Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 1985

Last change on this file since 1985 was 1981, checked in by suehring, 8 years ago
last commit documented
Property svn:keywords set to `Id`
File size: 77.5 KB

Rev	Line
[1682]	1	!> @file data_output_2d.f90
[1036]	2	!--------------------------------------------------------------------------------!
	3	! This file is part of PALM.
	4	!
	5	! PALM is free software: you can redistribute it and/or modify it under the terms
	6	! of the GNU General Public License as published by the Free Software Foundation,
	7	! either version 3 of the License, or (at your option) any later version.
	8	!
	9	! PALM is distributed in the hope that it will be useful, but WITHOUT ANY
	10	! WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
	11	! A PARTICULAR PURPOSE. See the GNU General Public License for more details.
	12	!
	13	! You should have received a copy of the GNU General Public License along with
	14	! PALM. If not, see <http://www.gnu.org/licenses/>.
	15	!
[1818]	16	! Copyright 1997-2016 Leibniz Universitaet Hannover
[1704]	17	!--------------------------------------------------------------------------------!
[1036]	18	!
[254]	19	! Current revisions:
[1]	20	! -----------------
[1961]	21	!
[1981]	22	!
[1552]	23	! Former revisions:
	24	! -----------------
	25	! $Id: data_output_2d.f90 1981 2016-07-29 15:53:00Z suehring $
	26	!
[1981]	27	! 1980 2016-07-29 15:51:57Z suehring
	28	! Bugfix, in order to steer user-defined output, setting flag found explicitly
	29	! to .F.
	30	!
[1977]	31	! 1976 2016-07-27 13:28:04Z maronga
	32	! Output of radiation quantities is now done directly in the respective module
	33	!
[1973]	34	! 1972 2016-07-26 07:52:02Z maronga
[1976]	35	! Output of land surface quantities is now done directly in the respective
	36	! module
[1973]	37	!
[1961]	38	! 1960 2016-07-12 16:34:24Z suehring
	39	! Scalar surface flux added
	40	! Rename INTEGER variable s into s_ind, as s is already assigned to scalar
	41	!
[1851]	42	! 1849 2016-04-08 11:33:18Z hoffmann
	43	! precipitation_amount, precipitation_rate, prr moved to arrays_3d
	44	!
[1823]	45	! 1822 2016-04-07 07:49:42Z hoffmann
	46	! Output of bulk cloud physics simplified.
	47	!
[1789]	48	! 1788 2016-03-10 11:01:04Z maronga
	49	! Added output of z0q
	50	!
[1784]	51	! 1783 2016-03-06 18:36:17Z raasch
	52	! name change of netcdf routines and module + related changes
	53	!
[1746]	54	! 1745 2016-02-05 13:06:51Z gronemeier
	55	! Bugfix: test if time axis limit exceeds moved to point after call of check_open
	56	!
[1704]	57	! 1703 2015-11-02 12:38:44Z raasch
	58	! bugfix for output of single (*) xy-sections in case of parallel netcdf I/O
	59	!
[1702]	60	! 1701 2015-11-02 07:43:04Z maronga
	61	! Bugfix in output of RRTGM data
	62	!
[1692]	63	! 1691 2015-10-26 16:17:44Z maronga
	64	! Added output of Obukhov length (ol) and radiative heating rates for RRTMG.
	65	! Formatting corrections.
	66	!
[1683]	67	! 1682 2015-10-07 23:56:08Z knoop
	68	! Code annotations made doxygen readable
	69	!
[1586]	70	! 1585 2015-04-30 07:05:52Z maronga
	71	! Added support for RRTMG
	72	!
[1556]	73	! 1555 2015-03-04 17:44:27Z maronga
	74	! Added output of r_a and r_s
	75	!
[1552]	76	! 1551 2015-03-03 14:18:16Z maronga
[1551]	77	! Added suppport for land surface model and radiation model output. In the course
	78	! of this action, the limits for vertical loops have been changed (from nzb and
	79	! nzt+1 to nzb_do and nzt_do, respectively in order to allow soil model output).
	80	! Moreover, a new vertical grid zs was introduced.
[1329]	81	!
[1360]	82	! 1359 2014-04-11 17:15:14Z hoffmann
	83	! New particle structure integrated.
	84	!
[1354]	85	! 1353 2014-04-08 15:21:23Z heinze
	86	! REAL constants provided with KIND-attribute
	87	!
[1329]	88	! 1327 2014-03-21 11:00:16Z raasch
	89	! parts concerning iso2d output removed,
	90	! -netcdf output queries
	91	!
[1321]	92	! 1320 2014-03-20 08:40:49Z raasch
[1320]	93	! ONLY-attribute added to USE-statements,
	94	! kind-parameters added to all INTEGER and REAL declaration statements,
	95	! kinds are defined in new module kinds,
	96	! revision history before 2012 removed,
	97	! comment fields (!:) to be used for variable explanations added to
	98	! all variable declaration statements
[1309]	99	!
[1319]	100	! 1318 2014-03-17 13:35:16Z raasch
	101	! barrier argument removed from cpu_log.
	102	! module interfaces removed
	103	!
[1312]	104	! 1311 2014-03-14 12:13:39Z heinze
	105	! bugfix: close #if defined( __netcdf )
	106	!
[1309]	107	! 1308 2014-03-13 14:58:42Z fricke
[1308]	108	! +local_2d_sections, local_2d_sections_l, ns
	109	! Check, if the limit of the time dimension is exceeded for parallel output
	110	! To increase the performance for parallel output, the following is done:
	111	! - Update of time axis is only done by PE0
	112	! - Cross sections are first stored on a local array and are written
	113	! collectively to the output file by all PEs.
[674]	114	!
[1116]	115	! 1115 2013-03-26 18:16:16Z hoffmann
	116	! ql is calculated by calc_liquid_water_content
	117	!
[1077]	118	! 1076 2012-12-05 08:30:18Z hoffmann
	119	! Bugfix in output of ql
	120	!
[1066]	121	! 1065 2012-11-22 17:42:36Z hoffmann
	122	! Bugfix: Output of cross sections of ql
	123	!
[1054]	124	! 1053 2012-11-13 17:11:03Z hoffmann
	125	! +qr, nr, qc and cross sections
	126	!
[1037]	127	! 1036 2012-10-22 13:43:42Z raasch
	128	! code put under GPL (PALM 3.9)
	129	!
[1035]	130	! 1031 2012-10-19 14:35:30Z raasch
	131	! netCDF4 without parallel file support implemented
	132	!
[1008]	133	! 1007 2012-09-19 14:30:36Z franke
	134	! Bugfix: missing calculation of ql_vp added
	135	!
[979]	136	! 978 2012-08-09 08:28:32Z fricke
	137	! +z0h
	138	!
[1]	139	! Revision 1.1 1997/08/11 06:24:09 raasch
	140	! Initial revision
	141	!
	142	!
	143	! Description:
	144	! ------------
[1682]	145	!> Data output of horizontal cross-sections in netCDF format or binary format
	146	!> compatible to old graphic software iso2d.
	147	!> Attention: The position of the sectional planes is still not always computed
	148	!> --------- correctly. (zu is used always)!
[1]	149	!------------------------------------------------------------------------------!
[1682]	150	SUBROUTINE data_output_2d( mode, av )
	151
[1]	152
[1320]	153	USE arrays_3d, &
[1849]	154	ONLY: dzw, e, nr, ol, p, pt, precipitation_amount, precipitation_rate,&
[1960]	155	prr,q, qc, ql, ql_c, ql_v, ql_vp, qr, qsws, rho, s, sa, shf, &
	156	ssws, tend, ts, u, us, v, vpt, w, z0, z0h, z0q, zu, zw
[1320]	157
[1]	158	USE averaging
[1320]	159
	160	USE cloud_parameters, &
[1849]	161	ONLY: hyrho, l_d_cp, pt_d_t
[1320]	162
	163	USE control_parameters, &
	164	ONLY: cloud_physics, data_output_2d_on_each_pe, data_output_xy, &
	165	data_output_xz, data_output_yz, do2d, &
	166	do2d_xy_last_time, do2d_xy_n, do2d_xy_time_count, &
	167	do2d_xz_last_time, do2d_xz_n, do2d_xz_time_count, &
	168	do2d_yz_last_time, do2d_yz_n, do2d_yz_time_count, &
[1822]	169	ibc_uv_b, io_blocks, io_group, message_string, &
	170	ntdim_2d_xy, ntdim_2d_xz, ntdim_2d_yz, &
	171	psolver, section, simulated_time, simulated_time_chr, &
	172	time_since_reference_point
[1320]	173
	174	USE cpulog, &
	175	ONLY: cpu_log, log_point
	176
	177	USE grid_variables, &
	178	ONLY: dx, dy
	179
	180	USE indices, &
	181	ONLY: nbgp, nx, nxl, nxlg, nxr, nxrg, ny, nyn, nyng, nys, nysg, &
	182	nz, nzb, nzt
	183
	184	USE kinds
[1551]	185
	186	USE land_surface_model_mod, &
[1972]	187	ONLY: land_surface, lsm_data_output_2d, zs
[1551]	188
[1783]	189	#if defined( __netcdf )
	190	USE NETCDF
	191	#endif
[1320]	192
[1783]	193	USE netcdf_interface, &
	194	ONLY: id_set_xy, id_set_xz, id_set_yz, id_var_do2d, id_var_time_xy, &
	195	id_var_time_xz, id_var_time_yz, nc_stat, netcdf_data_format, &
	196	netcdf_handle_error
	197
[1320]	198	USE particle_attributes, &
[1359]	199	ONLY: grid_particles, number_of_particles, particle_advection_start, &
	200	particles, prt_count
[1320]	201
[1]	202	USE pegrid
	203
[1551]	204	USE radiation_model_mod, &
[1976]	205	ONLY: radiation, radiation_data_output_2d
[1551]	206
[1]	207	IMPLICIT NONE
	208
[1682]	209	CHARACTER (LEN=2) :: do2d_mode !<
	210	CHARACTER (LEN=2) :: mode !<
	211	CHARACTER (LEN=4) :: grid !<
	212	CHARACTER (LEN=25) :: section_chr !<
	213	CHARACTER (LEN=50) :: rtext !<
[1320]	214
[1682]	215	INTEGER(iwp) :: av !<
	216	INTEGER(iwp) :: ngp !<
	217	INTEGER(iwp) :: file_id !<
	218	INTEGER(iwp) :: i !<
	219	INTEGER(iwp) :: if !<
	220	INTEGER(iwp) :: is !<
	221	INTEGER(iwp) :: iis !<
	222	INTEGER(iwp) :: j !<
	223	INTEGER(iwp) :: k !<
	224	INTEGER(iwp) :: l !<
	225	INTEGER(iwp) :: layer_xy !<
	226	INTEGER(iwp) :: n !<
[1703]	227	INTEGER(iwp) :: nis !<
[1682]	228	INTEGER(iwp) :: ns !<
	229	INTEGER(iwp) :: nzb_do !< lower limit of the data field (usually nzb)
	230	INTEGER(iwp) :: nzt_do !< upper limit of the data field (usually nzt+1)
	231	INTEGER(iwp) :: psi !<
[1960]	232	INTEGER(iwp) :: s_ind !<
[1682]	233	INTEGER(iwp) :: sender !<
	234	INTEGER(iwp) :: ind(4) !<
[1320]	235
[1682]	236	LOGICAL :: found !<
	237	LOGICAL :: resorted !<
	238	LOGICAL :: two_d !<
[1320]	239
[1682]	240	REAL(wp) :: mean_r !<
	241	REAL(wp) :: s_r2 !<
	242	REAL(wp) :: s_r3 !<
[1320]	243
[1682]	244	REAL(wp), DIMENSION(:), ALLOCATABLE :: level_z !<
	245	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d !<
	246	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d_l !<
	247	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !<
	248	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections !<
	249	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections_l !<
[1359]	250
[1]	251	#if defined( __parallel )
[1682]	252	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: total_2d !<
[1]	253	#endif
[1682]	254	REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !<
[1]	255
	256	NAMELIST /LOCAL/ rtext
	257
	258	!
	259	!-- Immediate return, if no output is requested (no respective sections
	260	!-- found in parameter data_output)
	261	IF ( mode == 'xy' .AND. .NOT. data_output_xy(av) ) RETURN
	262	IF ( mode == 'xz' .AND. .NOT. data_output_xz(av) ) RETURN
	263	IF ( mode == 'yz' .AND. .NOT. data_output_yz(av) ) RETURN
	264
[1308]	265	CALL cpu_log (log_point(3),'data_output_2d','start')
	266
[1]	267	two_d = .FALSE. ! local variable to distinguish between output of pure 2D
	268	! arrays and cross-sections of 3D arrays.
	269
	270	!
	271	!-- Depending on the orientation of the cross-section, the respective output
	272	!-- files have to be opened.
	273	SELECT CASE ( mode )
	274
	275	CASE ( 'xy' )
[1960]	276	s_ind = 1
[667]	277	ALLOCATE( level_z(nzb:nzt+1), local_2d(nxlg:nxrg,nysg:nyng) )
[1]	278
[1308]	279	IF ( netcdf_data_format > 4 ) THEN
	280	ns = 1
[1960]	281	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	282	ns = ns + 1
	283	ENDDO
	284	ns = ns - 1
	285	ALLOCATE( local_2d_sections(nxlg:nxrg,nysg:nyng,1:ns) )
[1353]	286	local_2d_sections = 0.0_wp
[1308]	287	ENDIF
	288
[493]	289	!
[1031]	290	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	291	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	292	CALL check_open( 101+av*10 )
	293	ENDIF
[1]	294	IF ( data_output_2d_on_each_pe ) THEN
	295	CALL check_open( 21 )
	296	ELSE
	297	IF ( myid == 0 ) THEN
	298	#if defined( __parallel )
[667]	299	ALLOCATE( total_2d(-nbgp:nx+nbgp,-nbgp:ny+nbgp) )
[1]	300	#endif
	301	ENDIF
	302	ENDIF
	303
	304	CASE ( 'xz' )
[1960]	305	s_ind = 2
[667]	306	ALLOCATE( local_2d(nxlg:nxrg,nzb:nzt+1) )
[1]	307
[1308]	308	IF ( netcdf_data_format > 4 ) THEN
	309	ns = 1
[1960]	310	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	311	ns = ns + 1
	312	ENDDO
	313	ns = ns - 1
	314	ALLOCATE( local_2d_sections(nxlg:nxrg,1:ns,nzb:nzt+1) )
	315	ALLOCATE( local_2d_sections_l(nxlg:nxrg,1:ns,nzb:nzt+1) )
[1353]	316	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
[1308]	317	ENDIF
	318
[493]	319	!
[1031]	320	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	321	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	322	CALL check_open( 102+av*10 )
	323	ENDIF
[1]	324
	325	IF ( data_output_2d_on_each_pe ) THEN
	326	CALL check_open( 22 )
	327	ELSE
	328	IF ( myid == 0 ) THEN
	329	#if defined( __parallel )
[667]	330	ALLOCATE( total_2d(-nbgp:nx+nbgp,nzb:nzt+1) )
[1]	331	#endif
	332	ENDIF
	333	ENDIF
	334
	335	CASE ( 'yz' )
[1960]	336	s_ind = 3
[667]	337	ALLOCATE( local_2d(nysg:nyng,nzb:nzt+1) )
[1]	338
[1308]	339	IF ( netcdf_data_format > 4 ) THEN
	340	ns = 1
[1960]	341	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	342	ns = ns + 1
	343	ENDDO
	344	ns = ns - 1
	345	ALLOCATE( local_2d_sections(1:ns,nysg:nyng,nzb:nzt+1) )
	346	ALLOCATE( local_2d_sections_l(1:ns,nysg:nyng,nzb:nzt+1) )
[1353]	347	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
[1308]	348	ENDIF
	349
[493]	350	!
[1031]	351	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	352	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	353	CALL check_open( 103+av*10 )
	354	ENDIF
[1]	355
	356	IF ( data_output_2d_on_each_pe ) THEN
	357	CALL check_open( 23 )
	358	ELSE
	359	IF ( myid == 0 ) THEN
	360	#if defined( __parallel )
[667]	361	ALLOCATE( total_2d(-nbgp:ny+nbgp,nzb:nzt+1) )
[1]	362	#endif
	363	ENDIF
	364	ENDIF
	365
	366	CASE DEFAULT
[254]	367	message_string = 'unknown cross-section: ' // TRIM( mode )
	368	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
[1]	369
	370	END SELECT
	371
	372	!
[1745]	373	!-- For parallel netcdf output the time axis must be limited. Return, if this
	374	!-- limit is exceeded. This could be the case, if the simulated time exceeds
	375	!-- the given end time by the length of the given output interval.
	376	IF ( netcdf_data_format > 4 ) THEN
	377	IF ( mode == 'xy' .AND. do2d_xy_time_count(av) + 1 > &
	378	ntdim_2d_xy(av) ) THEN
	379	WRITE ( message_string, * ) 'Output of xy cross-sections is not ', &
	380	'given at t=', simulated_time, '&because the', &
	381	' maximum number of output time levels is exceeded.'
	382	CALL message( 'data_output_2d', 'PA0384', 0, 1, 0, 6, 0 )
	383	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	384	RETURN
	385	ENDIF
	386	IF ( mode == 'xz' .AND. do2d_xz_time_count(av) + 1 > &
	387	ntdim_2d_xz(av) ) THEN
	388	WRITE ( message_string, * ) 'Output of xz cross-sections is not ', &
	389	'given at t=', simulated_time, '&because the', &
	390	' maximum number of output time levels is exceeded.'
	391	CALL message( 'data_output_2d', 'PA0385', 0, 1, 0, 6, 0 )
	392	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	393	RETURN
	394	ENDIF
	395	IF ( mode == 'yz' .AND. do2d_yz_time_count(av) + 1 > &
	396	ntdim_2d_yz(av) ) THEN
	397	WRITE ( message_string, * ) 'Output of yz cross-sections is not ', &
	398	'given at t=', simulated_time, '&because the', &
	399	' maximum number of output time levels is exceeded.'
	400	CALL message( 'data_output_2d', 'PA0386', 0, 1, 0, 6, 0 )
	401	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	402	RETURN
	403	ENDIF
	404	ENDIF
	405
	406	!
[1]	407	!-- Allocate a temporary array for resorting (kji -> ijk).
[667]	408	ALLOCATE( local_pf(nxlg:nxrg,nysg:nyng,nzb:nzt+1) )
[1]	409
	410	!
	411	!-- Loop of all variables to be written.
	412	!-- Output dimensions chosen
	413	if = 1
	414	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
	415	do2d_mode = do2d(av,if)(l-1:l)
	416
	417	DO WHILE ( do2d(av,if)(1:1) /= ' ' )
	418
	419	IF ( do2d_mode == mode ) THEN
[1980]	420	!
	421	!-- Set flag to steer output of radiation, land-surface, or user-defined
	422	!-- quantities
	423	found = .FALSE.
[1551]	424
	425	nzb_do = nzb
	426	nzt_do = nzt+1
[1]	427	!
	428	!-- Store the array chosen on the temporary array.
	429	resorted = .FALSE.
	430	SELECT CASE ( TRIM( do2d(av,if) ) )
	431
	432	CASE ( 'e_xy', 'e_xz', 'e_yz' )
	433	IF ( av == 0 ) THEN
	434	to_be_resorted => e
	435	ELSE
	436	to_be_resorted => e_av
	437	ENDIF
	438	IF ( mode == 'xy' ) level_z = zu
	439
[771]	440	CASE ( 'lpt_xy', 'lpt_xz', 'lpt_yz' )
	441	IF ( av == 0 ) THEN
	442	to_be_resorted => pt
	443	ELSE
	444	to_be_resorted => lpt_av
	445	ENDIF
	446	IF ( mode == 'xy' ) level_z = zu
	447
[1]	448	CASE ( 'lwp*_xy' ) ! 2d-array
	449	IF ( av == 0 ) THEN
[667]	450	DO i = nxlg, nxrg
	451	DO j = nysg, nyng
[1320]	452	local_pf(i,j,nzb+1) = SUM( ql(nzb:nzt,j,i) * &
[1]	453	dzw(1:nzt+1) )
	454	ENDDO
	455	ENDDO
	456	ELSE
[667]	457	DO i = nxlg, nxrg
	458	DO j = nysg, nyng
[1]	459	local_pf(i,j,nzb+1) = lwp_av(j,i)
	460	ENDDO
	461	ENDDO
	462	ENDIF
	463	resorted = .TRUE.
	464	two_d = .TRUE.
	465	level_z(nzb+1) = zu(nzb+1)
	466
[1053]	467	CASE ( 'nr_xy', 'nr_xz', 'nr_yz' )
	468	IF ( av == 0 ) THEN
	469	to_be_resorted => nr
	470	ELSE
	471	to_be_resorted => nr_av
	472	ENDIF
	473	IF ( mode == 'xy' ) level_z = zu
	474
[1691]	475	CASE ( 'ol*_xy' ) ! 2d-array
	476	IF ( av == 0 ) THEN
	477	DO i = nxlg, nxrg
	478	DO j = nysg, nyng
	479	local_pf(i,j,nzb+1) = ol(j,i)
	480	ENDDO
	481	ENDDO
	482	ELSE
	483	DO i = nxlg, nxrg
	484	DO j = nysg, nyng
	485	local_pf(i,j,nzb+1) = ol_av(j,i)
	486	ENDDO
	487	ENDDO
	488	ENDIF
	489	resorted = .TRUE.
	490	two_d = .TRUE.
	491	level_z(nzb+1) = zu(nzb+1)
	492
[1]	493	CASE ( 'p_xy', 'p_xz', 'p_yz' )
	494	IF ( av == 0 ) THEN
[729]	495	IF ( psolver /= 'sor' ) CALL exchange_horiz( p, nbgp )
[1]	496	to_be_resorted => p
	497	ELSE
[729]	498	IF ( psolver /= 'sor' ) CALL exchange_horiz( p_av, nbgp )
[1]	499	to_be_resorted => p_av
	500	ENDIF
	501	IF ( mode == 'xy' ) level_z = zu
	502
	503	CASE ( 'pc_xy', 'pc_xz', 'pc_yz' ) ! particle concentration
	504	IF ( av == 0 ) THEN
[215]	505	IF ( simulated_time >= particle_advection_start ) THEN
	506	tend = prt_count
[667]	507	CALL exchange_horiz( tend, nbgp )
[215]	508	ELSE
[1353]	509	tend = 0.0_wp
[215]	510	ENDIF
[667]	511	DO i = nxlg, nxrg
	512	DO j = nysg, nyng
[1]	513	DO k = nzb, nzt+1
	514	local_pf(i,j,k) = tend(k,j,i)
	515	ENDDO
	516	ENDDO
	517	ENDDO
	518	resorted = .TRUE.
	519	ELSE
[667]	520	CALL exchange_horiz( pc_av, nbgp )
[1]	521	to_be_resorted => pc_av
	522	ENDIF
	523
[1359]	524	CASE ( 'pr_xy', 'pr_xz', 'pr_yz' ) ! mean particle radius (effective radius)
[1]	525	IF ( av == 0 ) THEN
[215]	526	IF ( simulated_time >= particle_advection_start ) THEN
	527	DO i = nxl, nxr
	528	DO j = nys, nyn
	529	DO k = nzb, nzt+1
[1359]	530	number_of_particles = prt_count(k,j,i)
	531	IF (number_of_particles <= 0) CYCLE
	532	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
	533	s_r2 = 0.0_wp
[1353]	534	s_r3 = 0.0_wp
[1359]	535	DO n = 1, number_of_particles
	536	IF ( particles(n)%particle_mask ) THEN
	537	s_r2 = s_r2 + particles(n)%radius*2 &
	538	particles(n)%weight_factor
	539	s_r3 = s_r3 + particles(n)%radius*3 &
	540	particles(n)%weight_factor
	541	ENDIF
[215]	542	ENDDO
[1359]	543	IF ( s_r2 > 0.0_wp ) THEN
	544	mean_r = s_r3 / s_r2
[215]	545	ELSE
[1353]	546	mean_r = 0.0_wp
[215]	547	ENDIF
	548	tend(k,j,i) = mean_r
[1]	549	ENDDO
	550	ENDDO
	551	ENDDO
[667]	552	CALL exchange_horiz( tend, nbgp )
[215]	553	ELSE
[1353]	554	tend = 0.0_wp
[1359]	555	ENDIF
[667]	556	DO i = nxlg, nxrg
	557	DO j = nysg, nyng
[1]	558	DO k = nzb, nzt+1
	559	local_pf(i,j,k) = tend(k,j,i)
	560	ENDDO
	561	ENDDO
	562	ENDDO
	563	resorted = .TRUE.
	564	ELSE
[667]	565	CALL exchange_horiz( pr_av, nbgp )
[1]	566	to_be_resorted => pr_av
	567	ENDIF
	568
[72]	569	CASE ( 'pra*_xy' ) ! 2d-array / integral quantity => no av
	570	CALL exchange_horiz_2d( precipitation_amount )
[667]	571	DO i = nxlg, nxrg
	572	DO j = nysg, nyng
[72]	573	local_pf(i,j,nzb+1) = precipitation_amount(j,i)
	574	ENDDO
	575	ENDDO
[1353]	576	precipitation_amount = 0.0_wp ! reset for next integ. interval
[72]	577	resorted = .TRUE.
	578	two_d = .TRUE.
	579	level_z(nzb+1) = zu(nzb+1)
	580
	581	CASE ( 'prr*_xy' ) ! 2d-array
[1822]	582	IF ( av == 0 ) THEN
	583	CALL exchange_horiz_2d( prr(nzb+1,:,:) )
	584	DO i = nxlg, nxrg
	585	DO j = nysg, nyng
	586	local_pf(i,j,nzb+1) = prr(nzb+1,j,i) * hyrho(nzb+1)
[1053]	587	ENDDO
[1822]	588	ENDDO
[1053]	589	ELSE
[1822]	590	CALL exchange_horiz_2d( prr_av(nzb+1,:,:) )
	591	DO i = nxlg, nxrg
	592	DO j = nysg, nyng
	593	local_pf(i,j,nzb+1) = prr_av(nzb+1,j,i) * hyrho(nzb+1)
[1053]	594	ENDDO
[1822]	595	ENDDO
[1053]	596	ENDIF
	597	resorted = .TRUE.
	598	two_d = .TRUE.
	599	level_z(nzb+1) = zu(nzb+1)
	600
	601	CASE ( 'prr_xy', 'prr_xz', 'prr_yz' )
[72]	602	IF ( av == 0 ) THEN
[1053]	603	CALL exchange_horiz( prr, nbgp )
[667]	604	DO i = nxlg, nxrg
	605	DO j = nysg, nyng
[1053]	606	DO k = nzb, nzt+1
[1822]	607	local_pf(i,j,k) = prr(k,j,i) * hyrho(nzb+1)
[1053]	608	ENDDO
[72]	609	ENDDO
	610	ENDDO
	611	ELSE
[1053]	612	CALL exchange_horiz( prr_av, nbgp )
[667]	613	DO i = nxlg, nxrg
	614	DO j = nysg, nyng
[1053]	615	DO k = nzb, nzt+1
[1822]	616	local_pf(i,j,k) = prr_av(k,j,i) * hyrho(nzb+1)
[1053]	617	ENDDO
[72]	618	ENDDO
	619	ENDDO
	620	ENDIF
	621	resorted = .TRUE.
[1053]	622	IF ( mode == 'xy' ) level_z = zu
[72]	623
[1]	624	CASE ( 'pt_xy', 'pt_xz', 'pt_yz' )
	625	IF ( av == 0 ) THEN
	626	IF ( .NOT. cloud_physics ) THEN
	627	to_be_resorted => pt
	628	ELSE
[667]	629	DO i = nxlg, nxrg
	630	DO j = nysg, nyng
[1]	631	DO k = nzb, nzt+1
[1320]	632	local_pf(i,j,k) = pt(k,j,i) + l_d_cp * &
	633	pt_d_t(k) * &
[1]	634	ql(k,j,i)
	635	ENDDO
	636	ENDDO
	637	ENDDO
	638	resorted = .TRUE.
	639	ENDIF
	640	ELSE
	641	to_be_resorted => pt_av
	642	ENDIF
	643	IF ( mode == 'xy' ) level_z = zu
	644
	645	CASE ( 'q_xy', 'q_xz', 'q_yz' )
	646	IF ( av == 0 ) THEN
	647	to_be_resorted => q
	648	ELSE
	649	to_be_resorted => q_av
	650	ENDIF
	651	IF ( mode == 'xy' ) level_z = zu
	652
[1053]	653	CASE ( 'qc_xy', 'qc_xz', 'qc_yz' )
[1]	654	IF ( av == 0 ) THEN
[1115]	655	to_be_resorted => qc
[1]	656	ELSE
[1115]	657	to_be_resorted => qc_av
[1]	658	ENDIF
	659	IF ( mode == 'xy' ) level_z = zu
	660
[1053]	661	CASE ( 'ql_xy', 'ql_xz', 'ql_yz' )
	662	IF ( av == 0 ) THEN
[1115]	663	to_be_resorted => ql
[1053]	664	ELSE
[1115]	665	to_be_resorted => ql_av
[1053]	666	ENDIF
	667	IF ( mode == 'xy' ) level_z = zu
	668
[1]	669	CASE ( 'ql_c_xy', 'ql_c_xz', 'ql_c_yz' )
	670	IF ( av == 0 ) THEN
	671	to_be_resorted => ql_c
	672	ELSE
	673	to_be_resorted => ql_c_av
	674	ENDIF
	675	IF ( mode == 'xy' ) level_z = zu
	676
	677	CASE ( 'ql_v_xy', 'ql_v_xz', 'ql_v_yz' )
	678	IF ( av == 0 ) THEN
	679	to_be_resorted => ql_v
	680	ELSE
	681	to_be_resorted => ql_v_av
	682	ENDIF
	683	IF ( mode == 'xy' ) level_z = zu
	684
	685	CASE ( 'ql_vp_xy', 'ql_vp_xz', 'ql_vp_yz' )
	686	IF ( av == 0 ) THEN
[1007]	687	IF ( simulated_time >= particle_advection_start ) THEN
	688	DO i = nxl, nxr
	689	DO j = nys, nyn
	690	DO k = nzb, nzt+1
[1359]	691	number_of_particles = prt_count(k,j,i)
	692	IF (number_of_particles <= 0) CYCLE
	693	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
	694	DO n = 1, number_of_particles
	695	IF ( particles(n)%particle_mask ) THEN
	696	tend(k,j,i) = tend(k,j,i) + &
	697	particles(n)%weight_factor / &
	698	prt_count(k,j,i)
	699	ENDIF
[1007]	700	ENDDO
	701	ENDDO
	702	ENDDO
	703	ENDDO
	704	CALL exchange_horiz( tend, nbgp )
	705	ELSE
[1353]	706	tend = 0.0_wp
[1359]	707	ENDIF
[1007]	708	DO i = nxlg, nxrg
	709	DO j = nysg, nyng
	710	DO k = nzb, nzt+1
	711	local_pf(i,j,k) = tend(k,j,i)
	712	ENDDO
	713	ENDDO
	714	ENDDO
	715	resorted = .TRUE.
	716	ELSE
	717	CALL exchange_horiz( ql_vp_av, nbgp )
[1]	718	to_be_resorted => ql_vp
	719	ENDIF
	720	IF ( mode == 'xy' ) level_z = zu
	721
[1053]	722	CASE ( 'qr_xy', 'qr_xz', 'qr_yz' )
	723	IF ( av == 0 ) THEN
	724	to_be_resorted => qr
	725	ELSE
	726	to_be_resorted => qr_av
	727	ENDIF
	728	IF ( mode == 'xy' ) level_z = zu
	729
[354]	730	CASE ( 'qsws*_xy' ) ! 2d-array
	731	IF ( av == 0 ) THEN
[667]	732	DO i = nxlg, nxrg
	733	DO j = nysg, nyng
[354]	734	local_pf(i,j,nzb+1) = qsws(j,i)
	735	ENDDO
	736	ENDDO
	737	ELSE
[667]	738	DO i = nxlg, nxrg
	739	DO j = nysg, nyng
[354]	740	local_pf(i,j,nzb+1) = qsws_av(j,i)
	741	ENDDO
	742	ENDDO
	743	ENDIF
	744	resorted = .TRUE.
	745	two_d = .TRUE.
	746	level_z(nzb+1) = zu(nzb+1)
	747
[1]	748	CASE ( 'qv_xy', 'qv_xz', 'qv_yz' )
	749	IF ( av == 0 ) THEN
[667]	750	DO i = nxlg, nxrg
	751	DO j = nysg, nyng
[1]	752	DO k = nzb, nzt+1
	753	local_pf(i,j,k) = q(k,j,i) - ql(k,j,i)
	754	ENDDO
	755	ENDDO
	756	ENDDO
	757	resorted = .TRUE.
	758	ELSE
	759	to_be_resorted => qv_av
	760	ENDIF
	761	IF ( mode == 'xy' ) level_z = zu
	762
[1551]	763
[1585]	764
[96]	765	CASE ( 'rho_xy', 'rho_xz', 'rho_yz' )
	766	IF ( av == 0 ) THEN
	767	to_be_resorted => rho
	768	ELSE
	769	to_be_resorted => rho_av
	770	ENDIF
	771
[1]	772	CASE ( 's_xy', 's_xz', 's_yz' )
	773	IF ( av == 0 ) THEN
[1960]	774	to_be_resorted => s
[1]	775	ELSE
[355]	776	to_be_resorted => s_av
[1]	777	ENDIF
	778
[96]	779	CASE ( 'sa_xy', 'sa_xz', 'sa_yz' )
	780	IF ( av == 0 ) THEN
	781	to_be_resorted => sa
	782	ELSE
	783	to_be_resorted => sa_av
	784	ENDIF
	785
[354]	786	CASE ( 'shf*_xy' ) ! 2d-array
	787	IF ( av == 0 ) THEN
[667]	788	DO i = nxlg, nxrg
	789	DO j = nysg, nyng
[354]	790	local_pf(i,j,nzb+1) = shf(j,i)
	791	ENDDO
	792	ENDDO
	793	ELSE
[667]	794	DO i = nxlg, nxrg
	795	DO j = nysg, nyng
[354]	796	local_pf(i,j,nzb+1) = shf_av(j,i)
	797	ENDDO
	798	ENDDO
	799	ENDIF
	800	resorted = .TRUE.
	801	two_d = .TRUE.
	802	level_z(nzb+1) = zu(nzb+1)
[1960]	803
	804	CASE ( 'ssws*_xy' ) ! 2d-array
	805	IF ( av == 0 ) THEN
	806	DO i = nxlg, nxrg
	807	DO j = nysg, nyng
	808	local_pf(i,j,nzb+1) = ssws(j,i)
	809	ENDDO
	810	ENDDO
	811	ELSE
	812	DO i = nxlg, nxrg
	813	DO j = nysg, nyng
	814	local_pf(i,j,nzb+1) = ssws_av(j,i)
	815	ENDDO
	816	ENDDO
	817	ENDIF
	818	resorted = .TRUE.
	819	two_d = .TRUE.
	820	level_z(nzb+1) = zu(nzb+1)
[1551]	821
[1]	822	CASE ( 't*_xy' ) ! 2d-array
	823	IF ( av == 0 ) THEN
[667]	824	DO i = nxlg, nxrg
	825	DO j = nysg, nyng
[1]	826	local_pf(i,j,nzb+1) = ts(j,i)
	827	ENDDO
	828	ENDDO
	829	ELSE
[667]	830	DO i = nxlg, nxrg
	831	DO j = nysg, nyng
[1]	832	local_pf(i,j,nzb+1) = ts_av(j,i)
	833	ENDDO
	834	ENDDO
	835	ENDIF
	836	resorted = .TRUE.
	837	two_d = .TRUE.
	838	level_z(nzb+1) = zu(nzb+1)
	839
	840	CASE ( 'u_xy', 'u_xz', 'u_yz' )
	841	IF ( av == 0 ) THEN
	842	to_be_resorted => u
	843	ELSE
	844	to_be_resorted => u_av
	845	ENDIF
	846	IF ( mode == 'xy' ) level_z = zu
	847	!
	848	!-- Substitute the values generated by "mirror" boundary condition
	849	!-- at the bottom boundary by the real surface values.
	850	IF ( do2d(av,if) == 'u_xz' .OR. do2d(av,if) == 'u_yz' ) THEN
[1353]	851	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
[1]	852	ENDIF
	853
	854	CASE ( 'u*_xy' ) ! 2d-array
	855	IF ( av == 0 ) THEN
[667]	856	DO i = nxlg, nxrg
	857	DO j = nysg, nyng
[1]	858	local_pf(i,j,nzb+1) = us(j,i)
	859	ENDDO
	860	ENDDO
	861	ELSE
[667]	862	DO i = nxlg, nxrg
	863	DO j = nysg, nyng
[1]	864	local_pf(i,j,nzb+1) = us_av(j,i)
	865	ENDDO
	866	ENDDO
	867	ENDIF
	868	resorted = .TRUE.
	869	two_d = .TRUE.
	870	level_z(nzb+1) = zu(nzb+1)
	871
	872	CASE ( 'v_xy', 'v_xz', 'v_yz' )
	873	IF ( av == 0 ) THEN
	874	to_be_resorted => v
	875	ELSE
	876	to_be_resorted => v_av
	877	ENDIF
	878	IF ( mode == 'xy' ) level_z = zu
	879	!
	880	!-- Substitute the values generated by "mirror" boundary condition
	881	!-- at the bottom boundary by the real surface values.
	882	IF ( do2d(av,if) == 'v_xz' .OR. do2d(av,if) == 'v_yz' ) THEN
[1353]	883	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
[1]	884	ENDIF
	885
	886	CASE ( 'vpt_xy', 'vpt_xz', 'vpt_yz' )
	887	IF ( av == 0 ) THEN
	888	to_be_resorted => vpt
	889	ELSE
	890	to_be_resorted => vpt_av
	891	ENDIF
	892	IF ( mode == 'xy' ) level_z = zu
	893
	894	CASE ( 'w_xy', 'w_xz', 'w_yz' )
	895	IF ( av == 0 ) THEN
	896	to_be_resorted => w
	897	ELSE
	898	to_be_resorted => w_av
	899	ENDIF
	900	IF ( mode == 'xy' ) level_z = zw
	901
[72]	902	CASE ( 'z0*_xy' ) ! 2d-array
	903	IF ( av == 0 ) THEN
[667]	904	DO i = nxlg, nxrg
	905	DO j = nysg, nyng
[72]	906	local_pf(i,j,nzb+1) = z0(j,i)
	907	ENDDO
	908	ENDDO
	909	ELSE
[667]	910	DO i = nxlg, nxrg
	911	DO j = nysg, nyng
[72]	912	local_pf(i,j,nzb+1) = z0_av(j,i)
	913	ENDDO
	914	ENDDO
	915	ENDIF
	916	resorted = .TRUE.
	917	two_d = .TRUE.
	918	level_z(nzb+1) = zu(nzb+1)
	919
[978]	920	CASE ( 'z0h*_xy' ) ! 2d-array
	921	IF ( av == 0 ) THEN
	922	DO i = nxlg, nxrg
	923	DO j = nysg, nyng
	924	local_pf(i,j,nzb+1) = z0h(j,i)
	925	ENDDO
	926	ENDDO
	927	ELSE
	928	DO i = nxlg, nxrg
	929	DO j = nysg, nyng
	930	local_pf(i,j,nzb+1) = z0h_av(j,i)
	931	ENDDO
	932	ENDDO
	933	ENDIF
	934	resorted = .TRUE.
	935	two_d = .TRUE.
	936	level_z(nzb+1) = zu(nzb+1)
	937
[1788]	938	CASE ( 'z0q*_xy' ) ! 2d-array
	939	IF ( av == 0 ) THEN
	940	DO i = nxlg, nxrg
	941	DO j = nysg, nyng
	942	local_pf(i,j,nzb+1) = z0q(j,i)
	943	ENDDO
	944	ENDDO
	945	ELSE
	946	DO i = nxlg, nxrg
	947	DO j = nysg, nyng
	948	local_pf(i,j,nzb+1) = z0q_av(j,i)
	949	ENDDO
	950	ENDDO
	951	ENDIF
	952	resorted = .TRUE.
	953	two_d = .TRUE.
	954	level_z(nzb+1) = zu(nzb+1)
	955
[1]	956	CASE DEFAULT
[1972]	957
[1]	958	!
[1972]	959	!-- Land surface model quantity
	960	IF ( land_surface ) THEN
	961	CALL lsm_data_output_2d( av, do2d(av,if), found, grid, mode,&
	962	local_pf, two_d, nzb_do, nzt_do )
	963	ENDIF
	964
	965	!
[1976]	966	!-- Radiation quantity
	967	IF ( .NOT. found .AND. radiation ) THEN
	968	CALL radiation_data_output_2d( av, do2d(av,if), found, grid,&
	969	mode, local_pf, two_d )
	970	ENDIF
	971
	972	!
[1]	973	!-- User defined quantity
[1972]	974	IF ( .NOT. found ) THEN
	975	CALL user_data_output_2d( av, do2d(av,if), found, grid, &
	976	local_pf, two_d, nzb_do, nzt_do )
	977	ENDIF
	978
[1]	979	resorted = .TRUE.
	980
	981	IF ( grid == 'zu' ) THEN
	982	IF ( mode == 'xy' ) level_z = zu
	983	ELSEIF ( grid == 'zw' ) THEN
	984	IF ( mode == 'xy' ) level_z = zw
[343]	985	ELSEIF ( grid == 'zu1' ) THEN
	986	IF ( mode == 'xy' ) level_z(nzb+1) = zu(nzb+1)
[1551]	987	ELSEIF ( grid == 'zs' ) THEN
	988	IF ( mode == 'xy' ) level_z = zs
[1]	989	ENDIF
	990
	991	IF ( .NOT. found ) THEN
[1320]	992	message_string = 'no output provided for: ' // &
[274]	993	TRIM( do2d(av,if) )
[254]	994	CALL message( 'data_output_2d', 'PA0181', 0, 0, 0, 6, 0 )
[1]	995	ENDIF
	996
	997	END SELECT
	998
	999	!
	1000	!-- Resort the array to be output, if not done above
	1001	IF ( .NOT. resorted ) THEN
[667]	1002	DO i = nxlg, nxrg
	1003	DO j = nysg, nyng
[1551]	1004	DO k = nzb_do, nzt_do
[1]	1005	local_pf(i,j,k) = to_be_resorted(k,j,i)
	1006	ENDDO
	1007	ENDDO
	1008	ENDDO
	1009	ENDIF
	1010
	1011	!
	1012	!-- Output of the individual cross-sections, depending on the cross-
	1013	!-- section mode chosen.
	1014	is = 1
[1960]	1015	loop1: DO WHILE ( section(is,s_ind) /= -9999 .OR. two_d )
[1]	1016
	1017	SELECT CASE ( mode )
	1018
	1019	CASE ( 'xy' )
	1020	!
	1021	!-- Determine the cross section index
	1022	IF ( two_d ) THEN
	1023	layer_xy = nzb+1
	1024	ELSE
[1960]	1025	layer_xy = section(is,s_ind)
[1]	1026	ENDIF
	1027
	1028	!
[1551]	1029	!-- Exit the loop for layers beyond the data output domain
	1030	!-- (used for soil model)
[1691]	1031	IF ( layer_xy > nzt_do ) THEN
[1551]	1032	EXIT loop1
	1033	ENDIF
	1034
	1035	!
[1308]	1036	!-- Update the netCDF xy cross section time axis.
	1037	!-- In case of parallel output, this is only done by PE0
	1038	!-- to increase the performance.
	1039	IF ( simulated_time /= do2d_xy_last_time(av) ) THEN
	1040	do2d_xy_time_count(av) = do2d_xy_time_count(av) + 1
	1041	do2d_xy_last_time(av) = simulated_time
	1042	IF ( myid == 0 ) THEN
[1327]	1043	IF ( .NOT. data_output_2d_on_each_pe &
	1044	.OR. netcdf_data_format > 4 ) &
[493]	1045	THEN
[1]	1046	#if defined( __netcdf )
	1047	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1048	id_var_time_xy(av), &
[291]	1049	(/ time_since_reference_point /), &
[1]	1050	start = (/ do2d_xy_time_count(av) /), &
	1051	count = (/ 1 /) )
[1783]	1052	CALL netcdf_handle_error( 'data_output_2d', 53 )
[1]	1053	#endif
	1054	ENDIF
	1055	ENDIF
	1056	ENDIF
	1057	!
	1058	!-- If required, carry out averaging along z
[1960]	1059	IF ( section(is,s_ind) == -1 .AND. .NOT. two_d ) THEN
[1]	1060
[1353]	1061	local_2d = 0.0_wp
[1]	1062	!
	1063	!-- Carry out the averaging (all data are on the PE)
[1551]	1064	DO k = nzb_do, nzt_do
[667]	1065	DO j = nysg, nyng
	1066	DO i = nxlg, nxrg
[1]	1067	local_2d(i,j) = local_2d(i,j) + local_pf(i,j,k)
	1068	ENDDO
	1069	ENDDO
	1070	ENDDO
	1071
[1551]	1072	local_2d = local_2d / ( nzt_do - nzb_do + 1.0_wp)
[1]	1073
	1074	ELSE
	1075	!
	1076	!-- Just store the respective section on the local array
	1077	local_2d = local_pf(:,:,layer_xy)
	1078
	1079	ENDIF
	1080
	1081	#if defined( __parallel )
[1327]	1082	IF ( netcdf_data_format > 4 ) THEN
[1]	1083	!
[1031]	1084	!-- Parallel output in netCDF4/HDF5 format.
[493]	1085	IF ( two_d ) THEN
	1086	iis = 1
	1087	ELSE
	1088	iis = is
	1089	ENDIF
	1090
[1]	1091	#if defined( __netcdf )
[1308]	1092	!
	1093	!-- For parallel output, all cross sections are first stored
	1094	!-- here on a local array and will be written to the output
	1095	!-- file afterwards to increase the performance.
	1096	DO i = nxlg, nxrg
	1097	DO j = nysg, nyng
	1098	local_2d_sections(i,j,iis) = local_2d(i,j)
	1099	ENDDO
	1100	ENDDO
[1]	1101	#endif
[493]	1102	ELSE
[1]	1103
[493]	1104	IF ( data_output_2d_on_each_pe ) THEN
[1]	1105	!
[493]	1106	!-- Output of partial arrays on each PE
	1107	#if defined( __netcdf )
[1327]	1108	IF ( myid == 0 ) THEN
[1320]	1109	WRITE ( 21 ) time_since_reference_point, &
[493]	1110	do2d_xy_time_count(av), av
	1111	ENDIF
	1112	#endif
[759]	1113	DO i = 0, io_blocks-1
	1114	IF ( i == io_group ) THEN
[1551]	1115	WRITE ( 21 ) nxlg, nxrg, nysg, nyng, nysg, nyng
[759]	1116	WRITE ( 21 ) local_2d
	1117	ENDIF
	1118	#if defined( __parallel )
	1119	CALL MPI_BARRIER( comm2d, ierr )
	1120	#endif
	1121	ENDDO
[559]	1122
[493]	1123	ELSE
[1]	1124	!
[493]	1125	!-- PE0 receives partial arrays from all processors and
	1126	!-- then outputs them. Here a barrier has to be set,
	1127	!-- because otherwise "-MPI- FATAL: Remote protocol queue
	1128	!-- full" may occur.
	1129	CALL MPI_BARRIER( comm2d, ierr )
	1130
[667]	1131	ngp = ( nxrg-nxlg+1 ) * ( nyng-nysg+1 )
[493]	1132	IF ( myid == 0 ) THEN
[1]	1133	!
[493]	1134	!-- Local array can be relocated directly.
[667]	1135	total_2d(nxlg:nxrg,nysg:nyng) = local_2d
[1]	1136	!
[493]	1137	!-- Receive data from all other PEs.
	1138	DO n = 1, numprocs-1
[1]	1139	!
[493]	1140	!-- Receive index limits first, then array.
	1141	!-- Index limits are received in arbitrary order from
	1142	!-- the PEs.
[1320]	1143	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1144	MPI_ANY_SOURCE, 0, comm2d, &
[493]	1145	status, ierr )
	1146	sender = status(MPI_SOURCE)
	1147	DEALLOCATE( local_2d )
	1148	ALLOCATE( local_2d(ind(1):ind(2),ind(3):ind(4)) )
[1320]	1149	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1150	MPI_REAL, sender, 1, comm2d, &
[493]	1151	status, ierr )
	1152	total_2d(ind(1):ind(2),ind(3):ind(4)) = local_2d
	1153	ENDDO
[1]	1154	!
[493]	1155	!-- Relocate the local array for the next loop increment
	1156	DEALLOCATE( local_2d )
[667]	1157	ALLOCATE( local_2d(nxlg:nxrg,nysg:nyng) )
[1]	1158
	1159	#if defined( __netcdf )
[1327]	1160	IF ( two_d ) THEN
	1161	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1162	id_var_do2d(av,if), &
	1163	total_2d(0:nx+1,0:ny+1), &
	1164	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
	1165	count = (/ nx+2, ny+2, 1, 1 /) )
	1166	ELSE
	1167	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1168	id_var_do2d(av,if), &
	1169	total_2d(0:nx+1,0:ny+1), &
	1170	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
	1171	count = (/ nx+2, ny+2, 1, 1 /) )
[1]	1172	ENDIF
[1783]	1173	CALL netcdf_handle_error( 'data_output_2d', 54 )
[1]	1174	#endif
	1175
[493]	1176	ELSE
[1]	1177	!
[493]	1178	!-- First send the local index limits to PE0
[667]	1179	ind(1) = nxlg; ind(2) = nxrg
	1180	ind(3) = nysg; ind(4) = nyng
[1320]	1181	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1182	comm2d, ierr )
[1]	1183	!
[493]	1184	!-- Send data to PE0
[1320]	1185	CALL MPI_SEND( local_2d(nxlg,nysg), ngp, &
[493]	1186	MPI_REAL, 0, 1, comm2d, ierr )
	1187	ENDIF
	1188	!
	1189	!-- A barrier has to be set, because otherwise some PEs may
	1190	!-- proceed too fast so that PE0 may receive wrong data on
	1191	!-- tag 0
	1192	CALL MPI_BARRIER( comm2d, ierr )
[1]	1193	ENDIF
[493]	1194
[1]	1195	ENDIF
	1196	#else
	1197	#if defined( __netcdf )
[1327]	1198	IF ( two_d ) THEN
	1199	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1200	id_var_do2d(av,if), &
	1201	local_2d(nxl:nxr+1,nys:nyn+1), &
	1202	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
	1203	count = (/ nx+2, ny+2, 1, 1 /) )
	1204	ELSE
	1205	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1206	id_var_do2d(av,if), &
	1207	local_2d(nxl:nxr+1,nys:nyn+1), &
	1208	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
	1209	count = (/ nx+2, ny+2, 1, 1 /) )
[1]	1210	ENDIF
[1783]	1211	CALL netcdf_handle_error( 'data_output_2d', 447 )
[1]	1212	#endif
	1213	#endif
	1214	do2d_xy_n = do2d_xy_n + 1
	1215	!
	1216	!-- For 2D-arrays (e.g. u*) only one cross-section is available.
	1217	!-- Hence exit loop of output levels.
	1218	IF ( two_d ) THEN
[1703]	1219	IF ( netcdf_data_format < 5 ) two_d = .FALSE.
[1]	1220	EXIT loop1
	1221	ENDIF
	1222
	1223	CASE ( 'xz' )
	1224	!
[1308]	1225	!-- Update the netCDF xz cross section time axis.
	1226	!-- In case of parallel output, this is only done by PE0
	1227	!-- to increase the performance.
	1228	IF ( simulated_time /= do2d_xz_last_time(av) ) THEN
	1229	do2d_xz_time_count(av) = do2d_xz_time_count(av) + 1
	1230	do2d_xz_last_time(av) = simulated_time
	1231	IF ( myid == 0 ) THEN
[1327]	1232	IF ( .NOT. data_output_2d_on_each_pe &
	1233	.OR. netcdf_data_format > 4 ) &
[493]	1234	THEN
[1]	1235	#if defined( __netcdf )
	1236	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1237	id_var_time_xz(av), &
[291]	1238	(/ time_since_reference_point /), &
[1]	1239	start = (/ do2d_xz_time_count(av) /), &
	1240	count = (/ 1 /) )
[1783]	1241	CALL netcdf_handle_error( 'data_output_2d', 56 )
[1]	1242	#endif
	1243	ENDIF
	1244	ENDIF
	1245	ENDIF
[667]	1246
[1]	1247	!
	1248	!-- If required, carry out averaging along y
[1960]	1249	IF ( section(is,s_ind) == -1 ) THEN
[1]	1250
[1551]	1251	ALLOCATE( local_2d_l(nxlg:nxrg,nzb_do:nzt_do) )
[1353]	1252	local_2d_l = 0.0_wp
[1551]	1253	ngp = ( nxrg-nxlg + 1 ) * ( nzt_do-nzb_do + 1 )
[1]	1254	!
	1255	!-- First local averaging on the PE
[1551]	1256	DO k = nzb_do, nzt_do
[1]	1257	DO j = nys, nyn
[667]	1258	DO i = nxlg, nxrg
[1320]	1259	local_2d_l(i,k) = local_2d_l(i,k) + &
[1]	1260	local_pf(i,j,k)
	1261	ENDDO
	1262	ENDDO
	1263	ENDDO
	1264	#if defined( __parallel )
	1265	!
	1266	!-- Now do the averaging over all PEs along y
[622]	1267	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1268	CALL MPI_ALLREDUCE( local_2d_l(nxlg,nzb_do), &
	1269	local_2d(nxlg,nzb_do), ngp, MPI_REAL, &
[1]	1270	MPI_SUM, comm1dy, ierr )
	1271	#else
	1272	local_2d = local_2d_l
	1273	#endif
[1353]	1274	local_2d = local_2d / ( ny + 1.0_wp )
[1]	1275
	1276	DEALLOCATE( local_2d_l )
	1277
	1278	ELSE
	1279	!
	1280	!-- Just store the respective section on the local array
	1281	!-- (but only if it is available on this PE!)
[1960]	1282	IF ( section(is,s_ind) >= nys .AND. section(is,s_ind) <= nyn ) &
[1]	1283	THEN
[1960]	1284	local_2d = local_pf(:,section(is,s_ind),nzb_do:nzt_do)
[1]	1285	ENDIF
	1286
	1287	ENDIF
	1288
	1289	#if defined( __parallel )
[1327]	1290	IF ( netcdf_data_format > 4 ) THEN
[1]	1291	!
[1031]	1292	!-- Output in netCDF4/HDF5 format.
[493]	1293	!-- Output only on those PEs where the respective cross
	1294	!-- sections reside. Cross sections averaged along y are
	1295	!-- output on the respective first PE along y (myidy=0).
[1960]	1296	IF ( ( section(is,s_ind) >= nys .AND. &
	1297	section(is,s_ind) <= nyn ) .OR. &
	1298	( section(is,s_ind) == -1 .AND. myidy == 0 ) ) THEN
[1]	1299	#if defined( __netcdf )
[493]	1300	!
[1308]	1301	!-- For parallel output, all cross sections are first
	1302	!-- stored here on a local array and will be written to the
	1303	!-- output file afterwards to increase the performance.
	1304	DO i = nxlg, nxrg
[1551]	1305	DO k = nzb_do, nzt_do
[1308]	1306	local_2d_sections_l(i,is,k) = local_2d(i,k)
	1307	ENDDO
	1308	ENDDO
[1]	1309	#endif
	1310	ENDIF
	1311
	1312	ELSE
	1313
[493]	1314	IF ( data_output_2d_on_each_pe ) THEN
[1]	1315	!
[493]	1316	!-- Output of partial arrays on each PE. If the cross
	1317	!-- section does not reside on the PE, output special
	1318	!-- index values.
	1319	#if defined( __netcdf )
[1327]	1320	IF ( myid == 0 ) THEN
[1320]	1321	WRITE ( 22 ) time_since_reference_point, &
[493]	1322	do2d_xz_time_count(av), av
	1323	ENDIF
	1324	#endif
[759]	1325	DO i = 0, io_blocks-1
	1326	IF ( i == io_group ) THEN
[1960]	1327	IF ( ( section(is,s_ind) >= nys .AND. &
	1328	section(is,s_ind) <= nyn ) .OR. &
	1329	( section(is,s_ind) == -1 .AND. &
[1320]	1330	nys-1 == -1 ) ) &
[759]	1331	THEN
[1551]	1332	WRITE (22) nxlg, nxrg, nzb_do, nzt_do, nzb, nzt+1
[759]	1333	WRITE (22) local_2d
	1334	ELSE
[1551]	1335	WRITE (22) -1, -1, -1, -1, -1, -1
[759]	1336	ENDIF
	1337	ENDIF
	1338	#if defined( __parallel )
	1339	CALL MPI_BARRIER( comm2d, ierr )
	1340	#endif
	1341	ENDDO
[493]	1342
	1343	ELSE
[1]	1344	!
[493]	1345	!-- PE0 receives partial arrays from all processors of the
	1346	!-- respective cross section and outputs them. Here a
	1347	!-- barrier has to be set, because otherwise
	1348	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
	1349	CALL MPI_BARRIER( comm2d, ierr )
	1350
[1551]	1351	ngp = ( nxrg-nxlg + 1 ) * ( nzt_do-nzb_do + 1 )
[493]	1352	IF ( myid == 0 ) THEN
[1]	1353	!
[493]	1354	!-- Local array can be relocated directly.
[1960]	1355	IF ( ( section(is,s_ind) >= nys .AND. &
	1356	section(is,s_ind) <= nyn ) .OR. &
	1357	( section(is,s_ind) == -1 .AND. &
	1358	nys-1 == -1 ) ) THEN
[1551]	1359	total_2d(nxlg:nxrg,nzb_do:nzt_do) = local_2d
[493]	1360	ENDIF
[1]	1361	!
[493]	1362	!-- Receive data from all other PEs.
	1363	DO n = 1, numprocs-1
	1364	!
	1365	!-- Receive index limits first, then array.
	1366	!-- Index limits are received in arbitrary order from
	1367	!-- the PEs.
[1320]	1368	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1369	MPI_ANY_SOURCE, 0, comm2d, &
[1]	1370	status, ierr )
[493]	1371	!
	1372	!-- Not all PEs have data for XZ-cross-section.
	1373	IF ( ind(1) /= -9999 ) THEN
	1374	sender = status(MPI_SOURCE)
	1375	DEALLOCATE( local_2d )
[1320]	1376	ALLOCATE( local_2d(ind(1):ind(2), &
[493]	1377	ind(3):ind(4)) )
	1378	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1379	MPI_REAL, sender, 1, comm2d, &
	1380	status, ierr )
[1320]	1381	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
[493]	1382	local_2d
	1383	ENDIF
	1384	ENDDO
	1385	!
	1386	!-- Relocate the local array for the next loop increment
	1387	DEALLOCATE( local_2d )
[1551]	1388	ALLOCATE( local_2d(nxlg:nxrg,nzb_do:nzt_do) )
[1]	1389
	1390	#if defined( __netcdf )
[1327]	1391	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1392	id_var_do2d(av,if), &
[1551]	1393	total_2d(0:nx+1,nzb_do:nzt_do),&
[1327]	1394	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
[1551]	1395	count = (/ nx+2, 1, nzt_do-nzb_do+1, 1 /) )
[1783]	1396	CALL netcdf_handle_error( 'data_output_2d', 58 )
[1]	1397	#endif
	1398
[493]	1399	ELSE
[1]	1400	!
[493]	1401	!-- If the cross section resides on the PE, send the
	1402	!-- local index limits, otherwise send -9999 to PE0.
[1960]	1403	IF ( ( section(is,s_ind) >= nys .AND. &
	1404	section(is,s_ind) <= nyn ) .OR. &
	1405	( section(is,s_ind) == -1 .AND. nys-1 == -1 ) ) &
[493]	1406	THEN
[667]	1407	ind(1) = nxlg; ind(2) = nxrg
[1551]	1408	ind(3) = nzb_do; ind(4) = nzt_do
[493]	1409	ELSE
	1410	ind(1) = -9999; ind(2) = -9999
	1411	ind(3) = -9999; ind(4) = -9999
	1412	ENDIF
[1320]	1413	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1414	comm2d, ierr )
	1415	!
	1416	!-- If applicable, send data to PE0.
	1417	IF ( ind(1) /= -9999 ) THEN
[1551]	1418	CALL MPI_SEND( local_2d(nxlg,nzb_do), ngp, &
[493]	1419	MPI_REAL, 0, 1, comm2d, ierr )
	1420	ENDIF
[1]	1421	ENDIF
	1422	!
[493]	1423	!-- A barrier has to be set, because otherwise some PEs may
	1424	!-- proceed too fast so that PE0 may receive wrong data on
	1425	!-- tag 0
	1426	CALL MPI_BARRIER( comm2d, ierr )
[1]	1427	ENDIF
[493]	1428
[1]	1429	ENDIF
	1430	#else
	1431	#if defined( __netcdf )
[1327]	1432	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1433	id_var_do2d(av,if), &
[1551]	1434	local_2d(nxl:nxr+1,nzb_do:nzt_do), &
[1327]	1435	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
[1551]	1436	count = (/ nx+2, 1, nzt_do-nzb_do+1, 1 /) )
[1783]	1437	CALL netcdf_handle_error( 'data_output_2d', 451 )
[1]	1438	#endif
	1439	#endif
	1440	do2d_xz_n = do2d_xz_n + 1
	1441
	1442	CASE ( 'yz' )
	1443	!
[1308]	1444	!-- Update the netCDF yz cross section time axis.
	1445	!-- In case of parallel output, this is only done by PE0
	1446	!-- to increase the performance.
	1447	IF ( simulated_time /= do2d_yz_last_time(av) ) THEN
	1448	do2d_yz_time_count(av) = do2d_yz_time_count(av) + 1
	1449	do2d_yz_last_time(av) = simulated_time
	1450	IF ( myid == 0 ) THEN
[1327]	1451	IF ( .NOT. data_output_2d_on_each_pe &
	1452	.OR. netcdf_data_format > 4 ) &
[493]	1453	THEN
[1]	1454	#if defined( __netcdf )
	1455	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1456	id_var_time_yz(av), &
[291]	1457	(/ time_since_reference_point /), &
[1]	1458	start = (/ do2d_yz_time_count(av) /), &
	1459	count = (/ 1 /) )
[1783]	1460	CALL netcdf_handle_error( 'data_output_2d', 59 )
[1]	1461	#endif
	1462	ENDIF
	1463	ENDIF
[1308]	1464	ENDIF
[493]	1465
[1]	1466	!
	1467	!-- If required, carry out averaging along x
[1960]	1468	IF ( section(is,s_ind) == -1 ) THEN
[1]	1469
[1551]	1470	ALLOCATE( local_2d_l(nysg:nyng,nzb_do:nzt_do) )
[1353]	1471	local_2d_l = 0.0_wp
[1551]	1472	ngp = ( nyng-nysg+1 ) * ( nzt_do-nzb_do+1 )
[1]	1473	!
	1474	!-- First local averaging on the PE
[1551]	1475	DO k = nzb_do, nzt_do
[667]	1476	DO j = nysg, nyng
[1]	1477	DO i = nxl, nxr
[1320]	1478	local_2d_l(j,k) = local_2d_l(j,k) + &
[1]	1479	local_pf(i,j,k)
	1480	ENDDO
	1481	ENDDO
	1482	ENDDO
	1483	#if defined( __parallel )
	1484	!
	1485	!-- Now do the averaging over all PEs along x
[622]	1486	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1487	CALL MPI_ALLREDUCE( local_2d_l(nysg,nzb_do), &
	1488	local_2d(nysg,nzb_do), ngp, MPI_REAL, &
[1]	1489	MPI_SUM, comm1dx, ierr )
	1490	#else
	1491	local_2d = local_2d_l
	1492	#endif
[1353]	1493	local_2d = local_2d / ( nx + 1.0_wp )
[1]	1494
	1495	DEALLOCATE( local_2d_l )
	1496
	1497	ELSE
	1498	!
	1499	!-- Just store the respective section on the local array
	1500	!-- (but only if it is available on this PE!)
[1960]	1501	IF ( section(is,s_ind) >= nxl .AND. section(is,s_ind) <= nxr ) &
[1]	1502	THEN
[1960]	1503	local_2d = local_pf(section(is,s_ind),:,nzb_do:nzt_do)
[1]	1504	ENDIF
	1505
	1506	ENDIF
	1507
	1508	#if defined( __parallel )
[1327]	1509	IF ( netcdf_data_format > 4 ) THEN
[1]	1510	!
[1031]	1511	!-- Output in netCDF4/HDF5 format.
[493]	1512	!-- Output only on those PEs where the respective cross
	1513	!-- sections reside. Cross sections averaged along x are
	1514	!-- output on the respective first PE along x (myidx=0).
[1960]	1515	IF ( ( section(is,s_ind) >= nxl .AND. &
	1516	section(is,s_ind) <= nxr ) .OR. &
	1517	( section(is,s_ind) == -1 .AND. myidx == 0 ) ) THEN
[1]	1518	#if defined( __netcdf )
[493]	1519	!
[1308]	1520	!-- For parallel output, all cross sections are first
	1521	!-- stored here on a local array and will be written to the
	1522	!-- output file afterwards to increase the performance.
	1523	DO j = nysg, nyng
[1551]	1524	DO k = nzb_do, nzt_do
[1308]	1525	local_2d_sections_l(is,j,k) = local_2d(j,k)
	1526	ENDDO
	1527	ENDDO
[1]	1528	#endif
	1529	ENDIF
	1530
	1531	ELSE
	1532
[493]	1533	IF ( data_output_2d_on_each_pe ) THEN
[1]	1534	!
[493]	1535	!-- Output of partial arrays on each PE. If the cross
	1536	!-- section does not reside on the PE, output special
	1537	!-- index values.
	1538	#if defined( __netcdf )
[1327]	1539	IF ( myid == 0 ) THEN
[1320]	1540	WRITE ( 23 ) time_since_reference_point, &
[493]	1541	do2d_yz_time_count(av), av
	1542	ENDIF
	1543	#endif
[759]	1544	DO i = 0, io_blocks-1
	1545	IF ( i == io_group ) THEN
[1960]	1546	IF ( ( section(is,s_ind) >= nxl .AND. &
	1547	section(is,s_ind) <= nxr ) .OR. &
	1548	( section(is,s_ind) == -1 .AND. &
[1320]	1549	nxl-1 == -1 ) ) &
[759]	1550	THEN
[1551]	1551	WRITE (23) nysg, nyng, nzb_do, nzt_do, nzb, nzt+1
[759]	1552	WRITE (23) local_2d
	1553	ELSE
[1551]	1554	WRITE (23) -1, -1, -1, -1, -1, -1
[759]	1555	ENDIF
	1556	ENDIF
	1557	#if defined( __parallel )
	1558	CALL MPI_BARRIER( comm2d, ierr )
	1559	#endif
	1560	ENDDO
[493]	1561
	1562	ELSE
[1]	1563	!
[493]	1564	!-- PE0 receives partial arrays from all processors of the
	1565	!-- respective cross section and outputs them. Here a
	1566	!-- barrier has to be set, because otherwise
	1567	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
	1568	CALL MPI_BARRIER( comm2d, ierr )
	1569
[1551]	1570	ngp = ( nyng-nysg+1 ) * ( nzt_do-nzb_do+1 )
[493]	1571	IF ( myid == 0 ) THEN
[1]	1572	!
[493]	1573	!-- Local array can be relocated directly.
[1960]	1574	IF ( ( section(is,s_ind) >= nxl .AND. &
	1575	section(is,s_ind) <= nxr ) .OR. &
	1576	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
[493]	1577	THEN
[1551]	1578	total_2d(nysg:nyng,nzb_do:nzt_do) = local_2d
[493]	1579	ENDIF
[1]	1580	!
[493]	1581	!-- Receive data from all other PEs.
	1582	DO n = 1, numprocs-1
	1583	!
	1584	!-- Receive index limits first, then array.
	1585	!-- Index limits are received in arbitrary order from
	1586	!-- the PEs.
[1320]	1587	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1588	MPI_ANY_SOURCE, 0, comm2d, &
[1]	1589	status, ierr )
[493]	1590	!
	1591	!-- Not all PEs have data for YZ-cross-section.
	1592	IF ( ind(1) /= -9999 ) THEN
	1593	sender = status(MPI_SOURCE)
	1594	DEALLOCATE( local_2d )
[1320]	1595	ALLOCATE( local_2d(ind(1):ind(2), &
[493]	1596	ind(3):ind(4)) )
	1597	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1598	MPI_REAL, sender, 1, comm2d, &
	1599	status, ierr )
[1320]	1600	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
[493]	1601	local_2d
	1602	ENDIF
	1603	ENDDO
	1604	!
	1605	!-- Relocate the local array for the next loop increment
	1606	DEALLOCATE( local_2d )
[1551]	1607	ALLOCATE( local_2d(nysg:nyng,nzb_do:nzt_do) )
[1]	1608
	1609	#if defined( __netcdf )
[1327]	1610	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1611	id_var_do2d(av,if), &
[1551]	1612	total_2d(0:ny+1,nzb_do:nzt_do),&
[1327]	1613	start = (/ is, 1, 1, do2d_yz_time_count(av) /), &
[1551]	1614	count = (/ 1, ny+2, nzt_do-nzb_do+1, 1 /) )
[1783]	1615	CALL netcdf_handle_error( 'data_output_2d', 61 )
[1]	1616	#endif
	1617
[493]	1618	ELSE
[1]	1619	!
[493]	1620	!-- If the cross section resides on the PE, send the
	1621	!-- local index limits, otherwise send -9999 to PE0.
[1960]	1622	IF ( ( section(is,s_ind) >= nxl .AND. &
	1623	section(is,s_ind) <= nxr ) .OR. &
	1624	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
[493]	1625	THEN
[667]	1626	ind(1) = nysg; ind(2) = nyng
[1551]	1627	ind(3) = nzb_do; ind(4) = nzt_do
[493]	1628	ELSE
	1629	ind(1) = -9999; ind(2) = -9999
	1630	ind(3) = -9999; ind(4) = -9999
	1631	ENDIF
[1320]	1632	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1633	comm2d, ierr )
	1634	!
	1635	!-- If applicable, send data to PE0.
	1636	IF ( ind(1) /= -9999 ) THEN
[1551]	1637	CALL MPI_SEND( local_2d(nysg,nzb_do), ngp, &
[493]	1638	MPI_REAL, 0, 1, comm2d, ierr )
	1639	ENDIF
[1]	1640	ENDIF
	1641	!
[493]	1642	!-- A barrier has to be set, because otherwise some PEs may
	1643	!-- proceed too fast so that PE0 may receive wrong data on
	1644	!-- tag 0
	1645	CALL MPI_BARRIER( comm2d, ierr )
[1]	1646	ENDIF
[493]	1647
[1]	1648	ENDIF
	1649	#else
	1650	#if defined( __netcdf )
[1327]	1651	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1652	id_var_do2d(av,if), &
[1551]	1653	local_2d(nys:nyn+1,nzb_do:nzt_do), &
[1327]	1654	start = (/ is, 1, 1, do2d_xz_time_count(av) /), &
[1551]	1655	count = (/ 1, ny+2, nzt_do-nzb_do+1, 1 /) )
[1783]	1656	CALL netcdf_handle_error( 'data_output_2d', 452 )
[1]	1657	#endif
	1658	#endif
	1659	do2d_yz_n = do2d_yz_n + 1
	1660
	1661	END SELECT
	1662
	1663	is = is + 1
	1664	ENDDO loop1
	1665
[1308]	1666	!
	1667	!-- For parallel output, all data were collected before on a local array
	1668	!-- and are written now to the netcdf file. This must be done to increase
	1669	!-- the performance of the parallel output.
	1670	#if defined( __netcdf )
[1327]	1671	IF ( netcdf_data_format > 4 ) THEN
[1308]	1672
	1673	SELECT CASE ( mode )
	1674
	1675	CASE ( 'xy' )
	1676	IF ( two_d ) THEN
[1703]	1677	nis = 1
	1678	two_d = .FALSE.
[1308]	1679	ELSE
[1703]	1680	nis = ns
[1308]	1681	ENDIF
	1682	!
	1683	!-- Do not output redundant ghost point data except for the
	1684	!-- boundaries of the total domain.
	1685	IF ( nxr == nx .AND. nyn /= ny ) THEN
	1686	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1687	id_var_do2d(av,if), &
	1688	local_2d_sections(nxl:nxr+1, &
[1703]	1689	nys:nyn,1:nis), &
[1308]	1690	start = (/ nxl+1, nys+1, 1, &
	1691	do2d_xy_time_count(av) /), &
	1692	count = (/ nxr-nxl+2, &
[1703]	1693	nyn-nys+1, nis, 1 &
[1308]	1694	/) )
	1695	ELSEIF ( 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, &
[1703]	1699	nys:nyn+1,1:nis), &
[1308]	1700	start = (/ nxl+1, nys+1, 1, &
	1701	do2d_xy_time_count(av) /), &
	1702	count = (/ nxr-nxl+1, &
[1703]	1703	nyn-nys+2, 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+1, &
[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+2, &
[1703]	1713	nyn-nys+2, nis, 1 &
[1308]	1714	/) )
	1715	ELSE
	1716	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1717	id_var_do2d(av,if), &
	1718	local_2d_sections(nxl:nxr, &
[1703]	1719	nys:nyn,1:nis), &
[1308]	1720	start = (/ nxl+1, nys+1, 1, &
	1721	do2d_xy_time_count(av) /), &
	1722	count = (/ nxr-nxl+1, &
[1703]	1723	nyn-nys+1, nis, 1 &
[1308]	1724	/) )
	1725	ENDIF
	1726
[1783]	1727	CALL netcdf_handle_error( 'data_output_2d', 55 )
[1308]	1728
	1729	CASE ( 'xz' )
	1730	!
	1731	!-- First, all PEs get the information of all cross-sections.
	1732	!-- Then the data are written to the output file by all PEs
	1733	!-- while NF90_COLLECTIVE is set in subroutine
	1734	!-- define_netcdf_header. Although redundant information are
	1735	!-- written to the output file in that case, the performance
	1736	!-- is significantly better compared to the case where only
	1737	!-- the first row of PEs in x-direction (myidx = 0) is given
	1738	!-- the output while NF90_INDEPENDENT is set.
	1739	IF ( npey /= 1 ) THEN
	1740
	1741	#if defined( __parallel )
	1742	!
	1743	!-- Distribute data over all PEs along y
[1551]	1744	ngp = ( nxrg-nxlg+1 ) * ( nzt_do-nzb_do+1 ) * ns
[1308]	1745	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1746	CALL MPI_ALLREDUCE( local_2d_sections_l(nxlg,1,nzb_do), &
	1747	local_2d_sections(nxlg,1,nzb_do), &
[1308]	1748	ngp, MPI_REAL, MPI_SUM, comm1dy, &
	1749	ierr )
	1750	#else
	1751	local_2d_sections = local_2d_sections_l
	1752	#endif
	1753	ENDIF
	1754	!
	1755	!-- Do not output redundant ghost point data except for the
	1756	!-- boundaries of the total domain.
	1757	IF ( nxr == nx ) THEN
	1758	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1759	id_var_do2d(av,if), &
	1760	local_2d_sections(nxl:nxr+1,1:ns, &
[1551]	1761	nzb_do:nzt_do), &
[1308]	1762	start = (/ nxl+1, 1, 1, &
	1763	do2d_xz_time_count(av) /), &
[1551]	1764	count = (/ nxr-nxl+2, ns, nzt_do-nzb_do+1, &
[1308]	1765	1 /) )
	1766	ELSE
	1767	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1768	id_var_do2d(av,if), &
	1769	local_2d_sections(nxl:nxr,1:ns, &
[1551]	1770	nzb_do:nzt_do), &
[1308]	1771	start = (/ nxl+1, 1, 1, &
	1772	do2d_xz_time_count(av) /), &
[1551]	1773	count = (/ nxr-nxl+1, ns, nzt_do-nzb_do+1, &
[1308]	1774	1 /) )
	1775	ENDIF
	1776
[1783]	1777	CALL netcdf_handle_error( 'data_output_2d', 57 )
[1308]	1778
	1779	CASE ( 'yz' )
	1780	!
	1781	!-- First, all PEs get the information of all cross-sections.
	1782	!-- Then the data are written to the output file by all PEs
	1783	!-- while NF90_COLLECTIVE is set in subroutine
	1784	!-- define_netcdf_header. Although redundant information are
	1785	!-- written to the output file in that case, the performance
	1786	!-- is significantly better compared to the case where only
	1787	!-- the first row of PEs in y-direction (myidy = 0) is given
	1788	!-- the output while NF90_INDEPENDENT is set.
	1789	IF ( npex /= 1 ) THEN
	1790
	1791	#if defined( __parallel )
	1792	!
	1793	!-- Distribute data over all PEs along x
	1794	ngp = ( nyng-nysg+1 ) * ( nzt-nzb + 2 ) * ns
	1795	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1796	CALL MPI_ALLREDUCE( local_2d_sections_l(1,nysg,nzb_do), &
	1797	local_2d_sections(1,nysg,nzb_do), &
[1308]	1798	ngp, MPI_REAL, MPI_SUM, comm1dx, &
	1799	ierr )
	1800	#else
	1801	local_2d_sections = local_2d_sections_l
	1802	#endif
	1803	ENDIF
	1804	!
	1805	!-- Do not output redundant ghost point data except for the
	1806	!-- boundaries of the total domain.
	1807	IF ( nyn == ny ) THEN
	1808	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1809	id_var_do2d(av,if), &
	1810	local_2d_sections(1:ns, &
[1551]	1811	nys:nyn+1,nzb_do:nzt_do), &
[1308]	1812	start = (/ 1, nys+1, 1, &
	1813	do2d_yz_time_count(av) /), &
	1814	count = (/ ns, nyn-nys+2, &
[1551]	1815	nzt_do-nzb_do+1, 1 /) )
[1308]	1816	ELSE
	1817	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1818	id_var_do2d(av,if), &
	1819	local_2d_sections(1:ns,nys:nyn, &
[1551]	1820	nzb_do:nzt_do), &
[1308]	1821	start = (/ 1, nys+1, 1, &
	1822	do2d_yz_time_count(av) /), &
	1823	count = (/ ns, nyn-nys+1, &
[1551]	1824	nzt_do-nzb_do+1, 1 /) )
[1308]	1825	ENDIF
	1826
[1783]	1827	CALL netcdf_handle_error( 'data_output_2d', 60 )
[1308]	1828
	1829	CASE DEFAULT
	1830	message_string = 'unknown cross-section: ' // TRIM( mode )
	1831	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
	1832
	1833	END SELECT
	1834
	1835	ENDIF
[1311]	1836	#endif
[1]	1837	ENDIF
	1838
	1839	if = if + 1
	1840	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
	1841	do2d_mode = do2d(av,if)(l-1:l)
	1842
	1843	ENDDO
	1844
	1845	!
	1846	!-- Deallocate temporary arrays.
	1847	IF ( ALLOCATED( level_z ) ) DEALLOCATE( level_z )
[1308]	1848	IF ( netcdf_data_format > 4 ) THEN
	1849	DEALLOCATE( local_pf, local_2d, local_2d_sections )
	1850	IF( mode == 'xz' .OR. mode == 'yz' ) DEALLOCATE( local_2d_sections_l )
	1851	ENDIF
[1]	1852	#if defined( __parallel )
	1853	IF ( .NOT. data_output_2d_on_each_pe .AND. myid == 0 ) THEN
	1854	DEALLOCATE( total_2d )
	1855	ENDIF
	1856	#endif
	1857
	1858	!
	1859	!-- Close plot output file.
[1960]	1860	file_id = 20 + s_ind
[1]	1861
	1862	IF ( data_output_2d_on_each_pe ) THEN
[759]	1863	DO i = 0, io_blocks-1
	1864	IF ( i == io_group ) THEN
	1865	CALL close_file( file_id )
	1866	ENDIF
	1867	#if defined( __parallel )
	1868	CALL MPI_BARRIER( comm2d, ierr )
	1869	#endif
	1870	ENDDO
[1]	1871	ELSE
	1872	IF ( myid == 0 ) CALL close_file( file_id )
	1873	ENDIF
	1874
[1318]	1875	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
[1]	1876
	1877	END SUBROUTINE data_output_2d

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |