Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 3377

Last change on this file since 3377 was 3298, checked in by kanani, 6 years ago
Merge chemistry branch at r3297 to trunk
Property svn:keywords set to `Id`
File size: 94.7 KB

Line
1	!> @file data_output_2d.f90
2	!------------------------------------------------------------------------------!
3	! This file is part of the PALM model system.
4	!
5	! PALM is free software: you can redistribute it and/or modify it under the
6	! terms of the GNU General Public License as published by the Free Software
7	! Foundation, either version 3 of the License, or (at your option) any later
8	! version.
9	!
10	! PALM is distributed in the hope that it will be useful, but WITHOUT ANY
11	! WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
12	! A PARTICULAR PURPOSE. See the GNU General Public License for more details.
13	!
14	! You should have received a copy of the GNU General Public License along with
15	! PALM. If not, see <http://www.gnu.org/licenses/>.
16	!
17	! Copyright 1997-2018 Leibniz Universitaet Hannover
18	!------------------------------------------------------------------------------!
19	!
20	! Current revisions:
21	! -----------------
22	!
23	!
24	! Former revisions:
25	! -----------------
26	! $Id: data_output_2d.f90 3298 2018-10-02 12:21:11Z knoop $
27	! minor formatting (kanani)
28	! chem_data_output_2d subroutine added (basit)
29	!
30	! 3294 2018-10-01 02:37:10Z raasch
31	! changes concerning modularization of ocean option
32	!
33	! 3274 2018-09-24 15:42:55Z knoop
34	! Modularization of all bulk cloud physics code components
35	!
36	! 3241 2018-09-12 15:02:00Z raasch
37	! unused variables removed
38	!
39	! 3176 2018-07-26 17:12:48Z suehring
40	! Remove output of latent heat flux at urban-surfaces and set fill values
41	! instead
42	!
43	! 3052 2018-05-31 06:11:20Z raasch
44	! Do not open FORTRAN binary files in case of parallel netCDF I/O
45	!
46	! 3049 2018-05-29 13:52:36Z Giersch
47	! Error messages revised
48	!
49	! 3045 2018-05-28 07:55:41Z Giersch
50	! Error messages revised
51	!
52	! 3014 2018-05-09 08:42:38Z maronga
53	! Added nzb_do and nzt_do for some modules for 2d output
54	!
55	! 3004 2018-04-27 12:33:25Z Giersch
56	! precipitation_rate removed, case prr*_xy removed, to_be_resorted have to point
57	! to ql_vp_av and not to ql_vp, allocation checks implemented (averaged data
58	! will be assigned to fill values if no allocation happened so far)
59	!
60	! 2963 2018-04-12 14:47:44Z suehring
61	! Introduce index for vegetation/wall, pavement/green-wall and water/window
62	! surfaces, for clearer access of surface fraction, albedo, emissivity, etc. .
63	!
64	! 2817 2018-02-19 16:32:21Z knoop
65	! Preliminary gust module interface implemented
66	!
67	! 2805 2018-02-14 17:00:09Z suehring
68	! Consider also default-type surfaces for surface temperature output.
69	!
70	! 2797 2018-02-08 13:24:35Z suehring
71	! Enable output of ground-heat flux also at urban surfaces.
72	!
73	! 2743 2018-01-12 16:03:39Z suehring
74	! In case of natural- and urban-type surfaces output surfaces fluxes in W/m2.
75	!
76	! 2742 2018-01-12 14:59:47Z suehring
77	! Enable output of surface temperature
78	!
79	! 2735 2018-01-11 12:01:27Z suehring
80	! output of r_a moved from land-surface to consider also urban-type surfaces
81	!
82	! 2718 2018-01-02 08:49:38Z maronga
83	! Corrected "Former revisions" section
84	!
85	! 2696 2017-12-14 17:12:51Z kanani
86	! Change in file header (GPL part)
87	! Implementation of uv exposure model (FK)
88	! Implementation of turbulence_closure_mod (TG)
89	! Set fill values at topography grid points or e.g. non-natural-type surface
90	! in case of LSM output (MS)
91	!
92	! 2512 2017-10-04 08:26:59Z raasch
93	! upper bounds of cross section output changed from nx+1,ny+1 to nx,ny
94	! no output of ghost layer data
95	!
96	! 2292 2017-06-20 09:51:42Z schwenkel
97	! Implementation of new microphysic scheme: cloud_scheme = 'morrison'
98	! includes two more prognostic equations for cloud drop concentration (nc)
99	! and cloud water content (qc).
100	!
101	! 2277 2017-06-12 10:47:51Z kanani
102	! Removed unused variables do2d_xy_n, do2d_xz_n, do2d_yz_n
103	!
104	! 2233 2017-05-30 18:08:54Z suehring
105	!
106	! 2232 2017-05-30 17:47:52Z suehring
107	! Adjustments to new surface concept
108	!
109	!
110	! 2190 2017-03-21 12:16:43Z raasch
111	! bugfix for r2031: string rho replaced by rho_ocean
112	!
113	! 2031 2016-10-21 15:11:58Z knoop
114	! renamed variable rho to rho_ocean and rho_av to rho_ocean_av
115	!
116	! 2000 2016-08-20 18:09:15Z knoop
117	! Forced header and separation lines into 80 columns
118	!
119	! 1980 2016-07-29 15:51:57Z suehring
120	! Bugfix, in order to steer user-defined output, setting flag found explicitly
121	! to .F.
122	!
123	! 1976 2016-07-27 13:28:04Z maronga
124	! Output of radiation quantities is now done directly in the respective module
125	!
126	! 1972 2016-07-26 07:52:02Z maronga
127	! Output of land surface quantities is now done directly in the respective
128	! module
129	!
130	! 1960 2016-07-12 16:34:24Z suehring
131	! Scalar surface flux added
132	! Rename INTEGER variable s into s_ind, as s is already assigned to scalar
133	!
134	! 1849 2016-04-08 11:33:18Z hoffmann
135	! precipitation_amount, precipitation_rate, prr moved to arrays_3d
136	!
137	! 1822 2016-04-07 07:49:42Z hoffmann
138	! Output of bulk cloud physics simplified.
139	!
140	! 1788 2016-03-10 11:01:04Z maronga
141	! Added output of z0q
142	!
143	! 1783 2016-03-06 18:36:17Z raasch
144	! name change of netcdf routines and module + related changes
145	!
146	! 1745 2016-02-05 13:06:51Z gronemeier
147	! Bugfix: test if time axis limit exceeds moved to point after call of check_open
148	!
149	! 1703 2015-11-02 12:38:44Z raasch
150	! bugfix for output of single (*) xy-sections in case of parallel netcdf I/O
151	!
152	! 1701 2015-11-02 07:43:04Z maronga
153	! Bugfix in output of RRTGM data
154	!
155	! 1691 2015-10-26 16:17:44Z maronga
156	! Added output of Obukhov length (ol) and radiative heating rates for RRTMG.
157	! Formatting corrections.
158	!
159	! 1682 2015-10-07 23:56:08Z knoop
160	! Code annotations made doxygen readable
161	!
162	! 1585 2015-04-30 07:05:52Z maronga
163	! Added support for RRTMG
164	!
165	! 1555 2015-03-04 17:44:27Z maronga
166	! Added output of r_a and r_s
167	!
168	! 1551 2015-03-03 14:18:16Z maronga
169	! Added suppport for land surface model and radiation model output. In the course
170	! of this action, the limits for vertical loops have been changed (from nzb and
171	! nzt+1 to nzb_do and nzt_do, respectively in order to allow soil model output).
172	! Moreover, a new vertical grid zs was introduced.
173	!
174	! 1359 2014-04-11 17:15:14Z hoffmann
175	! New particle structure integrated.
176	!
177	! 1353 2014-04-08 15:21:23Z heinze
178	! REAL constants provided with KIND-attribute
179	!
180	! 1327 2014-03-21 11:00:16Z raasch
181	! parts concerning iso2d output removed,
182	! -netcdf output queries
183	!
184	! 1320 2014-03-20 08:40:49Z raasch
185	! ONLY-attribute added to USE-statements,
186	! kind-parameters added to all INTEGER and REAL declaration statements,
187	! kinds are defined in new module kinds,
188	! revision history before 2012 removed,
189	! comment fields (!:) to be used for variable explanations added to
190	! all variable declaration statements
191	!
192	! 1318 2014-03-17 13:35:16Z raasch
193	! barrier argument removed from cpu_log.
194	! module interfaces removed
195	!
196	! 1311 2014-03-14 12:13:39Z heinze
197	! bugfix: close #if defined( __netcdf )
198	!
199	! 1308 2014-03-13 14:58:42Z fricke
200	! +local_2d_sections, local_2d_sections_l, ns
201	! Check, if the limit of the time dimension is exceeded for parallel output
202	! To increase the performance for parallel output, the following is done:
203	! - Update of time axis is only done by PE0
204	! - Cross sections are first stored on a local array and are written
205	! collectively to the output file by all PEs.
206	!
207	! 1115 2013-03-26 18:16:16Z hoffmann
208	! ql is calculated by calc_liquid_water_content
209	!
210	! 1076 2012-12-05 08:30:18Z hoffmann
211	! Bugfix in output of ql
212	!
213	! 1065 2012-11-22 17:42:36Z hoffmann
214	! Bugfix: Output of cross sections of ql
215	!
216	! 1053 2012-11-13 17:11:03Z hoffmann
217	! +qr, nr, qc and cross sections
218	!
219	! 1036 2012-10-22 13:43:42Z raasch
220	! code put under GPL (PALM 3.9)
221	!
222	! 1031 2012-10-19 14:35:30Z raasch
223	! netCDF4 without parallel file support implemented
224	!
225	! 1007 2012-09-19 14:30:36Z franke
226	! Bugfix: missing calculation of ql_vp added
227	!
228	! 978 2012-08-09 08:28:32Z fricke
229	! +z0h
230	!
231	! Revision 1.1 1997/08/11 06:24:09 raasch
232	! Initial revision
233	!
234	!
235	! Description:
236	! ------------
237	!> Data output of cross-sections in netCDF format or binary format
238	!> to be later converted to NetCDF by helper routine combine_plot_fields.
239	!> Attention: The position of the sectional planes is still not always computed
240	!> --------- correctly. (zu is used always)!
241	!------------------------------------------------------------------------------!
242	SUBROUTINE data_output_2d( mode, av )
243
244
245	USE arrays_3d, &
246	ONLY: dzw, e, heatflux_output_conversion, nc, nr, p, pt, &
247	precipitation_amount, prr, q, qc, ql, ql_c, ql_v, qr, s, tend, &
248	u, v, vpt, w, zu, zw, waterflux_output_conversion, hyrho, d_exner
249
250	USE averaging
251
252	USE basic_constants_and_equations_mod, &
253	ONLY: c_p, lv_d_cp, l_v
254
255	USE bulk_cloud_model_mod, &
256	ONLY: bulk_cloud_model, bcm_data_output_2d
257
258	USE chemistry_model_mod, &
259	ONLY: chem_data_output_2d
260
261	USE control_parameters, &
262	ONLY: air_chemistry, data_output_2d_on_each_pe, data_output_xy, &
263	data_output_xz, data_output_yz, do2d, &
264	do2d_xy_last_time, do2d_xy_time_count, &
265	do2d_xz_last_time, do2d_xz_time_count, &
266	do2d_yz_last_time, do2d_yz_time_count, &
267	ibc_uv_b, io_blocks, io_group, land_surface, message_string, &
268	ntdim_2d_xy, ntdim_2d_xz, ntdim_2d_yz, &
269	ocean_mode, psolver, section, simulated_time, &
270	time_since_reference_point, uv_exposure
271
272	USE cpulog, &
273	ONLY: cpu_log, log_point
274
275	USE gust_mod, &
276	ONLY: gust_data_output_2d, gust_module_enabled
277
278	USE indices, &
279	ONLY: nbgp, nx, nxl, nxlg, nxr, nxrg, ny, nyn, nyng, nys, nysg, &
280	nzb, nzt, wall_flags_0
281
282	USE kinds
283
284	USE land_surface_model_mod, &
285	ONLY: lsm_data_output_2d, zs
286
287	#if defined( __netcdf )
288	USE NETCDF
289	#endif
290
291	USE netcdf_interface, &
292	ONLY: fill_value, id_set_xy, id_set_xz, id_set_yz, id_var_do2d, &
293	id_var_time_xy, id_var_time_xz, id_var_time_yz, nc_stat, &
294	netcdf_data_format, netcdf_handle_error
295
296	USE ocean_mod, &
297	ONLY: ocean_data_output_2d
298
299	USE particle_attributes, &
300	ONLY: grid_particles, number_of_particles, particle_advection_start, &
301	particles, prt_count
302
303	USE pegrid
304
305	USE radiation_model_mod, &
306	ONLY: radiation, radiation_data_output_2d
307
308	USE surface_mod, &
309	ONLY: ind_pav_green, ind_veg_wall, ind_wat_win, surf_def_h, &
310	surf_lsm_h, surf_usm_h
311
312	USE turbulence_closure_mod, &
313	ONLY: tcm_data_output_2d
314
315	USE uv_exposure_model_mod, &
316	ONLY: uvem_data_output_2d
317
318
319	IMPLICIT NONE
320
321	CHARACTER (LEN=2) :: do2d_mode !<
322	CHARACTER (LEN=2) :: mode !<
323	CHARACTER (LEN=4) :: grid !<
324	CHARACTER (LEN=50) :: rtext !<
325
326	INTEGER(iwp) :: av !<
327	INTEGER(iwp) :: ngp !<
328	INTEGER(iwp) :: file_id !<
329	INTEGER(iwp) :: flag_nr !< number of masking flag
330	INTEGER(iwp) :: i !<
331	INTEGER(iwp) :: if !<
332	INTEGER(iwp) :: is !<
333	INTEGER(iwp) :: iis !<
334	INTEGER(iwp) :: j !<
335	INTEGER(iwp) :: k !<
336	INTEGER(iwp) :: l !<
337	INTEGER(iwp) :: layer_xy !<
338	INTEGER(iwp) :: m !<
339	INTEGER(iwp) :: n !<
340	INTEGER(iwp) :: nis !<
341	INTEGER(iwp) :: ns !<
342	INTEGER(iwp) :: nzb_do !< lower limit of the data field (usually nzb)
343	INTEGER(iwp) :: nzt_do !< upper limit of the data field (usually nzt+1)
344	INTEGER(iwp) :: s_ind !<
345	INTEGER(iwp) :: sender !<
346	INTEGER(iwp) :: ind(4) !<
347
348	LOGICAL :: found !<
349	LOGICAL :: resorted !<
350	LOGICAL :: two_d !<
351
352	REAL(wp) :: mean_r !<
353	REAL(wp) :: s_r2 !<
354	REAL(wp) :: s_r3 !<
355
356	REAL(wp), DIMENSION(:), ALLOCATABLE :: level_z !<
357	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d !<
358	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d_l !<
359
360	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !<
361	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections !<
362	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections_l !<
363
364	#if defined( __parallel )
365	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: total_2d !<
366	#endif
367	REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !<
368
369	NAMELIST /LOCAL/ rtext
370
371	!
372	!-- Immediate return, if no output is requested (no respective sections
373	!-- found in parameter data_output)
374	IF ( mode == 'xy' .AND. .NOT. data_output_xy(av) ) RETURN
375	IF ( mode == 'xz' .AND. .NOT. data_output_xz(av) ) RETURN
376	IF ( mode == 'yz' .AND. .NOT. data_output_yz(av) ) RETURN
377
378	CALL cpu_log (log_point(3),'data_output_2d','start')
379
380	two_d = .FALSE. ! local variable to distinguish between output of pure 2D
381	! arrays and cross-sections of 3D arrays.
382
383	!
384	!-- Depending on the orientation of the cross-section, the respective output
385	!-- files have to be opened.
386	SELECT CASE ( mode )
387
388	CASE ( 'xy' )
389	s_ind = 1
390	ALLOCATE( level_z(nzb:nzt+1), local_2d(nxl:nxr,nys:nyn) )
391
392	IF ( netcdf_data_format > 4 ) THEN
393	ns = 1
394	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
395	ns = ns + 1
396	ENDDO
397	ns = ns - 1
398	ALLOCATE( local_2d_sections(nxl:nxr,nys:nyn,1:ns) )
399	local_2d_sections = 0.0_wp
400	ENDIF
401
402	!
403	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
404	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
405	CALL check_open( 101+av*10 )
406	ENDIF
407	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
408	CALL check_open( 21 )
409	ELSE
410	IF ( myid == 0 ) THEN
411	#if defined( __parallel )
412	ALLOCATE( total_2d(0:nx,0:ny) )
413	#endif
414	ENDIF
415	ENDIF
416
417	CASE ( 'xz' )
418	s_ind = 2
419	ALLOCATE( local_2d(nxl:nxr,nzb:nzt+1) )
420
421	IF ( netcdf_data_format > 4 ) THEN
422	ns = 1
423	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
424	ns = ns + 1
425	ENDDO
426	ns = ns - 1
427	ALLOCATE( local_2d_sections(nxl:nxr,1:ns,nzb:nzt+1) )
428	ALLOCATE( local_2d_sections_l(nxl:nxr,1:ns,nzb:nzt+1) )
429	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
430	ENDIF
431
432	!
433	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
434	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
435	CALL check_open( 102+av*10 )
436	ENDIF
437
438	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
439	CALL check_open( 22 )
440	ELSE
441	IF ( myid == 0 ) THEN
442	#if defined( __parallel )
443	ALLOCATE( total_2d(0:nx,nzb:nzt+1) )
444	#endif
445	ENDIF
446	ENDIF
447
448	CASE ( 'yz' )
449	s_ind = 3
450	ALLOCATE( local_2d(nys:nyn,nzb:nzt+1) )
451
452	IF ( netcdf_data_format > 4 ) THEN
453	ns = 1
454	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
455	ns = ns + 1
456	ENDDO
457	ns = ns - 1
458	ALLOCATE( local_2d_sections(1:ns,nys:nyn,nzb:nzt+1) )
459	ALLOCATE( local_2d_sections_l(1:ns,nys:nyn,nzb:nzt+1) )
460	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
461	ENDIF
462
463	!
464	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
465	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
466	CALL check_open( 103+av*10 )
467	ENDIF
468
469	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
470	CALL check_open( 23 )
471	ELSE
472	IF ( myid == 0 ) THEN
473	#if defined( __parallel )
474	ALLOCATE( total_2d(0:ny,nzb:nzt+1) )
475	#endif
476	ENDIF
477	ENDIF
478
479	CASE DEFAULT
480	message_string = 'unknown cross-section: ' // TRIM( mode )
481	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
482
483	END SELECT
484
485	!
486	!-- For parallel netcdf output the time axis must be limited. Return, if this
487	!-- limit is exceeded. This could be the case, if the simulated time exceeds
488	!-- the given end time by the length of the given output interval.
489	IF ( netcdf_data_format > 4 ) THEN
490	IF ( mode == 'xy' .AND. do2d_xy_time_count(av) + 1 > &
491	ntdim_2d_xy(av) ) THEN
492	WRITE ( message_string, * ) 'Output of xy cross-sections is not ', &
493	'given at t=', simulated_time, '&because the', &
494	' maximum number of output time levels is exceeded.'
495	CALL message( 'data_output_2d', 'PA0384', 0, 1, 0, 6, 0 )
496	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
497	RETURN
498	ENDIF
499	IF ( mode == 'xz' .AND. do2d_xz_time_count(av) + 1 > &
500	ntdim_2d_xz(av) ) THEN
501	WRITE ( message_string, * ) 'Output of xz cross-sections is not ', &
502	'given at t=', simulated_time, '&because the', &
503	' maximum number of output time levels is exceeded.'
504	CALL message( 'data_output_2d', 'PA0385', 0, 1, 0, 6, 0 )
505	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
506	RETURN
507	ENDIF
508	IF ( mode == 'yz' .AND. do2d_yz_time_count(av) + 1 > &
509	ntdim_2d_yz(av) ) THEN
510	WRITE ( message_string, * ) 'Output of yz cross-sections is not ', &
511	'given at t=', simulated_time, '&because the', &
512	' maximum number of output time levels is exceeded.'
513	CALL message( 'data_output_2d', 'PA0386', 0, 1, 0, 6, 0 )
514	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
515	RETURN
516	ENDIF
517	ENDIF
518
519	!
520	!-- Allocate a temporary array for resorting (kji -> ijk).
521	ALLOCATE( local_pf(nxl:nxr,nys:nyn,nzb:nzt+1) )
522	local_pf = 0.0
523
524	!
525	!-- Loop of all variables to be written.
526	!-- Output dimensions chosen
527	if = 1
528	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
529	do2d_mode = do2d(av,if)(l-1:l)
530
531	DO WHILE ( do2d(av,if)(1:1) /= ' ' )
532
533	IF ( do2d_mode == mode ) THEN
534	!
535	!-- Set flag to steer output of radiation, land-surface, or user-defined
536	!-- quantities
537	found = .FALSE.
538
539	nzb_do = nzb
540	nzt_do = nzt+1
541	!
542	!-- Before each output, set array local_pf to fill value
543	local_pf = fill_value
544	!
545	!-- Set masking flag for topography for not resorted arrays
546	flag_nr = 0
547
548	!
549	!-- Store the array chosen on the temporary array.
550	resorted = .FALSE.
551	SELECT CASE ( TRIM( do2d(av,if) ) )
552	CASE ( 'e_xy', 'e_xz', 'e_yz' )
553	IF ( av == 0 ) THEN
554	to_be_resorted => e
555	ELSE
556	IF ( .NOT. ALLOCATED( e_av ) ) THEN
557	ALLOCATE( e_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
558	e_av = REAL( fill_value, KIND = wp )
559	ENDIF
560	to_be_resorted => e_av
561	ENDIF
562	IF ( mode == 'xy' ) level_z = zu
563
564	CASE ( 'lpt_xy', 'lpt_xz', 'lpt_yz' )
565	IF ( av == 0 ) THEN
566	to_be_resorted => pt
567	ELSE
568	IF ( .NOT. ALLOCATED( lpt_av ) ) THEN
569	ALLOCATE( lpt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
570	lpt_av = REAL( fill_value, KIND = wp )
571	ENDIF
572	to_be_resorted => lpt_av
573	ENDIF
574	IF ( mode == 'xy' ) level_z = zu
575
576	CASE ( 'lwp*_xy' ) ! 2d-array
577	IF ( av == 0 ) THEN
578	DO i = nxl, nxr
579	DO j = nys, nyn
580	local_pf(i,j,nzb+1) = SUM( ql(nzb:nzt,j,i) * &
581	dzw(1:nzt+1) )
582	ENDDO
583	ENDDO
584	ELSE
585	IF ( .NOT. ALLOCATED( lwp_av ) ) THEN
586	ALLOCATE( lwp_av(nysg:nyng,nxlg:nxrg) )
587	lwp_av = REAL( fill_value, KIND = wp )
588	ENDIF
589	DO i = nxl, nxr
590	DO j = nys, nyn
591	local_pf(i,j,nzb+1) = lwp_av(j,i)
592	ENDDO
593	ENDDO
594	ENDIF
595	resorted = .TRUE.
596	two_d = .TRUE.
597	level_z(nzb+1) = zu(nzb+1)
598
599	CASE ( 'ghf*_xy' ) ! 2d-array
600	IF ( av == 0 ) THEN
601	DO m = 1, surf_lsm_h%ns
602	i = surf_lsm_h%i(m)
603	j = surf_lsm_h%j(m)
604	local_pf(i,j,nzb+1) = surf_lsm_h%ghf(m)
605	ENDDO
606	DO m = 1, surf_usm_h%ns
607	i = surf_usm_h%i(m)
608	j = surf_usm_h%j(m)
609	local_pf(i,j,nzb+1) = surf_usm_h%frac(ind_veg_wall,m) * &
610	surf_usm_h%wghf_eb(m) + &
611	surf_usm_h%frac(ind_pav_green,m) * &
612	surf_usm_h%wghf_eb_green(m) + &
613	surf_usm_h%frac(ind_wat_win,m) * &
614	surf_usm_h%wghf_eb_window(m)
615	ENDDO
616	ELSE
617	IF ( .NOT. ALLOCATED( ghf_av ) ) THEN
618	ALLOCATE( ghf_av(nysg:nyng,nxlg:nxrg) )
619	ghf_av = REAL( fill_value, KIND = wp )
620	ENDIF
621	DO i = nxl, nxr
622	DO j = nys, nyn
623	local_pf(i,j,nzb+1) = ghf_av(j,i)
624	ENDDO
625	ENDDO
626	ENDIF
627
628	resorted = .TRUE.
629	two_d = .TRUE.
630	level_z(nzb+1) = zu(nzb+1)
631
632	CASE ( 'ol*_xy' ) ! 2d-array
633	IF ( av == 0 ) THEN
634	DO m = 1, surf_def_h(0)%ns
635	i = surf_def_h(0)%i(m)
636	j = surf_def_h(0)%j(m)
637	local_pf(i,j,nzb+1) = surf_def_h(0)%ol(m)
638	ENDDO
639	DO m = 1, surf_lsm_h%ns
640	i = surf_lsm_h%i(m)
641	j = surf_lsm_h%j(m)
642	local_pf(i,j,nzb+1) = surf_lsm_h%ol(m)
643	ENDDO
644	DO m = 1, surf_usm_h%ns
645	i = surf_usm_h%i(m)
646	j = surf_usm_h%j(m)
647	local_pf(i,j,nzb+1) = surf_usm_h%ol(m)
648	ENDDO
649	ELSE
650	IF ( .NOT. ALLOCATED( ol_av ) ) THEN
651	ALLOCATE( ol_av(nysg:nyng,nxlg:nxrg) )
652	ol_av = REAL( fill_value, KIND = wp )
653	ENDIF
654	DO i = nxl, nxr
655	DO j = nys, nyn
656	local_pf(i,j,nzb+1) = ol_av(j,i)
657	ENDDO
658	ENDDO
659	ENDIF
660	resorted = .TRUE.
661	two_d = .TRUE.
662	level_z(nzb+1) = zu(nzb+1)
663
664	CASE ( 'p_xy', 'p_xz', 'p_yz' )
665	IF ( av == 0 ) THEN
666	IF ( psolver /= 'sor' ) CALL exchange_horiz( p, nbgp )
667	to_be_resorted => p
668	ELSE
669	IF ( .NOT. ALLOCATED( p_av ) ) THEN
670	ALLOCATE( p_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
671	p_av = REAL( fill_value, KIND = wp )
672	ENDIF
673	IF ( psolver /= 'sor' ) CALL exchange_horiz( p_av, nbgp )
674	to_be_resorted => p_av
675	ENDIF
676	IF ( mode == 'xy' ) level_z = zu
677
678	CASE ( 'pc_xy', 'pc_xz', 'pc_yz' ) ! particle concentration
679	IF ( av == 0 ) THEN
680	IF ( simulated_time >= particle_advection_start ) THEN
681	tend = prt_count
682	! CALL exchange_horiz( tend, nbgp )
683	ELSE
684	tend = 0.0_wp
685	ENDIF
686	DO i = nxl, nxr
687	DO j = nys, nyn
688	DO k = nzb, nzt+1
689	local_pf(i,j,k) = tend(k,j,i)
690	ENDDO
691	ENDDO
692	ENDDO
693	resorted = .TRUE.
694	ELSE
695	IF ( .NOT. ALLOCATED( pc_av ) ) THEN
696	ALLOCATE( pc_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
697	pc_av = REAL( fill_value, KIND = wp )
698	ENDIF
699	! CALL exchange_horiz( pc_av, nbgp )
700	to_be_resorted => pc_av
701	ENDIF
702
703	CASE ( 'pr_xy', 'pr_xz', 'pr_yz' ) ! mean particle radius (effective radius)
704	IF ( av == 0 ) THEN
705	IF ( simulated_time >= particle_advection_start ) THEN
706	DO i = nxl, nxr
707	DO j = nys, nyn
708	DO k = nzb, nzt+1
709	number_of_particles = prt_count(k,j,i)
710	IF (number_of_particles <= 0) CYCLE
711	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
712	s_r2 = 0.0_wp
713	s_r3 = 0.0_wp
714	DO n = 1, number_of_particles
715	IF ( particles(n)%particle_mask ) THEN
716	s_r2 = s_r2 + particles(n)%radius*2 &
717	particles(n)%weight_factor
718	s_r3 = s_r3 + particles(n)%radius*3 &
719	particles(n)%weight_factor
720	ENDIF
721	ENDDO
722	IF ( s_r2 > 0.0_wp ) THEN
723	mean_r = s_r3 / s_r2
724	ELSE
725	mean_r = 0.0_wp
726	ENDIF
727	tend(k,j,i) = mean_r
728	ENDDO
729	ENDDO
730	ENDDO
731	! CALL exchange_horiz( tend, nbgp )
732	ELSE
733	tend = 0.0_wp
734	ENDIF
735	DO i = nxl, nxr
736	DO j = nys, nyn
737	DO k = nzb, nzt+1
738	local_pf(i,j,k) = tend(k,j,i)
739	ENDDO
740	ENDDO
741	ENDDO
742	resorted = .TRUE.
743	ELSE
744	IF ( .NOT. ALLOCATED( pr_av ) ) THEN
745	ALLOCATE( pr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
746	pr_av = REAL( fill_value, KIND = wp )
747	ENDIF
748	! CALL exchange_horiz( pr_av, nbgp )
749	to_be_resorted => pr_av
750	ENDIF
751
752	CASE ( 'pt_xy', 'pt_xz', 'pt_yz' )
753	IF ( av == 0 ) THEN
754	IF ( .NOT. bulk_cloud_model ) THEN
755	to_be_resorted => pt
756	ELSE
757	DO i = nxl, nxr
758	DO j = nys, nyn
759	DO k = nzb, nzt+1
760	local_pf(i,j,k) = pt(k,j,i) + lv_d_cp * &
761	d_exner(k) * &
762	ql(k,j,i)
763	ENDDO
764	ENDDO
765	ENDDO
766	resorted = .TRUE.
767	ENDIF
768	ELSE
769	IF ( .NOT. ALLOCATED( pt_av ) ) THEN
770	ALLOCATE( pt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
771	pt_av = REAL( fill_value, KIND = wp )
772	ENDIF
773	to_be_resorted => pt_av
774	ENDIF
775	IF ( mode == 'xy' ) level_z = zu
776
777	CASE ( 'q_xy', 'q_xz', 'q_yz' )
778	IF ( av == 0 ) THEN
779	to_be_resorted => q
780	ELSE
781	IF ( .NOT. ALLOCATED( q_av ) ) THEN
782	ALLOCATE( q_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
783	q_av = REAL( fill_value, KIND = wp )
784	ENDIF
785	to_be_resorted => q_av
786	ENDIF
787	IF ( mode == 'xy' ) level_z = zu
788
789	CASE ( 'ql_xy', 'ql_xz', 'ql_yz' )
790	IF ( av == 0 ) THEN
791	to_be_resorted => ql
792	ELSE
793	IF ( .NOT. ALLOCATED( ql_av ) ) THEN
794	ALLOCATE( ql_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
795	ql_av = REAL( fill_value, KIND = wp )
796	ENDIF
797	to_be_resorted => ql_av
798	ENDIF
799	IF ( mode == 'xy' ) level_z = zu
800
801	CASE ( 'ql_c_xy', 'ql_c_xz', 'ql_c_yz' )
802	IF ( av == 0 ) THEN
803	to_be_resorted => ql_c
804	ELSE
805	IF ( .NOT. ALLOCATED( ql_c_av ) ) THEN
806	ALLOCATE( ql_c_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
807	ql_c_av = REAL( fill_value, KIND = wp )
808	ENDIF
809	to_be_resorted => ql_c_av
810	ENDIF
811	IF ( mode == 'xy' ) level_z = zu
812
813	CASE ( 'ql_v_xy', 'ql_v_xz', 'ql_v_yz' )
814	IF ( av == 0 ) THEN
815	to_be_resorted => ql_v
816	ELSE
817	IF ( .NOT. ALLOCATED( ql_v_av ) ) THEN
818	ALLOCATE( ql_v_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
819	ql_v_av = REAL( fill_value, KIND = wp )
820	ENDIF
821	to_be_resorted => ql_v_av
822	ENDIF
823	IF ( mode == 'xy' ) level_z = zu
824
825	CASE ( 'ql_vp_xy', 'ql_vp_xz', 'ql_vp_yz' )
826	IF ( av == 0 ) THEN
827	IF ( simulated_time >= particle_advection_start ) THEN
828	DO i = nxl, nxr
829	DO j = nys, nyn
830	DO k = nzb, nzt+1
831	number_of_particles = prt_count(k,j,i)
832	IF (number_of_particles <= 0) CYCLE
833	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
834	DO n = 1, number_of_particles
835	IF ( particles(n)%particle_mask ) THEN
836	tend(k,j,i) = tend(k,j,i) + &
837	particles(n)%weight_factor / &
838	prt_count(k,j,i)
839	ENDIF
840	ENDDO
841	ENDDO
842	ENDDO
843	ENDDO
844	! CALL exchange_horiz( tend, nbgp )
845	ELSE
846	tend = 0.0_wp
847	ENDIF
848	DO i = nxl, nxr
849	DO j = nys, nyn
850	DO k = nzb, nzt+1
851	local_pf(i,j,k) = tend(k,j,i)
852	ENDDO
853	ENDDO
854	ENDDO
855	resorted = .TRUE.
856	ELSE
857	IF ( .NOT. ALLOCATED( ql_vp_av ) ) THEN
858	ALLOCATE( ql_vp_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
859	ql_vp_av = REAL( fill_value, KIND = wp )
860	ENDIF
861	! CALL exchange_horiz( ql_vp_av, nbgp )
862	to_be_resorted => ql_vp_av
863	ENDIF
864	IF ( mode == 'xy' ) level_z = zu
865
866	CASE ( 'qsws*_xy' ) ! 2d-array
867	IF ( av == 0 ) THEN
868	local_pf(:,:,nzb+1) = REAL( fill_value, KIND = wp )
869	!
870	!-- In case of default surfaces, clean-up flux by density.
871	!-- In case of land-surfaces, convert fluxes into
872	!-- dynamic units
873	DO m = 1, surf_def_h(0)%ns
874	i = surf_def_h(0)%i(m)
875	j = surf_def_h(0)%j(m)
876	k = surf_def_h(0)%k(m)
877	local_pf(i,j,nzb+1) = surf_def_h(0)%qsws(m) * &
878	waterflux_output_conversion(k)
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	k = surf_lsm_h%k(m)
884	local_pf(i,j,nzb+1) = surf_lsm_h%qsws(m) * l_v
885	ENDDO
886	ELSE
887	IF ( .NOT. ALLOCATED( qsws_av ) ) THEN
888	ALLOCATE( qsws_av(nysg:nyng,nxlg:nxrg) )
889	qsws_av = REAL( fill_value, KIND = wp )
890	ENDIF
891	DO i = nxl, nxr
892	DO j = nys, nyn
893	local_pf(i,j,nzb+1) = qsws_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 ( 'qv_xy', 'qv_xz', 'qv_yz' )
902	IF ( av == 0 ) THEN
903	DO i = nxl, nxr
904	DO j = nys, nyn
905	DO k = nzb, nzt+1
906	local_pf(i,j,k) = q(k,j,i) - ql(k,j,i)
907	ENDDO
908	ENDDO
909	ENDDO
910	resorted = .TRUE.
911	ELSE
912	IF ( .NOT. ALLOCATED( qv_av ) ) THEN
913	ALLOCATE( qv_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
914	qv_av = REAL( fill_value, KIND = wp )
915	ENDIF
916	to_be_resorted => qv_av
917	ENDIF
918	IF ( mode == 'xy' ) level_z = zu
919
920	CASE ( 'r_a*_xy' ) ! 2d-array
921	IF ( av == 0 ) THEN
922	DO m = 1, surf_lsm_h%ns
923	i = surf_lsm_h%i(m)
924	j = surf_lsm_h%j(m)
925	local_pf(i,j,nzb+1) = surf_lsm_h%r_a(m)
926	ENDDO
927
928	DO m = 1, surf_usm_h%ns
929	i = surf_usm_h%i(m)
930	j = surf_usm_h%j(m)
931	local_pf(i,j,nzb+1) = &
932	( surf_usm_h%frac(ind_veg_wall,m) * &
933	surf_usm_h%r_a(m) + &
934	surf_usm_h%frac(ind_pav_green,m) * &
935	surf_usm_h%r_a_green(m) + &
936	surf_usm_h%frac(ind_wat_win,m) * &
937	surf_usm_h%r_a_window(m) )
938	ENDDO
939	ELSE
940	IF ( .NOT. ALLOCATED( r_a_av ) ) THEN
941	ALLOCATE( r_a_av(nysg:nyng,nxlg:nxrg) )
942	r_a_av = REAL( fill_value, KIND = wp )
943	ENDIF
944	DO i = nxl, nxr
945	DO j = nys, nyn
946	local_pf(i,j,nzb+1) = r_a_av(j,i)
947	ENDDO
948	ENDDO
949	ENDIF
950	resorted = .TRUE.
951	two_d = .TRUE.
952	level_z(nzb+1) = zu(nzb+1)
953
954	CASE ( 's_xy', 's_xz', 's_yz' )
955	IF ( av == 0 ) THEN
956	to_be_resorted => s
957	ELSE
958	IF ( .NOT. ALLOCATED( s_av ) ) THEN
959	ALLOCATE( s_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
960	s_av = REAL( fill_value, KIND = wp )
961	ENDIF
962	to_be_resorted => s_av
963	ENDIF
964
965	CASE ( 'shf*_xy' ) ! 2d-array
966	IF ( av == 0 ) THEN
967	!
968	!-- In case of default surfaces, clean-up flux by density.
969	!-- In case of land- and urban-surfaces, convert fluxes into
970	!-- dynamic units.
971	DO m = 1, surf_def_h(0)%ns
972	i = surf_def_h(0)%i(m)
973	j = surf_def_h(0)%j(m)
974	k = surf_def_h(0)%k(m)
975	local_pf(i,j,nzb+1) = surf_def_h(0)%shf(m) * &
976	heatflux_output_conversion(k)
977	ENDDO
978	DO m = 1, surf_lsm_h%ns
979	i = surf_lsm_h%i(m)
980	j = surf_lsm_h%j(m)
981	k = surf_lsm_h%k(m)
982	local_pf(i,j,nzb+1) = surf_lsm_h%shf(m) * c_p
983	ENDDO
984	DO m = 1, surf_usm_h%ns
985	i = surf_usm_h%i(m)
986	j = surf_usm_h%j(m)
987	k = surf_usm_h%k(m)
988	local_pf(i,j,nzb+1) = surf_usm_h%shf(m) * c_p
989	ENDDO
990	ELSE
991	IF ( .NOT. ALLOCATED( shf_av ) ) THEN
992	ALLOCATE( shf_av(nysg:nyng,nxlg:nxrg) )
993	shf_av = REAL( fill_value, KIND = wp )
994	ENDIF
995	DO i = nxl, nxr
996	DO j = nys, nyn
997	local_pf(i,j,nzb+1) = shf_av(j,i)
998	ENDDO
999	ENDDO
1000	ENDIF
1001	resorted = .TRUE.
1002	two_d = .TRUE.
1003	level_z(nzb+1) = zu(nzb+1)
1004
1005	CASE ( 'ssws*_xy' ) ! 2d-array
1006	IF ( av == 0 ) THEN
1007	DO m = 1, surf_def_h(0)%ns
1008	i = surf_def_h(0)%i(m)
1009	j = surf_def_h(0)%j(m)
1010	local_pf(i,j,nzb+1) = surf_def_h(0)%ssws(m)
1011	ENDDO
1012	DO m = 1, surf_lsm_h%ns
1013	i = surf_lsm_h%i(m)
1014	j = surf_lsm_h%j(m)
1015	local_pf(i,j,nzb+1) = surf_lsm_h%ssws(m)
1016	ENDDO
1017	DO m = 1, surf_usm_h%ns
1018	i = surf_usm_h%i(m)
1019	j = surf_usm_h%j(m)
1020	local_pf(i,j,nzb+1) = surf_usm_h%ssws(m)
1021	ENDDO
1022	ELSE
1023	IF ( .NOT. ALLOCATED( ssws_av ) ) THEN
1024	ALLOCATE( ssws_av(nysg:nyng,nxlg:nxrg) )
1025	ssws_av = REAL( fill_value, KIND = wp )
1026	ENDIF
1027	DO i = nxl, nxr
1028	DO j = nys, nyn
1029	local_pf(i,j,nzb+1) = ssws_av(j,i)
1030	ENDDO
1031	ENDDO
1032	ENDIF
1033	resorted = .TRUE.
1034	two_d = .TRUE.
1035	level_z(nzb+1) = zu(nzb+1)
1036
1037	CASE ( 't*_xy' ) ! 2d-array
1038	IF ( av == 0 ) THEN
1039	DO m = 1, surf_def_h(0)%ns
1040	i = surf_def_h(0)%i(m)
1041	j = surf_def_h(0)%j(m)
1042	local_pf(i,j,nzb+1) = surf_def_h(0)%ts(m)
1043	ENDDO
1044	DO m = 1, surf_lsm_h%ns
1045	i = surf_lsm_h%i(m)
1046	j = surf_lsm_h%j(m)
1047	local_pf(i,j,nzb+1) = surf_lsm_h%ts(m)
1048	ENDDO
1049	DO m = 1, surf_usm_h%ns
1050	i = surf_usm_h%i(m)
1051	j = surf_usm_h%j(m)
1052	local_pf(i,j,nzb+1) = surf_usm_h%ts(m)
1053	ENDDO
1054	ELSE
1055	IF ( .NOT. ALLOCATED( ts_av ) ) THEN
1056	ALLOCATE( ts_av(nysg:nyng,nxlg:nxrg) )
1057	ts_av = REAL( fill_value, KIND = wp )
1058	ENDIF
1059	DO i = nxl, nxr
1060	DO j = nys, nyn
1061	local_pf(i,j,nzb+1) = ts_av(j,i)
1062	ENDDO
1063	ENDDO
1064	ENDIF
1065	resorted = .TRUE.
1066	two_d = .TRUE.
1067	level_z(nzb+1) = zu(nzb+1)
1068
1069	CASE ( 'tsurf*_xy' ) ! 2d-array
1070	IF ( av == 0 ) THEN
1071	DO m = 1, surf_def_h(0)%ns
1072	i = surf_def_h(0)%i(m)
1073	j = surf_def_h(0)%j(m)
1074	local_pf(i,j,nzb+1) = surf_def_h(0)%pt_surface(m)
1075	ENDDO
1076
1077	DO m = 1, surf_lsm_h%ns
1078	i = surf_lsm_h%i(m)
1079	j = surf_lsm_h%j(m)
1080	local_pf(i,j,nzb+1) = surf_lsm_h%pt_surface(m)
1081	ENDDO
1082
1083	DO m = 1, surf_usm_h%ns
1084	i = surf_usm_h%i(m)
1085	j = surf_usm_h%j(m)
1086	local_pf(i,j,nzb+1) = surf_usm_h%pt_surface(m)
1087	ENDDO
1088
1089	ELSE
1090	IF ( .NOT. ALLOCATED( tsurf_av ) ) THEN
1091	ALLOCATE( tsurf_av(nysg:nyng,nxlg:nxrg) )
1092	tsurf_av = REAL( fill_value, KIND = wp )
1093	ENDIF
1094	DO i = nxl, nxr
1095	DO j = nys, nyn
1096	local_pf(i,j,nzb+1) = tsurf_av(j,i)
1097	ENDDO
1098	ENDDO
1099	ENDIF
1100	resorted = .TRUE.
1101	two_d = .TRUE.
1102	level_z(nzb+1) = zu(nzb+1)
1103
1104	CASE ( 'u_xy', 'u_xz', 'u_yz' )
1105	flag_nr = 1
1106	IF ( av == 0 ) THEN
1107	to_be_resorted => u
1108	ELSE
1109	IF ( .NOT. ALLOCATED( u_av ) ) THEN
1110	ALLOCATE( u_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1111	u_av = REAL( fill_value, KIND = wp )
1112	ENDIF
1113	to_be_resorted => u_av
1114	ENDIF
1115	IF ( mode == 'xy' ) level_z = zu
1116	!
1117	!-- Substitute the values generated by "mirror" boundary condition
1118	!-- at the bottom boundary by the real surface values.
1119	IF ( do2d(av,if) == 'u_xz' .OR. do2d(av,if) == 'u_yz' ) THEN
1120	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
1121	ENDIF
1122
1123	CASE ( 'u*_xy' ) ! 2d-array
1124	IF ( av == 0 ) THEN
1125	DO m = 1, surf_def_h(0)%ns
1126	i = surf_def_h(0)%i(m)
1127	j = surf_def_h(0)%j(m)
1128	local_pf(i,j,nzb+1) = surf_def_h(0)%us(m)
1129	ENDDO
1130	DO m = 1, surf_lsm_h%ns
1131	i = surf_lsm_h%i(m)
1132	j = surf_lsm_h%j(m)
1133	local_pf(i,j,nzb+1) = surf_lsm_h%us(m)
1134	ENDDO
1135	DO m = 1, surf_usm_h%ns
1136	i = surf_usm_h%i(m)
1137	j = surf_usm_h%j(m)
1138	local_pf(i,j,nzb+1) = surf_usm_h%us(m)
1139	ENDDO
1140	ELSE
1141	IF ( .NOT. ALLOCATED( us_av ) ) THEN
1142	ALLOCATE( us_av(nysg:nyng,nxlg:nxrg) )
1143	us_av = REAL( fill_value, KIND = wp )
1144	ENDIF
1145	DO i = nxl, nxr
1146	DO j = nys, nyn
1147	local_pf(i,j,nzb+1) = us_av(j,i)
1148	ENDDO
1149	ENDDO
1150	ENDIF
1151	resorted = .TRUE.
1152	two_d = .TRUE.
1153	level_z(nzb+1) = zu(nzb+1)
1154
1155	CASE ( 'v_xy', 'v_xz', 'v_yz' )
1156	flag_nr = 2
1157	IF ( av == 0 ) THEN
1158	to_be_resorted => v
1159	ELSE
1160	IF ( .NOT. ALLOCATED( v_av ) ) THEN
1161	ALLOCATE( v_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1162	v_av = REAL( fill_value, KIND = wp )
1163	ENDIF
1164	to_be_resorted => v_av
1165	ENDIF
1166	IF ( mode == 'xy' ) level_z = zu
1167	!
1168	!-- Substitute the values generated by "mirror" boundary condition
1169	!-- at the bottom boundary by the real surface values.
1170	IF ( do2d(av,if) == 'v_xz' .OR. do2d(av,if) == 'v_yz' ) THEN
1171	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
1172	ENDIF
1173
1174	CASE ( 'vpt_xy', 'vpt_xz', 'vpt_yz' )
1175	IF ( av == 0 ) THEN
1176	to_be_resorted => vpt
1177	ELSE
1178	IF ( .NOT. ALLOCATED( vpt_av ) ) THEN
1179	ALLOCATE( vpt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1180	vpt_av = REAL( fill_value, KIND = wp )
1181	ENDIF
1182	to_be_resorted => vpt_av
1183	ENDIF
1184	IF ( mode == 'xy' ) level_z = zu
1185
1186	CASE ( 'w_xy', 'w_xz', 'w_yz' )
1187	flag_nr = 3
1188	IF ( av == 0 ) THEN
1189	to_be_resorted => w
1190	ELSE
1191	IF ( .NOT. ALLOCATED( w_av ) ) THEN
1192	ALLOCATE( w_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1193	w_av = REAL( fill_value, KIND = wp )
1194	ENDIF
1195	to_be_resorted => w_av
1196	ENDIF
1197	IF ( mode == 'xy' ) level_z = zw
1198
1199	CASE ( 'z0*_xy' ) ! 2d-array
1200	IF ( av == 0 ) THEN
1201	DO m = 1, surf_def_h(0)%ns
1202	i = surf_def_h(0)%i(m)
1203	j = surf_def_h(0)%j(m)
1204	local_pf(i,j,nzb+1) = surf_def_h(0)%z0(m)
1205	ENDDO
1206	DO m = 1, surf_lsm_h%ns
1207	i = surf_lsm_h%i(m)
1208	j = surf_lsm_h%j(m)
1209	local_pf(i,j,nzb+1) = surf_lsm_h%z0(m)
1210	ENDDO
1211	DO m = 1, surf_usm_h%ns
1212	i = surf_usm_h%i(m)
1213	j = surf_usm_h%j(m)
1214	local_pf(i,j,nzb+1) = surf_usm_h%z0(m)
1215	ENDDO
1216	ELSE
1217	IF ( .NOT. ALLOCATED( z0_av ) ) THEN
1218	ALLOCATE( z0_av(nysg:nyng,nxlg:nxrg) )
1219	z0_av = REAL( fill_value, KIND = wp )
1220	ENDIF
1221	DO i = nxl, nxr
1222	DO j = nys, nyn
1223	local_pf(i,j,nzb+1) = z0_av(j,i)
1224	ENDDO
1225	ENDDO
1226	ENDIF
1227	resorted = .TRUE.
1228	two_d = .TRUE.
1229	level_z(nzb+1) = zu(nzb+1)
1230
1231	CASE ( 'z0h*_xy' ) ! 2d-array
1232	IF ( av == 0 ) THEN
1233	DO m = 1, surf_def_h(0)%ns
1234	i = surf_def_h(0)%i(m)
1235	j = surf_def_h(0)%j(m)
1236	local_pf(i,j,nzb+1) = surf_def_h(0)%z0h(m)
1237	ENDDO
1238	DO m = 1, surf_lsm_h%ns
1239	i = surf_lsm_h%i(m)
1240	j = surf_lsm_h%j(m)
1241	local_pf(i,j,nzb+1) = surf_lsm_h%z0h(m)
1242	ENDDO
1243	DO m = 1, surf_usm_h%ns
1244	i = surf_usm_h%i(m)
1245	j = surf_usm_h%j(m)
1246	local_pf(i,j,nzb+1) = surf_usm_h%z0h(m)
1247	ENDDO
1248	ELSE
1249	IF ( .NOT. ALLOCATED( z0h_av ) ) THEN
1250	ALLOCATE( z0h_av(nysg:nyng,nxlg:nxrg) )
1251	z0h_av = REAL( fill_value, KIND = wp )
1252	ENDIF
1253	DO i = nxl, nxr
1254	DO j = nys, nyn
1255	local_pf(i,j,nzb+1) = z0h_av(j,i)
1256	ENDDO
1257	ENDDO
1258	ENDIF
1259	resorted = .TRUE.
1260	two_d = .TRUE.
1261	level_z(nzb+1) = zu(nzb+1)
1262
1263	CASE ( 'z0q*_xy' ) ! 2d-array
1264	IF ( av == 0 ) THEN
1265	DO m = 1, surf_def_h(0)%ns
1266	i = surf_def_h(0)%i(m)
1267	j = surf_def_h(0)%j(m)
1268	local_pf(i,j,nzb+1) = surf_def_h(0)%z0q(m)
1269	ENDDO
1270	DO m = 1, surf_lsm_h%ns
1271	i = surf_lsm_h%i(m)
1272	j = surf_lsm_h%j(m)
1273	local_pf(i,j,nzb+1) = surf_lsm_h%z0q(m)
1274	ENDDO
1275	DO m = 1, surf_usm_h%ns
1276	i = surf_usm_h%i(m)
1277	j = surf_usm_h%j(m)
1278	local_pf(i,j,nzb+1) = surf_usm_h%z0q(m)
1279	ENDDO
1280	ELSE
1281	IF ( .NOT. ALLOCATED( z0q_av ) ) THEN
1282	ALLOCATE( z0q_av(nysg:nyng,nxlg:nxrg) )
1283	z0q_av = REAL( fill_value, KIND = wp )
1284	ENDIF
1285	DO i = nxl, nxr
1286	DO j = nys, nyn
1287	local_pf(i,j,nzb+1) = z0q_av(j,i)
1288	ENDDO
1289	ENDDO
1290	ENDIF
1291	resorted = .TRUE.
1292	two_d = .TRUE.
1293	level_z(nzb+1) = zu(nzb+1)
1294
1295	CASE DEFAULT
1296
1297	!
1298	!-- Quantities of other modules
1299	IF ( .NOT. found .AND. bulk_cloud_model ) THEN
1300	CALL bcm_data_output_2d( av, do2d(av,if), found, grid, mode,&
1301	local_pf, two_d, nzb_do, nzt_do )
1302	ENDIF
1303
1304	IF ( .NOT. found .AND. gust_module_enabled ) THEN
1305	CALL gust_data_output_2d( av, do2d(av,if), found, grid, &
1306	local_pf, two_d, nzb_do, nzt_do )
1307	ENDIF
1308
1309	IF ( .NOT. found .AND. land_surface ) THEN
1310	CALL lsm_data_output_2d( av, do2d(av,if), found, grid, mode,&
1311	local_pf, two_d, nzb_do, nzt_do )
1312	ENDIF
1313
1314	IF ( .NOT. found .AND. ocean_mode ) THEN
1315	CALL ocean_data_output_2d( av, do2d(av,if), found, grid, &
1316	mode, local_pf, nzb_do, nzt_do )
1317	ENDIF
1318
1319	IF ( .NOT. found .AND. radiation ) THEN
1320	CALL radiation_data_output_2d( av, do2d(av,if), found, grid,&
1321	mode, local_pf, two_d, &
1322	nzb_do, nzt_do )
1323	ENDIF
1324
1325	IF ( .NOT. found ) THEN
1326	CALL tcm_data_output_2d( av, do2d(av,if), found, grid, mode,&
1327	local_pf, nzb_do, nzt_do )
1328	ENDIF
1329
1330	IF ( .NOT. found .AND. uv_exposure ) THEN
1331	CALL uvem_data_output_2d( av, do2d(av,if), found, grid, &
1332	local_pf, two_d, nzb_do, nzt_do )
1333	ENDIF
1334
1335	IF ( .NOT. found .AND. air_chemistry ) THEN
1336	CALL chem_data_output_2d( av, do2d(av,if), found, grid, mode, &
1337	local_pf, two_d, nzb_do, nzt_do, fill_value )
1338	ENDIF
1339	!
1340	!-- User defined quantities
1341	IF ( .NOT. found ) THEN
1342	CALL user_data_output_2d( av, do2d(av,if), found, grid, &
1343	local_pf, two_d, nzb_do, nzt_do )
1344	ENDIF
1345
1346	resorted = .TRUE.
1347
1348	IF ( grid == 'zu' ) THEN
1349	IF ( mode == 'xy' ) level_z = zu
1350	ELSEIF ( grid == 'zw' ) THEN
1351	IF ( mode == 'xy' ) level_z = zw
1352	ELSEIF ( grid == 'zu1' ) THEN
1353	IF ( mode == 'xy' ) level_z(nzb+1) = zu(nzb+1)
1354	ELSEIF ( grid == 'zs' ) THEN
1355	IF ( mode == 'xy' ) level_z = zs
1356	ENDIF
1357
1358	IF ( .NOT. found ) THEN
1359	message_string = 'no output provided for: ' // &
1360	TRIM( do2d(av,if) )
1361	CALL message( 'data_output_2d', 'PA0181', 0, 0, 0, 6, 0 )
1362	ENDIF
1363
1364	END SELECT
1365
1366	!
1367	!-- Resort the array to be output, if not done above. Flag topography
1368	!-- grid points with fill values, using the corresponding maksing flag.
1369	IF ( .NOT. resorted ) THEN
1370	DO i = nxl, nxr
1371	DO j = nys, nyn
1372	DO k = nzb_do, nzt_do
1373	local_pf(i,j,k) = MERGE( to_be_resorted(k,j,i), &
1374	REAL( fill_value, KIND = wp ), &
1375	BTEST( wall_flags_0(k,j,i), &
1376	flag_nr ) )
1377	ENDDO
1378	ENDDO
1379	ENDDO
1380	ENDIF
1381
1382	!
1383	!-- Output of the individual cross-sections, depending on the cross-
1384	!-- section mode chosen.
1385	is = 1
1386	loop1: DO WHILE ( section(is,s_ind) /= -9999 .OR. two_d )
1387
1388	SELECT CASE ( mode )
1389
1390	CASE ( 'xy' )
1391	!
1392	!-- Determine the cross section index
1393	IF ( two_d ) THEN
1394	layer_xy = nzb+1
1395	ELSE
1396	layer_xy = section(is,s_ind)
1397	ENDIF
1398
1399	!
1400	!-- Exit the loop for layers beyond the data output domain
1401	!-- (used for soil model)
1402	IF ( layer_xy > nzt_do ) THEN
1403	EXIT loop1
1404	ENDIF
1405
1406	!
1407	!-- Update the netCDF xy cross section time axis.
1408	!-- In case of parallel output, this is only done by PE0
1409	!-- to increase the performance.
1410	IF ( simulated_time /= do2d_xy_last_time(av) ) THEN
1411	do2d_xy_time_count(av) = do2d_xy_time_count(av) + 1
1412	do2d_xy_last_time(av) = simulated_time
1413	IF ( myid == 0 ) THEN
1414	IF ( .NOT. data_output_2d_on_each_pe &
1415	.OR. netcdf_data_format > 4 ) &
1416	THEN
1417	#if defined( __netcdf )
1418	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1419	id_var_time_xy(av), &
1420	(/ time_since_reference_point /), &
1421	start = (/ do2d_xy_time_count(av) /), &
1422	count = (/ 1 /) )
1423	CALL netcdf_handle_error( 'data_output_2d', 53 )
1424	#endif
1425	ENDIF
1426	ENDIF
1427	ENDIF
1428	!
1429	!-- If required, carry out averaging along z
1430	IF ( section(is,s_ind) == -1 .AND. .NOT. two_d ) THEN
1431
1432	local_2d = 0.0_wp
1433	!
1434	!-- Carry out the averaging (all data are on the PE)
1435	DO k = nzb_do, nzt_do
1436	DO j = nys, nyn
1437	DO i = nxl, nxr
1438	local_2d(i,j) = local_2d(i,j) + local_pf(i,j,k)
1439	ENDDO
1440	ENDDO
1441	ENDDO
1442
1443	local_2d = local_2d / ( nzt_do - nzb_do + 1.0_wp)
1444
1445	ELSE
1446	!
1447	!-- Just store the respective section on the local array
1448	local_2d = local_pf(:,:,layer_xy)
1449
1450	ENDIF
1451
1452	#if defined( __parallel )
1453	IF ( netcdf_data_format > 4 ) THEN
1454	!
1455	!-- Parallel output in netCDF4/HDF5 format.
1456	IF ( two_d ) THEN
1457	iis = 1
1458	ELSE
1459	iis = is
1460	ENDIF
1461
1462	#if defined( __netcdf )
1463	!
1464	!-- For parallel output, all cross sections are first stored
1465	!-- here on a local array and will be written to the output
1466	!-- file afterwards to increase the performance.
1467	DO i = nxl, nxr
1468	DO j = nys, nyn
1469	local_2d_sections(i,j,iis) = local_2d(i,j)
1470	ENDDO
1471	ENDDO
1472	#endif
1473	ELSE
1474
1475	IF ( data_output_2d_on_each_pe ) THEN
1476	!
1477	!-- Output of partial arrays on each PE
1478	#if defined( __netcdf )
1479	IF ( myid == 0 ) THEN
1480	WRITE ( 21 ) time_since_reference_point, &
1481	do2d_xy_time_count(av), av
1482	ENDIF
1483	#endif
1484	DO i = 0, io_blocks-1
1485	IF ( i == io_group ) THEN
1486	WRITE ( 21 ) nxl, nxr, nys, nyn, nys, nyn
1487	WRITE ( 21 ) local_2d
1488	ENDIF
1489	#if defined( __parallel )
1490	CALL MPI_BARRIER( comm2d, ierr )
1491	#endif
1492	ENDDO
1493
1494	ELSE
1495	!
1496	!-- PE0 receives partial arrays from all processors and
1497	!-- then outputs them. Here a barrier has to be set,
1498	!-- because otherwise "-MPI- FATAL: Remote protocol queue
1499	!-- full" may occur.
1500	CALL MPI_BARRIER( comm2d, ierr )
1501
1502	ngp = ( nxr-nxl+1 ) * ( nyn-nys+1 )
1503	IF ( myid == 0 ) THEN
1504	!
1505	!-- Local array can be relocated directly.
1506	total_2d(nxl:nxr,nys:nyn) = local_2d
1507	!
1508	!-- Receive data from all other PEs.
1509	DO n = 1, numprocs-1
1510	!
1511	!-- Receive index limits first, then array.
1512	!-- Index limits are received in arbitrary order from
1513	!-- the PEs.
1514	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1515	MPI_ANY_SOURCE, 0, comm2d, &
1516	status, ierr )
1517	sender = status(MPI_SOURCE)
1518	DEALLOCATE( local_2d )
1519	ALLOCATE( local_2d(ind(1):ind(2),ind(3):ind(4)) )
1520	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1521	MPI_REAL, sender, 1, comm2d, &
1522	status, ierr )
1523	total_2d(ind(1):ind(2),ind(3):ind(4)) = local_2d
1524	ENDDO
1525	!
1526	!-- Relocate the local array for the next loop increment
1527	DEALLOCATE( local_2d )
1528	ALLOCATE( local_2d(nxl:nxr,nys:nyn) )
1529
1530	#if defined( __netcdf )
1531	IF ( two_d ) THEN
1532	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1533	id_var_do2d(av,if), &
1534	total_2d(0:nx,0:ny), &
1535	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
1536	count = (/ nx+1, ny+1, 1, 1 /) )
1537	ELSE
1538	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1539	id_var_do2d(av,if), &
1540	total_2d(0:nx,0:ny), &
1541	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
1542	count = (/ nx+1, ny+1, 1, 1 /) )
1543	ENDIF
1544	CALL netcdf_handle_error( 'data_output_2d', 54 )
1545	#endif
1546
1547	ELSE
1548	!
1549	!-- First send the local index limits to PE0
1550	ind(1) = nxl; ind(2) = nxr
1551	ind(3) = nys; ind(4) = nyn
1552	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1553	comm2d, ierr )
1554	!
1555	!-- Send data to PE0
1556	CALL MPI_SEND( local_2d(nxl,nys), ngp, &
1557	MPI_REAL, 0, 1, comm2d, ierr )
1558	ENDIF
1559	!
1560	!-- A barrier has to be set, because otherwise some PEs may
1561	!-- proceed too fast so that PE0 may receive wrong data on
1562	!-- tag 0
1563	CALL MPI_BARRIER( comm2d, ierr )
1564	ENDIF
1565
1566	ENDIF
1567	#else
1568	#if defined( __netcdf )
1569	IF ( two_d ) THEN
1570	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1571	id_var_do2d(av,if), &
1572	local_2d(nxl:nxr,nys:nyn), &
1573	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
1574	count = (/ nx+1, ny+1, 1, 1 /) )
1575	ELSE
1576	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1577	id_var_do2d(av,if), &
1578	local_2d(nxl:nxr,nys:nyn), &
1579	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
1580	count = (/ nx+1, ny+1, 1, 1 /) )
1581	ENDIF
1582	CALL netcdf_handle_error( 'data_output_2d', 447 )
1583	#endif
1584	#endif
1585
1586	!
1587	!-- For 2D-arrays (e.g. u*) only one cross-section is available.
1588	!-- Hence exit loop of output levels.
1589	IF ( two_d ) THEN
1590	IF ( netcdf_data_format < 5 ) two_d = .FALSE.
1591	EXIT loop1
1592	ENDIF
1593
1594	CASE ( 'xz' )
1595	!
1596	!-- Update the netCDF xz cross section time axis.
1597	!-- In case of parallel output, this is only done by PE0
1598	!-- to increase the performance.
1599	IF ( simulated_time /= do2d_xz_last_time(av) ) THEN
1600	do2d_xz_time_count(av) = do2d_xz_time_count(av) + 1
1601	do2d_xz_last_time(av) = simulated_time
1602	IF ( myid == 0 ) THEN
1603	IF ( .NOT. data_output_2d_on_each_pe &
1604	.OR. netcdf_data_format > 4 ) &
1605	THEN
1606	#if defined( __netcdf )
1607	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1608	id_var_time_xz(av), &
1609	(/ time_since_reference_point /), &
1610	start = (/ do2d_xz_time_count(av) /), &
1611	count = (/ 1 /) )
1612	CALL netcdf_handle_error( 'data_output_2d', 56 )
1613	#endif
1614	ENDIF
1615	ENDIF
1616	ENDIF
1617
1618	!
1619	!-- If required, carry out averaging along y
1620	IF ( section(is,s_ind) == -1 ) THEN
1621
1622	ALLOCATE( local_2d_l(nxl:nxr,nzb_do:nzt_do) )
1623	local_2d_l = 0.0_wp
1624	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
1625	!
1626	!-- First local averaging on the PE
1627	DO k = nzb_do, nzt_do
1628	DO j = nys, nyn
1629	DO i = nxl, nxr
1630	local_2d_l(i,k) = local_2d_l(i,k) + &
1631	local_pf(i,j,k)
1632	ENDDO
1633	ENDDO
1634	ENDDO
1635	#if defined( __parallel )
1636	!
1637	!-- Now do the averaging over all PEs along y
1638	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1639	CALL MPI_ALLREDUCE( local_2d_l(nxl,nzb_do), &
1640	local_2d(nxl,nzb_do), ngp, MPI_REAL, &
1641	MPI_SUM, comm1dy, ierr )
1642	#else
1643	local_2d = local_2d_l
1644	#endif
1645	local_2d = local_2d / ( ny + 1.0_wp )
1646
1647	DEALLOCATE( local_2d_l )
1648
1649	ELSE
1650	!
1651	!-- Just store the respective section on the local array
1652	!-- (but only if it is available on this PE!)
1653	IF ( section(is,s_ind) >= nys .AND. section(is,s_ind) <= nyn ) &
1654	THEN
1655	local_2d = local_pf(:,section(is,s_ind),nzb_do:nzt_do)
1656	ENDIF
1657
1658	ENDIF
1659
1660	#if defined( __parallel )
1661	IF ( netcdf_data_format > 4 ) THEN
1662	!
1663	!-- Output in netCDF4/HDF5 format.
1664	!-- Output only on those PEs where the respective cross
1665	!-- sections reside. Cross sections averaged along y are
1666	!-- output on the respective first PE along y (myidy=0).
1667	IF ( ( section(is,s_ind) >= nys .AND. &
1668	section(is,s_ind) <= nyn ) .OR. &
1669	( section(is,s_ind) == -1 .AND. myidy == 0 ) ) THEN
1670	#if defined( __netcdf )
1671	!
1672	!-- For parallel output, all cross sections are first
1673	!-- stored here on a local array and will be written to the
1674	!-- output file afterwards to increase the performance.
1675	DO i = nxl, nxr
1676	DO k = nzb_do, nzt_do
1677	local_2d_sections_l(i,is,k) = local_2d(i,k)
1678	ENDDO
1679	ENDDO
1680	#endif
1681	ENDIF
1682
1683	ELSE
1684
1685	IF ( data_output_2d_on_each_pe ) THEN
1686	!
1687	!-- Output of partial arrays on each PE. If the cross
1688	!-- section does not reside on the PE, output special
1689	!-- index values.
1690	#if defined( __netcdf )
1691	IF ( myid == 0 ) THEN
1692	WRITE ( 22 ) time_since_reference_point, &
1693	do2d_xz_time_count(av), av
1694	ENDIF
1695	#endif
1696	DO i = 0, io_blocks-1
1697	IF ( i == io_group ) THEN
1698	IF ( ( section(is,s_ind) >= nys .AND. &
1699	section(is,s_ind) <= nyn ) .OR. &
1700	( section(is,s_ind) == -1 .AND. &
1701	nys-1 == -1 ) ) &
1702	THEN
1703	WRITE (22) nxl, nxr, nzb_do, nzt_do, nzb, nzt+1
1704	WRITE (22) local_2d
1705	ELSE
1706	WRITE (22) -1, -1, -1, -1, -1, -1
1707	ENDIF
1708	ENDIF
1709	#if defined( __parallel )
1710	CALL MPI_BARRIER( comm2d, ierr )
1711	#endif
1712	ENDDO
1713
1714	ELSE
1715	!
1716	!-- PE0 receives partial arrays from all processors of the
1717	!-- respective cross section and outputs them. Here a
1718	!-- barrier has to be set, because otherwise
1719	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
1720	CALL MPI_BARRIER( comm2d, ierr )
1721
1722	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
1723	IF ( myid == 0 ) THEN
1724	!
1725	!-- Local array can be relocated directly.
1726	IF ( ( section(is,s_ind) >= nys .AND. &
1727	section(is,s_ind) <= nyn ) .OR. &
1728	( section(is,s_ind) == -1 .AND. &
1729	nys-1 == -1 ) ) THEN
1730	total_2d(nxl:nxr,nzb_do:nzt_do) = local_2d
1731	ENDIF
1732	!
1733	!-- Receive data from all other PEs.
1734	DO n = 1, numprocs-1
1735	!
1736	!-- Receive index limits first, then array.
1737	!-- Index limits are received in arbitrary order from
1738	!-- the PEs.
1739	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1740	MPI_ANY_SOURCE, 0, comm2d, &
1741	status, ierr )
1742	!
1743	!-- Not all PEs have data for XZ-cross-section.
1744	IF ( ind(1) /= -9999 ) THEN
1745	sender = status(MPI_SOURCE)
1746	DEALLOCATE( local_2d )
1747	ALLOCATE( local_2d(ind(1):ind(2), &
1748	ind(3):ind(4)) )
1749	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1750	MPI_REAL, sender, 1, comm2d, &
1751	status, ierr )
1752	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
1753	local_2d
1754	ENDIF
1755	ENDDO
1756	!
1757	!-- Relocate the local array for the next loop increment
1758	DEALLOCATE( local_2d )
1759	ALLOCATE( local_2d(nxl:nxr,nzb_do:nzt_do) )
1760
1761	#if defined( __netcdf )
1762	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1763	id_var_do2d(av,if), &
1764	total_2d(0:nx,nzb_do:nzt_do), &
1765	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
1766	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
1767	CALL netcdf_handle_error( 'data_output_2d', 58 )
1768	#endif
1769
1770	ELSE
1771	!
1772	!-- If the cross section resides on the PE, send the
1773	!-- local index limits, otherwise send -9999 to PE0.
1774	IF ( ( section(is,s_ind) >= nys .AND. &
1775	section(is,s_ind) <= nyn ) .OR. &
1776	( section(is,s_ind) == -1 .AND. nys-1 == -1 ) ) &
1777	THEN
1778	ind(1) = nxl; ind(2) = nxr
1779	ind(3) = nzb_do; ind(4) = nzt_do
1780	ELSE
1781	ind(1) = -9999; ind(2) = -9999
1782	ind(3) = -9999; ind(4) = -9999
1783	ENDIF
1784	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1785	comm2d, ierr )
1786	!
1787	!-- If applicable, send data to PE0.
1788	IF ( ind(1) /= -9999 ) THEN
1789	CALL MPI_SEND( local_2d(nxl,nzb_do), ngp, &
1790	MPI_REAL, 0, 1, comm2d, ierr )
1791	ENDIF
1792	ENDIF
1793	!
1794	!-- A barrier has to be set, because otherwise some PEs may
1795	!-- proceed too fast so that PE0 may receive wrong data on
1796	!-- tag 0
1797	CALL MPI_BARRIER( comm2d, ierr )
1798	ENDIF
1799
1800	ENDIF
1801	#else
1802	#if defined( __netcdf )
1803	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1804	id_var_do2d(av,if), &
1805	local_2d(nxl:nxr,nzb_do:nzt_do), &
1806	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
1807	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
1808	CALL netcdf_handle_error( 'data_output_2d', 451 )
1809	#endif
1810	#endif
1811
1812	CASE ( 'yz' )
1813	!
1814	!-- Update the netCDF yz cross section time axis.
1815	!-- In case of parallel output, this is only done by PE0
1816	!-- to increase the performance.
1817	IF ( simulated_time /= do2d_yz_last_time(av) ) THEN
1818	do2d_yz_time_count(av) = do2d_yz_time_count(av) + 1
1819	do2d_yz_last_time(av) = simulated_time
1820	IF ( myid == 0 ) THEN
1821	IF ( .NOT. data_output_2d_on_each_pe &
1822	.OR. netcdf_data_format > 4 ) &
1823	THEN
1824	#if defined( __netcdf )
1825	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
1826	id_var_time_yz(av), &
1827	(/ time_since_reference_point /), &
1828	start = (/ do2d_yz_time_count(av) /), &
1829	count = (/ 1 /) )
1830	CALL netcdf_handle_error( 'data_output_2d', 59 )
1831	#endif
1832	ENDIF
1833	ENDIF
1834	ENDIF
1835
1836	!
1837	!-- If required, carry out averaging along x
1838	IF ( section(is,s_ind) == -1 ) THEN
1839
1840	ALLOCATE( local_2d_l(nys:nyn,nzb_do:nzt_do) )
1841	local_2d_l = 0.0_wp
1842	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
1843	!
1844	!-- First local averaging on the PE
1845	DO k = nzb_do, nzt_do
1846	DO j = nys, nyn
1847	DO i = nxl, nxr
1848	local_2d_l(j,k) = local_2d_l(j,k) + &
1849	local_pf(i,j,k)
1850	ENDDO
1851	ENDDO
1852	ENDDO
1853	#if defined( __parallel )
1854	!
1855	!-- Now do the averaging over all PEs along x
1856	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1857	CALL MPI_ALLREDUCE( local_2d_l(nys,nzb_do), &
1858	local_2d(nys,nzb_do), ngp, MPI_REAL, &
1859	MPI_SUM, comm1dx, ierr )
1860	#else
1861	local_2d = local_2d_l
1862	#endif
1863	local_2d = local_2d / ( nx + 1.0_wp )
1864
1865	DEALLOCATE( local_2d_l )
1866
1867	ELSE
1868	!
1869	!-- Just store the respective section on the local array
1870	!-- (but only if it is available on this PE!)
1871	IF ( section(is,s_ind) >= nxl .AND. section(is,s_ind) <= nxr ) &
1872	THEN
1873	local_2d = local_pf(section(is,s_ind),:,nzb_do:nzt_do)
1874	ENDIF
1875
1876	ENDIF
1877
1878	#if defined( __parallel )
1879	IF ( netcdf_data_format > 4 ) THEN
1880	!
1881	!-- Output in netCDF4/HDF5 format.
1882	!-- Output only on those PEs where the respective cross
1883	!-- sections reside. Cross sections averaged along x are
1884	!-- output on the respective first PE along x (myidx=0).
1885	IF ( ( section(is,s_ind) >= nxl .AND. &
1886	section(is,s_ind) <= nxr ) .OR. &
1887	( section(is,s_ind) == -1 .AND. myidx == 0 ) ) THEN
1888	#if defined( __netcdf )
1889	!
1890	!-- For parallel output, all cross sections are first
1891	!-- stored here on a local array and will be written to the
1892	!-- output file afterwards to increase the performance.
1893	DO j = nys, nyn
1894	DO k = nzb_do, nzt_do
1895	local_2d_sections_l(is,j,k) = local_2d(j,k)
1896	ENDDO
1897	ENDDO
1898	#endif
1899	ENDIF
1900
1901	ELSE
1902
1903	IF ( data_output_2d_on_each_pe ) THEN
1904	!
1905	!-- Output of partial arrays on each PE. If the cross
1906	!-- section does not reside on the PE, output special
1907	!-- index values.
1908	#if defined( __netcdf )
1909	IF ( myid == 0 ) THEN
1910	WRITE ( 23 ) time_since_reference_point, &
1911	do2d_yz_time_count(av), av
1912	ENDIF
1913	#endif
1914	DO i = 0, io_blocks-1
1915	IF ( i == io_group ) THEN
1916	IF ( ( section(is,s_ind) >= nxl .AND. &
1917	section(is,s_ind) <= nxr ) .OR. &
1918	( section(is,s_ind) == -1 .AND. &
1919	nxl-1 == -1 ) ) &
1920	THEN
1921	WRITE (23) nys, nyn, nzb_do, nzt_do, nzb, nzt+1
1922	WRITE (23) local_2d
1923	ELSE
1924	WRITE (23) -1, -1, -1, -1, -1, -1
1925	ENDIF
1926	ENDIF
1927	#if defined( __parallel )
1928	CALL MPI_BARRIER( comm2d, ierr )
1929	#endif
1930	ENDDO
1931
1932	ELSE
1933	!
1934	!-- PE0 receives partial arrays from all processors of the
1935	!-- respective cross section and outputs them. Here a
1936	!-- barrier has to be set, because otherwise
1937	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
1938	CALL MPI_BARRIER( comm2d, ierr )
1939
1940	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
1941	IF ( myid == 0 ) THEN
1942	!
1943	!-- Local array can be relocated directly.
1944	IF ( ( section(is,s_ind) >= nxl .AND. &
1945	section(is,s_ind) <= nxr ) .OR. &
1946	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
1947	THEN
1948	total_2d(nys:nyn,nzb_do:nzt_do) = local_2d
1949	ENDIF
1950	!
1951	!-- Receive data from all other PEs.
1952	DO n = 1, numprocs-1
1953	!
1954	!-- Receive index limits first, then array.
1955	!-- Index limits are received in arbitrary order from
1956	!-- the PEs.
1957	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1958	MPI_ANY_SOURCE, 0, comm2d, &
1959	status, ierr )
1960	!
1961	!-- Not all PEs have data for YZ-cross-section.
1962	IF ( ind(1) /= -9999 ) THEN
1963	sender = status(MPI_SOURCE)
1964	DEALLOCATE( local_2d )
1965	ALLOCATE( local_2d(ind(1):ind(2), &
1966	ind(3):ind(4)) )
1967	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1968	MPI_REAL, sender, 1, comm2d, &
1969	status, ierr )
1970	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
1971	local_2d
1972	ENDIF
1973	ENDDO
1974	!
1975	!-- Relocate the local array for the next loop increment
1976	DEALLOCATE( local_2d )
1977	ALLOCATE( local_2d(nys:nyn,nzb_do:nzt_do) )
1978
1979	#if defined( __netcdf )
1980	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
1981	id_var_do2d(av,if), &
1982	total_2d(0:ny,nzb_do:nzt_do), &
1983	start = (/ is, 1, 1, do2d_yz_time_count(av) /), &
1984	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
1985	CALL netcdf_handle_error( 'data_output_2d', 61 )
1986	#endif
1987
1988	ELSE
1989	!
1990	!-- If the cross section resides on the PE, send the
1991	!-- local index limits, otherwise send -9999 to PE0.
1992	IF ( ( section(is,s_ind) >= nxl .AND. &
1993	section(is,s_ind) <= nxr ) .OR. &
1994	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
1995	THEN
1996	ind(1) = nys; ind(2) = nyn
1997	ind(3) = nzb_do; ind(4) = nzt_do
1998	ELSE
1999	ind(1) = -9999; ind(2) = -9999
2000	ind(3) = -9999; ind(4) = -9999
2001	ENDIF
2002	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
2003	comm2d, ierr )
2004	!
2005	!-- If applicable, send data to PE0.
2006	IF ( ind(1) /= -9999 ) THEN
2007	CALL MPI_SEND( local_2d(nys,nzb_do), ngp, &
2008	MPI_REAL, 0, 1, comm2d, ierr )
2009	ENDIF
2010	ENDIF
2011	!
2012	!-- A barrier has to be set, because otherwise some PEs may
2013	!-- proceed too fast so that PE0 may receive wrong data on
2014	!-- tag 0
2015	CALL MPI_BARRIER( comm2d, ierr )
2016	ENDIF
2017
2018	ENDIF
2019	#else
2020	#if defined( __netcdf )
2021	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2022	id_var_do2d(av,if), &
2023	local_2d(nys:nyn,nzb_do:nzt_do), &
2024	start = (/ is, 1, 1, do2d_xz_time_count(av) /), &
2025	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
2026	CALL netcdf_handle_error( 'data_output_2d', 452 )
2027	#endif
2028	#endif
2029
2030	END SELECT
2031
2032	is = is + 1
2033	ENDDO loop1
2034
2035	!
2036	!-- For parallel output, all data were collected before on a local array
2037	!-- and are written now to the netcdf file. This must be done to increase
2038	!-- the performance of the parallel output.
2039	#if defined( __netcdf )
2040	IF ( netcdf_data_format > 4 ) THEN
2041
2042	SELECT CASE ( mode )
2043
2044	CASE ( 'xy' )
2045	IF ( two_d ) THEN
2046	nis = 1
2047	two_d = .FALSE.
2048	ELSE
2049	nis = ns
2050	ENDIF
2051	!
2052	!-- Do not output redundant ghost point data except for the
2053	!-- boundaries of the total domain.
2054	! IF ( nxr == nx .AND. nyn /= ny ) THEN
2055	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2056	! id_var_do2d(av,if), &
2057	! local_2d_sections(nxl:nxr+1, &
2058	! nys:nyn,1:nis), &
2059	! start = (/ nxl+1, nys+1, 1, &
2060	! do2d_xy_time_count(av) /), &
2061	! count = (/ nxr-nxl+2, &
2062	! nyn-nys+1, nis, 1 &
2063	! /) )
2064	! ELSEIF ( nxr /= nx .AND. nyn == ny ) THEN
2065	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2066	! id_var_do2d(av,if), &
2067	! local_2d_sections(nxl:nxr, &
2068	! nys:nyn+1,1:nis), &
2069	! start = (/ nxl+1, nys+1, 1, &
2070	! do2d_xy_time_count(av) /), &
2071	! count = (/ nxr-nxl+1, &
2072	! nyn-nys+2, nis, 1 &
2073	! /) )
2074	! ELSEIF ( nxr == nx .AND. nyn == ny ) THEN
2075	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2076	! id_var_do2d(av,if), &
2077	! local_2d_sections(nxl:nxr+1, &
2078	! nys:nyn+1,1:nis), &
2079	! start = (/ nxl+1, nys+1, 1, &
2080	! do2d_xy_time_count(av) /), &
2081	! count = (/ nxr-nxl+2, &
2082	! nyn-nys+2, nis, 1 &
2083	! /) )
2084	! ELSE
2085	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2086	id_var_do2d(av,if), &
2087	local_2d_sections(nxl:nxr, &
2088	nys:nyn,1:nis), &
2089	start = (/ nxl+1, nys+1, 1, &
2090	do2d_xy_time_count(av) /), &
2091	count = (/ nxr-nxl+1, &
2092	nyn-nys+1, nis, 1 &
2093	/) )
2094	! ENDIF
2095
2096	CALL netcdf_handle_error( 'data_output_2d', 55 )
2097
2098	CASE ( 'xz' )
2099	!
2100	!-- First, all PEs get the information of all cross-sections.
2101	!-- Then the data are written to the output file by all PEs
2102	!-- while NF90_COLLECTIVE is set in subroutine
2103	!-- define_netcdf_header. Although redundant information are
2104	!-- written to the output file in that case, the performance
2105	!-- is significantly better compared to the case where only
2106	!-- the first row of PEs in x-direction (myidx = 0) is given
2107	!-- the output while NF90_INDEPENDENT is set.
2108	IF ( npey /= 1 ) THEN
2109
2110	#if defined( __parallel )
2111	!
2112	!-- Distribute data over all PEs along y
2113	ngp = ( nxr-nxl+1 ) * ( nzt_do-nzb_do+1 ) * ns
2114	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
2115	CALL MPI_ALLREDUCE( local_2d_sections_l(nxl,1,nzb_do), &
2116	local_2d_sections(nxl,1,nzb_do), &
2117	ngp, MPI_REAL, MPI_SUM, comm1dy, &
2118	ierr )
2119	#else
2120	local_2d_sections = local_2d_sections_l
2121	#endif
2122	ENDIF
2123	!
2124	!-- Do not output redundant ghost point data except for the
2125	!-- boundaries of the total domain.
2126	! IF ( nxr == nx ) THEN
2127	! nc_stat = NF90_PUT_VAR( id_set_xz(av), &
2128	! id_var_do2d(av,if), &
2129	! local_2d_sections(nxl:nxr+1,1:ns, &
2130	! nzb_do:nzt_do), &
2131	! start = (/ nxl+1, 1, 1, &
2132	! do2d_xz_time_count(av) /), &
2133	! count = (/ nxr-nxl+2, ns, nzt_do-nzb_do+1, &
2134	! 1 /) )
2135	! ELSE
2136	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
2137	id_var_do2d(av,if), &
2138	local_2d_sections(nxl:nxr,1:ns, &
2139	nzb_do:nzt_do), &
2140	start = (/ nxl+1, 1, 1, &
2141	do2d_xz_time_count(av) /), &
2142	count = (/ nxr-nxl+1, ns, nzt_do-nzb_do+1, &
2143	1 /) )
2144	! ENDIF
2145
2146	CALL netcdf_handle_error( 'data_output_2d', 57 )
2147
2148	CASE ( 'yz' )
2149	!
2150	!-- First, all PEs get the information of all cross-sections.
2151	!-- Then the data are written to the output file by all PEs
2152	!-- while NF90_COLLECTIVE is set in subroutine
2153	!-- define_netcdf_header. Although redundant information are
2154	!-- written to the output file in that case, the performance
2155	!-- is significantly better compared to the case where only
2156	!-- the first row of PEs in y-direction (myidy = 0) is given
2157	!-- the output while NF90_INDEPENDENT is set.
2158	IF ( npex /= 1 ) THEN
2159
2160	#if defined( __parallel )
2161	!
2162	!-- Distribute data over all PEs along x
2163	ngp = ( nyn-nys+1 ) * ( nzt-nzb + 2 ) * ns
2164	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
2165	CALL MPI_ALLREDUCE( local_2d_sections_l(1,nys,nzb_do), &
2166	local_2d_sections(1,nys,nzb_do), &
2167	ngp, MPI_REAL, MPI_SUM, comm1dx, &
2168	ierr )
2169	#else
2170	local_2d_sections = local_2d_sections_l
2171	#endif
2172	ENDIF
2173	!
2174	!-- Do not output redundant ghost point data except for the
2175	!-- boundaries of the total domain.
2176	! IF ( nyn == ny ) THEN
2177	! nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2178	! id_var_do2d(av,if), &
2179	! local_2d_sections(1:ns, &
2180	! nys:nyn+1,nzb_do:nzt_do), &
2181	! start = (/ 1, nys+1, 1, &
2182	! do2d_yz_time_count(av) /), &
2183	! count = (/ ns, nyn-nys+2, &
2184	! nzt_do-nzb_do+1, 1 /) )
2185	! ELSE
2186	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2187	id_var_do2d(av,if), &
2188	local_2d_sections(1:ns,nys:nyn, &
2189	nzb_do:nzt_do), &
2190	start = (/ 1, nys+1, 1, &
2191	do2d_yz_time_count(av) /), &
2192	count = (/ ns, nyn-nys+1, &
2193	nzt_do-nzb_do+1, 1 /) )
2194	! ENDIF
2195
2196	CALL netcdf_handle_error( 'data_output_2d', 60 )
2197
2198	CASE DEFAULT
2199	message_string = 'unknown cross-section: ' // TRIM( mode )
2200	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
2201
2202	END SELECT
2203
2204	ENDIF
2205	#endif
2206	ENDIF
2207
2208	if = if + 1
2209	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
2210	do2d_mode = do2d(av,if)(l-1:l)
2211
2212	ENDDO
2213
2214	!
2215	!-- Deallocate temporary arrays.
2216	IF ( ALLOCATED( level_z ) ) DEALLOCATE( level_z )
2217	IF ( netcdf_data_format > 4 ) THEN
2218	DEALLOCATE( local_pf, local_2d, local_2d_sections )
2219	IF( mode == 'xz' .OR. mode == 'yz' ) DEALLOCATE( local_2d_sections_l )
2220	ENDIF
2221	#if defined( __parallel )
2222	IF ( .NOT. data_output_2d_on_each_pe .AND. myid == 0 ) THEN
2223	DEALLOCATE( total_2d )
2224	ENDIF
2225	#endif
2226
2227	!
2228	!-- Close plot output file.
2229	file_id = 20 + s_ind
2230
2231	IF ( data_output_2d_on_each_pe ) THEN
2232	DO i = 0, io_blocks-1
2233	IF ( i == io_group ) THEN
2234	CALL close_file( file_id )
2235	ENDIF
2236	#if defined( __parallel )
2237	CALL MPI_BARRIER( comm2d, ierr )
2238	#endif
2239	ENDDO
2240	ELSE
2241	IF ( myid == 0 ) CALL close_file( file_id )
2242	ENDIF
2243
2244	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
2245
2246	END SUBROUTINE data_output_2d

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |