Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 2742

Last change on this file since 2742 was 2742, checked in by suehring, 7 years ago
Output of surface temperature tsurf* at urban- and natural-type surfaces
Property svn:keywords set to `Id`
File size: 86.6 KB

Line
1	!> @file data_output_2d.f90
2	!------------------------------------------------------------------------------!
3	! This file is part of the PALM model system.
4	!
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.
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	!
17	! Copyright 1997-2018 Leibniz Universitaet Hannover
18	!------------------------------------------------------------------------------!
19	!
20	! Current revisions:
21	! -----------------
22	!
23	!
24	! Former revisions:
25	! -----------------
26	! $Id: data_output_2d.f90 2742 2018-01-12 14:59:47Z suehring $
27	! Enable output of surface temperature
28	!
29	! 2735 2018-01-11 12:01:27Z suehring
30	! output of r_a moved from land-surface to consider also urban-type surfaces
31	!
32	! 2718 2018-01-02 08:49:38Z maronga
33	! Corrected "Former revisions" section
34	!
35	! 2696 2017-12-14 17:12:51Z kanani
36	! Change in file header (GPL part)
37	! Implementation of uv exposure model (FK)
38	! Implementation of turbulence_closure_mod (TG)
39	! Set fill values at topography grid points or e.g. non-natural-type surface
40	! in case of LSM output (MS)
41	!
42	! 2512 2017-10-04 08:26:59Z raasch
43	! upper bounds of cross section output changed from nx+1,ny+1 to nx,ny
44	! no output of ghost layer data
45	!
46	! 2292 2017-06-20 09:51:42Z schwenkel
47	! Implementation of new microphysic scheme: cloud_scheme = 'morrison'
48	! includes two more prognostic equations for cloud drop concentration (nc)
49	! and cloud water content (qc).
50	!
51	! 2277 2017-06-12 10:47:51Z kanani
52	! Removed unused variables do2d_xy_n, do2d_xz_n, do2d_yz_n
53	!
54	! 2233 2017-05-30 18:08:54Z suehring
55	!
56	! 2232 2017-05-30 17:47:52Z suehring
57	! Adjustments to new surface concept
58	!
59	!
60	! 2190 2017-03-21 12:16:43Z raasch
61	! bugfix for r2031: string rho replaced by rho_ocean
62	!
63	! 2031 2016-10-21 15:11:58Z knoop
64	! renamed variable rho to rho_ocean and rho_av to rho_ocean_av
65	!
66	! 2000 2016-08-20 18:09:15Z knoop
67	! Forced header and separation lines into 80 columns
68	!
69	! 1980 2016-07-29 15:51:57Z suehring
70	! Bugfix, in order to steer user-defined output, setting flag found explicitly
71	! to .F.
72	!
73	! 1976 2016-07-27 13:28:04Z maronga
74	! Output of radiation quantities is now done directly in the respective module
75	!
76	! 1972 2016-07-26 07:52:02Z maronga
77	! Output of land surface quantities is now done directly in the respective
78	! module
79	!
80	! 1960 2016-07-12 16:34:24Z suehring
81	! Scalar surface flux added
82	! Rename INTEGER variable s into s_ind, as s is already assigned to scalar
83	!
84	! 1849 2016-04-08 11:33:18Z hoffmann
85	! precipitation_amount, precipitation_rate, prr moved to arrays_3d
86	!
87	! 1822 2016-04-07 07:49:42Z hoffmann
88	! Output of bulk cloud physics simplified.
89	!
90	! 1788 2016-03-10 11:01:04Z maronga
91	! Added output of z0q
92	!
93	! 1783 2016-03-06 18:36:17Z raasch
94	! name change of netcdf routines and module + related changes
95	!
96	! 1745 2016-02-05 13:06:51Z gronemeier
97	! Bugfix: test if time axis limit exceeds moved to point after call of check_open
98	!
99	! 1703 2015-11-02 12:38:44Z raasch
100	! bugfix for output of single (*) xy-sections in case of parallel netcdf I/O
101	!
102	! 1701 2015-11-02 07:43:04Z maronga
103	! Bugfix in output of RRTGM data
104	!
105	! 1691 2015-10-26 16:17:44Z maronga
106	! Added output of Obukhov length (ol) and radiative heating rates for RRTMG.
107	! Formatting corrections.
108	!
109	! 1682 2015-10-07 23:56:08Z knoop
110	! Code annotations made doxygen readable
111	!
112	! 1585 2015-04-30 07:05:52Z maronga
113	! Added support for RRTMG
114	!
115	! 1555 2015-03-04 17:44:27Z maronga
116	! Added output of r_a and r_s
117	!
118	! 1551 2015-03-03 14:18:16Z maronga
119	! Added suppport for land surface model and radiation model output. In the course
120	! of this action, the limits for vertical loops have been changed (from nzb and
121	! nzt+1 to nzb_do and nzt_do, respectively in order to allow soil model output).
122	! Moreover, a new vertical grid zs was introduced.
123	!
124	! 1359 2014-04-11 17:15:14Z hoffmann
125	! New particle structure integrated.
126	!
127	! 1353 2014-04-08 15:21:23Z heinze
128	! REAL constants provided with KIND-attribute
129	!
130	! 1327 2014-03-21 11:00:16Z raasch
131	! parts concerning iso2d output removed,
132	! -netcdf output queries
133	!
134	! 1320 2014-03-20 08:40:49Z raasch
135	! ONLY-attribute added to USE-statements,
136	! kind-parameters added to all INTEGER and REAL declaration statements,
137	! kinds are defined in new module kinds,
138	! revision history before 2012 removed,
139	! comment fields (!:) to be used for variable explanations added to
140	! all variable declaration statements
141	!
142	! 1318 2014-03-17 13:35:16Z raasch
143	! barrier argument removed from cpu_log.
144	! module interfaces removed
145	!
146	! 1311 2014-03-14 12:13:39Z heinze
147	! bugfix: close #if defined( __netcdf )
148	!
149	! 1308 2014-03-13 14:58:42Z fricke
150	! +local_2d_sections, local_2d_sections_l, ns
151	! Check, if the limit of the time dimension is exceeded for parallel output
152	! To increase the performance for parallel output, the following is done:
153	! - Update of time axis is only done by PE0
154	! - Cross sections are first stored on a local array and are written
155	! collectively to the output file by all PEs.
156	!
157	! 1115 2013-03-26 18:16:16Z hoffmann
158	! ql is calculated by calc_liquid_water_content
159	!
160	! 1076 2012-12-05 08:30:18Z hoffmann
161	! Bugfix in output of ql
162	!
163	! 1065 2012-11-22 17:42:36Z hoffmann
164	! Bugfix: Output of cross sections of ql
165	!
166	! 1053 2012-11-13 17:11:03Z hoffmann
167	! +qr, nr, qc and cross sections
168	!
169	! 1036 2012-10-22 13:43:42Z raasch
170	! code put under GPL (PALM 3.9)
171	!
172	! 1031 2012-10-19 14:35:30Z raasch
173	! netCDF4 without parallel file support implemented
174	!
175	! 1007 2012-09-19 14:30:36Z franke
176	! Bugfix: missing calculation of ql_vp added
177	!
178	! 978 2012-08-09 08:28:32Z fricke
179	! +z0h
180	!
181	! Revision 1.1 1997/08/11 06:24:09 raasch
182	! Initial revision
183	!
184	!
185	! Description:
186	! ------------
187	!> Data output of cross-sections in netCDF format or binary format
188	!> to be later converted to NetCDF by helper routine combine_plot_fields.
189	!> Attention: The position of the sectional planes is still not always computed
190	!> --------- correctly. (zu is used always)!
191	!------------------------------------------------------------------------------!
192	SUBROUTINE data_output_2d( mode, av )
193
194
195	USE arrays_3d, &
196	ONLY: dzw, e, nc, nr, p, pt, precipitation_amount, precipitation_rate,&
197	prr, q, qc, ql, ql_c, ql_v, ql_vp, qr, rho_ocean, s, sa, &
198	tend, u, v, vpt, w, zu, zw
199
200	USE averaging
201
202	USE cloud_parameters, &
203	ONLY: hyrho, l_d_cp, pt_d_t
204
205	USE control_parameters, &
206	ONLY: cloud_physics, data_output_2d_on_each_pe, data_output_xy, &
207	data_output_xz, data_output_yz, do2d, &
208	do2d_xy_last_time, do2d_xy_time_count, &
209	do2d_xz_last_time, do2d_xz_time_count, &
210	do2d_yz_last_time, do2d_yz_time_count, &
211	ibc_uv_b, io_blocks, io_group, land_surface, message_string, &
212	ntdim_2d_xy, ntdim_2d_xz, ntdim_2d_yz, &
213	psolver, section, simulated_time, simulated_time_chr, &
214	time_since_reference_point, uv_exposure
215
216	USE cpulog, &
217	ONLY: cpu_log, log_point
218
219	USE grid_variables, &
220	ONLY: dx, dy
221
222	USE indices, &
223	ONLY: nbgp, nx, nxl, nxr, ny, nyn, nys, nz, nzb, nzt, wall_flags_0
224
225	USE kinds
226
227	USE land_surface_model_mod, &
228	ONLY: lsm_data_output_2d, zs
229
230	#if defined( __netcdf )
231	USE NETCDF
232	#endif
233
234	USE netcdf_interface, &
235	ONLY: fill_value, id_set_xy, id_set_xz, id_set_yz, id_var_do2d, &
236	id_var_time_xy, id_var_time_xz, id_var_time_yz, nc_stat, &
237	netcdf_data_format, netcdf_handle_error
238
239	USE particle_attributes, &
240	ONLY: grid_particles, number_of_particles, particle_advection_start, &
241	particles, prt_count
242
243	USE pegrid
244
245	USE radiation_model_mod, &
246	ONLY: radiation, radiation_data_output_2d
247
248	USE surface_mod, &
249	ONLY: surf_def_h, surf_lsm_h, surf_usm_h
250
251	USE turbulence_closure_mod, &
252	ONLY: tcm_data_output_2d
253
254	USE uv_exposure_model_mod, &
255	ONLY: uvem_data_output_2d
256
257
258	IMPLICIT NONE
259
260	CHARACTER (LEN=2) :: do2d_mode !<
261	CHARACTER (LEN=2) :: mode !<
262	CHARACTER (LEN=4) :: grid !<
263	CHARACTER (LEN=25) :: section_chr !<
264	CHARACTER (LEN=50) :: rtext !<
265
266	INTEGER(iwp) :: av !<
267	INTEGER(iwp) :: ngp !<
268	INTEGER(iwp) :: file_id !<
269	INTEGER(iwp) :: flag_nr !< number of masking flag
270	INTEGER(iwp) :: i !<
271	INTEGER(iwp) :: if !<
272	INTEGER(iwp) :: is !<
273	INTEGER(iwp) :: iis !<
274	INTEGER(iwp) :: j !<
275	INTEGER(iwp) :: k !<
276	INTEGER(iwp) :: l !<
277	INTEGER(iwp) :: layer_xy !<
278	INTEGER(iwp) :: m !<
279	INTEGER(iwp) :: n !<
280	INTEGER(iwp) :: nis !<
281	INTEGER(iwp) :: ns !<
282	INTEGER(iwp) :: nzb_do !< lower limit of the data field (usually nzb)
283	INTEGER(iwp) :: nzt_do !< upper limit of the data field (usually nzt+1)
284	INTEGER(iwp) :: psi !<
285	INTEGER(iwp) :: s_ind !<
286	INTEGER(iwp) :: sender !<
287	INTEGER(iwp) :: ind(4) !<
288
289	LOGICAL :: found !<
290	LOGICAL :: resorted !<
291	LOGICAL :: two_d !<
292
293	REAL(wp) :: mean_r !<
294	REAL(wp) :: s_r2 !<
295	REAL(wp) :: s_r3 !<
296
297	REAL(wp), DIMENSION(:), ALLOCATABLE :: level_z !<
298	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d !<
299	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d_l !<
300
301	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !<
302	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections !<
303	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections_l !<
304
305	#if defined( __parallel )
306	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: total_2d !<
307	#endif
308	REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !<
309
310	NAMELIST /LOCAL/ rtext
311
312	!
313	!-- Immediate return, if no output is requested (no respective sections
314	!-- found in parameter data_output)
315	IF ( mode == 'xy' .AND. .NOT. data_output_xy(av) ) RETURN
316	IF ( mode == 'xz' .AND. .NOT. data_output_xz(av) ) RETURN
317	IF ( mode == 'yz' .AND. .NOT. data_output_yz(av) ) RETURN
318
319	CALL cpu_log (log_point(3),'data_output_2d','start')
320
321	two_d = .FALSE. ! local variable to distinguish between output of pure 2D
322	! arrays and cross-sections of 3D arrays.
323
324	!
325	!-- Depending on the orientation of the cross-section, the respective output
326	!-- files have to be opened.
327	SELECT CASE ( mode )
328
329	CASE ( 'xy' )
330	s_ind = 1
331	ALLOCATE( level_z(nzb:nzt+1), local_2d(nxl:nxr,nys:nyn) )
332
333	IF ( netcdf_data_format > 4 ) THEN
334	ns = 1
335	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
336	ns = ns + 1
337	ENDDO
338	ns = ns - 1
339	ALLOCATE( local_2d_sections(nxl:nxr,nys:nyn,1:ns) )
340	local_2d_sections = 0.0_wp
341	ENDIF
342
343	!
344	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
345	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
346	CALL check_open( 101+av*10 )
347	ENDIF
348	IF ( data_output_2d_on_each_pe ) THEN
349	CALL check_open( 21 )
350	ELSE
351	IF ( myid == 0 ) THEN
352	#if defined( __parallel )
353	ALLOCATE( total_2d(0:nx,0:ny) )
354	#endif
355	ENDIF
356	ENDIF
357
358	CASE ( 'xz' )
359	s_ind = 2
360	ALLOCATE( local_2d(nxl:nxr,nzb:nzt+1) )
361
362	IF ( netcdf_data_format > 4 ) THEN
363	ns = 1
364	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
365	ns = ns + 1
366	ENDDO
367	ns = ns - 1
368	ALLOCATE( local_2d_sections(nxl:nxr,1:ns,nzb:nzt+1) )
369	ALLOCATE( local_2d_sections_l(nxl:nxr,1:ns,nzb:nzt+1) )
370	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
371	ENDIF
372
373	!
374	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
375	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
376	CALL check_open( 102+av*10 )
377	ENDIF
378
379	IF ( data_output_2d_on_each_pe ) THEN
380	CALL check_open( 22 )
381	ELSE
382	IF ( myid == 0 ) THEN
383	#if defined( __parallel )
384	ALLOCATE( total_2d(0:nx,nzb:nzt+1) )
385	#endif
386	ENDIF
387	ENDIF
388
389	CASE ( 'yz' )
390	s_ind = 3
391	ALLOCATE( local_2d(nys:nyn,nzb:nzt+1) )
392
393	IF ( netcdf_data_format > 4 ) THEN
394	ns = 1
395	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
396	ns = ns + 1
397	ENDDO
398	ns = ns - 1
399	ALLOCATE( local_2d_sections(1:ns,nys:nyn,nzb:nzt+1) )
400	ALLOCATE( local_2d_sections_l(1:ns,nys:nyn,nzb:nzt+1) )
401	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
402	ENDIF
403
404	!
405	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
406	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
407	CALL check_open( 103+av*10 )
408	ENDIF
409
410	IF ( data_output_2d_on_each_pe ) THEN
411	CALL check_open( 23 )
412	ELSE
413	IF ( myid == 0 ) THEN
414	#if defined( __parallel )
415	ALLOCATE( total_2d(0:ny,nzb:nzt+1) )
416	#endif
417	ENDIF
418	ENDIF
419
420	CASE DEFAULT
421	message_string = 'unknown cross-section: ' // TRIM( mode )
422	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
423
424	END SELECT
425
426	!
427	!-- For parallel netcdf output the time axis must be limited. Return, if this
428	!-- limit is exceeded. This could be the case, if the simulated time exceeds
429	!-- the given end time by the length of the given output interval.
430	IF ( netcdf_data_format > 4 ) THEN
431	IF ( mode == 'xy' .AND. do2d_xy_time_count(av) + 1 > &
432	ntdim_2d_xy(av) ) THEN
433	WRITE ( message_string, * ) 'Output of xy cross-sections is not ', &
434	'given at t=', simulated_time, '&because the', &
435	' maximum number of output time levels is exceeded.'
436	CALL message( 'data_output_2d', 'PA0384', 0, 1, 0, 6, 0 )
437	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
438	RETURN
439	ENDIF
440	IF ( mode == 'xz' .AND. do2d_xz_time_count(av) + 1 > &
441	ntdim_2d_xz(av) ) THEN
442	WRITE ( message_string, * ) 'Output of xz cross-sections is not ', &
443	'given at t=', simulated_time, '&because the', &
444	' maximum number of output time levels is exceeded.'
445	CALL message( 'data_output_2d', 'PA0385', 0, 1, 0, 6, 0 )
446	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
447	RETURN
448	ENDIF
449	IF ( mode == 'yz' .AND. do2d_yz_time_count(av) + 1 > &
450	ntdim_2d_yz(av) ) THEN
451	WRITE ( message_string, * ) 'Output of yz cross-sections is not ', &
452	'given at t=', simulated_time, '&because the', &
453	' maximum number of output time levels is exceeded.'
454	CALL message( 'data_output_2d', 'PA0386', 0, 1, 0, 6, 0 )
455	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
456	RETURN
457	ENDIF
458	ENDIF
459
460	!
461	!-- Allocate a temporary array for resorting (kji -> ijk).
462	ALLOCATE( local_pf(nxl:nxr,nys:nyn,nzb:nzt+1) )
463	local_pf = 0.0
464
465	!
466	!-- Loop of all variables to be written.
467	!-- Output dimensions chosen
468	if = 1
469	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
470	do2d_mode = do2d(av,if)(l-1:l)
471
472	DO WHILE ( do2d(av,if)(1:1) /= ' ' )
473
474	IF ( do2d_mode == mode ) THEN
475	!
476	!-- Set flag to steer output of radiation, land-surface, or user-defined
477	!-- quantities
478	found = .FALSE.
479
480	nzb_do = nzb
481	nzt_do = nzt+1
482	!
483	!-- Before each output, set array local_pf to fill value
484	local_pf = fill_value
485	!
486	!-- Set masking flag for topography for not resorted arrays
487	flag_nr = 0
488
489	!
490	!-- Store the array chosen on the temporary array.
491	resorted = .FALSE.
492	SELECT CASE ( TRIM( do2d(av,if) ) )
493	CASE ( 'e_xy', 'e_xz', 'e_yz' )
494	IF ( av == 0 ) THEN
495	to_be_resorted => e
496	ELSE
497	to_be_resorted => e_av
498	ENDIF
499	IF ( mode == 'xy' ) level_z = zu
500
501	CASE ( 'lpt_xy', 'lpt_xz', 'lpt_yz' )
502	IF ( av == 0 ) THEN
503	to_be_resorted => pt
504	ELSE
505	to_be_resorted => lpt_av
506	ENDIF
507	IF ( mode == 'xy' ) level_z = zu
508
509	CASE ( 'lwp*_xy' ) ! 2d-array
510	IF ( av == 0 ) THEN
511	DO i = nxl, nxr
512	DO j = nys, nyn
513	local_pf(i,j,nzb+1) = SUM( ql(nzb:nzt,j,i) * &
514	dzw(1:nzt+1) )
515	ENDDO
516	ENDDO
517	ELSE
518	DO i = nxl, nxr
519	DO j = nys, nyn
520	local_pf(i,j,nzb+1) = lwp_av(j,i)
521	ENDDO
522	ENDDO
523	ENDIF
524	resorted = .TRUE.
525	two_d = .TRUE.
526	level_z(nzb+1) = zu(nzb+1)
527
528	CASE ( 'nc_xy', 'nc_xz', 'nc_yz' )
529	IF ( av == 0 ) THEN
530	to_be_resorted => nc
531	ELSE
532	to_be_resorted => nc_av
533	ENDIF
534	IF ( mode == 'xy' ) level_z = zu
535
536	CASE ( 'nr_xy', 'nr_xz', 'nr_yz' )
537	IF ( av == 0 ) THEN
538	to_be_resorted => nr
539	ELSE
540	to_be_resorted => nr_av
541	ENDIF
542	IF ( mode == 'xy' ) level_z = zu
543
544	CASE ( 'ol*_xy' ) ! 2d-array
545	IF ( av == 0 ) THEN
546	DO m = 1, surf_def_h(0)%ns
547	i = surf_def_h(0)%i(m)
548	j = surf_def_h(0)%j(m)
549	local_pf(i,j,nzb+1) = surf_def_h(0)%ol(m)
550	ENDDO
551	DO m = 1, surf_lsm_h%ns
552	i = surf_lsm_h%i(m)
553	j = surf_lsm_h%j(m)
554	local_pf(i,j,nzb+1) = surf_lsm_h%ol(m)
555	ENDDO
556	DO m = 1, surf_usm_h%ns
557	i = surf_usm_h%i(m)
558	j = surf_usm_h%j(m)
559	local_pf(i,j,nzb+1) = surf_usm_h%ol(m)
560	ENDDO
561	ELSE
562	DO i = nxl, nxr
563	DO j = nys, nyn
564	local_pf(i,j,nzb+1) = ol_av(j,i)
565	ENDDO
566	ENDDO
567	ENDIF
568	resorted = .TRUE.
569	two_d = .TRUE.
570	level_z(nzb+1) = zu(nzb+1)
571
572	CASE ( 'p_xy', 'p_xz', 'p_yz' )
573	IF ( av == 0 ) THEN
574	IF ( psolver /= 'sor' ) CALL exchange_horiz( p, nbgp )
575	to_be_resorted => p
576	ELSE
577	IF ( psolver /= 'sor' ) CALL exchange_horiz( p_av, nbgp )
578	to_be_resorted => p_av
579	ENDIF
580	IF ( mode == 'xy' ) level_z = zu
581
582	CASE ( 'pc_xy', 'pc_xz', 'pc_yz' ) ! particle concentration
583	IF ( av == 0 ) THEN
584	IF ( simulated_time >= particle_advection_start ) THEN
585	tend = prt_count
586	! CALL exchange_horiz( tend, nbgp )
587	ELSE
588	tend = 0.0_wp
589	ENDIF
590	DO i = nxl, nxr
591	DO j = nys, nyn
592	DO k = nzb, nzt+1
593	local_pf(i,j,k) = tend(k,j,i)
594	ENDDO
595	ENDDO
596	ENDDO
597	resorted = .TRUE.
598	ELSE
599	! CALL exchange_horiz( pc_av, nbgp )
600	to_be_resorted => pc_av
601	ENDIF
602
603	CASE ( 'pr_xy', 'pr_xz', 'pr_yz' ) ! mean particle radius (effective radius)
604	IF ( av == 0 ) THEN
605	IF ( simulated_time >= particle_advection_start ) THEN
606	DO i = nxl, nxr
607	DO j = nys, nyn
608	DO k = nzb, nzt+1
609	number_of_particles = prt_count(k,j,i)
610	IF (number_of_particles <= 0) CYCLE
611	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
612	s_r2 = 0.0_wp
613	s_r3 = 0.0_wp
614	DO n = 1, number_of_particles
615	IF ( particles(n)%particle_mask ) THEN
616	s_r2 = s_r2 + particles(n)%radius*2 &
617	particles(n)%weight_factor
618	s_r3 = s_r3 + particles(n)%radius*3 &
619	particles(n)%weight_factor
620	ENDIF
621	ENDDO
622	IF ( s_r2 > 0.0_wp ) THEN
623	mean_r = s_r3 / s_r2
624	ELSE
625	mean_r = 0.0_wp
626	ENDIF
627	tend(k,j,i) = mean_r
628	ENDDO
629	ENDDO
630	ENDDO
631	! CALL exchange_horiz( tend, nbgp )
632	ELSE
633	tend = 0.0_wp
634	ENDIF
635	DO i = nxl, nxr
636	DO j = nys, nyn
637	DO k = nzb, nzt+1
638	local_pf(i,j,k) = tend(k,j,i)
639	ENDDO
640	ENDDO
641	ENDDO
642	resorted = .TRUE.
643	ELSE
644	! CALL exchange_horiz( pr_av, nbgp )
645	to_be_resorted => pr_av
646	ENDIF
647
648	CASE ( 'pra*_xy' ) ! 2d-array / integral quantity => no av
649	! CALL exchange_horiz_2d( precipitation_amount )
650	DO i = nxl, nxr
651	DO j = nys, nyn
652	local_pf(i,j,nzb+1) = precipitation_amount(j,i)
653	ENDDO
654	ENDDO
655	precipitation_amount = 0.0_wp ! reset for next integ. interval
656	resorted = .TRUE.
657	two_d = .TRUE.
658	level_z(nzb+1) = zu(nzb+1)
659
660	CASE ( 'prr*_xy' ) ! 2d-array
661	IF ( av == 0 ) THEN
662	! CALL exchange_horiz_2d( prr(nzb+1,:,:) )
663	DO i = nxl, nxr
664	DO j = nys, nyn
665	local_pf(i,j,nzb+1) = prr(nzb+1,j,i) * hyrho(nzb+1)
666	ENDDO
667	ENDDO
668	ELSE
669	! CALL exchange_horiz_2d( prr_av(nzb+1,:,:) )
670	DO i = nxl, nxr
671	DO j = nys, nyn
672	local_pf(i,j,nzb+1) = prr_av(nzb+1,j,i) * hyrho(nzb+1)
673	ENDDO
674	ENDDO
675	ENDIF
676	resorted = .TRUE.
677	two_d = .TRUE.
678	level_z(nzb+1) = zu(nzb+1)
679
680	CASE ( 'prr_xy', 'prr_xz', 'prr_yz' )
681	IF ( av == 0 ) THEN
682	! CALL exchange_horiz( prr, nbgp )
683	DO i = nxl, nxr
684	DO j = nys, nyn
685	DO k = nzb, nzt+1
686	local_pf(i,j,k) = prr(k,j,i) * hyrho(nzb+1)
687	ENDDO
688	ENDDO
689	ENDDO
690	ELSE
691	! CALL exchange_horiz( prr_av, nbgp )
692	DO i = nxl, nxr
693	DO j = nys, nyn
694	DO k = nzb, nzt+1
695	local_pf(i,j,k) = prr_av(k,j,i) * hyrho(nzb+1)
696	ENDDO
697	ENDDO
698	ENDDO
699	ENDIF
700	resorted = .TRUE.
701	IF ( mode == 'xy' ) level_z = zu
702
703	CASE ( 'pt_xy', 'pt_xz', 'pt_yz' )
704	IF ( av == 0 ) THEN
705	IF ( .NOT. cloud_physics ) THEN
706	to_be_resorted => pt
707	ELSE
708	DO i = nxl, nxr
709	DO j = nys, nyn
710	DO k = nzb, nzt+1
711	local_pf(i,j,k) = pt(k,j,i) + l_d_cp * &
712	pt_d_t(k) * &
713	ql(k,j,i)
714	ENDDO
715	ENDDO
716	ENDDO
717	resorted = .TRUE.
718	ENDIF
719	ELSE
720	to_be_resorted => pt_av
721	ENDIF
722	IF ( mode == 'xy' ) level_z = zu
723
724	CASE ( 'q_xy', 'q_xz', 'q_yz' )
725	IF ( av == 0 ) THEN
726	to_be_resorted => q
727	ELSE
728	to_be_resorted => q_av
729	ENDIF
730	IF ( mode == 'xy' ) level_z = zu
731
732	CASE ( 'qc_xy', 'qc_xz', 'qc_yz' )
733	IF ( av == 0 ) THEN
734	to_be_resorted => qc
735	ELSE
736	to_be_resorted => qc_av
737	ENDIF
738	IF ( mode == 'xy' ) level_z = zu
739
740	CASE ( 'ql_xy', 'ql_xz', 'ql_yz' )
741	IF ( av == 0 ) THEN
742	to_be_resorted => ql
743	ELSE
744	to_be_resorted => ql_av
745	ENDIF
746	IF ( mode == 'xy' ) level_z = zu
747
748	CASE ( 'ql_c_xy', 'ql_c_xz', 'ql_c_yz' )
749	IF ( av == 0 ) THEN
750	to_be_resorted => ql_c
751	ELSE
752	to_be_resorted => ql_c_av
753	ENDIF
754	IF ( mode == 'xy' ) level_z = zu
755
756	CASE ( 'ql_v_xy', 'ql_v_xz', 'ql_v_yz' )
757	IF ( av == 0 ) THEN
758	to_be_resorted => ql_v
759	ELSE
760	to_be_resorted => ql_v_av
761	ENDIF
762	IF ( mode == 'xy' ) level_z = zu
763
764	CASE ( 'ql_vp_xy', 'ql_vp_xz', 'ql_vp_yz' )
765	IF ( av == 0 ) THEN
766	IF ( simulated_time >= particle_advection_start ) THEN
767	DO i = nxl, nxr
768	DO j = nys, nyn
769	DO k = nzb, nzt+1
770	number_of_particles = prt_count(k,j,i)
771	IF (number_of_particles <= 0) CYCLE
772	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
773	DO n = 1, number_of_particles
774	IF ( particles(n)%particle_mask ) THEN
775	tend(k,j,i) = tend(k,j,i) + &
776	particles(n)%weight_factor / &
777	prt_count(k,j,i)
778	ENDIF
779	ENDDO
780	ENDDO
781	ENDDO
782	ENDDO
783	! CALL exchange_horiz( tend, nbgp )
784	ELSE
785	tend = 0.0_wp
786	ENDIF
787	DO i = nxl, nxr
788	DO j = nys, nyn
789	DO k = nzb, nzt+1
790	local_pf(i,j,k) = tend(k,j,i)
791	ENDDO
792	ENDDO
793	ENDDO
794	resorted = .TRUE.
795	ELSE
796	! CALL exchange_horiz( ql_vp_av, nbgp )
797	to_be_resorted => ql_vp
798	ENDIF
799	IF ( mode == 'xy' ) level_z = zu
800
801	CASE ( 'qr_xy', 'qr_xz', 'qr_yz' )
802	IF ( av == 0 ) THEN
803	to_be_resorted => qr
804	ELSE
805	to_be_resorted => qr_av
806	ENDIF
807	IF ( mode == 'xy' ) level_z = zu
808
809	CASE ( 'qsws*_xy' ) ! 2d-array
810	IF ( av == 0 ) THEN
811	DO m = 1, surf_def_h(0)%ns
812	i = surf_def_h(0)%i(m)
813	j = surf_def_h(0)%j(m)
814	local_pf(i,j,nzb+1) = surf_def_h(0)%qsws(m)
815	ENDDO
816	DO m = 1, surf_lsm_h%ns
817	i = surf_lsm_h%i(m)
818	j = surf_lsm_h%j(m)
819	local_pf(i,j,nzb+1) = surf_lsm_h%qsws(m)
820	ENDDO
821	DO m = 1, surf_usm_h%ns
822	i = surf_usm_h%i(m)
823	j = surf_usm_h%j(m)
824	local_pf(i,j,nzb+1) = surf_usm_h%qsws(m)
825	ENDDO
826	ELSE
827	DO i = nxl, nxr
828	DO j = nys, nyn
829	local_pf(i,j,nzb+1) = qsws_av(j,i)
830	ENDDO
831	ENDDO
832	ENDIF
833	resorted = .TRUE.
834	two_d = .TRUE.
835	level_z(nzb+1) = zu(nzb+1)
836
837	CASE ( 'qv_xy', 'qv_xz', 'qv_yz' )
838	IF ( av == 0 ) THEN
839	DO i = nxl, nxr
840	DO j = nys, nyn
841	DO k = nzb, nzt+1
842	local_pf(i,j,k) = q(k,j,i) - ql(k,j,i)
843	ENDDO
844	ENDDO
845	ENDDO
846	resorted = .TRUE.
847	ELSE
848	to_be_resorted => qv_av
849	ENDIF
850	IF ( mode == 'xy' ) level_z = zu
851
852	CASE ( 'r_a*_xy' ) ! 2d-array
853	IF ( av == 0 ) THEN
854	DO m = 1, surf_lsm_h%ns
855	i = surf_lsm_h%i(m)
856	j = surf_lsm_h%j(m)
857	local_pf(i,j,nzb+1) = surf_lsm_h%r_a(m)
858	ENDDO
859
860	DO m = 1, surf_usm_h%ns
861	i = surf_usm_h%i(m)
862	j = surf_usm_h%j(m)
863	local_pf(i,j,nzb+1) = &
864	( surf_usm_h%frac(0,m) * surf_usm_h%r_a(m) + &
865	surf_usm_h%frac(1,m) * surf_usm_h%r_a_green(m) + &
866	surf_usm_h%frac(2,m) * surf_usm_h%r_a_window(m) )
867	ENDDO
868
869	ELSE
870	DO i = nxl, nxr
871	DO j = nys, nyn
872	local_pf(i,j,nzb+1) = r_a_av(j,i)
873	ENDDO
874	ENDDO
875	ENDIF
876	resorted = .TRUE.
877	two_d = .TRUE.
878	level_z(nzb+1) = zu(nzb+1)
879
880	CASE ( 'rho_ocean_xy', 'rho_ocean_xz', 'rho_ocean_yz' )
881	IF ( av == 0 ) THEN
882	to_be_resorted => rho_ocean
883	ELSE
884	to_be_resorted => rho_ocean_av
885	ENDIF
886
887	CASE ( 's_xy', 's_xz', 's_yz' )
888	IF ( av == 0 ) THEN
889	to_be_resorted => s
890	ELSE
891	to_be_resorted => s_av
892	ENDIF
893
894	CASE ( 'sa_xy', 'sa_xz', 'sa_yz' )
895	IF ( av == 0 ) THEN
896	to_be_resorted => sa
897	ELSE
898	to_be_resorted => sa_av
899	ENDIF
900
901	CASE ( 'shf*_xy' ) ! 2d-array
902	IF ( av == 0 ) THEN
903	DO m = 1, surf_def_h(0)%ns
904	i = surf_def_h(0)%i(m)
905	j = surf_def_h(0)%j(m)
906	local_pf(i,j,nzb+1) = surf_def_h(0)%shf(m)
907	ENDDO
908	DO m = 1, surf_lsm_h%ns
909	i = surf_lsm_h%i(m)
910	j = surf_lsm_h%j(m)
911	local_pf(i,j,nzb+1) = surf_lsm_h%shf(m)
912	ENDDO
913	DO m = 1, surf_usm_h%ns
914	i = surf_usm_h%i(m)
915	j = surf_usm_h%j(m)
916	local_pf(i,j,nzb+1) = surf_usm_h%shf(m)
917	ENDDO
918	ELSE
919	DO i = nxl, nxr
920	DO j = nys, nyn
921	local_pf(i,j,nzb+1) = shf_av(j,i)
922	ENDDO
923	ENDDO
924	ENDIF
925	resorted = .TRUE.
926	two_d = .TRUE.
927	level_z(nzb+1) = zu(nzb+1)
928
929	CASE ( 'ssws*_xy' ) ! 2d-array
930	IF ( av == 0 ) THEN
931	DO m = 1, surf_def_h(0)%ns
932	i = surf_def_h(0)%i(m)
933	j = surf_def_h(0)%j(m)
934	local_pf(i,j,nzb+1) = surf_def_h(0)%ssws(m)
935	ENDDO
936	DO m = 1, surf_lsm_h%ns
937	i = surf_lsm_h%i(m)
938	j = surf_lsm_h%j(m)
939	local_pf(i,j,nzb+1) = surf_lsm_h%ssws(m)
940	ENDDO
941	DO m = 1, surf_usm_h%ns
942	i = surf_usm_h%i(m)
943	j = surf_usm_h%j(m)
944	local_pf(i,j,nzb+1) = surf_usm_h%ssws(m)
945	ENDDO
946	ELSE
947	DO i = nxl, nxr
948	DO j = nys, nyn
949	local_pf(i,j,nzb+1) = ssws_av(j,i)
950	ENDDO
951	ENDDO
952	ENDIF
953	resorted = .TRUE.
954	two_d = .TRUE.
955	level_z(nzb+1) = zu(nzb+1)
956
957	CASE ( 't*_xy' ) ! 2d-array
958	IF ( av == 0 ) THEN
959	DO m = 1, surf_def_h(0)%ns
960	i = surf_def_h(0)%i(m)
961	j = surf_def_h(0)%j(m)
962	local_pf(i,j,nzb+1) = surf_def_h(0)%ts(m)
963	ENDDO
964	DO m = 1, surf_lsm_h%ns
965	i = surf_lsm_h%i(m)
966	j = surf_lsm_h%j(m)
967	local_pf(i,j,nzb+1) = surf_lsm_h%ts(m)
968	ENDDO
969	DO m = 1, surf_usm_h%ns
970	i = surf_usm_h%i(m)
971	j = surf_usm_h%j(m)
972	local_pf(i,j,nzb+1) = surf_usm_h%ts(m)
973	ENDDO
974	ELSE
975	DO i = nxl, nxr
976	DO j = nys, nyn
977	local_pf(i,j,nzb+1) = ts_av(j,i)
978	ENDDO
979	ENDDO
980	ENDIF
981	resorted = .TRUE.
982	two_d = .TRUE.
983	level_z(nzb+1) = zu(nzb+1)
984
985	CASE ( 'tsurf*_xy' ) ! 2d-array
986	IF ( av == 0 ) THEN
987	DO m = 1, surf_lsm_h%ns
988	i = surf_lsm_h%i(m)
989	j = surf_lsm_h%j(m)
990	local_pf(i,j,nzb+1) = surf_lsm_h%pt_surface(m)
991	ENDDO
992
993	DO m = 1, surf_usm_h%ns
994	i = surf_usm_h%i(m)
995	j = surf_usm_h%j(m)
996	local_pf(i,j,nzb+1) = surf_usm_h%pt_surface(m)
997	ENDDO
998
999	ELSE
1000	DO i = nxl, nxr
1001	DO j = nys, nyn
1002	local_pf(i,j,nzb+1) = tsurf_av(j,i)
1003	ENDDO
1004	ENDDO
1005	ENDIF
1006	resorted = .TRUE.
1007	two_d = .TRUE.
1008	level_z(nzb+1) = zu(nzb+1)
1009
1010	CASE ( 'u_xy', 'u_xz', 'u_yz' )
1011	flag_nr = 1
1012	IF ( av == 0 ) THEN
1013	to_be_resorted => u
1014	ELSE
1015	to_be_resorted => u_av
1016	ENDIF
1017	IF ( mode == 'xy' ) level_z = zu
1018	!
1019	!-- Substitute the values generated by "mirror" boundary condition
1020	!-- at the bottom boundary by the real surface values.
1021	IF ( do2d(av,if) == 'u_xz' .OR. do2d(av,if) == 'u_yz' ) THEN
1022	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
1023	ENDIF
1024
1025	CASE ( 'u*_xy' ) ! 2d-array
1026	IF ( av == 0 ) THEN
1027	DO m = 1, surf_def_h(0)%ns
1028	i = surf_def_h(0)%i(m)
1029	j = surf_def_h(0)%j(m)
1030	local_pf(i,j,nzb+1) = surf_def_h(0)%us(m)
1031	ENDDO
1032	DO m = 1, surf_lsm_h%ns
1033	i = surf_lsm_h%i(m)
1034	j = surf_lsm_h%j(m)
1035	local_pf(i,j,nzb+1) = surf_lsm_h%us(m)
1036	ENDDO
1037	DO m = 1, surf_usm_h%ns
1038	i = surf_usm_h%i(m)
1039	j = surf_usm_h%j(m)
1040	local_pf(i,j,nzb+1) = surf_usm_h%us(m)
1041	ENDDO
1042	ELSE
1043	DO i = nxl, nxr
1044	DO j = nys, nyn
1045	local_pf(i,j,nzb+1) = us_av(j,i)
1046	ENDDO
1047	ENDDO
1048	ENDIF
1049	resorted = .TRUE.
1050	two_d = .TRUE.
1051	level_z(nzb+1) = zu(nzb+1)
1052
1053	CASE ( 'v_xy', 'v_xz', 'v_yz' )
1054	flag_nr = 2
1055	IF ( av == 0 ) THEN
1056	to_be_resorted => v
1057	ELSE
1058	to_be_resorted => v_av
1059	ENDIF
1060	IF ( mode == 'xy' ) level_z = zu
1061	!
1062	!-- Substitute the values generated by "mirror" boundary condition
1063	!-- at the bottom boundary by the real surface values.
1064	IF ( do2d(av,if) == 'v_xz' .OR. do2d(av,if) == 'v_yz' ) THEN
1065	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
1066	ENDIF
1067
1068	CASE ( 'vpt_xy', 'vpt_xz', 'vpt_yz' )
1069	IF ( av == 0 ) THEN
1070	to_be_resorted => vpt
1071	ELSE
1072	to_be_resorted => vpt_av
1073	ENDIF
1074	IF ( mode == 'xy' ) level_z = zu
1075
1076	CASE ( 'w_xy', 'w_xz', 'w_yz' )
1077	flag_nr = 3
1078	IF ( av == 0 ) THEN
1079	to_be_resorted => w
1080	ELSE
1081	to_be_resorted => w_av
1082	ENDIF
1083	IF ( mode == 'xy' ) level_z = zw
1084
1085	CASE ( 'z0*_xy' ) ! 2d-array
1086	IF ( av == 0 ) THEN
1087	DO m = 1, surf_def_h(0)%ns
1088	i = surf_def_h(0)%i(m)
1089	j = surf_def_h(0)%j(m)
1090	local_pf(i,j,nzb+1) = surf_def_h(0)%z0(m)
1091	ENDDO
1092	DO m = 1, surf_lsm_h%ns
1093	i = surf_lsm_h%i(m)
1094	j = surf_lsm_h%j(m)
1095	local_pf(i,j,nzb+1) = surf_lsm_h%z0(m)
1096	ENDDO
1097	DO m = 1, surf_usm_h%ns
1098	i = surf_usm_h%i(m)
1099	j = surf_usm_h%j(m)
1100	local_pf(i,j,nzb+1) = surf_usm_h%z0(m)
1101	ENDDO
1102	ELSE
1103	DO i = nxl, nxr
1104	DO j = nys, nyn
1105	local_pf(i,j,nzb+1) = z0_av(j,i)
1106	ENDDO
1107	ENDDO
1108	ENDIF
1109	resorted = .TRUE.
1110	two_d = .TRUE.
1111	level_z(nzb+1) = zu(nzb+1)
1112
1113	CASE ( 'z0h*_xy' ) ! 2d-array
1114	IF ( av == 0 ) THEN
1115	DO m = 1, surf_def_h(0)%ns
1116	i = surf_def_h(0)%i(m)
1117	j = surf_def_h(0)%j(m)
1118	local_pf(i,j,nzb+1) = surf_def_h(0)%z0h(m)
1119	ENDDO
1120	DO m = 1, surf_lsm_h%ns
1121	i = surf_lsm_h%i(m)
1122	j = surf_lsm_h%j(m)
1123	local_pf(i,j,nzb+1) = surf_lsm_h%z0h(m)
1124	ENDDO
1125	DO m = 1, surf_usm_h%ns
1126	i = surf_usm_h%i(m)
1127	j = surf_usm_h%j(m)
1128	local_pf(i,j,nzb+1) = surf_usm_h%z0h(m)
1129	ENDDO
1130	ELSE
1131	DO i = nxl, nxr
1132	DO j = nys, nyn
1133	local_pf(i,j,nzb+1) = z0h_av(j,i)
1134	ENDDO
1135	ENDDO
1136	ENDIF
1137	resorted = .TRUE.
1138	two_d = .TRUE.
1139	level_z(nzb+1) = zu(nzb+1)
1140
1141	CASE ( 'z0q*_xy' ) ! 2d-array
1142	IF ( av == 0 ) THEN
1143	DO m = 1, surf_def_h(0)%ns
1144	i = surf_def_h(0)%i(m)
1145	j = surf_def_h(0)%j(m)
1146	local_pf(i,j,nzb+1) = surf_def_h(0)%z0q(m)
1147	ENDDO
1148	DO m = 1, surf_lsm_h%ns
1149	i = surf_lsm_h%i(m)
1150	j = surf_lsm_h%j(m)
1151	local_pf(i,j,nzb+1) = surf_lsm_h%z0q(m)
1152	ENDDO
1153	DO m = 1, surf_usm_h%ns
1154	i = surf_usm_h%i(m)
1155	j = surf_usm_h%j(m)
1156	local_pf(i,j,nzb+1) = surf_usm_h%z0q(m)
1157	ENDDO
1158	ELSE
1159	DO i = nxl, nxr
1160	DO j = nys, nyn
1161	local_pf(i,j,nzb+1) = z0q_av(j,i)
1162	ENDDO
1163	ENDDO
1164	ENDIF
1165	resorted = .TRUE.
1166	two_d = .TRUE.
1167	level_z(nzb+1) = zu(nzb+1)
1168
1169	CASE DEFAULT
1170
1171	!
1172	!-- Land surface model quantity
1173	IF ( land_surface ) THEN
1174	CALL lsm_data_output_2d( av, do2d(av,if), found, grid, mode,&
1175	local_pf, two_d, nzb_do, nzt_do )
1176	ENDIF
1177
1178	!
1179	!-- Turbulence closure variables
1180	IF ( .NOT. found ) THEN
1181	CALL tcm_data_output_2d( av, do2d(av,if), found, grid, mode,&
1182	local_pf, two_d, nzb_do, nzt_do )
1183	ENDIF
1184
1185	!
1186	!-- Radiation quantity
1187	IF ( .NOT. found .AND. radiation ) THEN
1188	CALL radiation_data_output_2d( av, do2d(av,if), found, grid,&
1189	mode, local_pf, two_d )
1190	ENDIF
1191
1192	!
1193	!-- UV exposure model quantity
1194	IF ( uv_exposure ) THEN
1195	CALL uvem_data_output_2d( av, do2d(av,if), found, grid, mode,&
1196	local_pf, two_d, nzb_do, nzt_do )
1197	ENDIF
1198
1199	!
1200	!-- User defined quantity
1201	IF ( .NOT. found ) THEN
1202	CALL user_data_output_2d( av, do2d(av,if), found, grid, &
1203	local_pf, two_d, nzb_do, nzt_do )
1204	ENDIF
1205
1206	resorted = .TRUE.
1207
1208	IF ( grid == 'zu' ) THEN
1209	IF ( mode == 'xy' ) level_z = zu
1210	ELSEIF ( grid == 'zw' ) THEN
1211	IF ( mode == 'xy' ) level_z = zw
1212	ELSEIF ( grid == 'zu1' ) THEN
1213	IF ( mode == 'xy' ) level_z(nzb+1) = zu(nzb+1)
1214	ELSEIF ( grid == 'zs' ) THEN
1215	IF ( mode == 'xy' ) level_z = zs
1216	ENDIF
1217
1218	IF ( .NOT. found ) THEN
1219	message_string = 'no output provided for: ' // &
1220	TRIM( do2d(av,if) )
1221	CALL message( 'data_output_2d', 'PA0181', 0, 0, 0, 6, 0 )
1222	ENDIF
1223
1224	END SELECT
1225
1226	!
1227	!-- Resort the array to be output, if not done above. Flag topography
1228	!-- grid points with fill values, using the corresponding maksing flag.
1229	IF ( .NOT. resorted ) THEN
1230	DO i = nxl, nxr
1231	DO j = nys, nyn
1232	DO k = nzb_do, nzt_do
1233	local_pf(i,j,k) = MERGE( to_be_resorted(k,j,i), &
1234	REAL( fill_value, KIND = wp ), &
1235	BTEST( wall_flags_0(k,j,i), &
1236	flag_nr ) )
1237	ENDDO
1238	ENDDO
1239	ENDDO
1240	ENDIF
1241
1242	!
1243	!-- Output of the individual cross-sections, depending on the cross-
1244	!-- section mode chosen.
1245	is = 1
1246	loop1: DO WHILE ( section(is,s_ind) /= -9999 .OR. two_d )
1247
1248	SELECT CASE ( mode )
1249
1250	CASE ( 'xy' )
1251	!
1252	!-- Determine the cross section index
1253	IF ( two_d ) THEN
1254	layer_xy = nzb+1
1255	ELSE
1256	layer_xy = section(is,s_ind)
1257	ENDIF
1258
1259	!
1260	!-- Exit the loop for layers beyond the data output domain
1261	!-- (used for soil model)
1262	IF ( layer_xy > nzt_do ) THEN
1263	EXIT loop1
1264	ENDIF
1265
1266	!
1267	!-- Update the netCDF xy cross section time axis.
1268	!-- In case of parallel output, this is only done by PE0
1269	!-- to increase the performance.
1270	IF ( simulated_time /= do2d_xy_last_time(av) ) THEN
1271	do2d_xy_time_count(av) = do2d_xy_time_count(av) + 1
1272	do2d_xy_last_time(av) = simulated_time
1273	IF ( myid == 0 ) THEN
1274	IF ( .NOT. data_output_2d_on_each_pe &
1275	.OR. netcdf_data_format > 4 ) &
1276	THEN
1277	#if defined( __netcdf )
1278	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1279	id_var_time_xy(av), &
1280	(/ time_since_reference_point /), &
1281	start = (/ do2d_xy_time_count(av) /), &
1282	count = (/ 1 /) )
1283	CALL netcdf_handle_error( 'data_output_2d', 53 )
1284	#endif
1285	ENDIF
1286	ENDIF
1287	ENDIF
1288	!
1289	!-- If required, carry out averaging along z
1290	IF ( section(is,s_ind) == -1 .AND. .NOT. two_d ) THEN
1291
1292	local_2d = 0.0_wp
1293	!
1294	!-- Carry out the averaging (all data are on the PE)
1295	DO k = nzb_do, nzt_do
1296	DO j = nys, nyn
1297	DO i = nxl, nxr
1298	local_2d(i,j) = local_2d(i,j) + local_pf(i,j,k)
1299	ENDDO
1300	ENDDO
1301	ENDDO
1302
1303	local_2d = local_2d / ( nzt_do - nzb_do + 1.0_wp)
1304
1305	ELSE
1306	!
1307	!-- Just store the respective section on the local array
1308	local_2d = local_pf(:,:,layer_xy)
1309
1310	ENDIF
1311
1312	#if defined( __parallel )
1313	IF ( netcdf_data_format > 4 ) THEN
1314	!
1315	!-- Parallel output in netCDF4/HDF5 format.
1316	IF ( two_d ) THEN
1317	iis = 1
1318	ELSE
1319	iis = is
1320	ENDIF
1321
1322	#if defined( __netcdf )
1323	!
1324	!-- For parallel output, all cross sections are first stored
1325	!-- here on a local array and will be written to the output
1326	!-- file afterwards to increase the performance.
1327	DO i = nxl, nxr
1328	DO j = nys, nyn
1329	local_2d_sections(i,j,iis) = local_2d(i,j)
1330	ENDDO
1331	ENDDO
1332	#endif
1333	ELSE
1334
1335	IF ( data_output_2d_on_each_pe ) THEN
1336	!
1337	!-- Output of partial arrays on each PE
1338	#if defined( __netcdf )
1339	IF ( myid == 0 ) THEN
1340	WRITE ( 21 ) time_since_reference_point, &
1341	do2d_xy_time_count(av), av
1342	ENDIF
1343	#endif
1344	DO i = 0, io_blocks-1
1345	IF ( i == io_group ) THEN
1346	WRITE ( 21 ) nxl, nxr, nys, nyn, nys, nyn
1347	WRITE ( 21 ) local_2d
1348	ENDIF
1349	#if defined( __parallel )
1350	CALL MPI_BARRIER( comm2d, ierr )
1351	#endif
1352	ENDDO
1353
1354	ELSE
1355	!
1356	!-- PE0 receives partial arrays from all processors and
1357	!-- then outputs them. Here a barrier has to be set,
1358	!-- because otherwise "-MPI- FATAL: Remote protocol queue
1359	!-- full" may occur.
1360	CALL MPI_BARRIER( comm2d, ierr )
1361
1362	ngp = ( nxr-nxl+1 ) * ( nyn-nys+1 )
1363	IF ( myid == 0 ) THEN
1364	!
1365	!-- Local array can be relocated directly.
1366	total_2d(nxl:nxr,nys:nyn) = local_2d
1367	!
1368	!-- Receive data from all other PEs.
1369	DO n = 1, numprocs-1
1370	!
1371	!-- Receive index limits first, then array.
1372	!-- Index limits are received in arbitrary order from
1373	!-- the PEs.
1374	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1375	MPI_ANY_SOURCE, 0, comm2d, &
1376	status, ierr )
1377	sender = status(MPI_SOURCE)
1378	DEALLOCATE( local_2d )
1379	ALLOCATE( local_2d(ind(1):ind(2),ind(3):ind(4)) )
1380	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1381	MPI_REAL, sender, 1, comm2d, &
1382	status, ierr )
1383	total_2d(ind(1):ind(2),ind(3):ind(4)) = local_2d
1384	ENDDO
1385	!
1386	!-- Relocate the local array for the next loop increment
1387	DEALLOCATE( local_2d )
1388	ALLOCATE( local_2d(nxl:nxr,nys:nyn) )
1389
1390	#if defined( __netcdf )
1391	IF ( two_d ) THEN
1392	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1393	id_var_do2d(av,if), &
1394	total_2d(0:nx,0:ny), &
1395	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
1396	count = (/ nx+1, ny+1, 1, 1 /) )
1397	ELSE
1398	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1399	id_var_do2d(av,if), &
1400	total_2d(0:nx,0:ny), &
1401	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
1402	count = (/ nx+1, ny+1, 1, 1 /) )
1403	ENDIF
1404	CALL netcdf_handle_error( 'data_output_2d', 54 )
1405	#endif
1406
1407	ELSE
1408	!
1409	!-- First send the local index limits to PE0
1410	ind(1) = nxl; ind(2) = nxr
1411	ind(3) = nys; ind(4) = nyn
1412	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1413	comm2d, ierr )
1414	!
1415	!-- Send data to PE0
1416	CALL MPI_SEND( local_2d(nxl,nys), ngp, &
1417	MPI_REAL, 0, 1, comm2d, ierr )
1418	ENDIF
1419	!
1420	!-- A barrier has to be set, because otherwise some PEs may
1421	!-- proceed too fast so that PE0 may receive wrong data on
1422	!-- tag 0
1423	CALL MPI_BARRIER( comm2d, ierr )
1424	ENDIF
1425
1426	ENDIF
1427	#else
1428	#if defined( __netcdf )
1429	IF ( two_d ) THEN
1430	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1431	id_var_do2d(av,if), &
1432	local_2d(nxl:nxr,nys:nyn), &
1433	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
1434	count = (/ nx+1, ny+1, 1, 1 /) )
1435	ELSE
1436	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1437	id_var_do2d(av,if), &
1438	local_2d(nxl:nxr,nys:nyn), &
1439	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
1440	count = (/ nx+1, ny+1, 1, 1 /) )
1441	ENDIF
1442	CALL netcdf_handle_error( 'data_output_2d', 447 )
1443	#endif
1444	#endif
1445
1446	!
1447	!-- For 2D-arrays (e.g. u*) only one cross-section is available.
1448	!-- Hence exit loop of output levels.
1449	IF ( two_d ) THEN
1450	IF ( netcdf_data_format < 5 ) two_d = .FALSE.
1451	EXIT loop1
1452	ENDIF
1453
1454	CASE ( 'xz' )
1455	!
1456	!-- Update the netCDF xz cross section time axis.
1457	!-- In case of parallel output, this is only done by PE0
1458	!-- to increase the performance.
1459	IF ( simulated_time /= do2d_xz_last_time(av) ) THEN
1460	do2d_xz_time_count(av) = do2d_xz_time_count(av) + 1
1461	do2d_xz_last_time(av) = simulated_time
1462	IF ( myid == 0 ) THEN
1463	IF ( .NOT. data_output_2d_on_each_pe &
1464	.OR. netcdf_data_format > 4 ) &
1465	THEN
1466	#if defined( __netcdf )
1467	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1468	id_var_time_xz(av), &
1469	(/ time_since_reference_point /), &
1470	start = (/ do2d_xz_time_count(av) /), &
1471	count = (/ 1 /) )
1472	CALL netcdf_handle_error( 'data_output_2d', 56 )
1473	#endif
1474	ENDIF
1475	ENDIF
1476	ENDIF
1477
1478	!
1479	!-- If required, carry out averaging along y
1480	IF ( section(is,s_ind) == -1 ) THEN
1481
1482	ALLOCATE( local_2d_l(nxl:nxr,nzb_do:nzt_do) )
1483	local_2d_l = 0.0_wp
1484	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
1485	!
1486	!-- First local averaging on the PE
1487	DO k = nzb_do, nzt_do
1488	DO j = nys, nyn
1489	DO i = nxl, nxr
1490	local_2d_l(i,k) = local_2d_l(i,k) + &
1491	local_pf(i,j,k)
1492	ENDDO
1493	ENDDO
1494	ENDDO
1495	#if defined( __parallel )
1496	!
1497	!-- Now do the averaging over all PEs along y
1498	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1499	CALL MPI_ALLREDUCE( local_2d_l(nxl,nzb_do), &
1500	local_2d(nxl,nzb_do), ngp, MPI_REAL, &
1501	MPI_SUM, comm1dy, ierr )
1502	#else
1503	local_2d = local_2d_l
1504	#endif
1505	local_2d = local_2d / ( ny + 1.0_wp )
1506
1507	DEALLOCATE( local_2d_l )
1508
1509	ELSE
1510	!
1511	!-- Just store the respective section on the local array
1512	!-- (but only if it is available on this PE!)
1513	IF ( section(is,s_ind) >= nys .AND. section(is,s_ind) <= nyn ) &
1514	THEN
1515	local_2d = local_pf(:,section(is,s_ind),nzb_do:nzt_do)
1516	ENDIF
1517
1518	ENDIF
1519
1520	#if defined( __parallel )
1521	IF ( netcdf_data_format > 4 ) THEN
1522	!
1523	!-- Output in netCDF4/HDF5 format.
1524	!-- Output only on those PEs where the respective cross
1525	!-- sections reside. Cross sections averaged along y are
1526	!-- output on the respective first PE along y (myidy=0).
1527	IF ( ( section(is,s_ind) >= nys .AND. &
1528	section(is,s_ind) <= nyn ) .OR. &
1529	( section(is,s_ind) == -1 .AND. myidy == 0 ) ) THEN
1530	#if defined( __netcdf )
1531	!
1532	!-- For parallel output, all cross sections are first
1533	!-- stored here on a local array and will be written to the
1534	!-- output file afterwards to increase the performance.
1535	DO i = nxl, nxr
1536	DO k = nzb_do, nzt_do
1537	local_2d_sections_l(i,is,k) = local_2d(i,k)
1538	ENDDO
1539	ENDDO
1540	#endif
1541	ENDIF
1542
1543	ELSE
1544
1545	IF ( data_output_2d_on_each_pe ) THEN
1546	!
1547	!-- Output of partial arrays on each PE. If the cross
1548	!-- section does not reside on the PE, output special
1549	!-- index values.
1550	#if defined( __netcdf )
1551	IF ( myid == 0 ) THEN
1552	WRITE ( 22 ) time_since_reference_point, &
1553	do2d_xz_time_count(av), av
1554	ENDIF
1555	#endif
1556	DO i = 0, io_blocks-1
1557	IF ( i == io_group ) THEN
1558	IF ( ( section(is,s_ind) >= nys .AND. &
1559	section(is,s_ind) <= nyn ) .OR. &
1560	( section(is,s_ind) == -1 .AND. &
1561	nys-1 == -1 ) ) &
1562	THEN
1563	WRITE (22) nxl, nxr, nzb_do, nzt_do, nzb, nzt+1
1564	WRITE (22) local_2d
1565	ELSE
1566	WRITE (22) -1, -1, -1, -1, -1, -1
1567	ENDIF
1568	ENDIF
1569	#if defined( __parallel )
1570	CALL MPI_BARRIER( comm2d, ierr )
1571	#endif
1572	ENDDO
1573
1574	ELSE
1575	!
1576	!-- PE0 receives partial arrays from all processors of the
1577	!-- respective cross section and outputs them. Here a
1578	!-- barrier has to be set, because otherwise
1579	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
1580	CALL MPI_BARRIER( comm2d, ierr )
1581
1582	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
1583	IF ( myid == 0 ) THEN
1584	!
1585	!-- Local array can be relocated directly.
1586	IF ( ( section(is,s_ind) >= nys .AND. &
1587	section(is,s_ind) <= nyn ) .OR. &
1588	( section(is,s_ind) == -1 .AND. &
1589	nys-1 == -1 ) ) THEN
1590	total_2d(nxl:nxr,nzb_do:nzt_do) = local_2d
1591	ENDIF
1592	!
1593	!-- Receive data from all other PEs.
1594	DO n = 1, numprocs-1
1595	!
1596	!-- Receive index limits first, then array.
1597	!-- Index limits are received in arbitrary order from
1598	!-- the PEs.
1599	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1600	MPI_ANY_SOURCE, 0, comm2d, &
1601	status, ierr )
1602	!
1603	!-- Not all PEs have data for XZ-cross-section.
1604	IF ( ind(1) /= -9999 ) THEN
1605	sender = status(MPI_SOURCE)
1606	DEALLOCATE( local_2d )
1607	ALLOCATE( local_2d(ind(1):ind(2), &
1608	ind(3):ind(4)) )
1609	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1610	MPI_REAL, sender, 1, comm2d, &
1611	status, ierr )
1612	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
1613	local_2d
1614	ENDIF
1615	ENDDO
1616	!
1617	!-- Relocate the local array for the next loop increment
1618	DEALLOCATE( local_2d )
1619	ALLOCATE( local_2d(nxl:nxr,nzb_do:nzt_do) )
1620
1621	#if defined( __netcdf )
1622	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1623	id_var_do2d(av,if), &
1624	total_2d(0:nx,nzb_do:nzt_do), &
1625	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
1626	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
1627	CALL netcdf_handle_error( 'data_output_2d', 58 )
1628	#endif
1629
1630	ELSE
1631	!
1632	!-- If the cross section resides on the PE, send the
1633	!-- local index limits, otherwise send -9999 to PE0.
1634	IF ( ( section(is,s_ind) >= nys .AND. &
1635	section(is,s_ind) <= nyn ) .OR. &
1636	( section(is,s_ind) == -1 .AND. nys-1 == -1 ) ) &
1637	THEN
1638	ind(1) = nxl; ind(2) = nxr
1639	ind(3) = nzb_do; ind(4) = nzt_do
1640	ELSE
1641	ind(1) = -9999; ind(2) = -9999
1642	ind(3) = -9999; ind(4) = -9999
1643	ENDIF
1644	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1645	comm2d, ierr )
1646	!
1647	!-- If applicable, send data to PE0.
1648	IF ( ind(1) /= -9999 ) THEN
1649	CALL MPI_SEND( local_2d(nxl,nzb_do), ngp, &
1650	MPI_REAL, 0, 1, comm2d, ierr )
1651	ENDIF
1652	ENDIF
1653	!
1654	!-- A barrier has to be set, because otherwise some PEs may
1655	!-- proceed too fast so that PE0 may receive wrong data on
1656	!-- tag 0
1657	CALL MPI_BARRIER( comm2d, ierr )
1658	ENDIF
1659
1660	ENDIF
1661	#else
1662	#if defined( __netcdf )
1663	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1664	id_var_do2d(av,if), &
1665	local_2d(nxl:nxr,nzb_do:nzt_do), &
1666	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
1667	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
1668	CALL netcdf_handle_error( 'data_output_2d', 451 )
1669	#endif
1670	#endif
1671
1672	CASE ( 'yz' )
1673	!
1674	!-- Update the netCDF yz cross section time axis.
1675	!-- In case of parallel output, this is only done by PE0
1676	!-- to increase the performance.
1677	IF ( simulated_time /= do2d_yz_last_time(av) ) THEN
1678	do2d_yz_time_count(av) = do2d_yz_time_count(av) + 1
1679	do2d_yz_last_time(av) = simulated_time
1680	IF ( myid == 0 ) THEN
1681	IF ( .NOT. data_output_2d_on_each_pe &
1682	.OR. netcdf_data_format > 4 ) &
1683	THEN
1684	#if defined( __netcdf )
1685	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
1686	id_var_time_yz(av), &
1687	(/ time_since_reference_point /), &
1688	start = (/ do2d_yz_time_count(av) /), &
1689	count = (/ 1 /) )
1690	CALL netcdf_handle_error( 'data_output_2d', 59 )
1691	#endif
1692	ENDIF
1693	ENDIF
1694	ENDIF
1695
1696	!
1697	!-- If required, carry out averaging along x
1698	IF ( section(is,s_ind) == -1 ) THEN
1699
1700	ALLOCATE( local_2d_l(nys:nyn,nzb_do:nzt_do) )
1701	local_2d_l = 0.0_wp
1702	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
1703	!
1704	!-- First local averaging on the PE
1705	DO k = nzb_do, nzt_do
1706	DO j = nys, nyn
1707	DO i = nxl, nxr
1708	local_2d_l(j,k) = local_2d_l(j,k) + &
1709	local_pf(i,j,k)
1710	ENDDO
1711	ENDDO
1712	ENDDO
1713	#if defined( __parallel )
1714	!
1715	!-- Now do the averaging over all PEs along x
1716	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1717	CALL MPI_ALLREDUCE( local_2d_l(nys,nzb_do), &
1718	local_2d(nys,nzb_do), ngp, MPI_REAL, &
1719	MPI_SUM, comm1dx, ierr )
1720	#else
1721	local_2d = local_2d_l
1722	#endif
1723	local_2d = local_2d / ( nx + 1.0_wp )
1724
1725	DEALLOCATE( local_2d_l )
1726
1727	ELSE
1728	!
1729	!-- Just store the respective section on the local array
1730	!-- (but only if it is available on this PE!)
1731	IF ( section(is,s_ind) >= nxl .AND. section(is,s_ind) <= nxr ) &
1732	THEN
1733	local_2d = local_pf(section(is,s_ind),:,nzb_do:nzt_do)
1734	ENDIF
1735
1736	ENDIF
1737
1738	#if defined( __parallel )
1739	IF ( netcdf_data_format > 4 ) THEN
1740	!
1741	!-- Output in netCDF4/HDF5 format.
1742	!-- Output only on those PEs where the respective cross
1743	!-- sections reside. Cross sections averaged along x are
1744	!-- output on the respective first PE along x (myidx=0).
1745	IF ( ( section(is,s_ind) >= nxl .AND. &
1746	section(is,s_ind) <= nxr ) .OR. &
1747	( section(is,s_ind) == -1 .AND. myidx == 0 ) ) THEN
1748	#if defined( __netcdf )
1749	!
1750	!-- For parallel output, all cross sections are first
1751	!-- stored here on a local array and will be written to the
1752	!-- output file afterwards to increase the performance.
1753	DO j = nys, nyn
1754	DO k = nzb_do, nzt_do
1755	local_2d_sections_l(is,j,k) = local_2d(j,k)
1756	ENDDO
1757	ENDDO
1758	#endif
1759	ENDIF
1760
1761	ELSE
1762
1763	IF ( data_output_2d_on_each_pe ) THEN
1764	!
1765	!-- Output of partial arrays on each PE. If the cross
1766	!-- section does not reside on the PE, output special
1767	!-- index values.
1768	#if defined( __netcdf )
1769	IF ( myid == 0 ) THEN
1770	WRITE ( 23 ) time_since_reference_point, &
1771	do2d_yz_time_count(av), av
1772	ENDIF
1773	#endif
1774	DO i = 0, io_blocks-1
1775	IF ( i == io_group ) THEN
1776	IF ( ( section(is,s_ind) >= nxl .AND. &
1777	section(is,s_ind) <= nxr ) .OR. &
1778	( section(is,s_ind) == -1 .AND. &
1779	nxl-1 == -1 ) ) &
1780	THEN
1781	WRITE (23) nys, nyn, nzb_do, nzt_do, nzb, nzt+1
1782	WRITE (23) local_2d
1783	ELSE
1784	WRITE (23) -1, -1, -1, -1, -1, -1
1785	ENDIF
1786	ENDIF
1787	#if defined( __parallel )
1788	CALL MPI_BARRIER( comm2d, ierr )
1789	#endif
1790	ENDDO
1791
1792	ELSE
1793	!
1794	!-- PE0 receives partial arrays from all processors of the
1795	!-- respective cross section and outputs them. Here a
1796	!-- barrier has to be set, because otherwise
1797	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
1798	CALL MPI_BARRIER( comm2d, ierr )
1799
1800	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
1801	IF ( myid == 0 ) THEN
1802	!
1803	!-- Local array can be relocated directly.
1804	IF ( ( section(is,s_ind) >= nxl .AND. &
1805	section(is,s_ind) <= nxr ) .OR. &
1806	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
1807	THEN
1808	total_2d(nys:nyn,nzb_do:nzt_do) = local_2d
1809	ENDIF
1810	!
1811	!-- Receive data from all other PEs.
1812	DO n = 1, numprocs-1
1813	!
1814	!-- Receive index limits first, then array.
1815	!-- Index limits are received in arbitrary order from
1816	!-- the PEs.
1817	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1818	MPI_ANY_SOURCE, 0, comm2d, &
1819	status, ierr )
1820	!
1821	!-- Not all PEs have data for YZ-cross-section.
1822	IF ( ind(1) /= -9999 ) THEN
1823	sender = status(MPI_SOURCE)
1824	DEALLOCATE( local_2d )
1825	ALLOCATE( local_2d(ind(1):ind(2), &
1826	ind(3):ind(4)) )
1827	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1828	MPI_REAL, sender, 1, comm2d, &
1829	status, ierr )
1830	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
1831	local_2d
1832	ENDIF
1833	ENDDO
1834	!
1835	!-- Relocate the local array for the next loop increment
1836	DEALLOCATE( local_2d )
1837	ALLOCATE( local_2d(nys:nyn,nzb_do:nzt_do) )
1838
1839	#if defined( __netcdf )
1840	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
1841	id_var_do2d(av,if), &
1842	total_2d(0:ny,nzb_do:nzt_do), &
1843	start = (/ is, 1, 1, do2d_yz_time_count(av) /), &
1844	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
1845	CALL netcdf_handle_error( 'data_output_2d', 61 )
1846	#endif
1847
1848	ELSE
1849	!
1850	!-- If the cross section resides on the PE, send the
1851	!-- local index limits, otherwise send -9999 to PE0.
1852	IF ( ( section(is,s_ind) >= nxl .AND. &
1853	section(is,s_ind) <= nxr ) .OR. &
1854	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
1855	THEN
1856	ind(1) = nys; ind(2) = nyn
1857	ind(3) = nzb_do; ind(4) = nzt_do
1858	ELSE
1859	ind(1) = -9999; ind(2) = -9999
1860	ind(3) = -9999; ind(4) = -9999
1861	ENDIF
1862	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1863	comm2d, ierr )
1864	!
1865	!-- If applicable, send data to PE0.
1866	IF ( ind(1) /= -9999 ) THEN
1867	CALL MPI_SEND( local_2d(nys,nzb_do), ngp, &
1868	MPI_REAL, 0, 1, comm2d, ierr )
1869	ENDIF
1870	ENDIF
1871	!
1872	!-- A barrier has to be set, because otherwise some PEs may
1873	!-- proceed too fast so that PE0 may receive wrong data on
1874	!-- tag 0
1875	CALL MPI_BARRIER( comm2d, ierr )
1876	ENDIF
1877
1878	ENDIF
1879	#else
1880	#if defined( __netcdf )
1881	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
1882	id_var_do2d(av,if), &
1883	local_2d(nys:nyn,nzb_do:nzt_do), &
1884	start = (/ is, 1, 1, do2d_xz_time_count(av) /), &
1885	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
1886	CALL netcdf_handle_error( 'data_output_2d', 452 )
1887	#endif
1888	#endif
1889
1890	END SELECT
1891
1892	is = is + 1
1893	ENDDO loop1
1894
1895	!
1896	!-- For parallel output, all data were collected before on a local array
1897	!-- and are written now to the netcdf file. This must be done to increase
1898	!-- the performance of the parallel output.
1899	#if defined( __netcdf )
1900	IF ( netcdf_data_format > 4 ) THEN
1901
1902	SELECT CASE ( mode )
1903
1904	CASE ( 'xy' )
1905	IF ( two_d ) THEN
1906	nis = 1
1907	two_d = .FALSE.
1908	ELSE
1909	nis = ns
1910	ENDIF
1911	!
1912	!-- Do not output redundant ghost point data except for the
1913	!-- boundaries of the total domain.
1914	! IF ( nxr == nx .AND. nyn /= ny ) THEN
1915	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1916	! id_var_do2d(av,if), &
1917	! local_2d_sections(nxl:nxr+1, &
1918	! nys:nyn,1:nis), &
1919	! start = (/ nxl+1, nys+1, 1, &
1920	! do2d_xy_time_count(av) /), &
1921	! count = (/ nxr-nxl+2, &
1922	! nyn-nys+1, nis, 1 &
1923	! /) )
1924	! ELSEIF ( nxr /= nx .AND. nyn == ny ) THEN
1925	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1926	! id_var_do2d(av,if), &
1927	! local_2d_sections(nxl:nxr, &
1928	! nys:nyn+1,1:nis), &
1929	! start = (/ nxl+1, nys+1, 1, &
1930	! do2d_xy_time_count(av) /), &
1931	! count = (/ nxr-nxl+1, &
1932	! nyn-nys+2, nis, 1 &
1933	! /) )
1934	! ELSEIF ( nxr == nx .AND. nyn == ny ) THEN
1935	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1936	! id_var_do2d(av,if), &
1937	! local_2d_sections(nxl:nxr+1, &
1938	! nys:nyn+1,1:nis), &
1939	! start = (/ nxl+1, nys+1, 1, &
1940	! do2d_xy_time_count(av) /), &
1941	! count = (/ nxr-nxl+2, &
1942	! nyn-nys+2, nis, 1 &
1943	! /) )
1944	! ELSE
1945	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1946	id_var_do2d(av,if), &
1947	local_2d_sections(nxl:nxr, &
1948	nys:nyn,1:nis), &
1949	start = (/ nxl+1, nys+1, 1, &
1950	do2d_xy_time_count(av) /), &
1951	count = (/ nxr-nxl+1, &
1952	nyn-nys+1, nis, 1 &
1953	/) )
1954	! ENDIF
1955
1956	CALL netcdf_handle_error( 'data_output_2d', 55 )
1957
1958	CASE ( 'xz' )
1959	!
1960	!-- First, all PEs get the information of all cross-sections.
1961	!-- Then the data are written to the output file by all PEs
1962	!-- while NF90_COLLECTIVE is set in subroutine
1963	!-- define_netcdf_header. Although redundant information are
1964	!-- written to the output file in that case, the performance
1965	!-- is significantly better compared to the case where only
1966	!-- the first row of PEs in x-direction (myidx = 0) is given
1967	!-- the output while NF90_INDEPENDENT is set.
1968	IF ( npey /= 1 ) THEN
1969
1970	#if defined( __parallel )
1971	!
1972	!-- Distribute data over all PEs along y
1973	ngp = ( nxr-nxl+1 ) * ( nzt_do-nzb_do+1 ) * ns
1974	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1975	CALL MPI_ALLREDUCE( local_2d_sections_l(nxl,1,nzb_do), &
1976	local_2d_sections(nxl,1,nzb_do), &
1977	ngp, MPI_REAL, MPI_SUM, comm1dy, &
1978	ierr )
1979	#else
1980	local_2d_sections = local_2d_sections_l
1981	#endif
1982	ENDIF
1983	!
1984	!-- Do not output redundant ghost point data except for the
1985	!-- boundaries of the total domain.
1986	! IF ( nxr == nx ) THEN
1987	! nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1988	! id_var_do2d(av,if), &
1989	! local_2d_sections(nxl:nxr+1,1:ns, &
1990	! nzb_do:nzt_do), &
1991	! start = (/ nxl+1, 1, 1, &
1992	! do2d_xz_time_count(av) /), &
1993	! count = (/ nxr-nxl+2, ns, nzt_do-nzb_do+1, &
1994	! 1 /) )
1995	! ELSE
1996	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1997	id_var_do2d(av,if), &
1998	local_2d_sections(nxl:nxr,1:ns, &
1999	nzb_do:nzt_do), &
2000	start = (/ nxl+1, 1, 1, &
2001	do2d_xz_time_count(av) /), &
2002	count = (/ nxr-nxl+1, ns, nzt_do-nzb_do+1, &
2003	1 /) )
2004	! ENDIF
2005
2006	CALL netcdf_handle_error( 'data_output_2d', 57 )
2007
2008	CASE ( 'yz' )
2009	!
2010	!-- First, all PEs get the information of all cross-sections.
2011	!-- Then the data are written to the output file by all PEs
2012	!-- while NF90_COLLECTIVE is set in subroutine
2013	!-- define_netcdf_header. Although redundant information are
2014	!-- written to the output file in that case, the performance
2015	!-- is significantly better compared to the case where only
2016	!-- the first row of PEs in y-direction (myidy = 0) is given
2017	!-- the output while NF90_INDEPENDENT is set.
2018	IF ( npex /= 1 ) THEN
2019
2020	#if defined( __parallel )
2021	!
2022	!-- Distribute data over all PEs along x
2023	ngp = ( nyn-nys+1 ) * ( nzt-nzb + 2 ) * ns
2024	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
2025	CALL MPI_ALLREDUCE( local_2d_sections_l(1,nys,nzb_do), &
2026	local_2d_sections(1,nys,nzb_do), &
2027	ngp, MPI_REAL, MPI_SUM, comm1dx, &
2028	ierr )
2029	#else
2030	local_2d_sections = local_2d_sections_l
2031	#endif
2032	ENDIF
2033	!
2034	!-- Do not output redundant ghost point data except for the
2035	!-- boundaries of the total domain.
2036	! IF ( nyn == ny ) THEN
2037	! nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2038	! id_var_do2d(av,if), &
2039	! local_2d_sections(1:ns, &
2040	! nys:nyn+1,nzb_do:nzt_do), &
2041	! start = (/ 1, nys+1, 1, &
2042	! do2d_yz_time_count(av) /), &
2043	! count = (/ ns, nyn-nys+2, &
2044	! nzt_do-nzb_do+1, 1 /) )
2045	! ELSE
2046	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2047	id_var_do2d(av,if), &
2048	local_2d_sections(1:ns,nys:nyn, &
2049	nzb_do:nzt_do), &
2050	start = (/ 1, nys+1, 1, &
2051	do2d_yz_time_count(av) /), &
2052	count = (/ ns, nyn-nys+1, &
2053	nzt_do-nzb_do+1, 1 /) )
2054	! ENDIF
2055
2056	CALL netcdf_handle_error( 'data_output_2d', 60 )
2057
2058	CASE DEFAULT
2059	message_string = 'unknown cross-section: ' // TRIM( mode )
2060	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
2061
2062	END SELECT
2063
2064	ENDIF
2065	#endif
2066	ENDIF
2067
2068	if = if + 1
2069	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
2070	do2d_mode = do2d(av,if)(l-1:l)
2071
2072	ENDDO
2073
2074	!
2075	!-- Deallocate temporary arrays.
2076	IF ( ALLOCATED( level_z ) ) DEALLOCATE( level_z )
2077	IF ( netcdf_data_format > 4 ) THEN
2078	DEALLOCATE( local_pf, local_2d, local_2d_sections )
2079	IF( mode == 'xz' .OR. mode == 'yz' ) DEALLOCATE( local_2d_sections_l )
2080	ENDIF
2081	#if defined( __parallel )
2082	IF ( .NOT. data_output_2d_on_each_pe .AND. myid == 0 ) THEN
2083	DEALLOCATE( total_2d )
2084	ENDIF
2085	#endif
2086
2087	!
2088	!-- Close plot output file.
2089	file_id = 20 + s_ind
2090
2091	IF ( data_output_2d_on_each_pe ) THEN
2092	DO i = 0, io_blocks-1
2093	IF ( i == io_group ) THEN
2094	CALL close_file( file_id )
2095	ENDIF
2096	#if defined( __parallel )
2097	CALL MPI_BARRIER( comm2d, ierr )
2098	#endif
2099	ENDDO
2100	ELSE
2101	IF ( myid == 0 ) CALL close_file( file_id )
2102	ENDIF
2103
2104	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
2105
2106	END SUBROUTINE data_output_2d

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |