Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 2513

Last change on this file since 2513 was 2512, checked in by raasch, 7 years ago
upper bounds of cross section and 3d output changed from nx+1,ny+1 to nx,ny; no output if redundant ghost layer data to NetCDF files
Property svn:keywords set to `Id`
File size: 82.5 KB

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

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |