Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 3046

Last change on this file since 3046 was 3046, checked in by Giersch, 6 years ago
Remaining error messages revised, comments extended
Property svn:keywords set to `Id`
File size: 97.9 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	! Error messages revised
23	!
24	! Former revisions:
25	! -----------------
26	! $Id: data_output_2d.f90 3046 2018-05-29 08:02:15Z Giersch $
27	! Error messages revised
28	!
29	! 3014 2018-05-09 08:42:38Z maronga
30	! Added nzb_do and nzt_do for some modules for 2d output
31	!
32	! 3004 2018-04-27 12:33:25Z Giersch
33	! precipitation_rate removed, case prr*_xy removed, to_be_resorted have to point
34	! to ql_vp_av and not to ql_vp, allocation checks implemented (averaged data
35	! will be assigned to fill values if no allocation happened so far)
36	!
37	! 2963 2018-04-12 14:47:44Z suehring
38	! Introduce index for vegetation/wall, pavement/green-wall and water/window
39	! surfaces, for clearer access of surface fraction, albedo, emissivity, etc. .
40	!
41	! 2817 2018-02-19 16:32:21Z knoop
42	! Preliminary gust module interface implemented
43	!
44	! 2805 2018-02-14 17:00:09Z suehring
45	! Consider also default-type surfaces for surface temperature output.
46	!
47	! 2797 2018-02-08 13:24:35Z suehring
48	! Enable output of ground-heat flux also at urban surfaces.
49	!
50	! 2743 2018-01-12 16:03:39Z suehring
51	! In case of natural- and urban-type surfaces output surfaces fluxes in W/m2.
52	!
53	! 2742 2018-01-12 14:59:47Z suehring
54	! Enable output of surface temperature
55	!
56	! 2735 2018-01-11 12:01:27Z suehring
57	! output of r_a moved from land-surface to consider also urban-type surfaces
58	!
59	! 2718 2018-01-02 08:49:38Z maronga
60	! Corrected "Former revisions" section
61	!
62	! 2696 2017-12-14 17:12:51Z kanani
63	! Change in file header (GPL part)
64	! Implementation of uv exposure model (FK)
65	! Implementation of turbulence_closure_mod (TG)
66	! Set fill values at topography grid points or e.g. non-natural-type surface
67	! in case of LSM output (MS)
68	!
69	! 2512 2017-10-04 08:26:59Z raasch
70	! upper bounds of cross section output changed from nx+1,ny+1 to nx,ny
71	! no output of ghost layer data
72	!
73	! 2292 2017-06-20 09:51:42Z schwenkel
74	! Implementation of new microphysic scheme: cloud_scheme = 'morrison'
75	! includes two more prognostic equations for cloud drop concentration (nc)
76	! and cloud water content (qc).
77	!
78	! 2277 2017-06-12 10:47:51Z kanani
79	! Removed unused variables do2d_xy_n, do2d_xz_n, do2d_yz_n
80	!
81	! 2233 2017-05-30 18:08:54Z suehring
82	!
83	! 2232 2017-05-30 17:47:52Z suehring
84	! Adjustments to new surface concept
85	!
86	!
87	! 2190 2017-03-21 12:16:43Z raasch
88	! bugfix for r2031: string rho replaced by rho_ocean
89	!
90	! 2031 2016-10-21 15:11:58Z knoop
91	! renamed variable rho to rho_ocean and rho_av to rho_ocean_av
92	!
93	! 2000 2016-08-20 18:09:15Z knoop
94	! Forced header and separation lines into 80 columns
95	!
96	! 1980 2016-07-29 15:51:57Z suehring
97	! Bugfix, in order to steer user-defined output, setting flag found explicitly
98	! to .F.
99	!
100	! 1976 2016-07-27 13:28:04Z maronga
101	! Output of radiation quantities is now done directly in the respective module
102	!
103	! 1972 2016-07-26 07:52:02Z maronga
104	! Output of land surface quantities is now done directly in the respective
105	! module
106	!
107	! 1960 2016-07-12 16:34:24Z suehring
108	! Scalar surface flux added
109	! Rename INTEGER variable s into s_ind, as s is already assigned to scalar
110	!
111	! 1849 2016-04-08 11:33:18Z hoffmann
112	! precipitation_amount, precipitation_rate, prr moved to arrays_3d
113	!
114	! 1822 2016-04-07 07:49:42Z hoffmann
115	! Output of bulk cloud physics simplified.
116	!
117	! 1788 2016-03-10 11:01:04Z maronga
118	! Added output of z0q
119	!
120	! 1783 2016-03-06 18:36:17Z raasch
121	! name change of netcdf routines and module + related changes
122	!
123	! 1745 2016-02-05 13:06:51Z gronemeier
124	! Bugfix: test if time axis limit exceeds moved to point after call of check_open
125	!
126	! 1703 2015-11-02 12:38:44Z raasch
127	! bugfix for output of single (*) xy-sections in case of parallel netcdf I/O
128	!
129	! 1701 2015-11-02 07:43:04Z maronga
130	! Bugfix in output of RRTGM data
131	!
132	! 1691 2015-10-26 16:17:44Z maronga
133	! Added output of Obukhov length (ol) and radiative heating rates for RRTMG.
134	! Formatting corrections.
135	!
136	! 1682 2015-10-07 23:56:08Z knoop
137	! Code annotations made doxygen readable
138	!
139	! 1585 2015-04-30 07:05:52Z maronga
140	! Added support for RRTMG
141	!
142	! 1555 2015-03-04 17:44:27Z maronga
143	! Added output of r_a and r_s
144	!
145	! 1551 2015-03-03 14:18:16Z maronga
146	! Added suppport for land surface model and radiation model output. In the course
147	! of this action, the limits for vertical loops have been changed (from nzb and
148	! nzt+1 to nzb_do and nzt_do, respectively in order to allow soil model output).
149	! Moreover, a new vertical grid zs was introduced.
150	!
151	! 1359 2014-04-11 17:15:14Z hoffmann
152	! New particle structure integrated.
153	!
154	! 1353 2014-04-08 15:21:23Z heinze
155	! REAL constants provided with KIND-attribute
156	!
157	! 1327 2014-03-21 11:00:16Z raasch
158	! parts concerning iso2d output removed,
159	! -netcdf output queries
160	!
161	! 1320 2014-03-20 08:40:49Z raasch
162	! ONLY-attribute added to USE-statements,
163	! kind-parameters added to all INTEGER and REAL declaration statements,
164	! kinds are defined in new module kinds,
165	! revision history before 2012 removed,
166	! comment fields (!:) to be used for variable explanations added to
167	! all variable declaration statements
168	!
169	! 1318 2014-03-17 13:35:16Z raasch
170	! barrier argument removed from cpu_log.
171	! module interfaces removed
172	!
173	! 1311 2014-03-14 12:13:39Z heinze
174	! bugfix: close #if defined( __netcdf )
175	!
176	! 1308 2014-03-13 14:58:42Z fricke
177	! +local_2d_sections, local_2d_sections_l, ns
178	! Check, if the limit of the time dimension is exceeded for parallel output
179	! To increase the performance for parallel output, the following is done:
180	! - Update of time axis is only done by PE0
181	! - Cross sections are first stored on a local array and are written
182	! collectively to the output file by all PEs.
183	!
184	! 1115 2013-03-26 18:16:16Z hoffmann
185	! ql is calculated by calc_liquid_water_content
186	!
187	! 1076 2012-12-05 08:30:18Z hoffmann
188	! Bugfix in output of ql
189	!
190	! 1065 2012-11-22 17:42:36Z hoffmann
191	! Bugfix: Output of cross sections of ql
192	!
193	! 1053 2012-11-13 17:11:03Z hoffmann
194	! +qr, nr, qc and cross sections
195	!
196	! 1036 2012-10-22 13:43:42Z raasch
197	! code put under GPL (PALM 3.9)
198	!
199	! 1031 2012-10-19 14:35:30Z raasch
200	! netCDF4 without parallel file support implemented
201	!
202	! 1007 2012-09-19 14:30:36Z franke
203	! Bugfix: missing calculation of ql_vp added
204	!
205	! 978 2012-08-09 08:28:32Z fricke
206	! +z0h
207	!
208	! Revision 1.1 1997/08/11 06:24:09 raasch
209	! Initial revision
210	!
211	!
212	! Description:
213	! ------------
214	!> Data output of cross-sections in netCDF format or binary format
215	!> to be later converted to NetCDF by helper routine combine_plot_fields.
216	!> Attention: The position of the sectional planes is still not always computed
217	!> --------- correctly. (zu is used always)!
218	!------------------------------------------------------------------------------!
219	SUBROUTINE data_output_2d( mode, av )
220
221
222	USE arrays_3d, &
223	ONLY: dzw, e, heatflux_output_conversion, nc, nr, p, pt, &
224	precipitation_amount, prr, q, qc, ql, ql_c, ql_v, ql_vp, qr, &
225	rho_ocean, s, sa, tend, u, v, vpt, w, zu, zw, &
226	waterflux_output_conversion
227
228	USE averaging
229
230	USE cloud_parameters, &
231	ONLY: cp, hyrho, l_d_cp, l_v, pt_d_t
232
233	USE control_parameters, &
234	ONLY: cloud_physics, data_output_2d_on_each_pe, data_output_xy, &
235	data_output_xz, data_output_yz, do2d, &
236	do2d_xy_last_time, do2d_xy_time_count, &
237	do2d_xz_last_time, do2d_xz_time_count, &
238	do2d_yz_last_time, do2d_yz_time_count, &
239	ibc_uv_b, io_blocks, io_group, land_surface, message_string, &
240	ntdim_2d_xy, ntdim_2d_xz, ntdim_2d_yz, &
241	psolver, section, simulated_time, simulated_time_chr, &
242	time_since_reference_point, uv_exposure
243
244	USE cpulog, &
245	ONLY: cpu_log, log_point
246
247	USE grid_variables, &
248	ONLY: dx, dy
249
250	USE gust_mod, &
251	ONLY: gust_data_output_2d, gust_module_enabled
252
253	USE indices, &
254	ONLY: nbgp, nx, nxl, nxlg, nxr, nxrg, ny, nyn, nyng, nys, nysg, nz, &
255	nzb, nzt, wall_flags_0
256
257	USE kinds
258
259	USE land_surface_model_mod, &
260	ONLY: lsm_data_output_2d, zs
261
262	#if defined( __netcdf )
263	USE NETCDF
264	#endif
265
266	USE netcdf_interface, &
267	ONLY: fill_value, id_set_xy, id_set_xz, id_set_yz, id_var_do2d, &
268	id_var_time_xy, id_var_time_xz, id_var_time_yz, nc_stat, &
269	netcdf_data_format, netcdf_handle_error
270
271	USE particle_attributes, &
272	ONLY: grid_particles, number_of_particles, particle_advection_start, &
273	particles, prt_count
274
275	USE pegrid
276
277	USE radiation_model_mod, &
278	ONLY: radiation, radiation_data_output_2d
279
280	USE surface_mod, &
281	ONLY: ind_pav_green, ind_veg_wall, ind_wat_win, surf_def_h, &
282	surf_lsm_h, surf_usm_h
283
284	USE turbulence_closure_mod, &
285	ONLY: tcm_data_output_2d
286
287	USE uv_exposure_model_mod, &
288	ONLY: uvem_data_output_2d
289
290
291	IMPLICIT NONE
292
293	CHARACTER (LEN=2) :: do2d_mode !<
294	CHARACTER (LEN=2) :: mode !<
295	CHARACTER (LEN=4) :: grid !<
296	CHARACTER (LEN=25) :: section_chr !<
297	CHARACTER (LEN=50) :: rtext !<
298
299	INTEGER(iwp) :: av !<
300	INTEGER(iwp) :: ngp !<
301	INTEGER(iwp) :: file_id !<
302	INTEGER(iwp) :: flag_nr !< number of masking flag
303	INTEGER(iwp) :: i !<
304	INTEGER(iwp) :: if !<
305	INTEGER(iwp) :: is !<
306	INTEGER(iwp) :: iis !<
307	INTEGER(iwp) :: j !<
308	INTEGER(iwp) :: k !<
309	INTEGER(iwp) :: l !<
310	INTEGER(iwp) :: layer_xy !<
311	INTEGER(iwp) :: m !<
312	INTEGER(iwp) :: n !<
313	INTEGER(iwp) :: nis !<
314	INTEGER(iwp) :: ns !<
315	INTEGER(iwp) :: nzb_do !< lower limit of the data field (usually nzb)
316	INTEGER(iwp) :: nzt_do !< upper limit of the data field (usually nzt+1)
317	INTEGER(iwp) :: psi !<
318	INTEGER(iwp) :: s_ind !<
319	INTEGER(iwp) :: sender !<
320	INTEGER(iwp) :: ind(4) !<
321
322	LOGICAL :: found !<
323	LOGICAL :: resorted !<
324	LOGICAL :: two_d !<
325
326	REAL(wp) :: mean_r !<
327	REAL(wp) :: s_r2 !<
328	REAL(wp) :: s_r3 !<
329
330	REAL(wp), DIMENSION(:), ALLOCATABLE :: level_z !<
331	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d !<
332	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d_l !<
333
334	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !<
335	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections !<
336	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections_l !<
337
338	#if defined( __parallel )
339	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: total_2d !<
340	#endif
341	REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !<
342
343	NAMELIST /LOCAL/ rtext
344
345	!
346	!-- Immediate return, if no output is requested (no respective sections
347	!-- found in parameter data_output)
348	IF ( mode == 'xy' .AND. .NOT. data_output_xy(av) ) RETURN
349	IF ( mode == 'xz' .AND. .NOT. data_output_xz(av) ) RETURN
350	IF ( mode == 'yz' .AND. .NOT. data_output_yz(av) ) RETURN
351
352	CALL cpu_log (log_point(3),'data_output_2d','start')
353
354	two_d = .FALSE. ! local variable to distinguish between output of pure 2D
355	! arrays and cross-sections of 3D arrays.
356
357	!
358	!-- Depending on the orientation of the cross-section, the respective output
359	!-- files have to be opened.
360	SELECT CASE ( mode )
361
362	CASE ( 'xy' )
363	s_ind = 1
364	ALLOCATE( level_z(nzb:nzt+1), local_2d(nxl:nxr,nys:nyn) )
365
366	IF ( netcdf_data_format > 4 ) THEN
367	ns = 1
368	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
369	ns = ns + 1
370	ENDDO
371	ns = ns - 1
372	ALLOCATE( local_2d_sections(nxl:nxr,nys:nyn,1:ns) )
373	local_2d_sections = 0.0_wp
374	ENDIF
375
376	!
377	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
378	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
379	CALL check_open( 101+av*10 )
380	ENDIF
381	IF ( data_output_2d_on_each_pe ) THEN
382	CALL check_open( 21 )
383	ELSE
384	IF ( myid == 0 ) THEN
385	#if defined( __parallel )
386	ALLOCATE( total_2d(0:nx,0:ny) )
387	#endif
388	ENDIF
389	ENDIF
390
391	CASE ( 'xz' )
392	s_ind = 2
393	ALLOCATE( local_2d(nxl:nxr,nzb:nzt+1) )
394
395	IF ( netcdf_data_format > 4 ) THEN
396	ns = 1
397	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
398	ns = ns + 1
399	ENDDO
400	ns = ns - 1
401	ALLOCATE( local_2d_sections(nxl:nxr,1:ns,nzb:nzt+1) )
402	ALLOCATE( local_2d_sections_l(nxl:nxr,1:ns,nzb:nzt+1) )
403	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
404	ENDIF
405
406	!
407	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
408	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
409	CALL check_open( 102+av*10 )
410	ENDIF
411
412	IF ( data_output_2d_on_each_pe ) THEN
413	CALL check_open( 22 )
414	ELSE
415	IF ( myid == 0 ) THEN
416	#if defined( __parallel )
417	ALLOCATE( total_2d(0:nx,nzb:nzt+1) )
418	#endif
419	ENDIF
420	ENDIF
421
422	CASE ( 'yz' )
423	s_ind = 3
424	ALLOCATE( local_2d(nys:nyn,nzb:nzt+1) )
425
426	IF ( netcdf_data_format > 4 ) THEN
427	ns = 1
428	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
429	ns = ns + 1
430	ENDDO
431	ns = ns - 1
432	ALLOCATE( local_2d_sections(1:ns,nys:nyn,nzb:nzt+1) )
433	ALLOCATE( local_2d_sections_l(1:ns,nys:nyn,nzb:nzt+1) )
434	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
435	ENDIF
436
437	!
438	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
439	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
440	CALL check_open( 103+av*10 )
441	ENDIF
442
443	IF ( data_output_2d_on_each_pe ) THEN
444	CALL check_open( 23 )
445	ELSE
446	IF ( myid == 0 ) THEN
447	#if defined( __parallel )
448	ALLOCATE( total_2d(0:ny,nzb:nzt+1) )
449	#endif
450	ENDIF
451	ENDIF
452
453	CASE DEFAULT
454	message_string = 'unknown cross-section: ' // TRIM( mode )
455	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
456
457	END SELECT
458
459	!
460	!-- For parallel netcdf output the time axis must be limited. Return, if this
461	!-- limit is exceeded. This could be the case, if the simulated time exceeds
462	!-- the given end time by the length of the given output interval.
463	IF ( netcdf_data_format > 4 ) THEN
464	IF ( mode == 'xy' .AND. do2d_xy_time_count(av) + 1 > &
465	ntdim_2d_xy(av) ) THEN
466	WRITE ( message_string, * ) 'Output of xy cross-sections is not ', &
467	'given at t=', simulated_time, '&because the', &
468	' maximum number of output time levels is exceeded.'
469	CALL message( 'data_output_2d', 'PA0384', 0, 1, 0, 6, 0 )
470	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
471	RETURN
472	ENDIF
473	IF ( mode == 'xz' .AND. do2d_xz_time_count(av) + 1 > &
474	ntdim_2d_xz(av) ) THEN
475	WRITE ( message_string, * ) 'Output of xz cross-sections is not ', &
476	'given at t=', simulated_time, '&because the', &
477	' maximum number of output time levels is exceeded.'
478	CALL message( 'data_output_2d', 'PA0385', 0, 1, 0, 6, 0 )
479	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
480	RETURN
481	ENDIF
482	IF ( mode == 'yz' .AND. do2d_yz_time_count(av) + 1 > &
483	ntdim_2d_yz(av) ) THEN
484	WRITE ( message_string, * ) 'Output of yz cross-sections is not ', &
485	'given at t=', simulated_time, '&because the', &
486	' maximum number of output time levels is exceeded.'
487	CALL message( 'data_output_2d', 'PA0386', 0, 1, 0, 6, 0 )
488	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
489	RETURN
490	ENDIF
491	ENDIF
492
493	!
494	!-- Allocate a temporary array for resorting (kji -> ijk).
495	ALLOCATE( local_pf(nxl:nxr,nys:nyn,nzb:nzt+1) )
496	local_pf = 0.0
497
498	!
499	!-- Loop of all variables to be written.
500	!-- Output dimensions chosen
501	if = 1
502	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
503	do2d_mode = do2d(av,if)(l-1:l)
504
505	DO WHILE ( do2d(av,if)(1:1) /= ' ' )
506
507	IF ( do2d_mode == mode ) THEN
508	!
509	!-- Set flag to steer output of radiation, land-surface, or user-defined
510	!-- quantities
511	found = .FALSE.
512
513	nzb_do = nzb
514	nzt_do = nzt+1
515	!
516	!-- Before each output, set array local_pf to fill value
517	local_pf = fill_value
518	!
519	!-- Set masking flag for topography for not resorted arrays
520	flag_nr = 0
521
522	!
523	!-- Store the array chosen on the temporary array.
524	resorted = .FALSE.
525	SELECT CASE ( TRIM( do2d(av,if) ) )
526	CASE ( 'e_xy', 'e_xz', 'e_yz' )
527	IF ( av == 0 ) THEN
528	to_be_resorted => e
529	ELSE
530	IF ( .NOT. ALLOCATED( e_av ) ) THEN
531	ALLOCATE( e_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
532	e_av = REAL( fill_value, KIND = wp )
533	ENDIF
534	to_be_resorted => e_av
535	ENDIF
536	IF ( mode == 'xy' ) level_z = zu
537
538	CASE ( 'lpt_xy', 'lpt_xz', 'lpt_yz' )
539	IF ( av == 0 ) THEN
540	to_be_resorted => pt
541	ELSE
542	IF ( .NOT. ALLOCATED( lpt_av ) ) THEN
543	ALLOCATE( lpt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
544	lpt_av = REAL( fill_value, KIND = wp )
545	ENDIF
546	to_be_resorted => lpt_av
547	ENDIF
548	IF ( mode == 'xy' ) level_z = zu
549
550	CASE ( 'lwp*_xy' ) ! 2d-array
551	IF ( av == 0 ) THEN
552	DO i = nxl, nxr
553	DO j = nys, nyn
554	local_pf(i,j,nzb+1) = SUM( ql(nzb:nzt,j,i) * &
555	dzw(1:nzt+1) )
556	ENDDO
557	ENDDO
558	ELSE
559	IF ( .NOT. ALLOCATED( lwp_av ) ) THEN
560	ALLOCATE( lwp_av(nysg:nyng,nxlg:nxrg) )
561	lwp_av = REAL( fill_value, KIND = wp )
562	ENDIF
563	DO i = nxl, nxr
564	DO j = nys, nyn
565	local_pf(i,j,nzb+1) = lwp_av(j,i)
566	ENDDO
567	ENDDO
568	ENDIF
569	resorted = .TRUE.
570	two_d = .TRUE.
571	level_z(nzb+1) = zu(nzb+1)
572
573	CASE ( 'nc_xy', 'nc_xz', 'nc_yz' )
574	IF ( av == 0 ) THEN
575	to_be_resorted => nc
576	ELSE
577	IF ( .NOT. ALLOCATED( nc_av ) ) THEN
578	ALLOCATE( nc_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
579	nc_av = REAL( fill_value, KIND = wp )
580	ENDIF
581	to_be_resorted => nc_av
582	ENDIF
583	IF ( mode == 'xy' ) level_z = zu
584
585	CASE ( 'nr_xy', 'nr_xz', 'nr_yz' )
586	IF ( av == 0 ) THEN
587	to_be_resorted => nr
588	ELSE
589	IF ( .NOT. ALLOCATED( nr_av ) ) THEN
590	ALLOCATE( nr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
591	nr_av = REAL( fill_value, KIND = wp )
592	ENDIF
593	to_be_resorted => nr_av
594	ENDIF
595	IF ( mode == 'xy' ) level_z = zu
596
597	CASE ( 'ghf*_xy' ) ! 2d-array
598	IF ( av == 0 ) THEN
599	DO m = 1, surf_lsm_h%ns
600	i = surf_lsm_h%i(m)
601	j = surf_lsm_h%j(m)
602	local_pf(i,j,nzb+1) = surf_lsm_h%ghf(m)
603	ENDDO
604	DO m = 1, surf_usm_h%ns
605	i = surf_usm_h%i(m)
606	j = surf_usm_h%j(m)
607	local_pf(i,j,nzb+1) = surf_usm_h%frac(ind_veg_wall,m) * &
608	surf_usm_h%wghf_eb(m) + &
609	surf_usm_h%frac(ind_pav_green,m) * &
610	surf_usm_h%wghf_eb_green(m) + &
611	surf_usm_h%frac(ind_wat_win,m) * &
612	surf_usm_h%wghf_eb_window(m)
613	ENDDO
614	ELSE
615	IF ( .NOT. ALLOCATED( ghf_av ) ) THEN
616	ALLOCATE( ghf_av(nysg:nyng,nxlg:nxrg) )
617	ghf_av = REAL( fill_value, KIND = wp )
618	ENDIF
619	DO i = nxl, nxr
620	DO j = nys, nyn
621	local_pf(i,j,nzb+1) = ghf_av(j,i)
622	ENDDO
623	ENDDO
624	ENDIF
625
626	resorted = .TRUE.
627	two_d = .TRUE.
628	level_z(nzb+1) = zu(nzb+1)
629
630	CASE ( 'ol*_xy' ) ! 2d-array
631	IF ( av == 0 ) THEN
632	DO m = 1, surf_def_h(0)%ns
633	i = surf_def_h(0)%i(m)
634	j = surf_def_h(0)%j(m)
635	local_pf(i,j,nzb+1) = surf_def_h(0)%ol(m)
636	ENDDO
637	DO m = 1, surf_lsm_h%ns
638	i = surf_lsm_h%i(m)
639	j = surf_lsm_h%j(m)
640	local_pf(i,j,nzb+1) = surf_lsm_h%ol(m)
641	ENDDO
642	DO m = 1, surf_usm_h%ns
643	i = surf_usm_h%i(m)
644	j = surf_usm_h%j(m)
645	local_pf(i,j,nzb+1) = surf_usm_h%ol(m)
646	ENDDO
647	ELSE
648	IF ( .NOT. ALLOCATED( ol_av ) ) THEN
649	ALLOCATE( ol_av(nysg:nyng,nxlg:nxrg) )
650	ol_av = REAL( fill_value, KIND = wp )
651	ENDIF
652	DO i = nxl, nxr
653	DO j = nys, nyn
654	local_pf(i,j,nzb+1) = ol_av(j,i)
655	ENDDO
656	ENDDO
657	ENDIF
658	resorted = .TRUE.
659	two_d = .TRUE.
660	level_z(nzb+1) = zu(nzb+1)
661
662	CASE ( 'p_xy', 'p_xz', 'p_yz' )
663	IF ( av == 0 ) THEN
664	IF ( psolver /= 'sor' ) CALL exchange_horiz( p, nbgp )
665	to_be_resorted => p
666	ELSE
667	IF ( .NOT. ALLOCATED( p_av ) ) THEN
668	ALLOCATE( p_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
669	p_av = REAL( fill_value, KIND = wp )
670	ENDIF
671	IF ( psolver /= 'sor' ) CALL exchange_horiz( p_av, nbgp )
672	to_be_resorted => p_av
673	ENDIF
674	IF ( mode == 'xy' ) level_z = zu
675
676	CASE ( 'pc_xy', 'pc_xz', 'pc_yz' ) ! particle concentration
677	IF ( av == 0 ) THEN
678	IF ( simulated_time >= particle_advection_start ) THEN
679	tend = prt_count
680	! CALL exchange_horiz( tend, nbgp )
681	ELSE
682	tend = 0.0_wp
683	ENDIF
684	DO i = nxl, nxr
685	DO j = nys, nyn
686	DO k = nzb, nzt+1
687	local_pf(i,j,k) = tend(k,j,i)
688	ENDDO
689	ENDDO
690	ENDDO
691	resorted = .TRUE.
692	ELSE
693	IF ( .NOT. ALLOCATED( pc_av ) ) THEN
694	ALLOCATE( pc_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
695	pc_av = REAL( fill_value, KIND = wp )
696	ENDIF
697	! CALL exchange_horiz( pc_av, nbgp )
698	to_be_resorted => pc_av
699	ENDIF
700
701	CASE ( 'pr_xy', 'pr_xz', 'pr_yz' ) ! mean particle radius (effective radius)
702	IF ( av == 0 ) THEN
703	IF ( simulated_time >= particle_advection_start ) THEN
704	DO i = nxl, nxr
705	DO j = nys, nyn
706	DO k = nzb, nzt+1
707	number_of_particles = prt_count(k,j,i)
708	IF (number_of_particles <= 0) CYCLE
709	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
710	s_r2 = 0.0_wp
711	s_r3 = 0.0_wp
712	DO n = 1, number_of_particles
713	IF ( particles(n)%particle_mask ) THEN
714	s_r2 = s_r2 + particles(n)%radius*2 &
715	particles(n)%weight_factor
716	s_r3 = s_r3 + particles(n)%radius*3 &
717	particles(n)%weight_factor
718	ENDIF
719	ENDDO
720	IF ( s_r2 > 0.0_wp ) THEN
721	mean_r = s_r3 / s_r2
722	ELSE
723	mean_r = 0.0_wp
724	ENDIF
725	tend(k,j,i) = mean_r
726	ENDDO
727	ENDDO
728	ENDDO
729	! CALL exchange_horiz( tend, nbgp )
730	ELSE
731	tend = 0.0_wp
732	ENDIF
733	DO i = nxl, nxr
734	DO j = nys, nyn
735	DO k = nzb, nzt+1
736	local_pf(i,j,k) = tend(k,j,i)
737	ENDDO
738	ENDDO
739	ENDDO
740	resorted = .TRUE.
741	ELSE
742	IF ( .NOT. ALLOCATED( pr_av ) ) THEN
743	ALLOCATE( pr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
744	pr_av = REAL( fill_value, KIND = wp )
745	ENDIF
746	! CALL exchange_horiz( pr_av, nbgp )
747	to_be_resorted => pr_av
748	ENDIF
749
750	CASE ( 'pra*_xy' ) ! 2d-array / integral quantity => no av
751	! CALL exchange_horiz_2d( precipitation_amount )
752	DO i = nxl, nxr
753	DO j = nys, nyn
754	local_pf(i,j,nzb+1) = precipitation_amount(j,i)
755	ENDDO
756	ENDDO
757	precipitation_amount = 0.0_wp ! reset for next integ. interval
758	resorted = .TRUE.
759	two_d = .TRUE.
760	level_z(nzb+1) = zu(nzb+1)
761
762	CASE ( 'prr_xy', 'prr_xz', 'prr_yz' )
763	IF ( av == 0 ) THEN
764	! CALL exchange_horiz( prr, nbgp )
765	DO i = nxl, nxr
766	DO j = nys, nyn
767	DO k = nzb, nzt+1
768	local_pf(i,j,k) = prr(k,j,i) * hyrho(nzb+1)
769	ENDDO
770	ENDDO
771	ENDDO
772	ELSE
773	IF ( .NOT. ALLOCATED( prr_av ) ) THEN
774	ALLOCATE( prr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
775	prr_av = REAL( fill_value, KIND = wp )
776	ENDIF
777	! CALL exchange_horiz( prr_av, nbgp )
778	DO i = nxl, nxr
779	DO j = nys, nyn
780	DO k = nzb, nzt+1
781	local_pf(i,j,k) = prr_av(k,j,i) * hyrho(nzb+1)
782	ENDDO
783	ENDDO
784	ENDDO
785	ENDIF
786	resorted = .TRUE.
787	IF ( mode == 'xy' ) level_z = zu
788
789	CASE ( 'pt_xy', 'pt_xz', 'pt_yz' )
790	IF ( av == 0 ) THEN
791	IF ( .NOT. cloud_physics ) THEN
792	to_be_resorted => pt
793	ELSE
794	DO i = nxl, nxr
795	DO j = nys, nyn
796	DO k = nzb, nzt+1
797	local_pf(i,j,k) = pt(k,j,i) + l_d_cp * &
798	pt_d_t(k) * &
799	ql(k,j,i)
800	ENDDO
801	ENDDO
802	ENDDO
803	resorted = .TRUE.
804	ENDIF
805	ELSE
806	IF ( .NOT. ALLOCATED( pt_av ) ) THEN
807	ALLOCATE( pt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
808	pt_av = REAL( fill_value, KIND = wp )
809	ENDIF
810	to_be_resorted => pt_av
811	ENDIF
812	IF ( mode == 'xy' ) level_z = zu
813
814	CASE ( 'q_xy', 'q_xz', 'q_yz' )
815	IF ( av == 0 ) THEN
816	to_be_resorted => q
817	ELSE
818	IF ( .NOT. ALLOCATED( q_av ) ) THEN
819	ALLOCATE( q_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
820	q_av = REAL( fill_value, KIND = wp )
821	ENDIF
822	to_be_resorted => q_av
823	ENDIF
824	IF ( mode == 'xy' ) level_z = zu
825
826	CASE ( 'qc_xy', 'qc_xz', 'qc_yz' )
827	IF ( av == 0 ) THEN
828	to_be_resorted => qc
829	ELSE
830	IF ( .NOT. ALLOCATED( qc_av ) ) THEN
831	ALLOCATE( qc_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
832	qc_av = REAL( fill_value, KIND = wp )
833	ENDIF
834	to_be_resorted => qc_av
835	ENDIF
836	IF ( mode == 'xy' ) level_z = zu
837
838	CASE ( 'ql_xy', 'ql_xz', 'ql_yz' )
839	IF ( av == 0 ) THEN
840	to_be_resorted => ql
841	ELSE
842	IF ( .NOT. ALLOCATED( ql_av ) ) THEN
843	ALLOCATE( ql_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
844	ql_av = REAL( fill_value, KIND = wp )
845	ENDIF
846	to_be_resorted => ql_av
847	ENDIF
848	IF ( mode == 'xy' ) level_z = zu
849
850	CASE ( 'ql_c_xy', 'ql_c_xz', 'ql_c_yz' )
851	IF ( av == 0 ) THEN
852	to_be_resorted => ql_c
853	ELSE
854	IF ( .NOT. ALLOCATED( ql_c_av ) ) THEN
855	ALLOCATE( ql_c_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
856	ql_c_av = REAL( fill_value, KIND = wp )
857	ENDIF
858	to_be_resorted => ql_c_av
859	ENDIF
860	IF ( mode == 'xy' ) level_z = zu
861
862	CASE ( 'ql_v_xy', 'ql_v_xz', 'ql_v_yz' )
863	IF ( av == 0 ) THEN
864	to_be_resorted => ql_v
865	ELSE
866	IF ( .NOT. ALLOCATED( ql_v_av ) ) THEN
867	ALLOCATE( ql_v_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
868	ql_v_av = REAL( fill_value, KIND = wp )
869	ENDIF
870	to_be_resorted => ql_v_av
871	ENDIF
872	IF ( mode == 'xy' ) level_z = zu
873
874	CASE ( 'ql_vp_xy', 'ql_vp_xz', 'ql_vp_yz' )
875	IF ( av == 0 ) THEN
876	IF ( simulated_time >= particle_advection_start ) THEN
877	DO i = nxl, nxr
878	DO j = nys, nyn
879	DO k = nzb, nzt+1
880	number_of_particles = prt_count(k,j,i)
881	IF (number_of_particles <= 0) CYCLE
882	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
883	DO n = 1, number_of_particles
884	IF ( particles(n)%particle_mask ) THEN
885	tend(k,j,i) = tend(k,j,i) + &
886	particles(n)%weight_factor / &
887	prt_count(k,j,i)
888	ENDIF
889	ENDDO
890	ENDDO
891	ENDDO
892	ENDDO
893	! CALL exchange_horiz( tend, nbgp )
894	ELSE
895	tend = 0.0_wp
896	ENDIF
897	DO i = nxl, nxr
898	DO j = nys, nyn
899	DO k = nzb, nzt+1
900	local_pf(i,j,k) = tend(k,j,i)
901	ENDDO
902	ENDDO
903	ENDDO
904	resorted = .TRUE.
905	ELSE
906	IF ( .NOT. ALLOCATED( ql_vp_av ) ) THEN
907	ALLOCATE( ql_vp_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
908	ql_vp_av = REAL( fill_value, KIND = wp )
909	ENDIF
910	! CALL exchange_horiz( ql_vp_av, nbgp )
911	to_be_resorted => ql_vp_av
912	ENDIF
913	IF ( mode == 'xy' ) level_z = zu
914
915	CASE ( 'qr_xy', 'qr_xz', 'qr_yz' )
916	IF ( av == 0 ) THEN
917	to_be_resorted => qr
918	ELSE
919	IF ( .NOT. ALLOCATED( qr_av ) ) THEN
920	ALLOCATE( qr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
921	qr_av = REAL( fill_value, KIND = wp )
922	ENDIF
923	to_be_resorted => qr_av
924	ENDIF
925	IF ( mode == 'xy' ) level_z = zu
926
927	CASE ( 'qsws*_xy' ) ! 2d-array
928	IF ( av == 0 ) THEN
929	!
930	!-- In case of default surfaces, clean-up flux by density.
931	!-- In case of land- and urban-surfaces, convert fluxes into
932	!-- dynamic units
933	DO m = 1, surf_def_h(0)%ns
934	i = surf_def_h(0)%i(m)
935	j = surf_def_h(0)%j(m)
936	k = surf_def_h(0)%k(m)
937	local_pf(i,j,nzb+1) = surf_def_h(0)%qsws(m) * &
938	waterflux_output_conversion(k)
939	ENDDO
940	DO m = 1, surf_lsm_h%ns
941	i = surf_lsm_h%i(m)
942	j = surf_lsm_h%j(m)
943	k = surf_lsm_h%k(m)
944	local_pf(i,j,nzb+1) = surf_lsm_h%qsws(m) * l_v
945	ENDDO
946	DO m = 1, surf_usm_h%ns
947	i = surf_usm_h%i(m)
948	j = surf_usm_h%j(m)
949	k = surf_usm_h%k(m)
950	local_pf(i,j,nzb+1) = surf_usm_h%qsws(m) * l_v
951	ENDDO
952	ELSE
953	IF ( .NOT. ALLOCATED( qsws_av ) ) THEN
954	ALLOCATE( qsws_av(nysg:nyng,nxlg:nxrg) )
955	qsws_av = REAL( fill_value, KIND = wp )
956	ENDIF
957	DO i = nxl, nxr
958	DO j = nys, nyn
959	local_pf(i,j,nzb+1) = qsws_av(j,i)
960	ENDDO
961	ENDDO
962	ENDIF
963	resorted = .TRUE.
964	two_d = .TRUE.
965	level_z(nzb+1) = zu(nzb+1)
966
967	CASE ( 'qv_xy', 'qv_xz', 'qv_yz' )
968	IF ( av == 0 ) THEN
969	DO i = nxl, nxr
970	DO j = nys, nyn
971	DO k = nzb, nzt+1
972	local_pf(i,j,k) = q(k,j,i) - ql(k,j,i)
973	ENDDO
974	ENDDO
975	ENDDO
976	resorted = .TRUE.
977	ELSE
978	IF ( .NOT. ALLOCATED( qv_av ) ) THEN
979	ALLOCATE( qv_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
980	qv_av = REAL( fill_value, KIND = wp )
981	ENDIF
982	to_be_resorted => qv_av
983	ENDIF
984	IF ( mode == 'xy' ) level_z = zu
985
986	CASE ( 'r_a*_xy' ) ! 2d-array
987	IF ( av == 0 ) THEN
988	DO m = 1, surf_lsm_h%ns
989	i = surf_lsm_h%i(m)
990	j = surf_lsm_h%j(m)
991	local_pf(i,j,nzb+1) = surf_lsm_h%r_a(m)
992	ENDDO
993
994	DO m = 1, surf_usm_h%ns
995	i = surf_usm_h%i(m)
996	j = surf_usm_h%j(m)
997	local_pf(i,j,nzb+1) = &
998	( surf_usm_h%frac(ind_veg_wall,m) * &
999	surf_usm_h%r_a(m) + &
1000	surf_usm_h%frac(ind_pav_green,m) * &
1001	surf_usm_h%r_a_green(m) + &
1002	surf_usm_h%frac(ind_wat_win,m) * &
1003	surf_usm_h%r_a_window(m) )
1004	ENDDO
1005	ELSE
1006	IF ( .NOT. ALLOCATED( r_a_av ) ) THEN
1007	ALLOCATE( r_a_av(nysg:nyng,nxlg:nxrg) )
1008	r_a_av = REAL( fill_value, KIND = wp )
1009	ENDIF
1010	DO i = nxl, nxr
1011	DO j = nys, nyn
1012	local_pf(i,j,nzb+1) = r_a_av(j,i)
1013	ENDDO
1014	ENDDO
1015	ENDIF
1016	resorted = .TRUE.
1017	two_d = .TRUE.
1018	level_z(nzb+1) = zu(nzb+1)
1019
1020	CASE ( 'rho_ocean_xy', 'rho_ocean_xz', 'rho_ocean_yz' )
1021	IF ( av == 0 ) THEN
1022	to_be_resorted => rho_ocean
1023	ELSE
1024	IF ( .NOT. ALLOCATED( rho_ocean_av ) ) THEN
1025	ALLOCATE( rho_ocean_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1026	rho_ocean_av = REAL( fill_value, KIND = wp )
1027	ENDIF
1028	to_be_resorted => rho_ocean_av
1029	ENDIF
1030
1031	CASE ( 's_xy', 's_xz', 's_yz' )
1032	IF ( av == 0 ) THEN
1033	to_be_resorted => s
1034	ELSE
1035	IF ( .NOT. ALLOCATED( s_av ) ) THEN
1036	ALLOCATE( s_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1037	s_av = REAL( fill_value, KIND = wp )
1038	ENDIF
1039	to_be_resorted => s_av
1040	ENDIF
1041
1042	CASE ( 'sa_xy', 'sa_xz', 'sa_yz' )
1043	IF ( av == 0 ) THEN
1044	to_be_resorted => sa
1045	ELSE
1046	IF ( .NOT. ALLOCATED( sa_av ) ) THEN
1047	ALLOCATE( sa_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1048	sa_av = REAL( fill_value, KIND = wp )
1049	ENDIF
1050	to_be_resorted => sa_av
1051	ENDIF
1052
1053	CASE ( 'shf*_xy' ) ! 2d-array
1054	IF ( av == 0 ) THEN
1055	!
1056	!-- In case of default surfaces, clean-up flux by density.
1057	!-- In case of land- and urban-surfaces, convert fluxes into
1058	!-- dynamic units.
1059	DO m = 1, surf_def_h(0)%ns
1060	i = surf_def_h(0)%i(m)
1061	j = surf_def_h(0)%j(m)
1062	k = surf_def_h(0)%k(m)
1063	local_pf(i,j,nzb+1) = surf_def_h(0)%shf(m) * &
1064	heatflux_output_conversion(k)
1065	ENDDO
1066	DO m = 1, surf_lsm_h%ns
1067	i = surf_lsm_h%i(m)
1068	j = surf_lsm_h%j(m)
1069	k = surf_lsm_h%k(m)
1070	local_pf(i,j,nzb+1) = surf_lsm_h%shf(m) * cp
1071	ENDDO
1072	DO m = 1, surf_usm_h%ns
1073	i = surf_usm_h%i(m)
1074	j = surf_usm_h%j(m)
1075	k = surf_usm_h%k(m)
1076	local_pf(i,j,nzb+1) = surf_usm_h%shf(m) * cp
1077	ENDDO
1078	ELSE
1079	IF ( .NOT. ALLOCATED( shf_av ) ) THEN
1080	ALLOCATE( shf_av(nysg:nyng,nxlg:nxrg) )
1081	shf_av = REAL( fill_value, KIND = wp )
1082	ENDIF
1083	DO i = nxl, nxr
1084	DO j = nys, nyn
1085	local_pf(i,j,nzb+1) = shf_av(j,i)
1086	ENDDO
1087	ENDDO
1088	ENDIF
1089	resorted = .TRUE.
1090	two_d = .TRUE.
1091	level_z(nzb+1) = zu(nzb+1)
1092
1093	CASE ( 'ssws*_xy' ) ! 2d-array
1094	IF ( av == 0 ) THEN
1095	DO m = 1, surf_def_h(0)%ns
1096	i = surf_def_h(0)%i(m)
1097	j = surf_def_h(0)%j(m)
1098	local_pf(i,j,nzb+1) = surf_def_h(0)%ssws(m)
1099	ENDDO
1100	DO m = 1, surf_lsm_h%ns
1101	i = surf_lsm_h%i(m)
1102	j = surf_lsm_h%j(m)
1103	local_pf(i,j,nzb+1) = surf_lsm_h%ssws(m)
1104	ENDDO
1105	DO m = 1, surf_usm_h%ns
1106	i = surf_usm_h%i(m)
1107	j = surf_usm_h%j(m)
1108	local_pf(i,j,nzb+1) = surf_usm_h%ssws(m)
1109	ENDDO
1110	ELSE
1111	IF ( .NOT. ALLOCATED( ssws_av ) ) THEN
1112	ALLOCATE( ssws_av(nysg:nyng,nxlg:nxrg) )
1113	ssws_av = REAL( fill_value, KIND = wp )
1114	ENDIF
1115	DO i = nxl, nxr
1116	DO j = nys, nyn
1117	local_pf(i,j,nzb+1) = ssws_av(j,i)
1118	ENDDO
1119	ENDDO
1120	ENDIF
1121	resorted = .TRUE.
1122	two_d = .TRUE.
1123	level_z(nzb+1) = zu(nzb+1)
1124
1125	CASE ( 't*_xy' ) ! 2d-array
1126	IF ( av == 0 ) THEN
1127	DO m = 1, surf_def_h(0)%ns
1128	i = surf_def_h(0)%i(m)
1129	j = surf_def_h(0)%j(m)
1130	local_pf(i,j,nzb+1) = surf_def_h(0)%ts(m)
1131	ENDDO
1132	DO m = 1, surf_lsm_h%ns
1133	i = surf_lsm_h%i(m)
1134	j = surf_lsm_h%j(m)
1135	local_pf(i,j,nzb+1) = surf_lsm_h%ts(m)
1136	ENDDO
1137	DO m = 1, surf_usm_h%ns
1138	i = surf_usm_h%i(m)
1139	j = surf_usm_h%j(m)
1140	local_pf(i,j,nzb+1) = surf_usm_h%ts(m)
1141	ENDDO
1142	ELSE
1143	IF ( .NOT. ALLOCATED( ts_av ) ) THEN
1144	ALLOCATE( ts_av(nysg:nyng,nxlg:nxrg) )
1145	ts_av = REAL( fill_value, KIND = wp )
1146	ENDIF
1147	DO i = nxl, nxr
1148	DO j = nys, nyn
1149	local_pf(i,j,nzb+1) = ts_av(j,i)
1150	ENDDO
1151	ENDDO
1152	ENDIF
1153	resorted = .TRUE.
1154	two_d = .TRUE.
1155	level_z(nzb+1) = zu(nzb+1)
1156
1157	CASE ( 'tsurf*_xy' ) ! 2d-array
1158	IF ( av == 0 ) THEN
1159	DO m = 1, surf_def_h(0)%ns
1160	i = surf_def_h(0)%i(m)
1161	j = surf_def_h(0)%j(m)
1162	local_pf(i,j,nzb+1) = surf_def_h(0)%pt_surface(m)
1163	ENDDO
1164
1165	DO m = 1, surf_lsm_h%ns
1166	i = surf_lsm_h%i(m)
1167	j = surf_lsm_h%j(m)
1168	local_pf(i,j,nzb+1) = surf_lsm_h%pt_surface(m)
1169	ENDDO
1170
1171	DO m = 1, surf_usm_h%ns
1172	i = surf_usm_h%i(m)
1173	j = surf_usm_h%j(m)
1174	local_pf(i,j,nzb+1) = surf_usm_h%pt_surface(m)
1175	ENDDO
1176
1177	ELSE
1178	IF ( .NOT. ALLOCATED( tsurf_av ) ) THEN
1179	ALLOCATE( tsurf_av(nysg:nyng,nxlg:nxrg) )
1180	tsurf_av = REAL( fill_value, KIND = wp )
1181	ENDIF
1182	DO i = nxl, nxr
1183	DO j = nys, nyn
1184	local_pf(i,j,nzb+1) = tsurf_av(j,i)
1185	ENDDO
1186	ENDDO
1187	ENDIF
1188	resorted = .TRUE.
1189	two_d = .TRUE.
1190	level_z(nzb+1) = zu(nzb+1)
1191
1192	CASE ( 'u_xy', 'u_xz', 'u_yz' )
1193	flag_nr = 1
1194	IF ( av == 0 ) THEN
1195	to_be_resorted => u
1196	ELSE
1197	IF ( .NOT. ALLOCATED( u_av ) ) THEN
1198	ALLOCATE( u_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1199	u_av = REAL( fill_value, KIND = wp )
1200	ENDIF
1201	to_be_resorted => u_av
1202	ENDIF
1203	IF ( mode == 'xy' ) level_z = zu
1204	!
1205	!-- Substitute the values generated by "mirror" boundary condition
1206	!-- at the bottom boundary by the real surface values.
1207	IF ( do2d(av,if) == 'u_xz' .OR. do2d(av,if) == 'u_yz' ) THEN
1208	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
1209	ENDIF
1210
1211	CASE ( 'u*_xy' ) ! 2d-array
1212	IF ( av == 0 ) THEN
1213	DO m = 1, surf_def_h(0)%ns
1214	i = surf_def_h(0)%i(m)
1215	j = surf_def_h(0)%j(m)
1216	local_pf(i,j,nzb+1) = surf_def_h(0)%us(m)
1217	ENDDO
1218	DO m = 1, surf_lsm_h%ns
1219	i = surf_lsm_h%i(m)
1220	j = surf_lsm_h%j(m)
1221	local_pf(i,j,nzb+1) = surf_lsm_h%us(m)
1222	ENDDO
1223	DO m = 1, surf_usm_h%ns
1224	i = surf_usm_h%i(m)
1225	j = surf_usm_h%j(m)
1226	local_pf(i,j,nzb+1) = surf_usm_h%us(m)
1227	ENDDO
1228	ELSE
1229	IF ( .NOT. ALLOCATED( us_av ) ) THEN
1230	ALLOCATE( us_av(nysg:nyng,nxlg:nxrg) )
1231	us_av = REAL( fill_value, KIND = wp )
1232	ENDIF
1233	DO i = nxl, nxr
1234	DO j = nys, nyn
1235	local_pf(i,j,nzb+1) = us_av(j,i)
1236	ENDDO
1237	ENDDO
1238	ENDIF
1239	resorted = .TRUE.
1240	two_d = .TRUE.
1241	level_z(nzb+1) = zu(nzb+1)
1242
1243	CASE ( 'v_xy', 'v_xz', 'v_yz' )
1244	flag_nr = 2
1245	IF ( av == 0 ) THEN
1246	to_be_resorted => v
1247	ELSE
1248	IF ( .NOT. ALLOCATED( v_av ) ) THEN
1249	ALLOCATE( v_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1250	v_av = REAL( fill_value, KIND = wp )
1251	ENDIF
1252	to_be_resorted => v_av
1253	ENDIF
1254	IF ( mode == 'xy' ) level_z = zu
1255	!
1256	!-- Substitute the values generated by "mirror" boundary condition
1257	!-- at the bottom boundary by the real surface values.
1258	IF ( do2d(av,if) == 'v_xz' .OR. do2d(av,if) == 'v_yz' ) THEN
1259	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
1260	ENDIF
1261
1262	CASE ( 'vpt_xy', 'vpt_xz', 'vpt_yz' )
1263	IF ( av == 0 ) THEN
1264	to_be_resorted => vpt
1265	ELSE
1266	IF ( .NOT. ALLOCATED( vpt_av ) ) THEN
1267	ALLOCATE( vpt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1268	vpt_av = REAL( fill_value, KIND = wp )
1269	ENDIF
1270	to_be_resorted => vpt_av
1271	ENDIF
1272	IF ( mode == 'xy' ) level_z = zu
1273
1274	CASE ( 'w_xy', 'w_xz', 'w_yz' )
1275	flag_nr = 3
1276	IF ( av == 0 ) THEN
1277	to_be_resorted => w
1278	ELSE
1279	IF ( .NOT. ALLOCATED( w_av ) ) THEN
1280	ALLOCATE( w_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1281	w_av = REAL( fill_value, KIND = wp )
1282	ENDIF
1283	to_be_resorted => w_av
1284	ENDIF
1285	IF ( mode == 'xy' ) level_z = zw
1286
1287	CASE ( 'z0*_xy' ) ! 2d-array
1288	IF ( av == 0 ) THEN
1289	DO m = 1, surf_def_h(0)%ns
1290	i = surf_def_h(0)%i(m)
1291	j = surf_def_h(0)%j(m)
1292	local_pf(i,j,nzb+1) = surf_def_h(0)%z0(m)
1293	ENDDO
1294	DO m = 1, surf_lsm_h%ns
1295	i = surf_lsm_h%i(m)
1296	j = surf_lsm_h%j(m)
1297	local_pf(i,j,nzb+1) = surf_lsm_h%z0(m)
1298	ENDDO
1299	DO m = 1, surf_usm_h%ns
1300	i = surf_usm_h%i(m)
1301	j = surf_usm_h%j(m)
1302	local_pf(i,j,nzb+1) = surf_usm_h%z0(m)
1303	ENDDO
1304	ELSE
1305	IF ( .NOT. ALLOCATED( z0_av ) ) THEN
1306	ALLOCATE( z0_av(nysg:nyng,nxlg:nxrg) )
1307	z0_av = REAL( fill_value, KIND = wp )
1308	ENDIF
1309	DO i = nxl, nxr
1310	DO j = nys, nyn
1311	local_pf(i,j,nzb+1) = z0_av(j,i)
1312	ENDDO
1313	ENDDO
1314	ENDIF
1315	resorted = .TRUE.
1316	two_d = .TRUE.
1317	level_z(nzb+1) = zu(nzb+1)
1318
1319	CASE ( 'z0h*_xy' ) ! 2d-array
1320	IF ( av == 0 ) THEN
1321	DO m = 1, surf_def_h(0)%ns
1322	i = surf_def_h(0)%i(m)
1323	j = surf_def_h(0)%j(m)
1324	local_pf(i,j,nzb+1) = surf_def_h(0)%z0h(m)
1325	ENDDO
1326	DO m = 1, surf_lsm_h%ns
1327	i = surf_lsm_h%i(m)
1328	j = surf_lsm_h%j(m)
1329	local_pf(i,j,nzb+1) = surf_lsm_h%z0h(m)
1330	ENDDO
1331	DO m = 1, surf_usm_h%ns
1332	i = surf_usm_h%i(m)
1333	j = surf_usm_h%j(m)
1334	local_pf(i,j,nzb+1) = surf_usm_h%z0h(m)
1335	ENDDO
1336	ELSE
1337	IF ( .NOT. ALLOCATED( z0h_av ) ) THEN
1338	ALLOCATE( z0h_av(nysg:nyng,nxlg:nxrg) )
1339	z0h_av = REAL( fill_value, KIND = wp )
1340	ENDIF
1341	DO i = nxl, nxr
1342	DO j = nys, nyn
1343	local_pf(i,j,nzb+1) = z0h_av(j,i)
1344	ENDDO
1345	ENDDO
1346	ENDIF
1347	resorted = .TRUE.
1348	two_d = .TRUE.
1349	level_z(nzb+1) = zu(nzb+1)
1350
1351	CASE ( 'z0q*_xy' ) ! 2d-array
1352	IF ( av == 0 ) THEN
1353	DO m = 1, surf_def_h(0)%ns
1354	i = surf_def_h(0)%i(m)
1355	j = surf_def_h(0)%j(m)
1356	local_pf(i,j,nzb+1) = surf_def_h(0)%z0q(m)
1357	ENDDO
1358	DO m = 1, surf_lsm_h%ns
1359	i = surf_lsm_h%i(m)
1360	j = surf_lsm_h%j(m)
1361	local_pf(i,j,nzb+1) = surf_lsm_h%z0q(m)
1362	ENDDO
1363	DO m = 1, surf_usm_h%ns
1364	i = surf_usm_h%i(m)
1365	j = surf_usm_h%j(m)
1366	local_pf(i,j,nzb+1) = surf_usm_h%z0q(m)
1367	ENDDO
1368	ELSE
1369	IF ( .NOT. ALLOCATED( z0q_av ) ) THEN
1370	ALLOCATE( z0q_av(nysg:nyng,nxlg:nxrg) )
1371	z0q_av = REAL( fill_value, KIND = wp )
1372	ENDIF
1373	DO i = nxl, nxr
1374	DO j = nys, nyn
1375	local_pf(i,j,nzb+1) = z0q_av(j,i)
1376	ENDDO
1377	ENDDO
1378	ENDIF
1379	resorted = .TRUE.
1380	two_d = .TRUE.
1381	level_z(nzb+1) = zu(nzb+1)
1382
1383	CASE DEFAULT
1384
1385	!
1386	!-- Land surface model quantity
1387	IF ( land_surface ) THEN
1388	CALL lsm_data_output_2d( av, do2d(av,if), found, grid, mode,&
1389	local_pf, two_d, nzb_do, nzt_do )
1390	ENDIF
1391
1392	!
1393	!-- Turbulence closure variables
1394	IF ( .NOT. found ) THEN
1395	CALL tcm_data_output_2d( av, do2d(av,if), found, grid, mode,&
1396	local_pf, two_d, nzb_do, nzt_do )
1397	ENDIF
1398
1399	!
1400	!-- Radiation quantity
1401	IF ( .NOT. found .AND. radiation ) THEN
1402	CALL radiation_data_output_2d( av, do2d(av,if), found, grid,&
1403	mode, local_pf, two_d, &
1404	nzb_do, nzt_do )
1405	ENDIF
1406
1407	!
1408	!-- Gust module quantities
1409	IF ( .NOT. found .AND. gust_module_enabled ) THEN
1410	CALL gust_data_output_2d( av, do2d(av,if), found, grid, &
1411	local_pf, two_d, nzb_do, nzt_do )
1412	ENDIF
1413
1414	!
1415	!-- UV exposure model quantity
1416	IF ( uv_exposure ) THEN
1417	CALL uvem_data_output_2d( av, do2d(av,if), found, grid, mode,&
1418	local_pf, two_d, nzb_do, nzt_do )
1419	ENDIF
1420
1421	!
1422	!-- User defined quantity
1423	IF ( .NOT. found ) THEN
1424	CALL user_data_output_2d( av, do2d(av,if), found, grid, &
1425	local_pf, two_d, nzb_do, nzt_do )
1426	ENDIF
1427
1428	resorted = .TRUE.
1429
1430	IF ( grid == 'zu' ) THEN
1431	IF ( mode == 'xy' ) level_z = zu
1432	ELSEIF ( grid == 'zw' ) THEN
1433	IF ( mode == 'xy' ) level_z = zw
1434	ELSEIF ( grid == 'zu1' ) THEN
1435	IF ( mode == 'xy' ) level_z(nzb+1) = zu(nzb+1)
1436	ELSEIF ( grid == 'zs' ) THEN
1437	IF ( mode == 'xy' ) level_z = zs
1438	ENDIF
1439
1440	IF ( .NOT. found ) THEN
1441	message_string = 'no output provided for: ' // &
1442	TRIM( do2d(av,if) )
1443	CALL message( 'data_output_2d', 'PA0181', 0, 0, 0, 6, 0 )
1444	ENDIF
1445
1446	END SELECT
1447
1448	!
1449	!-- Resort the array to be output, if not done above. Flag topography
1450	!-- grid points with fill values, using the corresponding maksing flag.
1451	IF ( .NOT. resorted ) THEN
1452	DO i = nxl, nxr
1453	DO j = nys, nyn
1454	DO k = nzb_do, nzt_do
1455	local_pf(i,j,k) = MERGE( to_be_resorted(k,j,i), &
1456	REAL( fill_value, KIND = wp ), &
1457	BTEST( wall_flags_0(k,j,i), &
1458	flag_nr ) )
1459	ENDDO
1460	ENDDO
1461	ENDDO
1462	ENDIF
1463
1464	!
1465	!-- Output of the individual cross-sections, depending on the cross-
1466	!-- section mode chosen.
1467	is = 1
1468	loop1: DO WHILE ( section(is,s_ind) /= -9999 .OR. two_d )
1469
1470	SELECT CASE ( mode )
1471
1472	CASE ( 'xy' )
1473	!
1474	!-- Determine the cross section index
1475	IF ( two_d ) THEN
1476	layer_xy = nzb+1
1477	ELSE
1478	layer_xy = section(is,s_ind)
1479	ENDIF
1480
1481	!
1482	!-- Exit the loop for layers beyond the data output domain
1483	!-- (used for soil model)
1484	IF ( layer_xy > nzt_do ) THEN
1485	EXIT loop1
1486	ENDIF
1487
1488	!
1489	!-- Update the netCDF xy cross section time axis.
1490	!-- In case of parallel output, this is only done by PE0
1491	!-- to increase the performance.
1492	IF ( simulated_time /= do2d_xy_last_time(av) ) THEN
1493	do2d_xy_time_count(av) = do2d_xy_time_count(av) + 1
1494	do2d_xy_last_time(av) = simulated_time
1495	IF ( myid == 0 ) THEN
1496	IF ( .NOT. data_output_2d_on_each_pe &
1497	.OR. netcdf_data_format > 4 ) &
1498	THEN
1499	#if defined( __netcdf )
1500	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1501	id_var_time_xy(av), &
1502	(/ time_since_reference_point /), &
1503	start = (/ do2d_xy_time_count(av) /), &
1504	count = (/ 1 /) )
1505	CALL netcdf_handle_error( 'data_output_2d', 53 )
1506	#endif
1507	ENDIF
1508	ENDIF
1509	ENDIF
1510	!
1511	!-- If required, carry out averaging along z
1512	IF ( section(is,s_ind) == -1 .AND. .NOT. two_d ) THEN
1513
1514	local_2d = 0.0_wp
1515	!
1516	!-- Carry out the averaging (all data are on the PE)
1517	DO k = nzb_do, nzt_do
1518	DO j = nys, nyn
1519	DO i = nxl, nxr
1520	local_2d(i,j) = local_2d(i,j) + local_pf(i,j,k)
1521	ENDDO
1522	ENDDO
1523	ENDDO
1524
1525	local_2d = local_2d / ( nzt_do - nzb_do + 1.0_wp)
1526
1527	ELSE
1528	!
1529	!-- Just store the respective section on the local array
1530	local_2d = local_pf(:,:,layer_xy)
1531
1532	ENDIF
1533
1534	#if defined( __parallel )
1535	IF ( netcdf_data_format > 4 ) THEN
1536	!
1537	!-- Parallel output in netCDF4/HDF5 format.
1538	IF ( two_d ) THEN
1539	iis = 1
1540	ELSE
1541	iis = is
1542	ENDIF
1543
1544	#if defined( __netcdf )
1545	!
1546	!-- For parallel output, all cross sections are first stored
1547	!-- here on a local array and will be written to the output
1548	!-- file afterwards to increase the performance.
1549	DO i = nxl, nxr
1550	DO j = nys, nyn
1551	local_2d_sections(i,j,iis) = local_2d(i,j)
1552	ENDDO
1553	ENDDO
1554	#endif
1555	ELSE
1556
1557	IF ( data_output_2d_on_each_pe ) THEN
1558	!
1559	!-- Output of partial arrays on each PE
1560	#if defined( __netcdf )
1561	IF ( myid == 0 ) THEN
1562	WRITE ( 21 ) time_since_reference_point, &
1563	do2d_xy_time_count(av), av
1564	ENDIF
1565	#endif
1566	DO i = 0, io_blocks-1
1567	IF ( i == io_group ) THEN
1568	WRITE ( 21 ) nxl, nxr, nys, nyn, nys, nyn
1569	WRITE ( 21 ) local_2d
1570	ENDIF
1571	#if defined( __parallel )
1572	CALL MPI_BARRIER( comm2d, ierr )
1573	#endif
1574	ENDDO
1575
1576	ELSE
1577	!
1578	!-- PE0 receives partial arrays from all processors and
1579	!-- then outputs them. Here a barrier has to be set,
1580	!-- because otherwise "-MPI- FATAL: Remote protocol queue
1581	!-- full" may occur.
1582	CALL MPI_BARRIER( comm2d, ierr )
1583
1584	ngp = ( nxr-nxl+1 ) * ( nyn-nys+1 )
1585	IF ( myid == 0 ) THEN
1586	!
1587	!-- Local array can be relocated directly.
1588	total_2d(nxl:nxr,nys:nyn) = local_2d
1589	!
1590	!-- Receive data from all other PEs.
1591	DO n = 1, numprocs-1
1592	!
1593	!-- Receive index limits first, then array.
1594	!-- Index limits are received in arbitrary order from
1595	!-- the PEs.
1596	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1597	MPI_ANY_SOURCE, 0, comm2d, &
1598	status, ierr )
1599	sender = status(MPI_SOURCE)
1600	DEALLOCATE( local_2d )
1601	ALLOCATE( local_2d(ind(1):ind(2),ind(3):ind(4)) )
1602	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1603	MPI_REAL, sender, 1, comm2d, &
1604	status, ierr )
1605	total_2d(ind(1):ind(2),ind(3):ind(4)) = local_2d
1606	ENDDO
1607	!
1608	!-- Relocate the local array for the next loop increment
1609	DEALLOCATE( local_2d )
1610	ALLOCATE( local_2d(nxl:nxr,nys:nyn) )
1611
1612	#if defined( __netcdf )
1613	IF ( two_d ) THEN
1614	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1615	id_var_do2d(av,if), &
1616	total_2d(0:nx,0:ny), &
1617	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
1618	count = (/ nx+1, ny+1, 1, 1 /) )
1619	ELSE
1620	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1621	id_var_do2d(av,if), &
1622	total_2d(0:nx,0:ny), &
1623	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
1624	count = (/ nx+1, ny+1, 1, 1 /) )
1625	ENDIF
1626	CALL netcdf_handle_error( 'data_output_2d', 54 )
1627	#endif
1628
1629	ELSE
1630	!
1631	!-- First send the local index limits to PE0
1632	ind(1) = nxl; ind(2) = nxr
1633	ind(3) = nys; ind(4) = nyn
1634	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1635	comm2d, ierr )
1636	!
1637	!-- Send data to PE0
1638	CALL MPI_SEND( local_2d(nxl,nys), ngp, &
1639	MPI_REAL, 0, 1, comm2d, ierr )
1640	ENDIF
1641	!
1642	!-- A barrier has to be set, because otherwise some PEs may
1643	!-- proceed too fast so that PE0 may receive wrong data on
1644	!-- tag 0
1645	CALL MPI_BARRIER( comm2d, ierr )
1646	ENDIF
1647
1648	ENDIF
1649	#else
1650	#if defined( __netcdf )
1651	IF ( two_d ) THEN
1652	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1653	id_var_do2d(av,if), &
1654	local_2d(nxl:nxr,nys:nyn), &
1655	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
1656	count = (/ nx+1, ny+1, 1, 1 /) )
1657	ELSE
1658	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1659	id_var_do2d(av,if), &
1660	local_2d(nxl:nxr,nys:nyn), &
1661	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
1662	count = (/ nx+1, ny+1, 1, 1 /) )
1663	ENDIF
1664	CALL netcdf_handle_error( 'data_output_2d', 447 )
1665	#endif
1666	#endif
1667
1668	!
1669	!-- For 2D-arrays (e.g. u*) only one cross-section is available.
1670	!-- Hence exit loop of output levels.
1671	IF ( two_d ) THEN
1672	IF ( netcdf_data_format < 5 ) two_d = .FALSE.
1673	EXIT loop1
1674	ENDIF
1675
1676	CASE ( 'xz' )
1677	!
1678	!-- Update the netCDF xz cross section time axis.
1679	!-- In case of parallel output, this is only done by PE0
1680	!-- to increase the performance.
1681	IF ( simulated_time /= do2d_xz_last_time(av) ) THEN
1682	do2d_xz_time_count(av) = do2d_xz_time_count(av) + 1
1683	do2d_xz_last_time(av) = simulated_time
1684	IF ( myid == 0 ) THEN
1685	IF ( .NOT. data_output_2d_on_each_pe &
1686	.OR. netcdf_data_format > 4 ) &
1687	THEN
1688	#if defined( __netcdf )
1689	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1690	id_var_time_xz(av), &
1691	(/ time_since_reference_point /), &
1692	start = (/ do2d_xz_time_count(av) /), &
1693	count = (/ 1 /) )
1694	CALL netcdf_handle_error( 'data_output_2d', 56 )
1695	#endif
1696	ENDIF
1697	ENDIF
1698	ENDIF
1699
1700	!
1701	!-- If required, carry out averaging along y
1702	IF ( section(is,s_ind) == -1 ) THEN
1703
1704	ALLOCATE( local_2d_l(nxl:nxr,nzb_do:nzt_do) )
1705	local_2d_l = 0.0_wp
1706	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
1707	!
1708	!-- First local averaging on the PE
1709	DO k = nzb_do, nzt_do
1710	DO j = nys, nyn
1711	DO i = nxl, nxr
1712	local_2d_l(i,k) = local_2d_l(i,k) + &
1713	local_pf(i,j,k)
1714	ENDDO
1715	ENDDO
1716	ENDDO
1717	#if defined( __parallel )
1718	!
1719	!-- Now do the averaging over all PEs along y
1720	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1721	CALL MPI_ALLREDUCE( local_2d_l(nxl,nzb_do), &
1722	local_2d(nxl,nzb_do), ngp, MPI_REAL, &
1723	MPI_SUM, comm1dy, ierr )
1724	#else
1725	local_2d = local_2d_l
1726	#endif
1727	local_2d = local_2d / ( ny + 1.0_wp )
1728
1729	DEALLOCATE( local_2d_l )
1730
1731	ELSE
1732	!
1733	!-- Just store the respective section on the local array
1734	!-- (but only if it is available on this PE!)
1735	IF ( section(is,s_ind) >= nys .AND. section(is,s_ind) <= nyn ) &
1736	THEN
1737	local_2d = local_pf(:,section(is,s_ind),nzb_do:nzt_do)
1738	ENDIF
1739
1740	ENDIF
1741
1742	#if defined( __parallel )
1743	IF ( netcdf_data_format > 4 ) THEN
1744	!
1745	!-- Output in netCDF4/HDF5 format.
1746	!-- Output only on those PEs where the respective cross
1747	!-- sections reside. Cross sections averaged along y are
1748	!-- output on the respective first PE along y (myidy=0).
1749	IF ( ( section(is,s_ind) >= nys .AND. &
1750	section(is,s_ind) <= nyn ) .OR. &
1751	( section(is,s_ind) == -1 .AND. myidy == 0 ) ) THEN
1752	#if defined( __netcdf )
1753	!
1754	!-- For parallel output, all cross sections are first
1755	!-- stored here on a local array and will be written to the
1756	!-- output file afterwards to increase the performance.
1757	DO i = nxl, nxr
1758	DO k = nzb_do, nzt_do
1759	local_2d_sections_l(i,is,k) = local_2d(i,k)
1760	ENDDO
1761	ENDDO
1762	#endif
1763	ENDIF
1764
1765	ELSE
1766
1767	IF ( data_output_2d_on_each_pe ) THEN
1768	!
1769	!-- Output of partial arrays on each PE. If the cross
1770	!-- section does not reside on the PE, output special
1771	!-- index values.
1772	#if defined( __netcdf )
1773	IF ( myid == 0 ) THEN
1774	WRITE ( 22 ) time_since_reference_point, &
1775	do2d_xz_time_count(av), av
1776	ENDIF
1777	#endif
1778	DO i = 0, io_blocks-1
1779	IF ( i == io_group ) THEN
1780	IF ( ( section(is,s_ind) >= nys .AND. &
1781	section(is,s_ind) <= nyn ) .OR. &
1782	( section(is,s_ind) == -1 .AND. &
1783	nys-1 == -1 ) ) &
1784	THEN
1785	WRITE (22) nxl, nxr, nzb_do, nzt_do, nzb, nzt+1
1786	WRITE (22) local_2d
1787	ELSE
1788	WRITE (22) -1, -1, -1, -1, -1, -1
1789	ENDIF
1790	ENDIF
1791	#if defined( __parallel )
1792	CALL MPI_BARRIER( comm2d, ierr )
1793	#endif
1794	ENDDO
1795
1796	ELSE
1797	!
1798	!-- PE0 receives partial arrays from all processors of the
1799	!-- respective cross section and outputs them. Here a
1800	!-- barrier has to be set, because otherwise
1801	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
1802	CALL MPI_BARRIER( comm2d, ierr )
1803
1804	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
1805	IF ( myid == 0 ) THEN
1806	!
1807	!-- Local array can be relocated directly.
1808	IF ( ( section(is,s_ind) >= nys .AND. &
1809	section(is,s_ind) <= nyn ) .OR. &
1810	( section(is,s_ind) == -1 .AND. &
1811	nys-1 == -1 ) ) THEN
1812	total_2d(nxl:nxr,nzb_do:nzt_do) = local_2d
1813	ENDIF
1814	!
1815	!-- Receive data from all other PEs.
1816	DO n = 1, numprocs-1
1817	!
1818	!-- Receive index limits first, then array.
1819	!-- Index limits are received in arbitrary order from
1820	!-- the PEs.
1821	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1822	MPI_ANY_SOURCE, 0, comm2d, &
1823	status, ierr )
1824	!
1825	!-- Not all PEs have data for XZ-cross-section.
1826	IF ( ind(1) /= -9999 ) THEN
1827	sender = status(MPI_SOURCE)
1828	DEALLOCATE( local_2d )
1829	ALLOCATE( local_2d(ind(1):ind(2), &
1830	ind(3):ind(4)) )
1831	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1832	MPI_REAL, sender, 1, comm2d, &
1833	status, ierr )
1834	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
1835	local_2d
1836	ENDIF
1837	ENDDO
1838	!
1839	!-- Relocate the local array for the next loop increment
1840	DEALLOCATE( local_2d )
1841	ALLOCATE( local_2d(nxl:nxr,nzb_do:nzt_do) )
1842
1843	#if defined( __netcdf )
1844	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1845	id_var_do2d(av,if), &
1846	total_2d(0:nx,nzb_do:nzt_do), &
1847	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
1848	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
1849	CALL netcdf_handle_error( 'data_output_2d', 58 )
1850	#endif
1851
1852	ELSE
1853	!
1854	!-- If the cross section resides on the PE, send the
1855	!-- local index limits, otherwise send -9999 to PE0.
1856	IF ( ( section(is,s_ind) >= nys .AND. &
1857	section(is,s_ind) <= nyn ) .OR. &
1858	( section(is,s_ind) == -1 .AND. nys-1 == -1 ) ) &
1859	THEN
1860	ind(1) = nxl; ind(2) = nxr
1861	ind(3) = nzb_do; ind(4) = nzt_do
1862	ELSE
1863	ind(1) = -9999; ind(2) = -9999
1864	ind(3) = -9999; ind(4) = -9999
1865	ENDIF
1866	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1867	comm2d, ierr )
1868	!
1869	!-- If applicable, send data to PE0.
1870	IF ( ind(1) /= -9999 ) THEN
1871	CALL MPI_SEND( local_2d(nxl,nzb_do), ngp, &
1872	MPI_REAL, 0, 1, comm2d, ierr )
1873	ENDIF
1874	ENDIF
1875	!
1876	!-- A barrier has to be set, because otherwise some PEs may
1877	!-- proceed too fast so that PE0 may receive wrong data on
1878	!-- tag 0
1879	CALL MPI_BARRIER( comm2d, ierr )
1880	ENDIF
1881
1882	ENDIF
1883	#else
1884	#if defined( __netcdf )
1885	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1886	id_var_do2d(av,if), &
1887	local_2d(nxl:nxr,nzb_do:nzt_do), &
1888	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
1889	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
1890	CALL netcdf_handle_error( 'data_output_2d', 451 )
1891	#endif
1892	#endif
1893
1894	CASE ( 'yz' )
1895	!
1896	!-- Update the netCDF yz cross section time axis.
1897	!-- In case of parallel output, this is only done by PE0
1898	!-- to increase the performance.
1899	IF ( simulated_time /= do2d_yz_last_time(av) ) THEN
1900	do2d_yz_time_count(av) = do2d_yz_time_count(av) + 1
1901	do2d_yz_last_time(av) = simulated_time
1902	IF ( myid == 0 ) THEN
1903	IF ( .NOT. data_output_2d_on_each_pe &
1904	.OR. netcdf_data_format > 4 ) &
1905	THEN
1906	#if defined( __netcdf )
1907	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
1908	id_var_time_yz(av), &
1909	(/ time_since_reference_point /), &
1910	start = (/ do2d_yz_time_count(av) /), &
1911	count = (/ 1 /) )
1912	CALL netcdf_handle_error( 'data_output_2d', 59 )
1913	#endif
1914	ENDIF
1915	ENDIF
1916	ENDIF
1917
1918	!
1919	!-- If required, carry out averaging along x
1920	IF ( section(is,s_ind) == -1 ) THEN
1921
1922	ALLOCATE( local_2d_l(nys:nyn,nzb_do:nzt_do) )
1923	local_2d_l = 0.0_wp
1924	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
1925	!
1926	!-- First local averaging on the PE
1927	DO k = nzb_do, nzt_do
1928	DO j = nys, nyn
1929	DO i = nxl, nxr
1930	local_2d_l(j,k) = local_2d_l(j,k) + &
1931	local_pf(i,j,k)
1932	ENDDO
1933	ENDDO
1934	ENDDO
1935	#if defined( __parallel )
1936	!
1937	!-- Now do the averaging over all PEs along x
1938	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1939	CALL MPI_ALLREDUCE( local_2d_l(nys,nzb_do), &
1940	local_2d(nys,nzb_do), ngp, MPI_REAL, &
1941	MPI_SUM, comm1dx, ierr )
1942	#else
1943	local_2d = local_2d_l
1944	#endif
1945	local_2d = local_2d / ( nx + 1.0_wp )
1946
1947	DEALLOCATE( local_2d_l )
1948
1949	ELSE
1950	!
1951	!-- Just store the respective section on the local array
1952	!-- (but only if it is available on this PE!)
1953	IF ( section(is,s_ind) >= nxl .AND. section(is,s_ind) <= nxr ) &
1954	THEN
1955	local_2d = local_pf(section(is,s_ind),:,nzb_do:nzt_do)
1956	ENDIF
1957
1958	ENDIF
1959
1960	#if defined( __parallel )
1961	IF ( netcdf_data_format > 4 ) THEN
1962	!
1963	!-- Output in netCDF4/HDF5 format.
1964	!-- Output only on those PEs where the respective cross
1965	!-- sections reside. Cross sections averaged along x are
1966	!-- output on the respective first PE along x (myidx=0).
1967	IF ( ( section(is,s_ind) >= nxl .AND. &
1968	section(is,s_ind) <= nxr ) .OR. &
1969	( section(is,s_ind) == -1 .AND. myidx == 0 ) ) THEN
1970	#if defined( __netcdf )
1971	!
1972	!-- For parallel output, all cross sections are first
1973	!-- stored here on a local array and will be written to the
1974	!-- output file afterwards to increase the performance.
1975	DO j = nys, nyn
1976	DO k = nzb_do, nzt_do
1977	local_2d_sections_l(is,j,k) = local_2d(j,k)
1978	ENDDO
1979	ENDDO
1980	#endif
1981	ENDIF
1982
1983	ELSE
1984
1985	IF ( data_output_2d_on_each_pe ) THEN
1986	!
1987	!-- Output of partial arrays on each PE. If the cross
1988	!-- section does not reside on the PE, output special
1989	!-- index values.
1990	#if defined( __netcdf )
1991	IF ( myid == 0 ) THEN
1992	WRITE ( 23 ) time_since_reference_point, &
1993	do2d_yz_time_count(av), av
1994	ENDIF
1995	#endif
1996	DO i = 0, io_blocks-1
1997	IF ( i == io_group ) THEN
1998	IF ( ( section(is,s_ind) >= nxl .AND. &
1999	section(is,s_ind) <= nxr ) .OR. &
2000	( section(is,s_ind) == -1 .AND. &
2001	nxl-1 == -1 ) ) &
2002	THEN
2003	WRITE (23) nys, nyn, nzb_do, nzt_do, nzb, nzt+1
2004	WRITE (23) local_2d
2005	ELSE
2006	WRITE (23) -1, -1, -1, -1, -1, -1
2007	ENDIF
2008	ENDIF
2009	#if defined( __parallel )
2010	CALL MPI_BARRIER( comm2d, ierr )
2011	#endif
2012	ENDDO
2013
2014	ELSE
2015	!
2016	!-- PE0 receives partial arrays from all processors of the
2017	!-- respective cross section and outputs them. Here a
2018	!-- barrier has to be set, because otherwise
2019	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
2020	CALL MPI_BARRIER( comm2d, ierr )
2021
2022	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
2023	IF ( myid == 0 ) THEN
2024	!
2025	!-- Local array can be relocated directly.
2026	IF ( ( section(is,s_ind) >= nxl .AND. &
2027	section(is,s_ind) <= nxr ) .OR. &
2028	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
2029	THEN
2030	total_2d(nys:nyn,nzb_do:nzt_do) = local_2d
2031	ENDIF
2032	!
2033	!-- Receive data from all other PEs.
2034	DO n = 1, numprocs-1
2035	!
2036	!-- Receive index limits first, then array.
2037	!-- Index limits are received in arbitrary order from
2038	!-- the PEs.
2039	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
2040	MPI_ANY_SOURCE, 0, comm2d, &
2041	status, ierr )
2042	!
2043	!-- Not all PEs have data for YZ-cross-section.
2044	IF ( ind(1) /= -9999 ) THEN
2045	sender = status(MPI_SOURCE)
2046	DEALLOCATE( local_2d )
2047	ALLOCATE( local_2d(ind(1):ind(2), &
2048	ind(3):ind(4)) )
2049	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
2050	MPI_REAL, sender, 1, comm2d, &
2051	status, ierr )
2052	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
2053	local_2d
2054	ENDIF
2055	ENDDO
2056	!
2057	!-- Relocate the local array for the next loop increment
2058	DEALLOCATE( local_2d )
2059	ALLOCATE( local_2d(nys:nyn,nzb_do:nzt_do) )
2060
2061	#if defined( __netcdf )
2062	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2063	id_var_do2d(av,if), &
2064	total_2d(0:ny,nzb_do:nzt_do), &
2065	start = (/ is, 1, 1, do2d_yz_time_count(av) /), &
2066	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
2067	CALL netcdf_handle_error( 'data_output_2d', 61 )
2068	#endif
2069
2070	ELSE
2071	!
2072	!-- If the cross section resides on the PE, send the
2073	!-- local index limits, otherwise send -9999 to PE0.
2074	IF ( ( section(is,s_ind) >= nxl .AND. &
2075	section(is,s_ind) <= nxr ) .OR. &
2076	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
2077	THEN
2078	ind(1) = nys; ind(2) = nyn
2079	ind(3) = nzb_do; ind(4) = nzt_do
2080	ELSE
2081	ind(1) = -9999; ind(2) = -9999
2082	ind(3) = -9999; ind(4) = -9999
2083	ENDIF
2084	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
2085	comm2d, ierr )
2086	!
2087	!-- If applicable, send data to PE0.
2088	IF ( ind(1) /= -9999 ) THEN
2089	CALL MPI_SEND( local_2d(nys,nzb_do), ngp, &
2090	MPI_REAL, 0, 1, comm2d, ierr )
2091	ENDIF
2092	ENDIF
2093	!
2094	!-- A barrier has to be set, because otherwise some PEs may
2095	!-- proceed too fast so that PE0 may receive wrong data on
2096	!-- tag 0
2097	CALL MPI_BARRIER( comm2d, ierr )
2098	ENDIF
2099
2100	ENDIF
2101	#else
2102	#if defined( __netcdf )
2103	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2104	id_var_do2d(av,if), &
2105	local_2d(nys:nyn,nzb_do:nzt_do), &
2106	start = (/ is, 1, 1, do2d_xz_time_count(av) /), &
2107	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
2108	CALL netcdf_handle_error( 'data_output_2d', 452 )
2109	#endif
2110	#endif
2111
2112	END SELECT
2113
2114	is = is + 1
2115	ENDDO loop1
2116
2117	!
2118	!-- For parallel output, all data were collected before on a local array
2119	!-- and are written now to the netcdf file. This must be done to increase
2120	!-- the performance of the parallel output.
2121	#if defined( __netcdf )
2122	IF ( netcdf_data_format > 4 ) THEN
2123
2124	SELECT CASE ( mode )
2125
2126	CASE ( 'xy' )
2127	IF ( two_d ) THEN
2128	nis = 1
2129	two_d = .FALSE.
2130	ELSE
2131	nis = ns
2132	ENDIF
2133	!
2134	!-- Do not output redundant ghost point data except for the
2135	!-- boundaries of the total domain.
2136	! IF ( nxr == nx .AND. nyn /= ny ) THEN
2137	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2138	! id_var_do2d(av,if), &
2139	! local_2d_sections(nxl:nxr+1, &
2140	! nys:nyn,1:nis), &
2141	! start = (/ nxl+1, nys+1, 1, &
2142	! do2d_xy_time_count(av) /), &
2143	! count = (/ nxr-nxl+2, &
2144	! nyn-nys+1, nis, 1 &
2145	! /) )
2146	! ELSEIF ( nxr /= nx .AND. nyn == ny ) THEN
2147	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2148	! id_var_do2d(av,if), &
2149	! local_2d_sections(nxl:nxr, &
2150	! nys:nyn+1,1:nis), &
2151	! start = (/ nxl+1, nys+1, 1, &
2152	! do2d_xy_time_count(av) /), &
2153	! count = (/ nxr-nxl+1, &
2154	! nyn-nys+2, nis, 1 &
2155	! /) )
2156	! ELSEIF ( nxr == nx .AND. nyn == ny ) THEN
2157	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2158	! id_var_do2d(av,if), &
2159	! local_2d_sections(nxl:nxr+1, &
2160	! nys:nyn+1,1:nis), &
2161	! start = (/ nxl+1, nys+1, 1, &
2162	! do2d_xy_time_count(av) /), &
2163	! count = (/ nxr-nxl+2, &
2164	! nyn-nys+2, nis, 1 &
2165	! /) )
2166	! ELSE
2167	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2168	id_var_do2d(av,if), &
2169	local_2d_sections(nxl:nxr, &
2170	nys:nyn,1:nis), &
2171	start = (/ nxl+1, nys+1, 1, &
2172	do2d_xy_time_count(av) /), &
2173	count = (/ nxr-nxl+1, &
2174	nyn-nys+1, nis, 1 &
2175	/) )
2176	! ENDIF
2177
2178	CALL netcdf_handle_error( 'data_output_2d', 55 )
2179
2180	CASE ( 'xz' )
2181	!
2182	!-- First, all PEs get the information of all cross-sections.
2183	!-- Then the data are written to the output file by all PEs
2184	!-- while NF90_COLLECTIVE is set in subroutine
2185	!-- define_netcdf_header. Although redundant information are
2186	!-- written to the output file in that case, the performance
2187	!-- is significantly better compared to the case where only
2188	!-- the first row of PEs in x-direction (myidx = 0) is given
2189	!-- the output while NF90_INDEPENDENT is set.
2190	IF ( npey /= 1 ) THEN
2191
2192	#if defined( __parallel )
2193	!
2194	!-- Distribute data over all PEs along y
2195	ngp = ( nxr-nxl+1 ) * ( nzt_do-nzb_do+1 ) * ns
2196	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
2197	CALL MPI_ALLREDUCE( local_2d_sections_l(nxl,1,nzb_do), &
2198	local_2d_sections(nxl,1,nzb_do), &
2199	ngp, MPI_REAL, MPI_SUM, comm1dy, &
2200	ierr )
2201	#else
2202	local_2d_sections = local_2d_sections_l
2203	#endif
2204	ENDIF
2205	!
2206	!-- Do not output redundant ghost point data except for the
2207	!-- boundaries of the total domain.
2208	! IF ( nxr == nx ) THEN
2209	! nc_stat = NF90_PUT_VAR( id_set_xz(av), &
2210	! id_var_do2d(av,if), &
2211	! local_2d_sections(nxl:nxr+1,1:ns, &
2212	! nzb_do:nzt_do), &
2213	! start = (/ nxl+1, 1, 1, &
2214	! do2d_xz_time_count(av) /), &
2215	! count = (/ nxr-nxl+2, ns, nzt_do-nzb_do+1, &
2216	! 1 /) )
2217	! ELSE
2218	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
2219	id_var_do2d(av,if), &
2220	local_2d_sections(nxl:nxr,1:ns, &
2221	nzb_do:nzt_do), &
2222	start = (/ nxl+1, 1, 1, &
2223	do2d_xz_time_count(av) /), &
2224	count = (/ nxr-nxl+1, ns, nzt_do-nzb_do+1, &
2225	1 /) )
2226	! ENDIF
2227
2228	CALL netcdf_handle_error( 'data_output_2d', 57 )
2229
2230	CASE ( 'yz' )
2231	!
2232	!-- First, all PEs get the information of all cross-sections.
2233	!-- Then the data are written to the output file by all PEs
2234	!-- while NF90_COLLECTIVE is set in subroutine
2235	!-- define_netcdf_header. Although redundant information are
2236	!-- written to the output file in that case, the performance
2237	!-- is significantly better compared to the case where only
2238	!-- the first row of PEs in y-direction (myidy = 0) is given
2239	!-- the output while NF90_INDEPENDENT is set.
2240	IF ( npex /= 1 ) THEN
2241
2242	#if defined( __parallel )
2243	!
2244	!-- Distribute data over all PEs along x
2245	ngp = ( nyn-nys+1 ) * ( nzt-nzb + 2 ) * ns
2246	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
2247	CALL MPI_ALLREDUCE( local_2d_sections_l(1,nys,nzb_do), &
2248	local_2d_sections(1,nys,nzb_do), &
2249	ngp, MPI_REAL, MPI_SUM, comm1dx, &
2250	ierr )
2251	#else
2252	local_2d_sections = local_2d_sections_l
2253	#endif
2254	ENDIF
2255	!
2256	!-- Do not output redundant ghost point data except for the
2257	!-- boundaries of the total domain.
2258	! IF ( nyn == ny ) THEN
2259	! nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2260	! id_var_do2d(av,if), &
2261	! local_2d_sections(1:ns, &
2262	! nys:nyn+1,nzb_do:nzt_do), &
2263	! start = (/ 1, nys+1, 1, &
2264	! do2d_yz_time_count(av) /), &
2265	! count = (/ ns, nyn-nys+2, &
2266	! nzt_do-nzb_do+1, 1 /) )
2267	! ELSE
2268	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2269	id_var_do2d(av,if), &
2270	local_2d_sections(1:ns,nys:nyn, &
2271	nzb_do:nzt_do), &
2272	start = (/ 1, nys+1, 1, &
2273	do2d_yz_time_count(av) /), &
2274	count = (/ ns, nyn-nys+1, &
2275	nzt_do-nzb_do+1, 1 /) )
2276	! ENDIF
2277
2278	CALL netcdf_handle_error( 'data_output_2d', 60 )
2279
2280	CASE DEFAULT
2281	message_string = 'unknown cross-section: ' // TRIM( mode )
2282	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
2283
2284	END SELECT
2285
2286	ENDIF
2287	#endif
2288	ENDIF
2289
2290	if = if + 1
2291	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
2292	do2d_mode = do2d(av,if)(l-1:l)
2293
2294	ENDDO
2295
2296	!
2297	!-- Deallocate temporary arrays.
2298	IF ( ALLOCATED( level_z ) ) DEALLOCATE( level_z )
2299	IF ( netcdf_data_format > 4 ) THEN
2300	DEALLOCATE( local_pf, local_2d, local_2d_sections )
2301	IF( mode == 'xz' .OR. mode == 'yz' ) DEALLOCATE( local_2d_sections_l )
2302	ENDIF
2303	#if defined( __parallel )
2304	IF ( .NOT. data_output_2d_on_each_pe .AND. myid == 0 ) THEN
2305	DEALLOCATE( total_2d )
2306	ENDIF
2307	#endif
2308
2309	!
2310	!-- Close plot output file.
2311	file_id = 20 + s_ind
2312
2313	IF ( data_output_2d_on_each_pe ) THEN
2314	DO i = 0, io_blocks-1
2315	IF ( i == io_group ) THEN
2316	CALL close_file( file_id )
2317	ENDIF
2318	#if defined( __parallel )
2319	CALL MPI_BARRIER( comm2d, ierr )
2320	#endif
2321	ENDDO
2322	ELSE
2323	IF ( myid == 0 ) CALL close_file( file_id )
2324	ENDIF
2325
2326	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
2327
2328	END SUBROUTINE data_output_2d

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |