Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 3036

Last change on this file since 3036 was 3014, checked in by maronga, 7 years ago
series of bugfixes
Property svn:keywords set to `Id`
File size: 97.8 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 3014 2018-05-09 08:42:38Z gronemeier $
27	! Added nzb_do and nzt_do for some modules for 2d output
28	!
29	! 3004 2018-04-27 12:33:25Z Giersch
30	! precipitation_rate removed, case prr*_xy removed, to_be_resorted have to point
31	! to ql_vp_av and not to ql_vp, allocation checks implemented (averaged data
32	! will be assigned to fill values if no allocation happened so far)
33	!
34	! 2963 2018-04-12 14:47:44Z suehring
35	! Introduce index for vegetation/wall, pavement/green-wall and water/window
36	! surfaces, for clearer access of surface fraction, albedo, emissivity, etc. .
37	!
38	! 2817 2018-02-19 16:32:21Z knoop
39	! Preliminary gust module interface implemented
40	!
41	! 2805 2018-02-14 17:00:09Z suehring
42	! Consider also default-type surfaces for surface temperature output.
43	!
44	! 2797 2018-02-08 13:24:35Z suehring
45	! Enable output of ground-heat flux also at urban surfaces.
46	!
47	! 2743 2018-01-12 16:03:39Z suehring
48	! In case of natural- and urban-type surfaces output surfaces fluxes in W/m2.
49	!
50	! 2742 2018-01-12 14:59:47Z suehring
51	! Enable output of surface temperature
52	!
53	! 2735 2018-01-11 12:01:27Z suehring
54	! output of r_a moved from land-surface to consider also urban-type surfaces
55	!
56	! 2718 2018-01-02 08:49:38Z maronga
57	! Corrected "Former revisions" section
58	!
59	! 2696 2017-12-14 17:12:51Z kanani
60	! Change in file header (GPL part)
61	! Implementation of uv exposure model (FK)
62	! Implementation of turbulence_closure_mod (TG)
63	! Set fill values at topography grid points or e.g. non-natural-type surface
64	! in case of LSM output (MS)
65	!
66	! 2512 2017-10-04 08:26:59Z raasch
67	! upper bounds of cross section output changed from nx+1,ny+1 to nx,ny
68	! no output of ghost layer data
69	!
70	! 2292 2017-06-20 09:51:42Z schwenkel
71	! Implementation of new microphysic scheme: cloud_scheme = 'morrison'
72	! includes two more prognostic equations for cloud drop concentration (nc)
73	! and cloud water content (qc).
74	!
75	! 2277 2017-06-12 10:47:51Z kanani
76	! Removed unused variables do2d_xy_n, do2d_xz_n, do2d_yz_n
77	!
78	! 2233 2017-05-30 18:08:54Z suehring
79	!
80	! 2232 2017-05-30 17:47:52Z suehring
81	! Adjustments to new surface concept
82	!
83	!
84	! 2190 2017-03-21 12:16:43Z raasch
85	! bugfix for r2031: string rho replaced by rho_ocean
86	!
87	! 2031 2016-10-21 15:11:58Z knoop
88	! renamed variable rho to rho_ocean and rho_av to rho_ocean_av
89	!
90	! 2000 2016-08-20 18:09:15Z knoop
91	! Forced header and separation lines into 80 columns
92	!
93	! 1980 2016-07-29 15:51:57Z suehring
94	! Bugfix, in order to steer user-defined output, setting flag found explicitly
95	! to .F.
96	!
97	! 1976 2016-07-27 13:28:04Z maronga
98	! Output of radiation quantities is now done directly in the respective module
99	!
100	! 1972 2016-07-26 07:52:02Z maronga
101	! Output of land surface quantities is now done directly in the respective
102	! module
103	!
104	! 1960 2016-07-12 16:34:24Z suehring
105	! Scalar surface flux added
106	! Rename INTEGER variable s into s_ind, as s is already assigned to scalar
107	!
108	! 1849 2016-04-08 11:33:18Z hoffmann
109	! precipitation_amount, precipitation_rate, prr moved to arrays_3d
110	!
111	! 1822 2016-04-07 07:49:42Z hoffmann
112	! Output of bulk cloud physics simplified.
113	!
114	! 1788 2016-03-10 11:01:04Z maronga
115	! Added output of z0q
116	!
117	! 1783 2016-03-06 18:36:17Z raasch
118	! name change of netcdf routines and module + related changes
119	!
120	! 1745 2016-02-05 13:06:51Z gronemeier
121	! Bugfix: test if time axis limit exceeds moved to point after call of check_open
122	!
123	! 1703 2015-11-02 12:38:44Z raasch
124	! bugfix for output of single (*) xy-sections in case of parallel netcdf I/O
125	!
126	! 1701 2015-11-02 07:43:04Z maronga
127	! Bugfix in output of RRTGM data
128	!
129	! 1691 2015-10-26 16:17:44Z maronga
130	! Added output of Obukhov length (ol) and radiative heating rates for RRTMG.
131	! Formatting corrections.
132	!
133	! 1682 2015-10-07 23:56:08Z knoop
134	! Code annotations made doxygen readable
135	!
136	! 1585 2015-04-30 07:05:52Z maronga
137	! Added support for RRTMG
138	!
139	! 1555 2015-03-04 17:44:27Z maronga
140	! Added output of r_a and r_s
141	!
142	! 1551 2015-03-03 14:18:16Z maronga
143	! Added suppport for land surface model and radiation model output. In the course
144	! of this action, the limits for vertical loops have been changed (from nzb and
145	! nzt+1 to nzb_do and nzt_do, respectively in order to allow soil model output).
146	! Moreover, a new vertical grid zs was introduced.
147	!
148	! 1359 2014-04-11 17:15:14Z hoffmann
149	! New particle structure integrated.
150	!
151	! 1353 2014-04-08 15:21:23Z heinze
152	! REAL constants provided with KIND-attribute
153	!
154	! 1327 2014-03-21 11:00:16Z raasch
155	! parts concerning iso2d output removed,
156	! -netcdf output queries
157	!
158	! 1320 2014-03-20 08:40:49Z raasch
159	! ONLY-attribute added to USE-statements,
160	! kind-parameters added to all INTEGER and REAL declaration statements,
161	! kinds are defined in new module kinds,
162	! revision history before 2012 removed,
163	! comment fields (!:) to be used for variable explanations added to
164	! all variable declaration statements
165	!
166	! 1318 2014-03-17 13:35:16Z raasch
167	! barrier argument removed from cpu_log.
168	! module interfaces removed
169	!
170	! 1311 2014-03-14 12:13:39Z heinze
171	! bugfix: close #if defined( __netcdf )
172	!
173	! 1308 2014-03-13 14:58:42Z fricke
174	! +local_2d_sections, local_2d_sections_l, ns
175	! Check, if the limit of the time dimension is exceeded for parallel output
176	! To increase the performance for parallel output, the following is done:
177	! - Update of time axis is only done by PE0
178	! - Cross sections are first stored on a local array and are written
179	! collectively to the output file by all PEs.
180	!
181	! 1115 2013-03-26 18:16:16Z hoffmann
182	! ql is calculated by calc_liquid_water_content
183	!
184	! 1076 2012-12-05 08:30:18Z hoffmann
185	! Bugfix in output of ql
186	!
187	! 1065 2012-11-22 17:42:36Z hoffmann
188	! Bugfix: Output of cross sections of ql
189	!
190	! 1053 2012-11-13 17:11:03Z hoffmann
191	! +qr, nr, qc and cross sections
192	!
193	! 1036 2012-10-22 13:43:42Z raasch
194	! code put under GPL (PALM 3.9)
195	!
196	! 1031 2012-10-19 14:35:30Z raasch
197	! netCDF4 without parallel file support implemented
198	!
199	! 1007 2012-09-19 14:30:36Z franke
200	! Bugfix: missing calculation of ql_vp added
201	!
202	! 978 2012-08-09 08:28:32Z fricke
203	! +z0h
204	!
205	! Revision 1.1 1997/08/11 06:24:09 raasch
206	! Initial revision
207	!
208	!
209	! Description:
210	! ------------
211	!> Data output of cross-sections in netCDF format or binary format
212	!> to be later converted to NetCDF by helper routine combine_plot_fields.
213	!> Attention: The position of the sectional planes is still not always computed
214	!> --------- correctly. (zu is used always)!
215	!------------------------------------------------------------------------------!
216	SUBROUTINE data_output_2d( mode, av )
217
218
219	USE arrays_3d, &
220	ONLY: dzw, e, heatflux_output_conversion, nc, nr, p, pt, &
221	precipitation_amount, prr, q, qc, ql, ql_c, ql_v, ql_vp, qr, &
222	rho_ocean, s, sa, tend, u, v, vpt, w, zu, zw, &
223	waterflux_output_conversion
224
225	USE averaging
226
227	USE cloud_parameters, &
228	ONLY: cp, hyrho, l_d_cp, l_v, pt_d_t
229
230	USE control_parameters, &
231	ONLY: cloud_physics, data_output_2d_on_each_pe, data_output_xy, &
232	data_output_xz, data_output_yz, do2d, &
233	do2d_xy_last_time, do2d_xy_time_count, &
234	do2d_xz_last_time, do2d_xz_time_count, &
235	do2d_yz_last_time, do2d_yz_time_count, &
236	ibc_uv_b, io_blocks, io_group, land_surface, message_string, &
237	ntdim_2d_xy, ntdim_2d_xz, ntdim_2d_yz, &
238	psolver, section, simulated_time, simulated_time_chr, &
239	time_since_reference_point, uv_exposure
240
241	USE cpulog, &
242	ONLY: cpu_log, log_point
243
244	USE grid_variables, &
245	ONLY: dx, dy
246
247	USE gust_mod, &
248	ONLY: gust_data_output_2d, gust_module_enabled
249
250	USE indices, &
251	ONLY: nbgp, nx, nxl, nxlg, nxr, nxrg, ny, nyn, nyng, nys, nysg, nz, &
252	nzb, nzt, wall_flags_0
253
254	USE kinds
255
256	USE land_surface_model_mod, &
257	ONLY: lsm_data_output_2d, zs
258
259	#if defined( __netcdf )
260	USE NETCDF
261	#endif
262
263	USE netcdf_interface, &
264	ONLY: fill_value, id_set_xy, id_set_xz, id_set_yz, id_var_do2d, &
265	id_var_time_xy, id_var_time_xz, id_var_time_yz, nc_stat, &
266	netcdf_data_format, netcdf_handle_error
267
268	USE particle_attributes, &
269	ONLY: grid_particles, number_of_particles, particle_advection_start, &
270	particles, prt_count
271
272	USE pegrid
273
274	USE radiation_model_mod, &
275	ONLY: radiation, radiation_data_output_2d
276
277	USE surface_mod, &
278	ONLY: ind_pav_green, ind_veg_wall, ind_wat_win, surf_def_h, &
279	surf_lsm_h, surf_usm_h
280
281	USE turbulence_closure_mod, &
282	ONLY: tcm_data_output_2d
283
284	USE uv_exposure_model_mod, &
285	ONLY: uvem_data_output_2d
286
287
288	IMPLICIT NONE
289
290	CHARACTER (LEN=2) :: do2d_mode !<
291	CHARACTER (LEN=2) :: mode !<
292	CHARACTER (LEN=4) :: grid !<
293	CHARACTER (LEN=25) :: section_chr !<
294	CHARACTER (LEN=50) :: rtext !<
295
296	INTEGER(iwp) :: av !<
297	INTEGER(iwp) :: ngp !<
298	INTEGER(iwp) :: file_id !<
299	INTEGER(iwp) :: flag_nr !< number of masking flag
300	INTEGER(iwp) :: i !<
301	INTEGER(iwp) :: if !<
302	INTEGER(iwp) :: is !<
303	INTEGER(iwp) :: iis !<
304	INTEGER(iwp) :: j !<
305	INTEGER(iwp) :: k !<
306	INTEGER(iwp) :: l !<
307	INTEGER(iwp) :: layer_xy !<
308	INTEGER(iwp) :: m !<
309	INTEGER(iwp) :: n !<
310	INTEGER(iwp) :: nis !<
311	INTEGER(iwp) :: ns !<
312	INTEGER(iwp) :: nzb_do !< lower limit of the data field (usually nzb)
313	INTEGER(iwp) :: nzt_do !< upper limit of the data field (usually nzt+1)
314	INTEGER(iwp) :: psi !<
315	INTEGER(iwp) :: s_ind !<
316	INTEGER(iwp) :: sender !<
317	INTEGER(iwp) :: ind(4) !<
318
319	LOGICAL :: found !<
320	LOGICAL :: resorted !<
321	LOGICAL :: two_d !<
322
323	REAL(wp) :: mean_r !<
324	REAL(wp) :: s_r2 !<
325	REAL(wp) :: s_r3 !<
326
327	REAL(wp), DIMENSION(:), ALLOCATABLE :: level_z !<
328	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d !<
329	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d_l !<
330
331	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !<
332	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections !<
333	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections_l !<
334
335	#if defined( __parallel )
336	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: total_2d !<
337	#endif
338	REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !<
339
340	NAMELIST /LOCAL/ rtext
341
342	!
343	!-- Immediate return, if no output is requested (no respective sections
344	!-- found in parameter data_output)
345	IF ( mode == 'xy' .AND. .NOT. data_output_xy(av) ) RETURN
346	IF ( mode == 'xz' .AND. .NOT. data_output_xz(av) ) RETURN
347	IF ( mode == 'yz' .AND. .NOT. data_output_yz(av) ) RETURN
348
349	CALL cpu_log (log_point(3),'data_output_2d','start')
350
351	two_d = .FALSE. ! local variable to distinguish between output of pure 2D
352	! arrays and cross-sections of 3D arrays.
353
354	!
355	!-- Depending on the orientation of the cross-section, the respective output
356	!-- files have to be opened.
357	SELECT CASE ( mode )
358
359	CASE ( 'xy' )
360	s_ind = 1
361	ALLOCATE( level_z(nzb:nzt+1), local_2d(nxl:nxr,nys:nyn) )
362
363	IF ( netcdf_data_format > 4 ) THEN
364	ns = 1
365	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
366	ns = ns + 1
367	ENDDO
368	ns = ns - 1
369	ALLOCATE( local_2d_sections(nxl:nxr,nys:nyn,1:ns) )
370	local_2d_sections = 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( 101+av*10 )
377	ENDIF
378	IF ( data_output_2d_on_each_pe ) THEN
379	CALL check_open( 21 )
380	ELSE
381	IF ( myid == 0 ) THEN
382	#if defined( __parallel )
383	ALLOCATE( total_2d(0:nx,0:ny) )
384	#endif
385	ENDIF
386	ENDIF
387
388	CASE ( 'xz' )
389	s_ind = 2
390	ALLOCATE( local_2d(nxl:nxr,nzb:nzt+1) )
391
392	IF ( netcdf_data_format > 4 ) THEN
393	ns = 1
394	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
395	ns = ns + 1
396	ENDDO
397	ns = ns - 1
398	ALLOCATE( local_2d_sections(nxl:nxr,1:ns,nzb:nzt+1) )
399	ALLOCATE( local_2d_sections_l(nxl:nxr,1:ns,nzb:nzt+1) )
400	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
401	ENDIF
402
403	!
404	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
405	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
406	CALL check_open( 102+av*10 )
407	ENDIF
408
409	IF ( data_output_2d_on_each_pe ) THEN
410	CALL check_open( 22 )
411	ELSE
412	IF ( myid == 0 ) THEN
413	#if defined( __parallel )
414	ALLOCATE( total_2d(0:nx,nzb:nzt+1) )
415	#endif
416	ENDIF
417	ENDIF
418
419	CASE ( 'yz' )
420	s_ind = 3
421	ALLOCATE( local_2d(nys:nyn,nzb:nzt+1) )
422
423	IF ( netcdf_data_format > 4 ) THEN
424	ns = 1
425	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
426	ns = ns + 1
427	ENDDO
428	ns = ns - 1
429	ALLOCATE( local_2d_sections(1:ns,nys:nyn,nzb:nzt+1) )
430	ALLOCATE( local_2d_sections_l(1:ns,nys:nyn,nzb:nzt+1) )
431	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
432	ENDIF
433
434	!
435	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
436	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
437	CALL check_open( 103+av*10 )
438	ENDIF
439
440	IF ( data_output_2d_on_each_pe ) THEN
441	CALL check_open( 23 )
442	ELSE
443	IF ( myid == 0 ) THEN
444	#if defined( __parallel )
445	ALLOCATE( total_2d(0:ny,nzb:nzt+1) )
446	#endif
447	ENDIF
448	ENDIF
449
450	CASE DEFAULT
451	message_string = 'unknown cross-section: ' // TRIM( mode )
452	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
453
454	END SELECT
455
456	!
457	!-- For parallel netcdf output the time axis must be limited. Return, if this
458	!-- limit is exceeded. This could be the case, if the simulated time exceeds
459	!-- the given end time by the length of the given output interval.
460	IF ( netcdf_data_format > 4 ) THEN
461	IF ( mode == 'xy' .AND. do2d_xy_time_count(av) + 1 > &
462	ntdim_2d_xy(av) ) THEN
463	WRITE ( message_string, * ) 'Output of xy cross-sections is not ', &
464	'given at t=', simulated_time, '&because the', &
465	' maximum number of output time levels is exceeded.'
466	CALL message( 'data_output_2d', 'PA0384', 0, 1, 0, 6, 0 )
467	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
468	RETURN
469	ENDIF
470	IF ( mode == 'xz' .AND. do2d_xz_time_count(av) + 1 > &
471	ntdim_2d_xz(av) ) THEN
472	WRITE ( message_string, * ) 'Output of xz cross-sections is not ', &
473	'given at t=', simulated_time, '&because the', &
474	' maximum number of output time levels is exceeded.'
475	CALL message( 'data_output_2d', 'PA0385', 0, 1, 0, 6, 0 )
476	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
477	RETURN
478	ENDIF
479	IF ( mode == 'yz' .AND. do2d_yz_time_count(av) + 1 > &
480	ntdim_2d_yz(av) ) THEN
481	WRITE ( message_string, * ) 'Output of yz cross-sections is not ', &
482	'given at t=', simulated_time, '&because the', &
483	' maximum number of output time levels is exceeded.'
484	CALL message( 'data_output_2d', 'PA0386', 0, 1, 0, 6, 0 )
485	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
486	RETURN
487	ENDIF
488	ENDIF
489
490	!
491	!-- Allocate a temporary array for resorting (kji -> ijk).
492	ALLOCATE( local_pf(nxl:nxr,nys:nyn,nzb:nzt+1) )
493	local_pf = 0.0
494
495	!
496	!-- Loop of all variables to be written.
497	!-- Output dimensions chosen
498	if = 1
499	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
500	do2d_mode = do2d(av,if)(l-1:l)
501
502	DO WHILE ( do2d(av,if)(1:1) /= ' ' )
503
504	IF ( do2d_mode == mode ) THEN
505	!
506	!-- Set flag to steer output of radiation, land-surface, or user-defined
507	!-- quantities
508	found = .FALSE.
509
510	nzb_do = nzb
511	nzt_do = nzt+1
512	!
513	!-- Before each output, set array local_pf to fill value
514	local_pf = fill_value
515	!
516	!-- Set masking flag for topography for not resorted arrays
517	flag_nr = 0
518
519	!
520	!-- Store the array chosen on the temporary array.
521	resorted = .FALSE.
522	SELECT CASE ( TRIM( do2d(av,if) ) )
523	CASE ( 'e_xy', 'e_xz', 'e_yz' )
524	IF ( av == 0 ) THEN
525	to_be_resorted => e
526	ELSE
527	IF ( .NOT. ALLOCATED( e_av ) ) THEN
528	ALLOCATE( e_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
529	e_av = REAL( fill_value, KIND = wp )
530	ENDIF
531	to_be_resorted => e_av
532	ENDIF
533	IF ( mode == 'xy' ) level_z = zu
534
535	CASE ( 'lpt_xy', 'lpt_xz', 'lpt_yz' )
536	IF ( av == 0 ) THEN
537	to_be_resorted => pt
538	ELSE
539	IF ( .NOT. ALLOCATED( lpt_av ) ) THEN
540	ALLOCATE( lpt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
541	lpt_av = REAL( fill_value, KIND = wp )
542	ENDIF
543	to_be_resorted => lpt_av
544	ENDIF
545	IF ( mode == 'xy' ) level_z = zu
546
547	CASE ( 'lwp*_xy' ) ! 2d-array
548	IF ( av == 0 ) THEN
549	DO i = nxl, nxr
550	DO j = nys, nyn
551	local_pf(i,j,nzb+1) = SUM( ql(nzb:nzt,j,i) * &
552	dzw(1:nzt+1) )
553	ENDDO
554	ENDDO
555	ELSE
556	IF ( .NOT. ALLOCATED( lwp_av ) ) THEN
557	ALLOCATE( lwp_av(nysg:nyng,nxlg:nxrg) )
558	lwp_av = REAL( fill_value, KIND = wp )
559	ENDIF
560	DO i = nxl, nxr
561	DO j = nys, nyn
562	local_pf(i,j,nzb+1) = lwp_av(j,i)
563	ENDDO
564	ENDDO
565	ENDIF
566	resorted = .TRUE.
567	two_d = .TRUE.
568	level_z(nzb+1) = zu(nzb+1)
569
570	CASE ( 'nc_xy', 'nc_xz', 'nc_yz' )
571	IF ( av == 0 ) THEN
572	to_be_resorted => nc
573	ELSE
574	IF ( .NOT. ALLOCATED( nc_av ) ) THEN
575	ALLOCATE( nc_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
576	nc_av = REAL( fill_value, KIND = wp )
577	ENDIF
578	to_be_resorted => nc_av
579	ENDIF
580	IF ( mode == 'xy' ) level_z = zu
581
582	CASE ( 'nr_xy', 'nr_xz', 'nr_yz' )
583	IF ( av == 0 ) THEN
584	to_be_resorted => nr
585	ELSE
586	IF ( .NOT. ALLOCATED( nr_av ) ) THEN
587	ALLOCATE( nr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
588	nr_av = REAL( fill_value, KIND = wp )
589	ENDIF
590	to_be_resorted => nr_av
591	ENDIF
592	IF ( mode == 'xy' ) level_z = zu
593
594	CASE ( 'ghf*_xy' ) ! 2d-array
595	IF ( av == 0 ) THEN
596	DO m = 1, surf_lsm_h%ns
597	i = surf_lsm_h%i(m)
598	j = surf_lsm_h%j(m)
599	local_pf(i,j,nzb+1) = surf_lsm_h%ghf(m)
600	ENDDO
601	DO m = 1, surf_usm_h%ns
602	i = surf_usm_h%i(m)
603	j = surf_usm_h%j(m)
604	local_pf(i,j,nzb+1) = surf_usm_h%frac(ind_veg_wall,m) * &
605	surf_usm_h%wghf_eb(m) + &
606	surf_usm_h%frac(ind_pav_green,m) * &
607	surf_usm_h%wghf_eb_green(m) + &
608	surf_usm_h%frac(ind_wat_win,m) * &
609	surf_usm_h%wghf_eb_window(m)
610	ENDDO
611	ELSE
612	IF ( .NOT. ALLOCATED( ghf_av ) ) THEN
613	ALLOCATE( ghf_av(nysg:nyng,nxlg:nxrg) )
614	ghf_av = REAL( fill_value, KIND = wp )
615	ENDIF
616	DO i = nxl, nxr
617	DO j = nys, nyn
618	local_pf(i,j,nzb+1) = ghf_av(j,i)
619	ENDDO
620	ENDDO
621	ENDIF
622
623	resorted = .TRUE.
624	two_d = .TRUE.
625	level_z(nzb+1) = zu(nzb+1)
626
627	CASE ( 'ol*_xy' ) ! 2d-array
628	IF ( av == 0 ) THEN
629	DO m = 1, surf_def_h(0)%ns
630	i = surf_def_h(0)%i(m)
631	j = surf_def_h(0)%j(m)
632	local_pf(i,j,nzb+1) = surf_def_h(0)%ol(m)
633	ENDDO
634	DO m = 1, surf_lsm_h%ns
635	i = surf_lsm_h%i(m)
636	j = surf_lsm_h%j(m)
637	local_pf(i,j,nzb+1) = surf_lsm_h%ol(m)
638	ENDDO
639	DO m = 1, surf_usm_h%ns
640	i = surf_usm_h%i(m)
641	j = surf_usm_h%j(m)
642	local_pf(i,j,nzb+1) = surf_usm_h%ol(m)
643	ENDDO
644	ELSE
645	IF ( .NOT. ALLOCATED( ol_av ) ) THEN
646	ALLOCATE( ol_av(nysg:nyng,nxlg:nxrg) )
647	ol_av = REAL( fill_value, KIND = wp )
648	ENDIF
649	DO i = nxl, nxr
650	DO j = nys, nyn
651	local_pf(i,j,nzb+1) = ol_av(j,i)
652	ENDDO
653	ENDDO
654	ENDIF
655	resorted = .TRUE.
656	two_d = .TRUE.
657	level_z(nzb+1) = zu(nzb+1)
658
659	CASE ( 'p_xy', 'p_xz', 'p_yz' )
660	IF ( av == 0 ) THEN
661	IF ( psolver /= 'sor' ) CALL exchange_horiz( p, nbgp )
662	to_be_resorted => p
663	ELSE
664	IF ( .NOT. ALLOCATED( p_av ) ) THEN
665	ALLOCATE( p_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
666	p_av = REAL( fill_value, KIND = wp )
667	ENDIF
668	IF ( psolver /= 'sor' ) CALL exchange_horiz( p_av, nbgp )
669	to_be_resorted => p_av
670	ENDIF
671	IF ( mode == 'xy' ) level_z = zu
672
673	CASE ( 'pc_xy', 'pc_xz', 'pc_yz' ) ! particle concentration
674	IF ( av == 0 ) THEN
675	IF ( simulated_time >= particle_advection_start ) THEN
676	tend = prt_count
677	! CALL exchange_horiz( tend, nbgp )
678	ELSE
679	tend = 0.0_wp
680	ENDIF
681	DO i = nxl, nxr
682	DO j = nys, nyn
683	DO k = nzb, nzt+1
684	local_pf(i,j,k) = tend(k,j,i)
685	ENDDO
686	ENDDO
687	ENDDO
688	resorted = .TRUE.
689	ELSE
690	IF ( .NOT. ALLOCATED( pc_av ) ) THEN
691	ALLOCATE( pc_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
692	pc_av = REAL( fill_value, KIND = wp )
693	ENDIF
694	! CALL exchange_horiz( pc_av, nbgp )
695	to_be_resorted => pc_av
696	ENDIF
697
698	CASE ( 'pr_xy', 'pr_xz', 'pr_yz' ) ! mean particle radius (effective radius)
699	IF ( av == 0 ) THEN
700	IF ( simulated_time >= particle_advection_start ) THEN
701	DO i = nxl, nxr
702	DO j = nys, nyn
703	DO k = nzb, nzt+1
704	number_of_particles = prt_count(k,j,i)
705	IF (number_of_particles <= 0) CYCLE
706	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
707	s_r2 = 0.0_wp
708	s_r3 = 0.0_wp
709	DO n = 1, number_of_particles
710	IF ( particles(n)%particle_mask ) THEN
711	s_r2 = s_r2 + particles(n)%radius*2 &
712	particles(n)%weight_factor
713	s_r3 = s_r3 + particles(n)%radius*3 &
714	particles(n)%weight_factor
715	ENDIF
716	ENDDO
717	IF ( s_r2 > 0.0_wp ) THEN
718	mean_r = s_r3 / s_r2
719	ELSE
720	mean_r = 0.0_wp
721	ENDIF
722	tend(k,j,i) = mean_r
723	ENDDO
724	ENDDO
725	ENDDO
726	! CALL exchange_horiz( tend, nbgp )
727	ELSE
728	tend = 0.0_wp
729	ENDIF
730	DO i = nxl, nxr
731	DO j = nys, nyn
732	DO k = nzb, nzt+1
733	local_pf(i,j,k) = tend(k,j,i)
734	ENDDO
735	ENDDO
736	ENDDO
737	resorted = .TRUE.
738	ELSE
739	IF ( .NOT. ALLOCATED( pr_av ) ) THEN
740	ALLOCATE( pr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
741	pr_av = REAL( fill_value, KIND = wp )
742	ENDIF
743	! CALL exchange_horiz( pr_av, nbgp )
744	to_be_resorted => pr_av
745	ENDIF
746
747	CASE ( 'pra*_xy' ) ! 2d-array / integral quantity => no av
748	! CALL exchange_horiz_2d( precipitation_amount )
749	DO i = nxl, nxr
750	DO j = nys, nyn
751	local_pf(i,j,nzb+1) = precipitation_amount(j,i)
752	ENDDO
753	ENDDO
754	precipitation_amount = 0.0_wp ! reset for next integ. interval
755	resorted = .TRUE.
756	two_d = .TRUE.
757	level_z(nzb+1) = zu(nzb+1)
758
759	CASE ( 'prr_xy', 'prr_xz', 'prr_yz' )
760	IF ( av == 0 ) THEN
761	! CALL exchange_horiz( prr, nbgp )
762	DO i = nxl, nxr
763	DO j = nys, nyn
764	DO k = nzb, nzt+1
765	local_pf(i,j,k) = prr(k,j,i) * hyrho(nzb+1)
766	ENDDO
767	ENDDO
768	ENDDO
769	ELSE
770	IF ( .NOT. ALLOCATED( prr_av ) ) THEN
771	ALLOCATE( prr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
772	prr_av = REAL( fill_value, KIND = wp )
773	ENDIF
774	! CALL exchange_horiz( prr_av, nbgp )
775	DO i = nxl, nxr
776	DO j = nys, nyn
777	DO k = nzb, nzt+1
778	local_pf(i,j,k) = prr_av(k,j,i) * hyrho(nzb+1)
779	ENDDO
780	ENDDO
781	ENDDO
782	ENDIF
783	resorted = .TRUE.
784	IF ( mode == 'xy' ) level_z = zu
785
786	CASE ( 'pt_xy', 'pt_xz', 'pt_yz' )
787	IF ( av == 0 ) THEN
788	IF ( .NOT. cloud_physics ) THEN
789	to_be_resorted => pt
790	ELSE
791	DO i = nxl, nxr
792	DO j = nys, nyn
793	DO k = nzb, nzt+1
794	local_pf(i,j,k) = pt(k,j,i) + l_d_cp * &
795	pt_d_t(k) * &
796	ql(k,j,i)
797	ENDDO
798	ENDDO
799	ENDDO
800	resorted = .TRUE.
801	ENDIF
802	ELSE
803	IF ( .NOT. ALLOCATED( pt_av ) ) THEN
804	ALLOCATE( pt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
805	pt_av = REAL( fill_value, KIND = wp )
806	ENDIF
807	to_be_resorted => pt_av
808	ENDIF
809	IF ( mode == 'xy' ) level_z = zu
810
811	CASE ( 'q_xy', 'q_xz', 'q_yz' )
812	IF ( av == 0 ) THEN
813	to_be_resorted => q
814	ELSE
815	IF ( .NOT. ALLOCATED( q_av ) ) THEN
816	ALLOCATE( q_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
817	q_av = REAL( fill_value, KIND = wp )
818	ENDIF
819	to_be_resorted => q_av
820	ENDIF
821	IF ( mode == 'xy' ) level_z = zu
822
823	CASE ( 'qc_xy', 'qc_xz', 'qc_yz' )
824	IF ( av == 0 ) THEN
825	to_be_resorted => qc
826	ELSE
827	IF ( .NOT. ALLOCATED( qc_av ) ) THEN
828	ALLOCATE( qc_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
829	qc_av = REAL( fill_value, KIND = wp )
830	ENDIF
831	to_be_resorted => qc_av
832	ENDIF
833	IF ( mode == 'xy' ) level_z = zu
834
835	CASE ( 'ql_xy', 'ql_xz', 'ql_yz' )
836	IF ( av == 0 ) THEN
837	to_be_resorted => ql
838	ELSE
839	IF ( .NOT. ALLOCATED( ql_av ) ) THEN
840	ALLOCATE( ql_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
841	ql_av = REAL( fill_value, KIND = wp )
842	ENDIF
843	to_be_resorted => ql_av
844	ENDIF
845	IF ( mode == 'xy' ) level_z = zu
846
847	CASE ( 'ql_c_xy', 'ql_c_xz', 'ql_c_yz' )
848	IF ( av == 0 ) THEN
849	to_be_resorted => ql_c
850	ELSE
851	IF ( .NOT. ALLOCATED( ql_c_av ) ) THEN
852	ALLOCATE( ql_c_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
853	ql_c_av = REAL( fill_value, KIND = wp )
854	ENDIF
855	to_be_resorted => ql_c_av
856	ENDIF
857	IF ( mode == 'xy' ) level_z = zu
858
859	CASE ( 'ql_v_xy', 'ql_v_xz', 'ql_v_yz' )
860	IF ( av == 0 ) THEN
861	to_be_resorted => ql_v
862	ELSE
863	IF ( .NOT. ALLOCATED( ql_v_av ) ) THEN
864	ALLOCATE( ql_v_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
865	ql_v_av = REAL( fill_value, KIND = wp )
866	ENDIF
867	to_be_resorted => ql_v_av
868	ENDIF
869	IF ( mode == 'xy' ) level_z = zu
870
871	CASE ( 'ql_vp_xy', 'ql_vp_xz', 'ql_vp_yz' )
872	IF ( av == 0 ) THEN
873	IF ( simulated_time >= particle_advection_start ) THEN
874	DO i = nxl, nxr
875	DO j = nys, nyn
876	DO k = nzb, nzt+1
877	number_of_particles = prt_count(k,j,i)
878	IF (number_of_particles <= 0) CYCLE
879	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
880	DO n = 1, number_of_particles
881	IF ( particles(n)%particle_mask ) THEN
882	tend(k,j,i) = tend(k,j,i) + &
883	particles(n)%weight_factor / &
884	prt_count(k,j,i)
885	ENDIF
886	ENDDO
887	ENDDO
888	ENDDO
889	ENDDO
890	! CALL exchange_horiz( tend, nbgp )
891	ELSE
892	tend = 0.0_wp
893	ENDIF
894	DO i = nxl, nxr
895	DO j = nys, nyn
896	DO k = nzb, nzt+1
897	local_pf(i,j,k) = tend(k,j,i)
898	ENDDO
899	ENDDO
900	ENDDO
901	resorted = .TRUE.
902	ELSE
903	IF ( .NOT. ALLOCATED( ql_vp_av ) ) THEN
904	ALLOCATE( ql_vp_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
905	ql_vp_av = REAL( fill_value, KIND = wp )
906	ENDIF
907	! CALL exchange_horiz( ql_vp_av, nbgp )
908	to_be_resorted => ql_vp_av
909	ENDIF
910	IF ( mode == 'xy' ) level_z = zu
911
912	CASE ( 'qr_xy', 'qr_xz', 'qr_yz' )
913	IF ( av == 0 ) THEN
914	to_be_resorted => qr
915	ELSE
916	IF ( .NOT. ALLOCATED( qr_av ) ) THEN
917	ALLOCATE( qr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
918	qr_av = REAL( fill_value, KIND = wp )
919	ENDIF
920	to_be_resorted => qr_av
921	ENDIF
922	IF ( mode == 'xy' ) level_z = zu
923
924	CASE ( 'qsws*_xy' ) ! 2d-array
925	IF ( av == 0 ) THEN
926	!
927	!-- In case of default surfaces, clean-up flux by density.
928	!-- In case of land- and urban-surfaces, convert fluxes into
929	!-- dynamic units
930	DO m = 1, surf_def_h(0)%ns
931	i = surf_def_h(0)%i(m)
932	j = surf_def_h(0)%j(m)
933	k = surf_def_h(0)%k(m)
934	local_pf(i,j,nzb+1) = surf_def_h(0)%qsws(m) * &
935	waterflux_output_conversion(k)
936	ENDDO
937	DO m = 1, surf_lsm_h%ns
938	i = surf_lsm_h%i(m)
939	j = surf_lsm_h%j(m)
940	k = surf_lsm_h%k(m)
941	local_pf(i,j,nzb+1) = surf_lsm_h%qsws(m) * l_v
942	ENDDO
943	DO m = 1, surf_usm_h%ns
944	i = surf_usm_h%i(m)
945	j = surf_usm_h%j(m)
946	k = surf_usm_h%k(m)
947	local_pf(i,j,nzb+1) = surf_usm_h%qsws(m) * l_v
948	ENDDO
949	ELSE
950	IF ( .NOT. ALLOCATED( qsws_av ) ) THEN
951	ALLOCATE( qsws_av(nysg:nyng,nxlg:nxrg) )
952	qsws_av = REAL( fill_value, KIND = wp )
953	ENDIF
954	DO i = nxl, nxr
955	DO j = nys, nyn
956	local_pf(i,j,nzb+1) = qsws_av(j,i)
957	ENDDO
958	ENDDO
959	ENDIF
960	resorted = .TRUE.
961	two_d = .TRUE.
962	level_z(nzb+1) = zu(nzb+1)
963
964	CASE ( 'qv_xy', 'qv_xz', 'qv_yz' )
965	IF ( av == 0 ) THEN
966	DO i = nxl, nxr
967	DO j = nys, nyn
968	DO k = nzb, nzt+1
969	local_pf(i,j,k) = q(k,j,i) - ql(k,j,i)
970	ENDDO
971	ENDDO
972	ENDDO
973	resorted = .TRUE.
974	ELSE
975	IF ( .NOT. ALLOCATED( qv_av ) ) THEN
976	ALLOCATE( qv_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
977	qv_av = REAL( fill_value, KIND = wp )
978	ENDIF
979	to_be_resorted => qv_av
980	ENDIF
981	IF ( mode == 'xy' ) level_z = zu
982
983	CASE ( 'r_a*_xy' ) ! 2d-array
984	IF ( av == 0 ) THEN
985	DO m = 1, surf_lsm_h%ns
986	i = surf_lsm_h%i(m)
987	j = surf_lsm_h%j(m)
988	local_pf(i,j,nzb+1) = surf_lsm_h%r_a(m)
989	ENDDO
990
991	DO m = 1, surf_usm_h%ns
992	i = surf_usm_h%i(m)
993	j = surf_usm_h%j(m)
994	local_pf(i,j,nzb+1) = &
995	( surf_usm_h%frac(ind_veg_wall,m) * &
996	surf_usm_h%r_a(m) + &
997	surf_usm_h%frac(ind_pav_green,m) * &
998	surf_usm_h%r_a_green(m) + &
999	surf_usm_h%frac(ind_wat_win,m) * &
1000	surf_usm_h%r_a_window(m) )
1001	ENDDO
1002	ELSE
1003	IF ( .NOT. ALLOCATED( r_a_av ) ) THEN
1004	ALLOCATE( r_a_av(nysg:nyng,nxlg:nxrg) )
1005	r_a_av = REAL( fill_value, KIND = wp )
1006	ENDIF
1007	DO i = nxl, nxr
1008	DO j = nys, nyn
1009	local_pf(i,j,nzb+1) = r_a_av(j,i)
1010	ENDDO
1011	ENDDO
1012	ENDIF
1013	resorted = .TRUE.
1014	two_d = .TRUE.
1015	level_z(nzb+1) = zu(nzb+1)
1016
1017	CASE ( 'rho_ocean_xy', 'rho_ocean_xz', 'rho_ocean_yz' )
1018	IF ( av == 0 ) THEN
1019	to_be_resorted => rho_ocean
1020	ELSE
1021	IF ( .NOT. ALLOCATED( rho_ocean_av ) ) THEN
1022	ALLOCATE( rho_ocean_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1023	rho_ocean_av = REAL( fill_value, KIND = wp )
1024	ENDIF
1025	to_be_resorted => rho_ocean_av
1026	ENDIF
1027
1028	CASE ( 's_xy', 's_xz', 's_yz' )
1029	IF ( av == 0 ) THEN
1030	to_be_resorted => s
1031	ELSE
1032	IF ( .NOT. ALLOCATED( s_av ) ) THEN
1033	ALLOCATE( s_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1034	s_av = REAL( fill_value, KIND = wp )
1035	ENDIF
1036	to_be_resorted => s_av
1037	ENDIF
1038
1039	CASE ( 'sa_xy', 'sa_xz', 'sa_yz' )
1040	IF ( av == 0 ) THEN
1041	to_be_resorted => sa
1042	ELSE
1043	IF ( .NOT. ALLOCATED( sa_av ) ) THEN
1044	ALLOCATE( sa_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1045	sa_av = REAL( fill_value, KIND = wp )
1046	ENDIF
1047	to_be_resorted => sa_av
1048	ENDIF
1049
1050	CASE ( 'shf*_xy' ) ! 2d-array
1051	IF ( av == 0 ) THEN
1052	!
1053	!-- In case of default surfaces, clean-up flux by density.
1054	!-- In case of land- and urban-surfaces, convert fluxes into
1055	!-- dynamic units.
1056	DO m = 1, surf_def_h(0)%ns
1057	i = surf_def_h(0)%i(m)
1058	j = surf_def_h(0)%j(m)
1059	k = surf_def_h(0)%k(m)
1060	local_pf(i,j,nzb+1) = surf_def_h(0)%shf(m) * &
1061	heatflux_output_conversion(k)
1062	ENDDO
1063	DO m = 1, surf_lsm_h%ns
1064	i = surf_lsm_h%i(m)
1065	j = surf_lsm_h%j(m)
1066	k = surf_lsm_h%k(m)
1067	local_pf(i,j,nzb+1) = surf_lsm_h%shf(m) * cp
1068	ENDDO
1069	DO m = 1, surf_usm_h%ns
1070	i = surf_usm_h%i(m)
1071	j = surf_usm_h%j(m)
1072	k = surf_usm_h%k(m)
1073	local_pf(i,j,nzb+1) = surf_usm_h%shf(m) * cp
1074	ENDDO
1075	ELSE
1076	IF ( .NOT. ALLOCATED( shf_av ) ) THEN
1077	ALLOCATE( shf_av(nysg:nyng,nxlg:nxrg) )
1078	shf_av = REAL( fill_value, KIND = wp )
1079	ENDIF
1080	DO i = nxl, nxr
1081	DO j = nys, nyn
1082	local_pf(i,j,nzb+1) = shf_av(j,i)
1083	ENDDO
1084	ENDDO
1085	ENDIF
1086	resorted = .TRUE.
1087	two_d = .TRUE.
1088	level_z(nzb+1) = zu(nzb+1)
1089
1090	CASE ( 'ssws*_xy' ) ! 2d-array
1091	IF ( av == 0 ) THEN
1092	DO m = 1, surf_def_h(0)%ns
1093	i = surf_def_h(0)%i(m)
1094	j = surf_def_h(0)%j(m)
1095	local_pf(i,j,nzb+1) = surf_def_h(0)%ssws(m)
1096	ENDDO
1097	DO m = 1, surf_lsm_h%ns
1098	i = surf_lsm_h%i(m)
1099	j = surf_lsm_h%j(m)
1100	local_pf(i,j,nzb+1) = surf_lsm_h%ssws(m)
1101	ENDDO
1102	DO m = 1, surf_usm_h%ns
1103	i = surf_usm_h%i(m)
1104	j = surf_usm_h%j(m)
1105	local_pf(i,j,nzb+1) = surf_usm_h%ssws(m)
1106	ENDDO
1107	ELSE
1108	IF ( .NOT. ALLOCATED( ssws_av ) ) THEN
1109	ALLOCATE( ssws_av(nysg:nyng,nxlg:nxrg) )
1110	ssws_av = REAL( fill_value, KIND = wp )
1111	ENDIF
1112	DO i = nxl, nxr
1113	DO j = nys, nyn
1114	local_pf(i,j,nzb+1) = ssws_av(j,i)
1115	ENDDO
1116	ENDDO
1117	ENDIF
1118	resorted = .TRUE.
1119	two_d = .TRUE.
1120	level_z(nzb+1) = zu(nzb+1)
1121
1122	CASE ( 't*_xy' ) ! 2d-array
1123	IF ( av == 0 ) THEN
1124	DO m = 1, surf_def_h(0)%ns
1125	i = surf_def_h(0)%i(m)
1126	j = surf_def_h(0)%j(m)
1127	local_pf(i,j,nzb+1) = surf_def_h(0)%ts(m)
1128	ENDDO
1129	DO m = 1, surf_lsm_h%ns
1130	i = surf_lsm_h%i(m)
1131	j = surf_lsm_h%j(m)
1132	local_pf(i,j,nzb+1) = surf_lsm_h%ts(m)
1133	ENDDO
1134	DO m = 1, surf_usm_h%ns
1135	i = surf_usm_h%i(m)
1136	j = surf_usm_h%j(m)
1137	local_pf(i,j,nzb+1) = surf_usm_h%ts(m)
1138	ENDDO
1139	ELSE
1140	IF ( .NOT. ALLOCATED( ts_av ) ) THEN
1141	ALLOCATE( ts_av(nysg:nyng,nxlg:nxrg) )
1142	ts_av = REAL( fill_value, KIND = wp )
1143	ENDIF
1144	DO i = nxl, nxr
1145	DO j = nys, nyn
1146	local_pf(i,j,nzb+1) = ts_av(j,i)
1147	ENDDO
1148	ENDDO
1149	ENDIF
1150	resorted = .TRUE.
1151	two_d = .TRUE.
1152	level_z(nzb+1) = zu(nzb+1)
1153
1154	CASE ( 'tsurf*_xy' ) ! 2d-array
1155	IF ( av == 0 ) THEN
1156	DO m = 1, surf_def_h(0)%ns
1157	i = surf_def_h(0)%i(m)
1158	j = surf_def_h(0)%j(m)
1159	local_pf(i,j,nzb+1) = surf_def_h(0)%pt_surface(m)
1160	ENDDO
1161
1162	DO m = 1, surf_lsm_h%ns
1163	i = surf_lsm_h%i(m)
1164	j = surf_lsm_h%j(m)
1165	local_pf(i,j,nzb+1) = surf_lsm_h%pt_surface(m)
1166	ENDDO
1167
1168	DO m = 1, surf_usm_h%ns
1169	i = surf_usm_h%i(m)
1170	j = surf_usm_h%j(m)
1171	local_pf(i,j,nzb+1) = surf_usm_h%pt_surface(m)
1172	ENDDO
1173
1174	ELSE
1175	IF ( .NOT. ALLOCATED( tsurf_av ) ) THEN
1176	ALLOCATE( tsurf_av(nysg:nyng,nxlg:nxrg) )
1177	tsurf_av = REAL( fill_value, KIND = wp )
1178	ENDIF
1179	DO i = nxl, nxr
1180	DO j = nys, nyn
1181	local_pf(i,j,nzb+1) = tsurf_av(j,i)
1182	ENDDO
1183	ENDDO
1184	ENDIF
1185	resorted = .TRUE.
1186	two_d = .TRUE.
1187	level_z(nzb+1) = zu(nzb+1)
1188
1189	CASE ( 'u_xy', 'u_xz', 'u_yz' )
1190	flag_nr = 1
1191	IF ( av == 0 ) THEN
1192	to_be_resorted => u
1193	ELSE
1194	IF ( .NOT. ALLOCATED( u_av ) ) THEN
1195	ALLOCATE( u_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1196	u_av = REAL( fill_value, KIND = wp )
1197	ENDIF
1198	to_be_resorted => u_av
1199	ENDIF
1200	IF ( mode == 'xy' ) level_z = zu
1201	!
1202	!-- Substitute the values generated by "mirror" boundary condition
1203	!-- at the bottom boundary by the real surface values.
1204	IF ( do2d(av,if) == 'u_xz' .OR. do2d(av,if) == 'u_yz' ) THEN
1205	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
1206	ENDIF
1207
1208	CASE ( 'u*_xy' ) ! 2d-array
1209	IF ( av == 0 ) THEN
1210	DO m = 1, surf_def_h(0)%ns
1211	i = surf_def_h(0)%i(m)
1212	j = surf_def_h(0)%j(m)
1213	local_pf(i,j,nzb+1) = surf_def_h(0)%us(m)
1214	ENDDO
1215	DO m = 1, surf_lsm_h%ns
1216	i = surf_lsm_h%i(m)
1217	j = surf_lsm_h%j(m)
1218	local_pf(i,j,nzb+1) = surf_lsm_h%us(m)
1219	ENDDO
1220	DO m = 1, surf_usm_h%ns
1221	i = surf_usm_h%i(m)
1222	j = surf_usm_h%j(m)
1223	local_pf(i,j,nzb+1) = surf_usm_h%us(m)
1224	ENDDO
1225	ELSE
1226	IF ( .NOT. ALLOCATED( us_av ) ) THEN
1227	ALLOCATE( us_av(nysg:nyng,nxlg:nxrg) )
1228	us_av = REAL( fill_value, KIND = wp )
1229	ENDIF
1230	DO i = nxl, nxr
1231	DO j = nys, nyn
1232	local_pf(i,j,nzb+1) = us_av(j,i)
1233	ENDDO
1234	ENDDO
1235	ENDIF
1236	resorted = .TRUE.
1237	two_d = .TRUE.
1238	level_z(nzb+1) = zu(nzb+1)
1239
1240	CASE ( 'v_xy', 'v_xz', 'v_yz' )
1241	flag_nr = 2
1242	IF ( av == 0 ) THEN
1243	to_be_resorted => v
1244	ELSE
1245	IF ( .NOT. ALLOCATED( v_av ) ) THEN
1246	ALLOCATE( v_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1247	v_av = REAL( fill_value, KIND = wp )
1248	ENDIF
1249	to_be_resorted => v_av
1250	ENDIF
1251	IF ( mode == 'xy' ) level_z = zu
1252	!
1253	!-- Substitute the values generated by "mirror" boundary condition
1254	!-- at the bottom boundary by the real surface values.
1255	IF ( do2d(av,if) == 'v_xz' .OR. do2d(av,if) == 'v_yz' ) THEN
1256	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
1257	ENDIF
1258
1259	CASE ( 'vpt_xy', 'vpt_xz', 'vpt_yz' )
1260	IF ( av == 0 ) THEN
1261	to_be_resorted => vpt
1262	ELSE
1263	IF ( .NOT. ALLOCATED( vpt_av ) ) THEN
1264	ALLOCATE( vpt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1265	vpt_av = REAL( fill_value, KIND = wp )
1266	ENDIF
1267	to_be_resorted => vpt_av
1268	ENDIF
1269	IF ( mode == 'xy' ) level_z = zu
1270
1271	CASE ( 'w_xy', 'w_xz', 'w_yz' )
1272	flag_nr = 3
1273	IF ( av == 0 ) THEN
1274	to_be_resorted => w
1275	ELSE
1276	IF ( .NOT. ALLOCATED( w_av ) ) THEN
1277	ALLOCATE( w_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1278	w_av = REAL( fill_value, KIND = wp )
1279	ENDIF
1280	to_be_resorted => w_av
1281	ENDIF
1282	IF ( mode == 'xy' ) level_z = zw
1283
1284	CASE ( 'z0*_xy' ) ! 2d-array
1285	IF ( av == 0 ) THEN
1286	DO m = 1, surf_def_h(0)%ns
1287	i = surf_def_h(0)%i(m)
1288	j = surf_def_h(0)%j(m)
1289	local_pf(i,j,nzb+1) = surf_def_h(0)%z0(m)
1290	ENDDO
1291	DO m = 1, surf_lsm_h%ns
1292	i = surf_lsm_h%i(m)
1293	j = surf_lsm_h%j(m)
1294	local_pf(i,j,nzb+1) = surf_lsm_h%z0(m)
1295	ENDDO
1296	DO m = 1, surf_usm_h%ns
1297	i = surf_usm_h%i(m)
1298	j = surf_usm_h%j(m)
1299	local_pf(i,j,nzb+1) = surf_usm_h%z0(m)
1300	ENDDO
1301	ELSE
1302	IF ( .NOT. ALLOCATED( z0_av ) ) THEN
1303	ALLOCATE( z0_av(nysg:nyng,nxlg:nxrg) )
1304	z0_av = REAL( fill_value, KIND = wp )
1305	ENDIF
1306	DO i = nxl, nxr
1307	DO j = nys, nyn
1308	local_pf(i,j,nzb+1) = z0_av(j,i)
1309	ENDDO
1310	ENDDO
1311	ENDIF
1312	resorted = .TRUE.
1313	two_d = .TRUE.
1314	level_z(nzb+1) = zu(nzb+1)
1315
1316	CASE ( 'z0h*_xy' ) ! 2d-array
1317	IF ( av == 0 ) THEN
1318	DO m = 1, surf_def_h(0)%ns
1319	i = surf_def_h(0)%i(m)
1320	j = surf_def_h(0)%j(m)
1321	local_pf(i,j,nzb+1) = surf_def_h(0)%z0h(m)
1322	ENDDO
1323	DO m = 1, surf_lsm_h%ns
1324	i = surf_lsm_h%i(m)
1325	j = surf_lsm_h%j(m)
1326	local_pf(i,j,nzb+1) = surf_lsm_h%z0h(m)
1327	ENDDO
1328	DO m = 1, surf_usm_h%ns
1329	i = surf_usm_h%i(m)
1330	j = surf_usm_h%j(m)
1331	local_pf(i,j,nzb+1) = surf_usm_h%z0h(m)
1332	ENDDO
1333	ELSE
1334	IF ( .NOT. ALLOCATED( z0h_av ) ) THEN
1335	ALLOCATE( z0h_av(nysg:nyng,nxlg:nxrg) )
1336	z0h_av = REAL( fill_value, KIND = wp )
1337	ENDIF
1338	DO i = nxl, nxr
1339	DO j = nys, nyn
1340	local_pf(i,j,nzb+1) = z0h_av(j,i)
1341	ENDDO
1342	ENDDO
1343	ENDIF
1344	resorted = .TRUE.
1345	two_d = .TRUE.
1346	level_z(nzb+1) = zu(nzb+1)
1347
1348	CASE ( 'z0q*_xy' ) ! 2d-array
1349	IF ( av == 0 ) THEN
1350	DO m = 1, surf_def_h(0)%ns
1351	i = surf_def_h(0)%i(m)
1352	j = surf_def_h(0)%j(m)
1353	local_pf(i,j,nzb+1) = surf_def_h(0)%z0q(m)
1354	ENDDO
1355	DO m = 1, surf_lsm_h%ns
1356	i = surf_lsm_h%i(m)
1357	j = surf_lsm_h%j(m)
1358	local_pf(i,j,nzb+1) = surf_lsm_h%z0q(m)
1359	ENDDO
1360	DO m = 1, surf_usm_h%ns
1361	i = surf_usm_h%i(m)
1362	j = surf_usm_h%j(m)
1363	local_pf(i,j,nzb+1) = surf_usm_h%z0q(m)
1364	ENDDO
1365	ELSE
1366	IF ( .NOT. ALLOCATED( z0q_av ) ) THEN
1367	ALLOCATE( z0q_av(nysg:nyng,nxlg:nxrg) )
1368	z0q_av = REAL( fill_value, KIND = wp )
1369	ENDIF
1370	DO i = nxl, nxr
1371	DO j = nys, nyn
1372	local_pf(i,j,nzb+1) = z0q_av(j,i)
1373	ENDDO
1374	ENDDO
1375	ENDIF
1376	resorted = .TRUE.
1377	two_d = .TRUE.
1378	level_z(nzb+1) = zu(nzb+1)
1379
1380	CASE DEFAULT
1381
1382	!
1383	!-- Land surface model quantity
1384	IF ( land_surface ) THEN
1385	CALL lsm_data_output_2d( av, do2d(av,if), found, grid, mode,&
1386	local_pf, two_d, nzb_do, nzt_do )
1387	ENDIF
1388
1389	!
1390	!-- Turbulence closure variables
1391	IF ( .NOT. found ) THEN
1392	CALL tcm_data_output_2d( av, do2d(av,if), found, grid, mode,&
1393	local_pf, two_d, nzb_do, nzt_do )
1394	ENDIF
1395
1396	!
1397	!-- Radiation quantity
1398	IF ( .NOT. found .AND. radiation ) THEN
1399	CALL radiation_data_output_2d( av, do2d(av,if), found, grid,&
1400	mode, local_pf, two_d, &
1401	nzb_do, nzt_do )
1402	ENDIF
1403
1404	!
1405	!-- Gust module quantities
1406	IF ( .NOT. found .AND. gust_module_enabled ) THEN
1407	CALL gust_data_output_2d( av, do2d(av,if), found, grid, &
1408	local_pf, two_d, nzb_do, nzt_do )
1409	ENDIF
1410
1411	!
1412	!-- UV exposure model quantity
1413	IF ( uv_exposure ) THEN
1414	CALL uvem_data_output_2d( av, do2d(av,if), found, grid, mode,&
1415	local_pf, two_d, nzb_do, nzt_do )
1416	ENDIF
1417
1418	!
1419	!-- User defined quantity
1420	IF ( .NOT. found ) THEN
1421	CALL user_data_output_2d( av, do2d(av,if), found, grid, &
1422	local_pf, two_d, nzb_do, nzt_do )
1423	ENDIF
1424
1425	resorted = .TRUE.
1426
1427	IF ( grid == 'zu' ) THEN
1428	IF ( mode == 'xy' ) level_z = zu
1429	ELSEIF ( grid == 'zw' ) THEN
1430	IF ( mode == 'xy' ) level_z = zw
1431	ELSEIF ( grid == 'zu1' ) THEN
1432	IF ( mode == 'xy' ) level_z(nzb+1) = zu(nzb+1)
1433	ELSEIF ( grid == 'zs' ) THEN
1434	IF ( mode == 'xy' ) level_z = zs
1435	ENDIF
1436
1437	IF ( .NOT. found ) THEN
1438	message_string = 'no output provided for: ' // &
1439	TRIM( do2d(av,if) )
1440	CALL message( 'data_output_2d', 'PA0181', 0, 0, 0, 6, 0 )
1441	ENDIF
1442
1443	END SELECT
1444
1445	!
1446	!-- Resort the array to be output, if not done above. Flag topography
1447	!-- grid points with fill values, using the corresponding maksing flag.
1448	IF ( .NOT. resorted ) THEN
1449	DO i = nxl, nxr
1450	DO j = nys, nyn
1451	DO k = nzb_do, nzt_do
1452	local_pf(i,j,k) = MERGE( to_be_resorted(k,j,i), &
1453	REAL( fill_value, KIND = wp ), &
1454	BTEST( wall_flags_0(k,j,i), &
1455	flag_nr ) )
1456	ENDDO
1457	ENDDO
1458	ENDDO
1459	ENDIF
1460
1461	!
1462	!-- Output of the individual cross-sections, depending on the cross-
1463	!-- section mode chosen.
1464	is = 1
1465	loop1: DO WHILE ( section(is,s_ind) /= -9999 .OR. two_d )
1466
1467	SELECT CASE ( mode )
1468
1469	CASE ( 'xy' )
1470	!
1471	!-- Determine the cross section index
1472	IF ( two_d ) THEN
1473	layer_xy = nzb+1
1474	ELSE
1475	layer_xy = section(is,s_ind)
1476	ENDIF
1477
1478	!
1479	!-- Exit the loop for layers beyond the data output domain
1480	!-- (used for soil model)
1481	IF ( layer_xy > nzt_do ) THEN
1482	EXIT loop1
1483	ENDIF
1484
1485	!
1486	!-- Update the netCDF xy cross section time axis.
1487	!-- In case of parallel output, this is only done by PE0
1488	!-- to increase the performance.
1489	IF ( simulated_time /= do2d_xy_last_time(av) ) THEN
1490	do2d_xy_time_count(av) = do2d_xy_time_count(av) + 1
1491	do2d_xy_last_time(av) = simulated_time
1492	IF ( myid == 0 ) THEN
1493	IF ( .NOT. data_output_2d_on_each_pe &
1494	.OR. netcdf_data_format > 4 ) &
1495	THEN
1496	#if defined( __netcdf )
1497	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1498	id_var_time_xy(av), &
1499	(/ time_since_reference_point /), &
1500	start = (/ do2d_xy_time_count(av) /), &
1501	count = (/ 1 /) )
1502	CALL netcdf_handle_error( 'data_output_2d', 53 )
1503	#endif
1504	ENDIF
1505	ENDIF
1506	ENDIF
1507	!
1508	!-- If required, carry out averaging along z
1509	IF ( section(is,s_ind) == -1 .AND. .NOT. two_d ) THEN
1510
1511	local_2d = 0.0_wp
1512	!
1513	!-- Carry out the averaging (all data are on the PE)
1514	DO k = nzb_do, nzt_do
1515	DO j = nys, nyn
1516	DO i = nxl, nxr
1517	local_2d(i,j) = local_2d(i,j) + local_pf(i,j,k)
1518	ENDDO
1519	ENDDO
1520	ENDDO
1521
1522	local_2d = local_2d / ( nzt_do - nzb_do + 1.0_wp)
1523
1524	ELSE
1525	!
1526	!-- Just store the respective section on the local array
1527	local_2d = local_pf(:,:,layer_xy)
1528
1529	ENDIF
1530
1531	#if defined( __parallel )
1532	IF ( netcdf_data_format > 4 ) THEN
1533	!
1534	!-- Parallel output in netCDF4/HDF5 format.
1535	IF ( two_d ) THEN
1536	iis = 1
1537	ELSE
1538	iis = is
1539	ENDIF
1540
1541	#if defined( __netcdf )
1542	!
1543	!-- For parallel output, all cross sections are first stored
1544	!-- here on a local array and will be written to the output
1545	!-- file afterwards to increase the performance.
1546	DO i = nxl, nxr
1547	DO j = nys, nyn
1548	local_2d_sections(i,j,iis) = local_2d(i,j)
1549	ENDDO
1550	ENDDO
1551	#endif
1552	ELSE
1553
1554	IF ( data_output_2d_on_each_pe ) THEN
1555	!
1556	!-- Output of partial arrays on each PE
1557	#if defined( __netcdf )
1558	IF ( myid == 0 ) THEN
1559	WRITE ( 21 ) time_since_reference_point, &
1560	do2d_xy_time_count(av), av
1561	ENDIF
1562	#endif
1563	DO i = 0, io_blocks-1
1564	IF ( i == io_group ) THEN
1565	WRITE ( 21 ) nxl, nxr, nys, nyn, nys, nyn
1566	WRITE ( 21 ) local_2d
1567	ENDIF
1568	#if defined( __parallel )
1569	CALL MPI_BARRIER( comm2d, ierr )
1570	#endif
1571	ENDDO
1572
1573	ELSE
1574	!
1575	!-- PE0 receives partial arrays from all processors and
1576	!-- then outputs them. Here a barrier has to be set,
1577	!-- because otherwise "-MPI- FATAL: Remote protocol queue
1578	!-- full" may occur.
1579	CALL MPI_BARRIER( comm2d, ierr )
1580
1581	ngp = ( nxr-nxl+1 ) * ( nyn-nys+1 )
1582	IF ( myid == 0 ) THEN
1583	!
1584	!-- Local array can be relocated directly.
1585	total_2d(nxl:nxr,nys:nyn) = local_2d
1586	!
1587	!-- Receive data from all other PEs.
1588	DO n = 1, numprocs-1
1589	!
1590	!-- Receive index limits first, then array.
1591	!-- Index limits are received in arbitrary order from
1592	!-- the PEs.
1593	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1594	MPI_ANY_SOURCE, 0, comm2d, &
1595	status, ierr )
1596	sender = status(MPI_SOURCE)
1597	DEALLOCATE( local_2d )
1598	ALLOCATE( local_2d(ind(1):ind(2),ind(3):ind(4)) )
1599	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1600	MPI_REAL, sender, 1, comm2d, &
1601	status, ierr )
1602	total_2d(ind(1):ind(2),ind(3):ind(4)) = local_2d
1603	ENDDO
1604	!
1605	!-- Relocate the local array for the next loop increment
1606	DEALLOCATE( local_2d )
1607	ALLOCATE( local_2d(nxl:nxr,nys:nyn) )
1608
1609	#if defined( __netcdf )
1610	IF ( two_d ) THEN
1611	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1612	id_var_do2d(av,if), &
1613	total_2d(0:nx,0:ny), &
1614	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
1615	count = (/ nx+1, ny+1, 1, 1 /) )
1616	ELSE
1617	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1618	id_var_do2d(av,if), &
1619	total_2d(0:nx,0:ny), &
1620	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
1621	count = (/ nx+1, ny+1, 1, 1 /) )
1622	ENDIF
1623	CALL netcdf_handle_error( 'data_output_2d', 54 )
1624	#endif
1625
1626	ELSE
1627	!
1628	!-- First send the local index limits to PE0
1629	ind(1) = nxl; ind(2) = nxr
1630	ind(3) = nys; ind(4) = nyn
1631	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1632	comm2d, ierr )
1633	!
1634	!-- Send data to PE0
1635	CALL MPI_SEND( local_2d(nxl,nys), ngp, &
1636	MPI_REAL, 0, 1, comm2d, ierr )
1637	ENDIF
1638	!
1639	!-- A barrier has to be set, because otherwise some PEs may
1640	!-- proceed too fast so that PE0 may receive wrong data on
1641	!-- tag 0
1642	CALL MPI_BARRIER( comm2d, ierr )
1643	ENDIF
1644
1645	ENDIF
1646	#else
1647	#if defined( __netcdf )
1648	IF ( two_d ) THEN
1649	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1650	id_var_do2d(av,if), &
1651	local_2d(nxl:nxr,nys:nyn), &
1652	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
1653	count = (/ nx+1, ny+1, 1, 1 /) )
1654	ELSE
1655	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1656	id_var_do2d(av,if), &
1657	local_2d(nxl:nxr,nys:nyn), &
1658	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
1659	count = (/ nx+1, ny+1, 1, 1 /) )
1660	ENDIF
1661	CALL netcdf_handle_error( 'data_output_2d', 447 )
1662	#endif
1663	#endif
1664
1665	!
1666	!-- For 2D-arrays (e.g. u*) only one cross-section is available.
1667	!-- Hence exit loop of output levels.
1668	IF ( two_d ) THEN
1669	IF ( netcdf_data_format < 5 ) two_d = .FALSE.
1670	EXIT loop1
1671	ENDIF
1672
1673	CASE ( 'xz' )
1674	!
1675	!-- Update the netCDF xz cross section time axis.
1676	!-- In case of parallel output, this is only done by PE0
1677	!-- to increase the performance.
1678	IF ( simulated_time /= do2d_xz_last_time(av) ) THEN
1679	do2d_xz_time_count(av) = do2d_xz_time_count(av) + 1
1680	do2d_xz_last_time(av) = simulated_time
1681	IF ( myid == 0 ) THEN
1682	IF ( .NOT. data_output_2d_on_each_pe &
1683	.OR. netcdf_data_format > 4 ) &
1684	THEN
1685	#if defined( __netcdf )
1686	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1687	id_var_time_xz(av), &
1688	(/ time_since_reference_point /), &
1689	start = (/ do2d_xz_time_count(av) /), &
1690	count = (/ 1 /) )
1691	CALL netcdf_handle_error( 'data_output_2d', 56 )
1692	#endif
1693	ENDIF
1694	ENDIF
1695	ENDIF
1696
1697	!
1698	!-- If required, carry out averaging along y
1699	IF ( section(is,s_ind) == -1 ) THEN
1700
1701	ALLOCATE( local_2d_l(nxl:nxr,nzb_do:nzt_do) )
1702	local_2d_l = 0.0_wp
1703	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
1704	!
1705	!-- First local averaging on the PE
1706	DO k = nzb_do, nzt_do
1707	DO j = nys, nyn
1708	DO i = nxl, nxr
1709	local_2d_l(i,k) = local_2d_l(i,k) + &
1710	local_pf(i,j,k)
1711	ENDDO
1712	ENDDO
1713	ENDDO
1714	#if defined( __parallel )
1715	!
1716	!-- Now do the averaging over all PEs along y
1717	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1718	CALL MPI_ALLREDUCE( local_2d_l(nxl,nzb_do), &
1719	local_2d(nxl,nzb_do), ngp, MPI_REAL, &
1720	MPI_SUM, comm1dy, ierr )
1721	#else
1722	local_2d = local_2d_l
1723	#endif
1724	local_2d = local_2d / ( ny + 1.0_wp )
1725
1726	DEALLOCATE( local_2d_l )
1727
1728	ELSE
1729	!
1730	!-- Just store the respective section on the local array
1731	!-- (but only if it is available on this PE!)
1732	IF ( section(is,s_ind) >= nys .AND. section(is,s_ind) <= nyn ) &
1733	THEN
1734	local_2d = local_pf(:,section(is,s_ind),nzb_do:nzt_do)
1735	ENDIF
1736
1737	ENDIF
1738
1739	#if defined( __parallel )
1740	IF ( netcdf_data_format > 4 ) THEN
1741	!
1742	!-- Output in netCDF4/HDF5 format.
1743	!-- Output only on those PEs where the respective cross
1744	!-- sections reside. Cross sections averaged along y are
1745	!-- output on the respective first PE along y (myidy=0).
1746	IF ( ( section(is,s_ind) >= nys .AND. &
1747	section(is,s_ind) <= nyn ) .OR. &
1748	( section(is,s_ind) == -1 .AND. myidy == 0 ) ) THEN
1749	#if defined( __netcdf )
1750	!
1751	!-- For parallel output, all cross sections are first
1752	!-- stored here on a local array and will be written to the
1753	!-- output file afterwards to increase the performance.
1754	DO i = nxl, nxr
1755	DO k = nzb_do, nzt_do
1756	local_2d_sections_l(i,is,k) = local_2d(i,k)
1757	ENDDO
1758	ENDDO
1759	#endif
1760	ENDIF
1761
1762	ELSE
1763
1764	IF ( data_output_2d_on_each_pe ) THEN
1765	!
1766	!-- Output of partial arrays on each PE. If the cross
1767	!-- section does not reside on the PE, output special
1768	!-- index values.
1769	#if defined( __netcdf )
1770	IF ( myid == 0 ) THEN
1771	WRITE ( 22 ) time_since_reference_point, &
1772	do2d_xz_time_count(av), av
1773	ENDIF
1774	#endif
1775	DO i = 0, io_blocks-1
1776	IF ( i == io_group ) THEN
1777	IF ( ( section(is,s_ind) >= nys .AND. &
1778	section(is,s_ind) <= nyn ) .OR. &
1779	( section(is,s_ind) == -1 .AND. &
1780	nys-1 == -1 ) ) &
1781	THEN
1782	WRITE (22) nxl, nxr, nzb_do, nzt_do, nzb, nzt+1
1783	WRITE (22) local_2d
1784	ELSE
1785	WRITE (22) -1, -1, -1, -1, -1, -1
1786	ENDIF
1787	ENDIF
1788	#if defined( __parallel )
1789	CALL MPI_BARRIER( comm2d, ierr )
1790	#endif
1791	ENDDO
1792
1793	ELSE
1794	!
1795	!-- PE0 receives partial arrays from all processors of the
1796	!-- respective cross section and outputs them. Here a
1797	!-- barrier has to be set, because otherwise
1798	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
1799	CALL MPI_BARRIER( comm2d, ierr )
1800
1801	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
1802	IF ( myid == 0 ) THEN
1803	!
1804	!-- Local array can be relocated directly.
1805	IF ( ( section(is,s_ind) >= nys .AND. &
1806	section(is,s_ind) <= nyn ) .OR. &
1807	( section(is,s_ind) == -1 .AND. &
1808	nys-1 == -1 ) ) THEN
1809	total_2d(nxl:nxr,nzb_do:nzt_do) = local_2d
1810	ENDIF
1811	!
1812	!-- Receive data from all other PEs.
1813	DO n = 1, numprocs-1
1814	!
1815	!-- Receive index limits first, then array.
1816	!-- Index limits are received in arbitrary order from
1817	!-- the PEs.
1818	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1819	MPI_ANY_SOURCE, 0, comm2d, &
1820	status, ierr )
1821	!
1822	!-- Not all PEs have data for XZ-cross-section.
1823	IF ( ind(1) /= -9999 ) THEN
1824	sender = status(MPI_SOURCE)
1825	DEALLOCATE( local_2d )
1826	ALLOCATE( local_2d(ind(1):ind(2), &
1827	ind(3):ind(4)) )
1828	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1829	MPI_REAL, sender, 1, comm2d, &
1830	status, ierr )
1831	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
1832	local_2d
1833	ENDIF
1834	ENDDO
1835	!
1836	!-- Relocate the local array for the next loop increment
1837	DEALLOCATE( local_2d )
1838	ALLOCATE( local_2d(nxl:nxr,nzb_do:nzt_do) )
1839
1840	#if defined( __netcdf )
1841	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1842	id_var_do2d(av,if), &
1843	total_2d(0:nx,nzb_do:nzt_do), &
1844	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
1845	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
1846	CALL netcdf_handle_error( 'data_output_2d', 58 )
1847	#endif
1848
1849	ELSE
1850	!
1851	!-- If the cross section resides on the PE, send the
1852	!-- local index limits, otherwise send -9999 to PE0.
1853	IF ( ( section(is,s_ind) >= nys .AND. &
1854	section(is,s_ind) <= nyn ) .OR. &
1855	( section(is,s_ind) == -1 .AND. nys-1 == -1 ) ) &
1856	THEN
1857	ind(1) = nxl; ind(2) = nxr
1858	ind(3) = nzb_do; ind(4) = nzt_do
1859	ELSE
1860	ind(1) = -9999; ind(2) = -9999
1861	ind(3) = -9999; ind(4) = -9999
1862	ENDIF
1863	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1864	comm2d, ierr )
1865	!
1866	!-- If applicable, send data to PE0.
1867	IF ( ind(1) /= -9999 ) THEN
1868	CALL MPI_SEND( local_2d(nxl,nzb_do), ngp, &
1869	MPI_REAL, 0, 1, comm2d, ierr )
1870	ENDIF
1871	ENDIF
1872	!
1873	!-- A barrier has to be set, because otherwise some PEs may
1874	!-- proceed too fast so that PE0 may receive wrong data on
1875	!-- tag 0
1876	CALL MPI_BARRIER( comm2d, ierr )
1877	ENDIF
1878
1879	ENDIF
1880	#else
1881	#if defined( __netcdf )
1882	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1883	id_var_do2d(av,if), &
1884	local_2d(nxl:nxr,nzb_do:nzt_do), &
1885	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
1886	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
1887	CALL netcdf_handle_error( 'data_output_2d', 451 )
1888	#endif
1889	#endif
1890
1891	CASE ( 'yz' )
1892	!
1893	!-- Update the netCDF yz cross section time axis.
1894	!-- In case of parallel output, this is only done by PE0
1895	!-- to increase the performance.
1896	IF ( simulated_time /= do2d_yz_last_time(av) ) THEN
1897	do2d_yz_time_count(av) = do2d_yz_time_count(av) + 1
1898	do2d_yz_last_time(av) = simulated_time
1899	IF ( myid == 0 ) THEN
1900	IF ( .NOT. data_output_2d_on_each_pe &
1901	.OR. netcdf_data_format > 4 ) &
1902	THEN
1903	#if defined( __netcdf )
1904	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
1905	id_var_time_yz(av), &
1906	(/ time_since_reference_point /), &
1907	start = (/ do2d_yz_time_count(av) /), &
1908	count = (/ 1 /) )
1909	CALL netcdf_handle_error( 'data_output_2d', 59 )
1910	#endif
1911	ENDIF
1912	ENDIF
1913	ENDIF
1914
1915	!
1916	!-- If required, carry out averaging along x
1917	IF ( section(is,s_ind) == -1 ) THEN
1918
1919	ALLOCATE( local_2d_l(nys:nyn,nzb_do:nzt_do) )
1920	local_2d_l = 0.0_wp
1921	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
1922	!
1923	!-- First local averaging on the PE
1924	DO k = nzb_do, nzt_do
1925	DO j = nys, nyn
1926	DO i = nxl, nxr
1927	local_2d_l(j,k) = local_2d_l(j,k) + &
1928	local_pf(i,j,k)
1929	ENDDO
1930	ENDDO
1931	ENDDO
1932	#if defined( __parallel )
1933	!
1934	!-- Now do the averaging over all PEs along x
1935	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1936	CALL MPI_ALLREDUCE( local_2d_l(nys,nzb_do), &
1937	local_2d(nys,nzb_do), ngp, MPI_REAL, &
1938	MPI_SUM, comm1dx, ierr )
1939	#else
1940	local_2d = local_2d_l
1941	#endif
1942	local_2d = local_2d / ( nx + 1.0_wp )
1943
1944	DEALLOCATE( local_2d_l )
1945
1946	ELSE
1947	!
1948	!-- Just store the respective section on the local array
1949	!-- (but only if it is available on this PE!)
1950	IF ( section(is,s_ind) >= nxl .AND. section(is,s_ind) <= nxr ) &
1951	THEN
1952	local_2d = local_pf(section(is,s_ind),:,nzb_do:nzt_do)
1953	ENDIF
1954
1955	ENDIF
1956
1957	#if defined( __parallel )
1958	IF ( netcdf_data_format > 4 ) THEN
1959	!
1960	!-- Output in netCDF4/HDF5 format.
1961	!-- Output only on those PEs where the respective cross
1962	!-- sections reside. Cross sections averaged along x are
1963	!-- output on the respective first PE along x (myidx=0).
1964	IF ( ( section(is,s_ind) >= nxl .AND. &
1965	section(is,s_ind) <= nxr ) .OR. &
1966	( section(is,s_ind) == -1 .AND. myidx == 0 ) ) THEN
1967	#if defined( __netcdf )
1968	!
1969	!-- For parallel output, all cross sections are first
1970	!-- stored here on a local array and will be written to the
1971	!-- output file afterwards to increase the performance.
1972	DO j = nys, nyn
1973	DO k = nzb_do, nzt_do
1974	local_2d_sections_l(is,j,k) = local_2d(j,k)
1975	ENDDO
1976	ENDDO
1977	#endif
1978	ENDIF
1979
1980	ELSE
1981
1982	IF ( data_output_2d_on_each_pe ) THEN
1983	!
1984	!-- Output of partial arrays on each PE. If the cross
1985	!-- section does not reside on the PE, output special
1986	!-- index values.
1987	#if defined( __netcdf )
1988	IF ( myid == 0 ) THEN
1989	WRITE ( 23 ) time_since_reference_point, &
1990	do2d_yz_time_count(av), av
1991	ENDIF
1992	#endif
1993	DO i = 0, io_blocks-1
1994	IF ( i == io_group ) THEN
1995	IF ( ( section(is,s_ind) >= nxl .AND. &
1996	section(is,s_ind) <= nxr ) .OR. &
1997	( section(is,s_ind) == -1 .AND. &
1998	nxl-1 == -1 ) ) &
1999	THEN
2000	WRITE (23) nys, nyn, nzb_do, nzt_do, nzb, nzt+1
2001	WRITE (23) local_2d
2002	ELSE
2003	WRITE (23) -1, -1, -1, -1, -1, -1
2004	ENDIF
2005	ENDIF
2006	#if defined( __parallel )
2007	CALL MPI_BARRIER( comm2d, ierr )
2008	#endif
2009	ENDDO
2010
2011	ELSE
2012	!
2013	!-- PE0 receives partial arrays from all processors of the
2014	!-- respective cross section and outputs them. Here a
2015	!-- barrier has to be set, because otherwise
2016	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
2017	CALL MPI_BARRIER( comm2d, ierr )
2018
2019	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
2020	IF ( myid == 0 ) THEN
2021	!
2022	!-- Local array can be relocated directly.
2023	IF ( ( section(is,s_ind) >= nxl .AND. &
2024	section(is,s_ind) <= nxr ) .OR. &
2025	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
2026	THEN
2027	total_2d(nys:nyn,nzb_do:nzt_do) = local_2d
2028	ENDIF
2029	!
2030	!-- Receive data from all other PEs.
2031	DO n = 1, numprocs-1
2032	!
2033	!-- Receive index limits first, then array.
2034	!-- Index limits are received in arbitrary order from
2035	!-- the PEs.
2036	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
2037	MPI_ANY_SOURCE, 0, comm2d, &
2038	status, ierr )
2039	!
2040	!-- Not all PEs have data for YZ-cross-section.
2041	IF ( ind(1) /= -9999 ) THEN
2042	sender = status(MPI_SOURCE)
2043	DEALLOCATE( local_2d )
2044	ALLOCATE( local_2d(ind(1):ind(2), &
2045	ind(3):ind(4)) )
2046	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
2047	MPI_REAL, sender, 1, comm2d, &
2048	status, ierr )
2049	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
2050	local_2d
2051	ENDIF
2052	ENDDO
2053	!
2054	!-- Relocate the local array for the next loop increment
2055	DEALLOCATE( local_2d )
2056	ALLOCATE( local_2d(nys:nyn,nzb_do:nzt_do) )
2057
2058	#if defined( __netcdf )
2059	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2060	id_var_do2d(av,if), &
2061	total_2d(0:ny,nzb_do:nzt_do), &
2062	start = (/ is, 1, 1, do2d_yz_time_count(av) /), &
2063	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
2064	CALL netcdf_handle_error( 'data_output_2d', 61 )
2065	#endif
2066
2067	ELSE
2068	!
2069	!-- If the cross section resides on the PE, send the
2070	!-- local index limits, otherwise send -9999 to PE0.
2071	IF ( ( section(is,s_ind) >= nxl .AND. &
2072	section(is,s_ind) <= nxr ) .OR. &
2073	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
2074	THEN
2075	ind(1) = nys; ind(2) = nyn
2076	ind(3) = nzb_do; ind(4) = nzt_do
2077	ELSE
2078	ind(1) = -9999; ind(2) = -9999
2079	ind(3) = -9999; ind(4) = -9999
2080	ENDIF
2081	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
2082	comm2d, ierr )
2083	!
2084	!-- If applicable, send data to PE0.
2085	IF ( ind(1) /= -9999 ) THEN
2086	CALL MPI_SEND( local_2d(nys,nzb_do), ngp, &
2087	MPI_REAL, 0, 1, comm2d, ierr )
2088	ENDIF
2089	ENDIF
2090	!
2091	!-- A barrier has to be set, because otherwise some PEs may
2092	!-- proceed too fast so that PE0 may receive wrong data on
2093	!-- tag 0
2094	CALL MPI_BARRIER( comm2d, ierr )
2095	ENDIF
2096
2097	ENDIF
2098	#else
2099	#if defined( __netcdf )
2100	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2101	id_var_do2d(av,if), &
2102	local_2d(nys:nyn,nzb_do:nzt_do), &
2103	start = (/ is, 1, 1, do2d_xz_time_count(av) /), &
2104	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
2105	CALL netcdf_handle_error( 'data_output_2d', 452 )
2106	#endif
2107	#endif
2108
2109	END SELECT
2110
2111	is = is + 1
2112	ENDDO loop1
2113
2114	!
2115	!-- For parallel output, all data were collected before on a local array
2116	!-- and are written now to the netcdf file. This must be done to increase
2117	!-- the performance of the parallel output.
2118	#if defined( __netcdf )
2119	IF ( netcdf_data_format > 4 ) THEN
2120
2121	SELECT CASE ( mode )
2122
2123	CASE ( 'xy' )
2124	IF ( two_d ) THEN
2125	nis = 1
2126	two_d = .FALSE.
2127	ELSE
2128	nis = ns
2129	ENDIF
2130	!
2131	!-- Do not output redundant ghost point data except for the
2132	!-- boundaries of the total domain.
2133	! IF ( nxr == nx .AND. nyn /= ny ) THEN
2134	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2135	! id_var_do2d(av,if), &
2136	! local_2d_sections(nxl:nxr+1, &
2137	! nys:nyn,1:nis), &
2138	! start = (/ nxl+1, nys+1, 1, &
2139	! do2d_xy_time_count(av) /), &
2140	! count = (/ nxr-nxl+2, &
2141	! nyn-nys+1, nis, 1 &
2142	! /) )
2143	! ELSEIF ( nxr /= nx .AND. nyn == ny ) THEN
2144	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2145	! id_var_do2d(av,if), &
2146	! local_2d_sections(nxl:nxr, &
2147	! nys:nyn+1,1:nis), &
2148	! start = (/ nxl+1, nys+1, 1, &
2149	! do2d_xy_time_count(av) /), &
2150	! count = (/ nxr-nxl+1, &
2151	! nyn-nys+2, nis, 1 &
2152	! /) )
2153	! ELSEIF ( nxr == nx .AND. nyn == ny ) THEN
2154	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2155	! id_var_do2d(av,if), &
2156	! local_2d_sections(nxl:nxr+1, &
2157	! nys:nyn+1,1:nis), &
2158	! start = (/ nxl+1, nys+1, 1, &
2159	! do2d_xy_time_count(av) /), &
2160	! count = (/ nxr-nxl+2, &
2161	! nyn-nys+2, nis, 1 &
2162	! /) )
2163	! ELSE
2164	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2165	id_var_do2d(av,if), &
2166	local_2d_sections(nxl:nxr, &
2167	nys:nyn,1:nis), &
2168	start = (/ nxl+1, nys+1, 1, &
2169	do2d_xy_time_count(av) /), &
2170	count = (/ nxr-nxl+1, &
2171	nyn-nys+1, nis, 1 &
2172	/) )
2173	! ENDIF
2174
2175	CALL netcdf_handle_error( 'data_output_2d', 55 )
2176
2177	CASE ( 'xz' )
2178	!
2179	!-- First, all PEs get the information of all cross-sections.
2180	!-- Then the data are written to the output file by all PEs
2181	!-- while NF90_COLLECTIVE is set in subroutine
2182	!-- define_netcdf_header. Although redundant information are
2183	!-- written to the output file in that case, the performance
2184	!-- is significantly better compared to the case where only
2185	!-- the first row of PEs in x-direction (myidx = 0) is given
2186	!-- the output while NF90_INDEPENDENT is set.
2187	IF ( npey /= 1 ) THEN
2188
2189	#if defined( __parallel )
2190	!
2191	!-- Distribute data over all PEs along y
2192	ngp = ( nxr-nxl+1 ) * ( nzt_do-nzb_do+1 ) * ns
2193	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
2194	CALL MPI_ALLREDUCE( local_2d_sections_l(nxl,1,nzb_do), &
2195	local_2d_sections(nxl,1,nzb_do), &
2196	ngp, MPI_REAL, MPI_SUM, comm1dy, &
2197	ierr )
2198	#else
2199	local_2d_sections = local_2d_sections_l
2200	#endif
2201	ENDIF
2202	!
2203	!-- Do not output redundant ghost point data except for the
2204	!-- boundaries of the total domain.
2205	! IF ( nxr == nx ) THEN
2206	! nc_stat = NF90_PUT_VAR( id_set_xz(av), &
2207	! id_var_do2d(av,if), &
2208	! local_2d_sections(nxl:nxr+1,1:ns, &
2209	! nzb_do:nzt_do), &
2210	! start = (/ nxl+1, 1, 1, &
2211	! do2d_xz_time_count(av) /), &
2212	! count = (/ nxr-nxl+2, ns, nzt_do-nzb_do+1, &
2213	! 1 /) )
2214	! ELSE
2215	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
2216	id_var_do2d(av,if), &
2217	local_2d_sections(nxl:nxr,1:ns, &
2218	nzb_do:nzt_do), &
2219	start = (/ nxl+1, 1, 1, &
2220	do2d_xz_time_count(av) /), &
2221	count = (/ nxr-nxl+1, ns, nzt_do-nzb_do+1, &
2222	1 /) )
2223	! ENDIF
2224
2225	CALL netcdf_handle_error( 'data_output_2d', 57 )
2226
2227	CASE ( 'yz' )
2228	!
2229	!-- First, all PEs get the information of all cross-sections.
2230	!-- Then the data are written to the output file by all PEs
2231	!-- while NF90_COLLECTIVE is set in subroutine
2232	!-- define_netcdf_header. Although redundant information are
2233	!-- written to the output file in that case, the performance
2234	!-- is significantly better compared to the case where only
2235	!-- the first row of PEs in y-direction (myidy = 0) is given
2236	!-- the output while NF90_INDEPENDENT is set.
2237	IF ( npex /= 1 ) THEN
2238
2239	#if defined( __parallel )
2240	!
2241	!-- Distribute data over all PEs along x
2242	ngp = ( nyn-nys+1 ) * ( nzt-nzb + 2 ) * ns
2243	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
2244	CALL MPI_ALLREDUCE( local_2d_sections_l(1,nys,nzb_do), &
2245	local_2d_sections(1,nys,nzb_do), &
2246	ngp, MPI_REAL, MPI_SUM, comm1dx, &
2247	ierr )
2248	#else
2249	local_2d_sections = local_2d_sections_l
2250	#endif
2251	ENDIF
2252	!
2253	!-- Do not output redundant ghost point data except for the
2254	!-- boundaries of the total domain.
2255	! IF ( nyn == ny ) THEN
2256	! nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2257	! id_var_do2d(av,if), &
2258	! local_2d_sections(1:ns, &
2259	! nys:nyn+1,nzb_do:nzt_do), &
2260	! start = (/ 1, nys+1, 1, &
2261	! do2d_yz_time_count(av) /), &
2262	! count = (/ ns, nyn-nys+2, &
2263	! nzt_do-nzb_do+1, 1 /) )
2264	! ELSE
2265	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2266	id_var_do2d(av,if), &
2267	local_2d_sections(1:ns,nys:nyn, &
2268	nzb_do:nzt_do), &
2269	start = (/ 1, nys+1, 1, &
2270	do2d_yz_time_count(av) /), &
2271	count = (/ ns, nyn-nys+1, &
2272	nzt_do-nzb_do+1, 1 /) )
2273	! ENDIF
2274
2275	CALL netcdf_handle_error( 'data_output_2d', 60 )
2276
2277	CASE DEFAULT
2278	message_string = 'unknown cross-section: ' // TRIM( mode )
2279	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
2280
2281	END SELECT
2282
2283	ENDIF
2284	#endif
2285	ENDIF
2286
2287	if = if + 1
2288	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
2289	do2d_mode = do2d(av,if)(l-1:l)
2290
2291	ENDDO
2292
2293	!
2294	!-- Deallocate temporary arrays.
2295	IF ( ALLOCATED( level_z ) ) DEALLOCATE( level_z )
2296	IF ( netcdf_data_format > 4 ) THEN
2297	DEALLOCATE( local_pf, local_2d, local_2d_sections )
2298	IF( mode == 'xz' .OR. mode == 'yz' ) DEALLOCATE( local_2d_sections_l )
2299	ENDIF
2300	#if defined( __parallel )
2301	IF ( .NOT. data_output_2d_on_each_pe .AND. myid == 0 ) THEN
2302	DEALLOCATE( total_2d )
2303	ENDIF
2304	#endif
2305
2306	!
2307	!-- Close plot output file.
2308	file_id = 20 + s_ind
2309
2310	IF ( data_output_2d_on_each_pe ) THEN
2311	DO i = 0, io_blocks-1
2312	IF ( i == io_group ) THEN
2313	CALL close_file( file_id )
2314	ENDIF
2315	#if defined( __parallel )
2316	CALL MPI_BARRIER( comm2d, ierr )
2317	#endif
2318	ENDDO
2319	ELSE
2320	IF ( myid == 0 ) CALL close_file( file_id )
2321	ENDIF
2322
2323	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
2324
2325	END SUBROUTINE data_output_2d

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |