760	ENDIF
	1761	!
	1762	!-- Do not output redundant ghost point data except for the
	1763	!-- boundaries of the total domain.
	1764	IF ( nxr == nx ) THEN
	1765	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1766	id_var_do2d(av,if), &
	1767	local_2d_sections(nxl:nxr+1,1:ns, &
[1551]	1768	nzb_do:nzt_do), &
[1308]	1769	start = (/ nxl+1, 1, 1, &
	1770	do2d_xz_time_count(av) /), &
[1551]	1771	count = (/ nxr-nxl+2, ns, nzt_do-nzb_do+1, &
[1308]	1772	1 /) )
	1773	ELSE
	1774	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1775	id_var_do2d(av,if), &
	1776	local_2d_sections(nxl:nxr,1:ns, &
[1551]	1777	nzb_do:nzt_do), &
[1308]	1778	start = (/ nxl+1, 1, 1, &
	1779	do2d_xz_time_count(av) /), &
[1551]	1780	count = (/ nxr-nxl+1, ns, nzt_do-nzb_do+1, &
[1308]	1781	1 /) )
	1782	ENDIF
	1783
[1783]	1784	CALL netcdf_handle_error( 'data_output_2d', 57 )
[1308]	1785
	1786	CASE ( 'yz' )
	1787	!
	1788	!-- First, all PEs get the information of all cross-sections.
	1789	!-- Then the data are written to the output file by all PEs
	1790	!-- while NF90_COLLECTIVE is set in subroutine
	1791	!-- define_netcdf_header. Although redundant information are
	1792	!-- written to the output file in that case, the performance
	1793	!-- is significantly better compared to the case where only
	1794	!-- the first row of PEs in y-direction (myidy = 0) is given
	1795	!-- the output while NF90_INDEPENDENT is set.
	1796	IF ( npex /= 1 ) THEN
	1797
	1798	#if defined( __parallel )
	1799	!
	1800	!-- Distribute data over all PEs along x
	1801	ngp = ( nyng-nysg+1 ) * ( nzt-nzb + 2 ) * ns
	1802	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1803	CALL MPI_ALLREDUCE( local_2d_sections_l(1,nysg,nzb_do), &
	1804	local_2d_sections(1,nysg,nzb_do), &
[1308]	1805	ngp, MPI_REAL, MPI_SUM, comm1dx, &
	1806	ierr )
	1807	#else
	1808	local_2d_sections = local_2d_sections_l
	1809	#endif
	1810	ENDIF
	1811	!
	1812	!-- Do not output redundant ghost point data except for the
	1813	!-- boundaries of the total domain.
	1814	IF ( nyn == ny ) THEN
	1815	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1816	id_var_do2d(av,if), &
	1817	local_2d_sections(1:ns, &
[1551]	1818	nys:nyn+1,nzb_do:nzt_do), &
[1308]	1819	start = (/ 1, nys+1, 1, &
	1820	do2d_yz_time_count(av) /), &
	1821	count = (/ ns, nyn-nys+2, &
[1551]	1822	nzt_do-nzb_do+1, 1 /) )
[1308]	1823	ELSE
	1824	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1825	id_var_do2d(av,if), &
	1826	local_2d_sections(1:ns,nys:nyn, &
[1551]	1827	nzb_do:nzt_do), &
[1308]	1828	start = (/ 1, nys+1, 1, &
	1829	do2d_yz_time_count(av) /), &
	1830	count = (/ ns, nyn-nys+1, &
[1551]	1831	nzt_do-nzb_do+1, 1 /) )
[1308]	1832	ENDIF
	1833
[1783]	1834	CALL netcdf_handle_error( 'data_output_2d', 60 )
[1308]	1835
	1836	CASE DEFAULT
	1837	message_string = 'unknown cross-section: ' // TRIM( mode )
	1838	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
	1839
	1840	END SELECT
	1841
	1842	ENDIF
[1311]	1843	#endif
[1]	1844	ENDIF
	1845
	1846	if = if + 1
	1847	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
	1848	do2d_mode = do2d(av,if)(l-1:l)
	1849
	1850	ENDDO
	1851
	1852	!
	1853	!-- Deallocate temporary arrays.
	1854	IF ( ALLOCATED( level_z ) ) DEALLOCATE( level_z )
[1308]	1855	IF ( netcdf_data_format > 4 ) THEN
	1856	DEALLOCATE( local_pf, local_2d, local_2d_sections )
	1857	IF( mode == 'xz' .OR. mode == 'yz' ) DEALLOCATE( local_2d_sections_l )
	1858	ENDIF
[1]	1859	#if defined( __parallel )
	1860	IF ( .NOT. data_output_2d_on_each_pe .AND. myid == 0 ) THEN
	1861	DEALLOCATE( total_2d )
	1862	ENDIF
	1863	#endif
	1864
	1865	!
	1866	!-- Close plot output file.
[1960]	1867	file_id = 20 + s_ind
[1]	1868
	1869	IF ( data_output_2d_on_each_pe ) THEN
[759]	1870	DO i = 0, io_blocks-1
	1871	IF ( i == io_group ) THEN
	1872	CALL close_file( file_id )
	1873	ENDIF
	1874	#if defined( __parallel )
	1875	CALL MPI_BARRIER( comm2d, ierr )
	1876	#endif
	1877	ENDDO
[1]	1878	ELSE
	1879	IF ( myid == 0 ) CALL close_file( file_id )
	1880	ENDIF
	1881
[1318]	1882	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
[1]	1883
	1884	END SUBROUTINE data_output_2d

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |