Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 3541

Last change on this file since 3541 was 3525, checked in by kanani, 6 years ago
Changes related to clean-up of biometeorology (by dom_dwd_user)
Property svn:keywords set to `Id`
File size: 95.7 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 3525 2018-11-14 16:06:14Z gronemeier $
27	! Changes related to clean-up of biometeorology (dom_dwd_user)
28	!
29	! 3467 2018-10-30 19:05:21Z suehring
30	! Implementation of a new aerosol module salsa.
31	!
32	! 3448 2018-10-29 18:14:31Z kanani
33	! Add biometeorology
34	!
35	! 3421 2018-10-24 18:39:32Z gronemeier
36	! Renamed output variables
37	!
38	! 3419 2018-10-24 17:27:31Z gronemeier
39	! minor formatting (kanani)
40	! chem_data_output_2d subroutine added (basit)
41	!
42	! 3294 2018-10-01 02:37:10Z raasch
43	! changes concerning modularization of ocean option
44	!
45	! 3274 2018-09-24 15:42:55Z knoop
46	! Modularization of all bulk cloud physics code components
47	!
48	! 3241 2018-09-12 15:02:00Z raasch
49	! unused variables removed
50	!
51	! 3176 2018-07-26 17:12:48Z suehring
52	! Remove output of latent heat flux at urban-surfaces and set fill values
53	! instead
54	!
55	! 3052 2018-05-31 06:11:20Z raasch
56	! Do not open FORTRAN binary files in case of parallel netCDF I/O
57	!
58	! 3049 2018-05-29 13:52:36Z Giersch
59	! Error messages revised
60	!
61	! 3045 2018-05-28 07:55:41Z Giersch
62	! Error messages revised
63	!
64	! 3014 2018-05-09 08:42:38Z maronga
65	! Added nzb_do and nzt_do for some modules for 2d output
66	!
67	! 3004 2018-04-27 12:33:25Z Giersch
68	! precipitation_rate removed, case prr*_xy removed, to_be_resorted have to point
69	! to ql_vp_av and not to ql_vp, allocation checks implemented (averaged data
70	! will be assigned to fill values if no allocation happened so far)
71	!
72	! 2963 2018-04-12 14:47:44Z suehring
73	! Introduce index for vegetation/wall, pavement/green-wall and water/window
74	! surfaces, for clearer access of surface fraction, albedo, emissivity, etc. .
75	!
76	! 2817 2018-02-19 16:32:21Z knoop
77	! Preliminary gust module interface implemented
78	!
79	! 2805 2018-02-14 17:00:09Z suehring
80	! Consider also default-type surfaces for surface temperature output.
81	!
82	! 2797 2018-02-08 13:24:35Z suehring
83	! Enable output of ground-heat flux also at urban surfaces.
84	!
85	! 2743 2018-01-12 16:03:39Z suehring
86	! In case of natural- and urban-type surfaces output surfaces fluxes in W/m2.
87	!
88	! 2742 2018-01-12 14:59:47Z suehring
89	! Enable output of surface temperature
90	!
91	! 2735 2018-01-11 12:01:27Z suehring
92	! output of r_a moved from land-surface to consider also urban-type surfaces
93	!
94	! 2718 2018-01-02 08:49:38Z maronga
95	! Corrected "Former revisions" section
96	!
97	! 2696 2017-12-14 17:12:51Z kanani
98	! Change in file header (GPL part)
99	! Implementation of uv exposure model (FK)
100	! Implementation of turbulence_closure_mod (TG)
101	! Set fill values at topography grid points or e.g. non-natural-type surface
102	! in case of LSM output (MS)
103	!
104	! 2512 2017-10-04 08:26:59Z raasch
105	! upper bounds of cross section output changed from nx+1,ny+1 to nx,ny
106	! no output of ghost layer data
107	!
108	! 2292 2017-06-20 09:51:42Z schwenkel
109	! Implementation of new microphysic scheme: cloud_scheme = 'morrison'
110	! includes two more prognostic equations for cloud drop concentration (nc)
111	! and cloud water content (qc).
112	!
113	! 2277 2017-06-12 10:47:51Z kanani
114	! Removed unused variables do2d_xy_n, do2d_xz_n, do2d_yz_n
115	!
116	! 2233 2017-05-30 18:08:54Z suehring
117	!
118	! 2232 2017-05-30 17:47:52Z suehring
119	! Adjustments to new surface concept
120	!
121	!
122	! 2190 2017-03-21 12:16:43Z raasch
123	! bugfix for r2031: string rho replaced by rho_ocean
124	!
125	! 2031 2016-10-21 15:11:58Z knoop
126	! renamed variable rho to rho_ocean and rho_av to rho_ocean_av
127	!
128	! 2000 2016-08-20 18:09:15Z knoop
129	! Forced header and separation lines into 80 columns
130	!
131	! 1980 2016-07-29 15:51:57Z suehring
132	! Bugfix, in order to steer user-defined output, setting flag found explicitly
133	! to .F.
134	!
135	! 1976 2016-07-27 13:28:04Z maronga
136	! Output of radiation quantities is now done directly in the respective module
137	!
138	! 1972 2016-07-26 07:52:02Z maronga
139	! Output of land surface quantities is now done directly in the respective
140	! module
141	!
142	! 1960 2016-07-12 16:34:24Z suehring
143	! Scalar surface flux added
144	! Rename INTEGER variable s into s_ind, as s is already assigned to scalar
145	!
146	! 1849 2016-04-08 11:33:18Z hoffmann
147	! precipitation_amount, precipitation_rate, prr moved to arrays_3d
148	!
149	! 1822 2016-04-07 07:49:42Z hoffmann
150	! Output of bulk cloud physics simplified.
151	!
152	! 1788 2016-03-10 11:01:04Z maronga
153	! Added output of z0q
154	!
155	! 1783 2016-03-06 18:36:17Z raasch
156	! name change of netcdf routines and module + related changes
157	!
158	! 1745 2016-02-05 13:06:51Z gronemeier
159	! Bugfix: test if time axis limit exceeds moved to point after call of check_open
160	!
161	! 1703 2015-11-02 12:38:44Z raasch
162	! bugfix for output of single (*) xy-sections in case of parallel netcdf I/O
163	!
164	! 1701 2015-11-02 07:43:04Z maronga
165	! Bugfix in output of RRTGM data
166	!
167	! 1691 2015-10-26 16:17:44Z maronga
168	! Added output of Obukhov length (ol) and radiative heating rates for RRTMG.
169	! Formatting corrections.
170	!
171	! 1682 2015-10-07 23:56:08Z knoop
172	! Code annotations made doxygen readable
173	!
174	! 1585 2015-04-30 07:05:52Z maronga
175	! Added support for RRTMG
176	!
177	! 1555 2015-03-04 17:44:27Z maronga
178	! Added output of r_a and r_s
179	!
180	! 1551 2015-03-03 14:18:16Z maronga
181	! Added suppport for land surface model and radiation model output. In the course
182	! of this action, the limits for vertical loops have been changed (from nzb and
183	! nzt+1 to nzb_do and nzt_do, respectively in order to allow soil model output).
184	! Moreover, a new vertical grid zs was introduced.
185	!
186	! 1359 2014-04-11 17:15:14Z hoffmann
187	! New particle structure integrated.
188	!
189	! 1353 2014-04-08 15:21:23Z heinze
190	! REAL constants provided with KIND-attribute
191	!
192	! 1327 2014-03-21 11:00:16Z raasch
193	! parts concerning iso2d output removed,
194	! -netcdf output queries
195	!
196	! 1320 2014-03-20 08:40:49Z raasch
197	! ONLY-attribute added to USE-statements,
198	! kind-parameters added to all INTEGER and REAL declaration statements,
199	! kinds are defined in new module kinds,
200	! revision history before 2012 removed,
201	! comment fields (!:) to be used for variable explanations added to
202	! all variable declaration statements
203	!
204	! 1318 2014-03-17 13:35:16Z raasch
205	! barrier argument removed from cpu_log.
206	! module interfaces removed
207	!
208	! 1311 2014-03-14 12:13:39Z heinze
209	! bugfix: close #if defined( __netcdf )
210	!
211	! 1308 2014-03-13 14:58:42Z fricke
212	! +local_2d_sections, local_2d_sections_l, ns
213	! Check, if the limit of the time dimension is exceeded for parallel output
214	! To increase the performance for parallel output, the following is done:
215	! - Update of time axis is only done by PE0
216	! - Cross sections are first stored on a local array and are written
217	! collectively to the output file by all PEs.
218	!
219	! 1115 2013-03-26 18:16:16Z hoffmann
220	! ql is calculated by calc_liquid_water_content
221	!
222	! 1076 2012-12-05 08:30:18Z hoffmann
223	! Bugfix in output of ql
224	!
225	! 1065 2012-11-22 17:42:36Z hoffmann
226	! Bugfix: Output of cross sections of ql
227	!
228	! 1053 2012-11-13 17:11:03Z hoffmann
229	! +qr, nr, qc and cross sections
230	!
231	! 1036 2012-10-22 13:43:42Z raasch
232	! code put under GPL (PALM 3.9)
233	!
234	! 1031 2012-10-19 14:35:30Z raasch
235	! netCDF4 without parallel file support implemented
236	!
237	! 1007 2012-09-19 14:30:36Z franke
238	! Bugfix: missing calculation of ql_vp added
239	!
240	! 978 2012-08-09 08:28:32Z fricke
241	! +z0h
242	!
243	! Revision 1.1 1997/08/11 06:24:09 raasch
244	! Initial revision
245	!
246	!
247	! Description:
248	! ------------
249	!> Data output of cross-sections in netCDF format or binary format
250	!> to be later converted to NetCDF by helper routine combine_plot_fields.
251	!> Attention: The position of the sectional planes is still not always computed
252	!> --------- correctly. (zu is used always)!
253	!------------------------------------------------------------------------------!
254	SUBROUTINE data_output_2d( mode, av )
255
256
257	USE arrays_3d, &
258	ONLY: dzw, e, heatflux_output_conversion, nc, nr, p, pt, &
259	precipitation_amount, prr, q, qc, ql, ql_c, ql_v, qr, s, tend, &
260	u, v, vpt, w, zu, zw, waterflux_output_conversion, hyrho, d_exner
261
262	USE averaging
263
264	USE basic_constants_and_equations_mod, &
265	ONLY: c_p, lv_d_cp, l_v
266
267	USE biometeorology_mod, &
268	ONLY: bio_data_output_2d
269
270	USE bulk_cloud_model_mod, &
271	ONLY: bulk_cloud_model, bcm_data_output_2d
272
273	USE chemistry_model_mod, &
274	ONLY: chem_data_output_2d
275
276	USE control_parameters, &
277	ONLY: air_chemistry, biometeorology, data_output_2d_on_each_pe, &
278	data_output_xy, data_output_xz, data_output_yz, do2d, &
279	do2d_xy_last_time, do2d_xy_time_count, &
280	do2d_xz_last_time, do2d_xz_time_count, &
281	do2d_yz_last_time, do2d_yz_time_count, &
282	ibc_uv_b, io_blocks, io_group, land_surface, message_string, &
283	ntdim_2d_xy, ntdim_2d_xz, ntdim_2d_yz, &
284	ocean_mode, psolver, section, simulated_time, &
285	time_since_reference_point, uv_exposure
286
287	USE cpulog, &
288	ONLY: cpu_log, log_point
289
290	USE gust_mod, &
291	ONLY: gust_data_output_2d, gust_module_enabled
292
293	USE indices, &
294	ONLY: nbgp, nx, nxl, nxlg, nxr, nxrg, ny, nyn, nyng, nys, nysg, &
295	nzb, nzt, wall_flags_0
296
297	USE kinds
298
299	USE land_surface_model_mod, &
300	ONLY: lsm_data_output_2d, zs
301
302	#if defined( __netcdf )
303	USE NETCDF
304	#endif
305
306	USE netcdf_interface, &
307	ONLY: fill_value, id_set_xy, id_set_xz, id_set_yz, id_var_do2d, &
308	id_var_time_xy, id_var_time_xz, id_var_time_yz, nc_stat, &
309	netcdf_data_format, netcdf_handle_error
310
311	USE ocean_mod, &
312	ONLY: ocean_data_output_2d
313
314	USE particle_attributes, &
315	ONLY: grid_particles, number_of_particles, particle_advection_start, &
316	particles, prt_count
317
318	USE pegrid
319
320	USE radiation_model_mod, &
321	ONLY: radiation, radiation_data_output_2d
322
323	USE salsa_mod, &
324	ONLY: salsa, salsa_data_output_2d
325
326	USE surface_mod, &
327	ONLY: ind_pav_green, ind_veg_wall, ind_wat_win, surf_def_h, &
328	surf_lsm_h, surf_usm_h
329
330	USE turbulence_closure_mod, &
331	ONLY: tcm_data_output_2d
332
333	USE uv_exposure_model_mod, &
334	ONLY: uvem_data_output_2d
335
336
337	IMPLICIT NONE
338
339	CHARACTER (LEN=2) :: do2d_mode !<
340	CHARACTER (LEN=2) :: mode !<
341	CHARACTER (LEN=4) :: grid !<
342	CHARACTER (LEN=50) :: rtext !<
343
344	INTEGER(iwp) :: av !<
345	INTEGER(iwp) :: ngp !<
346	INTEGER(iwp) :: file_id !<
347	INTEGER(iwp) :: flag_nr !< number of masking flag
348	INTEGER(iwp) :: i !<
349	INTEGER(iwp) :: if !<
350	INTEGER(iwp) :: is !<
351	INTEGER(iwp) :: iis !<
352	INTEGER(iwp) :: j !<
353	INTEGER(iwp) :: k !<
354	INTEGER(iwp) :: l !<
355	INTEGER(iwp) :: layer_xy !<
356	INTEGER(iwp) :: m !<
357	INTEGER(iwp) :: n !<
358	INTEGER(iwp) :: nis !<
359	INTEGER(iwp) :: ns !<
360	INTEGER(iwp) :: nzb_do !< lower limit of the data field (usually nzb)
361	INTEGER(iwp) :: nzt_do !< upper limit of the data field (usually nzt+1)
362	INTEGER(iwp) :: s_ind !<
363	INTEGER(iwp) :: sender !<
364	INTEGER(iwp) :: ind(4) !<
365
366	LOGICAL :: found !<
367	LOGICAL :: resorted !<
368	LOGICAL :: two_d !<
369
370	REAL(wp) :: mean_r !<
371	REAL(wp) :: s_r2 !<
372	REAL(wp) :: s_r3 !<
373
374	REAL(wp), DIMENSION(:), ALLOCATABLE :: level_z !<
375	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d !<
376	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d_l !<
377
378	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !<
379	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections !<
380	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections_l !<
381
382	#if defined( __parallel )
383	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: total_2d !<
384	#endif
385	REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !<
386
387	NAMELIST /LOCAL/ rtext
388
389	!
390	!-- Immediate return, if no output is requested (no respective sections
391	!-- found in parameter data_output)
392	IF ( mode == 'xy' .AND. .NOT. data_output_xy(av) ) RETURN
393	IF ( mode == 'xz' .AND. .NOT. data_output_xz(av) ) RETURN
394	IF ( mode == 'yz' .AND. .NOT. data_output_yz(av) ) RETURN
395
396	CALL cpu_log (log_point(3),'data_output_2d','start')
397
398	two_d = .FALSE. ! local variable to distinguish between output of pure 2D
399	! arrays and cross-sections of 3D arrays.
400
401	!
402	!-- Depending on the orientation of the cross-section, the respective output
403	!-- files have to be opened.
404	SELECT CASE ( mode )
405
406	CASE ( 'xy' )
407	s_ind = 1
408	ALLOCATE( level_z(nzb:nzt+1), local_2d(nxl:nxr,nys:nyn) )
409
410	IF ( netcdf_data_format > 4 ) THEN
411	ns = 1
412	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
413	ns = ns + 1
414	ENDDO
415	ns = ns - 1
416	ALLOCATE( local_2d_sections(nxl:nxr,nys:nyn,1:ns) )
417	local_2d_sections = 0.0_wp
418	ENDIF
419
420	!
421	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
422	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
423	CALL check_open( 101+av*10 )
424	ENDIF
425	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
426	CALL check_open( 21 )
427	ELSE
428	IF ( myid == 0 ) THEN
429	#if defined( __parallel )
430	ALLOCATE( total_2d(0:nx,0:ny) )
431	#endif
432	ENDIF
433	ENDIF
434
435	CASE ( 'xz' )
436	s_ind = 2
437	ALLOCATE( local_2d(nxl:nxr,nzb:nzt+1) )
438
439	IF ( netcdf_data_format > 4 ) THEN
440	ns = 1
441	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
442	ns = ns + 1
443	ENDDO
444	ns = ns - 1
445	ALLOCATE( local_2d_sections(nxl:nxr,1:ns,nzb:nzt+1) )
446	ALLOCATE( local_2d_sections_l(nxl:nxr,1:ns,nzb:nzt+1) )
447	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
448	ENDIF
449
450	!
451	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
452	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
453	CALL check_open( 102+av*10 )
454	ENDIF
455
456	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
457	CALL check_open( 22 )
458	ELSE
459	IF ( myid == 0 ) THEN
460	#if defined( __parallel )
461	ALLOCATE( total_2d(0:nx,nzb:nzt+1) )
462	#endif
463	ENDIF
464	ENDIF
465
466	CASE ( 'yz' )
467	s_ind = 3
468	ALLOCATE( local_2d(nys:nyn,nzb:nzt+1) )
469
470	IF ( netcdf_data_format > 4 ) THEN
471	ns = 1
472	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
473	ns = ns + 1
474	ENDDO
475	ns = ns - 1
476	ALLOCATE( local_2d_sections(1:ns,nys:nyn,nzb:nzt+1) )
477	ALLOCATE( local_2d_sections_l(1:ns,nys:nyn,nzb:nzt+1) )
478	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
479	ENDIF
480
481	!
482	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
483	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
484	CALL check_open( 103+av*10 )
485	ENDIF
486
487	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
488	CALL check_open( 23 )
489	ELSE
490	IF ( myid == 0 ) THEN
491	#if defined( __parallel )
492	ALLOCATE( total_2d(0:ny,nzb:nzt+1) )
493	#endif
494	ENDIF
495	ENDIF
496
497	CASE DEFAULT
498	message_string = 'unknown cross-section: ' // TRIM( mode )
499	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
500
501	END SELECT
502
503	!
504	!-- For parallel netcdf output the time axis must be limited. Return, if this
505	!-- limit is exceeded. This could be the case, if the simulated time exceeds
506	!-- the given end time by the length of the given output interval.
507	IF ( netcdf_data_format > 4 ) THEN
508	IF ( mode == 'xy' .AND. do2d_xy_time_count(av) + 1 > &
509	ntdim_2d_xy(av) ) THEN
510	WRITE ( message_string, * ) 'Output of xy 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', 'PA0384', 0, 1, 0, 6, 0 )
514	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
515	RETURN
516	ENDIF
517	IF ( mode == 'xz' .AND. do2d_xz_time_count(av) + 1 > &
518	ntdim_2d_xz(av) ) THEN
519	WRITE ( message_string, * ) 'Output of xz cross-sections is not ', &
520	'given at t=', simulated_time, '&because the', &
521	' maximum number of output time levels is exceeded.'
522	CALL message( 'data_output_2d', 'PA0385', 0, 1, 0, 6, 0 )
523	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
524	RETURN
525	ENDIF
526	IF ( mode == 'yz' .AND. do2d_yz_time_count(av) + 1 > &
527	ntdim_2d_yz(av) ) THEN
528	WRITE ( message_string, * ) 'Output of yz cross-sections is not ', &
529	'given at t=', simulated_time, '&because the', &
530	' maximum number of output time levels is exceeded.'
531	CALL message( 'data_output_2d', 'PA0386', 0, 1, 0, 6, 0 )
532	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
533	RETURN
534	ENDIF
535	ENDIF
536
537	!
538	!-- Allocate a temporary array for resorting (kji -> ijk).
539	ALLOCATE( local_pf(nxl:nxr,nys:nyn,nzb:nzt+1) )
540	local_pf = 0.0
541
542	!
543	!-- Loop of all variables to be written.
544	!-- Output dimensions chosen
545	if = 1
546	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
547	do2d_mode = do2d(av,if)(l-1:l)
548
549	DO WHILE ( do2d(av,if)(1:1) /= ' ' )
550
551	IF ( do2d_mode == mode ) THEN
552	!
553	!-- Set flag to steer output of radiation, land-surface, or user-defined
554	!-- quantities
555	found = .FALSE.
556
557	nzb_do = nzb
558	nzt_do = nzt+1
559	!
560	!-- Before each output, set array local_pf to fill value
561	local_pf = fill_value
562	!
563	!-- Set masking flag for topography for not resorted arrays
564	flag_nr = 0
565
566	!
567	!-- Store the array chosen on the temporary array.
568	resorted = .FALSE.
569	SELECT CASE ( TRIM( do2d(av,if) ) )
570	CASE ( 'e_xy', 'e_xz', 'e_yz' )
571	IF ( av == 0 ) THEN
572	to_be_resorted => e
573	ELSE
574	IF ( .NOT. ALLOCATED( e_av ) ) THEN
575	ALLOCATE( e_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
576	e_av = REAL( fill_value, KIND = wp )
577	ENDIF
578	to_be_resorted => e_av
579	ENDIF
580	IF ( mode == 'xy' ) level_z = zu
581
582	CASE ( 'thetal_xy', 'thetal_xz', 'thetal_yz' )
583	IF ( av == 0 ) THEN
584	to_be_resorted => pt
585	ELSE
586	IF ( .NOT. ALLOCATED( lpt_av ) ) THEN
587	ALLOCATE( lpt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
588	lpt_av = REAL( fill_value, KIND = wp )
589	ENDIF
590	to_be_resorted => lpt_av
591	ENDIF
592	IF ( mode == 'xy' ) level_z = zu
593
594	CASE ( 'lwp*_xy' ) ! 2d-array
595	IF ( av == 0 ) THEN
596	DO i = nxl, nxr
597	DO j = nys, nyn
598	local_pf(i,j,nzb+1) = SUM( ql(nzb:nzt,j,i) * &
599	dzw(1:nzt+1) )
600	ENDDO
601	ENDDO
602	ELSE
603	IF ( .NOT. ALLOCATED( lwp_av ) ) THEN
604	ALLOCATE( lwp_av(nysg:nyng,nxlg:nxrg) )
605	lwp_av = REAL( fill_value, KIND = wp )
606	ENDIF
607	DO i = nxl, nxr
608	DO j = nys, nyn
609	local_pf(i,j,nzb+1) = lwp_av(j,i)
610	ENDDO
611	ENDDO
612	ENDIF
613	resorted = .TRUE.
614	two_d = .TRUE.
615	level_z(nzb+1) = zu(nzb+1)
616
617	CASE ( 'ghf*_xy' ) ! 2d-array
618	IF ( av == 0 ) THEN
619	DO m = 1, surf_lsm_h%ns
620	i = surf_lsm_h%i(m)
621	j = surf_lsm_h%j(m)
622	local_pf(i,j,nzb+1) = surf_lsm_h%ghf(m)
623	ENDDO
624	DO m = 1, surf_usm_h%ns
625	i = surf_usm_h%i(m)
626	j = surf_usm_h%j(m)
627	local_pf(i,j,nzb+1) = surf_usm_h%frac(ind_veg_wall,m) * &
628	surf_usm_h%wghf_eb(m) + &
629	surf_usm_h%frac(ind_pav_green,m) * &
630	surf_usm_h%wghf_eb_green(m) + &
631	surf_usm_h%frac(ind_wat_win,m) * &
632	surf_usm_h%wghf_eb_window(m)
633	ENDDO
634	ELSE
635	IF ( .NOT. ALLOCATED( ghf_av ) ) THEN
636	ALLOCATE( ghf_av(nysg:nyng,nxlg:nxrg) )
637	ghf_av = REAL( fill_value, KIND = wp )
638	ENDIF
639	DO i = nxl, nxr
640	DO j = nys, nyn
641	local_pf(i,j,nzb+1) = ghf_av(j,i)
642	ENDDO
643	ENDDO
644	ENDIF
645
646	resorted = .TRUE.
647	two_d = .TRUE.
648	level_z(nzb+1) = zu(nzb+1)
649
650	CASE ( 'ol*_xy' ) ! 2d-array
651	IF ( av == 0 ) THEN
652	DO m = 1, surf_def_h(0)%ns
653	i = surf_def_h(0)%i(m)
654	j = surf_def_h(0)%j(m)
655	local_pf(i,j,nzb+1) = surf_def_h(0)%ol(m)
656	ENDDO
657	DO m = 1, surf_lsm_h%ns
658	i = surf_lsm_h%i(m)
659	j = surf_lsm_h%j(m)
660	local_pf(i,j,nzb+1) = surf_lsm_h%ol(m)
661	ENDDO
662	DO m = 1, surf_usm_h%ns
663	i = surf_usm_h%i(m)
664	j = surf_usm_h%j(m)
665	local_pf(i,j,nzb+1) = surf_usm_h%ol(m)
666	ENDDO
667	ELSE
668	IF ( .NOT. ALLOCATED( ol_av ) ) THEN
669	ALLOCATE( ol_av(nysg:nyng,nxlg:nxrg) )
670	ol_av = REAL( fill_value, KIND = wp )
671	ENDIF
672	DO i = nxl, nxr
673	DO j = nys, nyn
674	local_pf(i,j,nzb+1) = ol_av(j,i)
675	ENDDO
676	ENDDO
677	ENDIF
678	resorted = .TRUE.
679	two_d = .TRUE.
680	level_z(nzb+1) = zu(nzb+1)
681
682	CASE ( 'p_xy', 'p_xz', 'p_yz' )
683	IF ( av == 0 ) THEN
684	IF ( psolver /= 'sor' ) CALL exchange_horiz( p, nbgp )
685	to_be_resorted => p
686	ELSE
687	IF ( .NOT. ALLOCATED( p_av ) ) THEN
688	ALLOCATE( p_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
689	p_av = REAL( fill_value, KIND = wp )
690	ENDIF
691	IF ( psolver /= 'sor' ) CALL exchange_horiz( p_av, nbgp )
692	to_be_resorted => p_av
693	ENDIF
694	IF ( mode == 'xy' ) level_z = zu
695
696	CASE ( 'pc_xy', 'pc_xz', 'pc_yz' ) ! particle concentration
697	IF ( av == 0 ) THEN
698	IF ( simulated_time >= particle_advection_start ) THEN
699	tend = prt_count
700	! CALL exchange_horiz( tend, nbgp )
701	ELSE
702	tend = 0.0_wp
703	ENDIF
704	DO i = nxl, nxr
705	DO j = nys, nyn
706	DO k = nzb, nzt+1
707	local_pf(i,j,k) = tend(k,j,i)
708	ENDDO
709	ENDDO
710	ENDDO
711	resorted = .TRUE.
712	ELSE
713	IF ( .NOT. ALLOCATED( pc_av ) ) THEN
714	ALLOCATE( pc_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
715	pc_av = REAL( fill_value, KIND = wp )
716	ENDIF
717	! CALL exchange_horiz( pc_av, nbgp )
718	to_be_resorted => pc_av
719	ENDIF
720
721	CASE ( 'pr_xy', 'pr_xz', 'pr_yz' ) ! mean particle radius (effective radius)
722	IF ( av == 0 ) THEN
723	IF ( simulated_time >= particle_advection_start ) THEN
724	DO i = nxl, nxr
725	DO j = nys, nyn
726	DO k = nzb, nzt+1
727	number_of_particles = prt_count(k,j,i)
728	IF (number_of_particles <= 0) CYCLE
729	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
730	s_r2 = 0.0_wp
731	s_r3 = 0.0_wp
732	DO n = 1, number_of_particles
733	IF ( particles(n)%particle_mask ) THEN
734	s_r2 = s_r2 + particles(n)%radius*2 &
735	particles(n)%weight_factor
736	s_r3 = s_r3 + particles(n)%radius*3 &
737	particles(n)%weight_factor
738	ENDIF
739	ENDDO
740	IF ( s_r2 > 0.0_wp ) THEN
741	mean_r = s_r3 / s_r2
742	ELSE
743	mean_r = 0.0_wp
744	ENDIF
745	tend(k,j,i) = mean_r
746	ENDDO
747	ENDDO
748	ENDDO
749	! CALL exchange_horiz( tend, nbgp )
750	ELSE
751	tend = 0.0_wp
752	ENDIF
753	DO i = nxl, nxr
754	DO j = nys, nyn
755	DO k = nzb, nzt+1
756	local_pf(i,j,k) = tend(k,j,i)
757	ENDDO
758	ENDDO
759	ENDDO
760	resorted = .TRUE.
761	ELSE
762	IF ( .NOT. ALLOCATED( pr_av ) ) THEN
763	ALLOCATE( pr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
764	pr_av = REAL( fill_value, KIND = wp )
765	ENDIF
766	! CALL exchange_horiz( pr_av, nbgp )
767	to_be_resorted => pr_av
768	ENDIF
769
770	CASE ( 'theta_xy', 'theta_xz', 'theta_yz' )
771	IF ( av == 0 ) THEN
772	IF ( .NOT. bulk_cloud_model ) THEN
773	to_be_resorted => pt
774	ELSE
775	DO i = nxl, nxr
776	DO j = nys, nyn
777	DO k = nzb, nzt+1
778	local_pf(i,j,k) = pt(k,j,i) + lv_d_cp * &
779	d_exner(k) * &
780	ql(k,j,i)
781	ENDDO
782	ENDDO
783	ENDDO
784	resorted = .TRUE.
785	ENDIF
786	ELSE
787	IF ( .NOT. ALLOCATED( pt_av ) ) THEN
788	ALLOCATE( pt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
789	pt_av = REAL( fill_value, KIND = wp )
790	ENDIF
791	to_be_resorted => pt_av
792	ENDIF
793	IF ( mode == 'xy' ) level_z = zu
794
795	CASE ( 'q_xy', 'q_xz', 'q_yz' )
796	IF ( av == 0 ) THEN
797	to_be_resorted => q
798	ELSE
799	IF ( .NOT. ALLOCATED( q_av ) ) THEN
800	ALLOCATE( q_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
801	q_av = REAL( fill_value, KIND = wp )
802	ENDIF
803	to_be_resorted => q_av
804	ENDIF
805	IF ( mode == 'xy' ) level_z = zu
806
807	CASE ( 'ql_xy', 'ql_xz', 'ql_yz' )
808	IF ( av == 0 ) THEN
809	to_be_resorted => ql
810	ELSE
811	IF ( .NOT. ALLOCATED( ql_av ) ) THEN
812	ALLOCATE( ql_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
813	ql_av = REAL( fill_value, KIND = wp )
814	ENDIF
815	to_be_resorted => ql_av
816	ENDIF
817	IF ( mode == 'xy' ) level_z = zu
818
819	CASE ( 'ql_c_xy', 'ql_c_xz', 'ql_c_yz' )
820	IF ( av == 0 ) THEN
821	to_be_resorted => ql_c
822	ELSE
823	IF ( .NOT. ALLOCATED( ql_c_av ) ) THEN
824	ALLOCATE( ql_c_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
825	ql_c_av = REAL( fill_value, KIND = wp )
826	ENDIF
827	to_be_resorted => ql_c_av
828	ENDIF
829	IF ( mode == 'xy' ) level_z = zu
830
831	CASE ( 'ql_v_xy', 'ql_v_xz', 'ql_v_yz' )
832	IF ( av == 0 ) THEN
833	to_be_resorted => ql_v
834	ELSE
835	IF ( .NOT. ALLOCATED( ql_v_av ) ) THEN
836	ALLOCATE( ql_v_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
837	ql_v_av = REAL( fill_value, KIND = wp )
838	ENDIF
839	to_be_resorted => ql_v_av
840	ENDIF
841	IF ( mode == 'xy' ) level_z = zu
842
843	CASE ( 'ql_vp_xy', 'ql_vp_xz', 'ql_vp_yz' )
844	IF ( av == 0 ) THEN
845	IF ( simulated_time >= particle_advection_start ) THEN
846	DO i = nxl, nxr
847	DO j = nys, nyn
848	DO k = nzb, nzt+1
849	number_of_particles = prt_count(k,j,i)
850	IF (number_of_particles <= 0) CYCLE
851	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
852	DO n = 1, number_of_particles
853	IF ( particles(n)%particle_mask ) THEN
854	tend(k,j,i) = tend(k,j,i) + &
855	particles(n)%weight_factor / &
856	prt_count(k,j,i)
857	ENDIF
858	ENDDO
859	ENDDO
860	ENDDO
861	ENDDO
862	! CALL exchange_horiz( tend, nbgp )
863	ELSE
864	tend = 0.0_wp
865	ENDIF
866	DO i = nxl, nxr
867	DO j = nys, nyn
868	DO k = nzb, nzt+1
869	local_pf(i,j,k) = tend(k,j,i)
870	ENDDO
871	ENDDO
872	ENDDO
873	resorted = .TRUE.
874	ELSE
875	IF ( .NOT. ALLOCATED( ql_vp_av ) ) THEN
876	ALLOCATE( ql_vp_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
877	ql_vp_av = REAL( fill_value, KIND = wp )
878	ENDIF
879	! CALL exchange_horiz( ql_vp_av, nbgp )
880	to_be_resorted => ql_vp_av
881	ENDIF
882	IF ( mode == 'xy' ) level_z = zu
883
884	CASE ( 'qsws*_xy' ) ! 2d-array
885	IF ( av == 0 ) THEN
886	local_pf(:,:,nzb+1) = REAL( fill_value, KIND = wp )
887	!
888	!-- In case of default surfaces, clean-up flux by density.
889	!-- In case of land-surfaces, convert fluxes into
890	!-- dynamic units
891	DO m = 1, surf_def_h(0)%ns
892	i = surf_def_h(0)%i(m)
893	j = surf_def_h(0)%j(m)
894	k = surf_def_h(0)%k(m)
895	local_pf(i,j,nzb+1) = surf_def_h(0)%qsws(m) * &
896	waterflux_output_conversion(k)
897	ENDDO
898	DO m = 1, surf_lsm_h%ns
899	i = surf_lsm_h%i(m)
900	j = surf_lsm_h%j(m)
901	k = surf_lsm_h%k(m)
902	local_pf(i,j,nzb+1) = surf_lsm_h%qsws(m) * l_v
903	ENDDO
904	ELSE
905	IF ( .NOT. ALLOCATED( qsws_av ) ) THEN
906	ALLOCATE( qsws_av(nysg:nyng,nxlg:nxrg) )
907	qsws_av = REAL( fill_value, KIND = wp )
908	ENDIF
909	DO i = nxl, nxr
910	DO j = nys, nyn
911	local_pf(i,j,nzb+1) = qsws_av(j,i)
912	ENDDO
913	ENDDO
914	ENDIF
915	resorted = .TRUE.
916	two_d = .TRUE.
917	level_z(nzb+1) = zu(nzb+1)
918
919	CASE ( 'qv_xy', 'qv_xz', 'qv_yz' )
920	IF ( av == 0 ) THEN
921	DO i = nxl, nxr
922	DO j = nys, nyn
923	DO k = nzb, nzt+1
924	local_pf(i,j,k) = q(k,j,i) - ql(k,j,i)
925	ENDDO
926	ENDDO
927	ENDDO
928	resorted = .TRUE.
929	ELSE
930	IF ( .NOT. ALLOCATED( qv_av ) ) THEN
931	ALLOCATE( qv_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
932	qv_av = REAL( fill_value, KIND = wp )
933	ENDIF
934	to_be_resorted => qv_av
935	ENDIF
936	IF ( mode == 'xy' ) level_z = zu
937
938	CASE ( 'r_a*_xy' ) ! 2d-array
939	IF ( av == 0 ) THEN
940	DO m = 1, surf_lsm_h%ns
941	i = surf_lsm_h%i(m)
942	j = surf_lsm_h%j(m)
943	local_pf(i,j,nzb+1) = surf_lsm_h%r_a(m)
944	ENDDO
945
946	DO m = 1, surf_usm_h%ns
947	i = surf_usm_h%i(m)
948	j = surf_usm_h%j(m)
949	local_pf(i,j,nzb+1) = &
950	( surf_usm_h%frac(ind_veg_wall,m) * &
951	surf_usm_h%r_a(m) + &
952	surf_usm_h%frac(ind_pav_green,m) * &
953	surf_usm_h%r_a_green(m) + &
954	surf_usm_h%frac(ind_wat_win,m) * &
955	surf_usm_h%r_a_window(m) )
956	ENDDO
957	ELSE
958	IF ( .NOT. ALLOCATED( r_a_av ) ) THEN
959	ALLOCATE( r_a_av(nysg:nyng,nxlg:nxrg) )
960	r_a_av = REAL( fill_value, KIND = wp )
961	ENDIF
962	DO i = nxl, nxr
963	DO j = nys, nyn
964	local_pf(i,j,nzb+1) = r_a_av(j,i)
965	ENDDO
966	ENDDO
967	ENDIF
968	resorted = .TRUE.
969	two_d = .TRUE.
970	level_z(nzb+1) = zu(nzb+1)
971
972	CASE ( 's_xy', 's_xz', 's_yz' )
973	IF ( av == 0 ) THEN
974	to_be_resorted => s
975	ELSE
976	IF ( .NOT. ALLOCATED( s_av ) ) THEN
977	ALLOCATE( s_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
978	s_av = REAL( fill_value, KIND = wp )
979	ENDIF
980	to_be_resorted => s_av
981	ENDIF
982
983	CASE ( 'shf*_xy' ) ! 2d-array
984	IF ( av == 0 ) THEN
985	!
986	!-- In case of default surfaces, clean-up flux by density.
987	!-- In case of land- and urban-surfaces, convert fluxes into
988	!-- dynamic units.
989	DO m = 1, surf_def_h(0)%ns
990	i = surf_def_h(0)%i(m)
991	j = surf_def_h(0)%j(m)
992	k = surf_def_h(0)%k(m)
993	local_pf(i,j,nzb+1) = surf_def_h(0)%shf(m) * &
994	heatflux_output_conversion(k)
995	ENDDO
996	DO m = 1, surf_lsm_h%ns
997	i = surf_lsm_h%i(m)
998	j = surf_lsm_h%j(m)
999	k = surf_lsm_h%k(m)
1000	local_pf(i,j,nzb+1) = surf_lsm_h%shf(m) * c_p
1001	ENDDO
1002	DO m = 1, surf_usm_h%ns
1003	i = surf_usm_h%i(m)
1004	j = surf_usm_h%j(m)
1005	k = surf_usm_h%k(m)
1006	local_pf(i,j,nzb+1) = surf_usm_h%shf(m) * c_p
1007	ENDDO
1008	ELSE
1009	IF ( .NOT. ALLOCATED( shf_av ) ) THEN
1010	ALLOCATE( shf_av(nysg:nyng,nxlg:nxrg) )
1011	shf_av = REAL( fill_value, KIND = wp )
1012	ENDIF
1013	DO i = nxl, nxr
1014	DO j = nys, nyn
1015	local_pf(i,j,nzb+1) = shf_av(j,i)
1016	ENDDO
1017	ENDDO
1018	ENDIF
1019	resorted = .TRUE.
1020	two_d = .TRUE.
1021	level_z(nzb+1) = zu(nzb+1)
1022
1023	CASE ( 'ssws*_xy' ) ! 2d-array
1024	IF ( av == 0 ) THEN
1025	DO m = 1, surf_def_h(0)%ns
1026	i = surf_def_h(0)%i(m)
1027	j = surf_def_h(0)%j(m)
1028	local_pf(i,j,nzb+1) = surf_def_h(0)%ssws(m)
1029	ENDDO
1030	DO m = 1, surf_lsm_h%ns
1031	i = surf_lsm_h%i(m)
1032	j = surf_lsm_h%j(m)
1033	local_pf(i,j,nzb+1) = surf_lsm_h%ssws(m)
1034	ENDDO
1035	DO m = 1, surf_usm_h%ns
1036	i = surf_usm_h%i(m)
1037	j = surf_usm_h%j(m)
1038	local_pf(i,j,nzb+1) = surf_usm_h%ssws(m)
1039	ENDDO
1040	ELSE
1041	IF ( .NOT. ALLOCATED( ssws_av ) ) THEN
1042	ALLOCATE( ssws_av(nysg:nyng,nxlg:nxrg) )
1043	ssws_av = REAL( fill_value, KIND = wp )
1044	ENDIF
1045	DO i = nxl, nxr
1046	DO j = nys, nyn
1047	local_pf(i,j,nzb+1) = ssws_av(j,i)
1048	ENDDO
1049	ENDDO
1050	ENDIF
1051	resorted = .TRUE.
1052	two_d = .TRUE.
1053	level_z(nzb+1) = zu(nzb+1)
1054
1055	CASE ( 't*_xy' ) ! 2d-array
1056	IF ( av == 0 ) THEN
1057	DO m = 1, surf_def_h(0)%ns
1058	i = surf_def_h(0)%i(m)
1059	j = surf_def_h(0)%j(m)
1060	local_pf(i,j,nzb+1) = surf_def_h(0)%ts(m)
1061	ENDDO
1062	DO m = 1, surf_lsm_h%ns
1063	i = surf_lsm_h%i(m)
1064	j = surf_lsm_h%j(m)
1065	local_pf(i,j,nzb+1) = surf_lsm_h%ts(m)
1066	ENDDO
1067	DO m = 1, surf_usm_h%ns
1068	i = surf_usm_h%i(m)
1069	j = surf_usm_h%j(m)
1070	local_pf(i,j,nzb+1) = surf_usm_h%ts(m)
1071	ENDDO
1072	ELSE
1073	IF ( .NOT. ALLOCATED( ts_av ) ) THEN
1074	ALLOCATE( ts_av(nysg:nyng,nxlg:nxrg) )
1075	ts_av = REAL( fill_value, KIND = wp )
1076	ENDIF
1077	DO i = nxl, nxr
1078	DO j = nys, nyn
1079	local_pf(i,j,nzb+1) = ts_av(j,i)
1080	ENDDO
1081	ENDDO
1082	ENDIF
1083	resorted = .TRUE.
1084	two_d = .TRUE.
1085	level_z(nzb+1) = zu(nzb+1)
1086
1087	CASE ( 'tsurf*_xy' ) ! 2d-array
1088	IF ( av == 0 ) THEN
1089	DO m = 1, surf_def_h(0)%ns
1090	i = surf_def_h(0)%i(m)
1091	j = surf_def_h(0)%j(m)
1092	local_pf(i,j,nzb+1) = surf_def_h(0)%pt_surface(m)
1093	ENDDO
1094
1095	DO m = 1, surf_lsm_h%ns
1096	i = surf_lsm_h%i(m)
1097	j = surf_lsm_h%j(m)
1098	local_pf(i,j,nzb+1) = surf_lsm_h%pt_surface(m)
1099	ENDDO
1100
1101	DO m = 1, surf_usm_h%ns
1102	i = surf_usm_h%i(m)
1103	j = surf_usm_h%j(m)
1104	local_pf(i,j,nzb+1) = surf_usm_h%pt_surface(m)
1105	ENDDO
1106
1107	ELSE
1108	IF ( .NOT. ALLOCATED( tsurf_av ) ) THEN
1109	ALLOCATE( tsurf_av(nysg:nyng,nxlg:nxrg) )
1110	tsurf_av = REAL( fill_value, KIND = wp )
1111	ENDIF
1112	DO i = nxl, nxr
1113	DO j = nys, nyn
1114	local_pf(i,j,nzb+1) = tsurf_av(j,i)
1115	ENDDO
1116	ENDDO
1117	ENDIF
1118	resorted = .TRUE.
1119	two_d = .TRUE.
1120	level_z(nzb+1) = zu(nzb+1)
1121
1122	CASE ( 'u_xy', 'u_xz', 'u_yz' )
1123	flag_nr = 1
1124	IF ( av == 0 ) THEN
1125	to_be_resorted => u
1126	ELSE
1127	IF ( .NOT. ALLOCATED( u_av ) ) THEN
1128	ALLOCATE( u_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1129	u_av = REAL( fill_value, KIND = wp )
1130	ENDIF
1131	to_be_resorted => u_av
1132	ENDIF
1133	IF ( mode == 'xy' ) level_z = zu
1134	!
1135	!-- Substitute the values generated by "mirror" boundary condition
1136	!-- at the bottom boundary by the real surface values.
1137	IF ( do2d(av,if) == 'u_xz' .OR. do2d(av,if) == 'u_yz' ) THEN
1138	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
1139	ENDIF
1140
1141	CASE ( 'us*_xy' ) ! 2d-array
1142	IF ( av == 0 ) THEN
1143	DO m = 1, surf_def_h(0)%ns
1144	i = surf_def_h(0)%i(m)
1145	j = surf_def_h(0)%j(m)
1146	local_pf(i,j,nzb+1) = surf_def_h(0)%us(m)
1147	ENDDO
1148	DO m = 1, surf_lsm_h%ns
1149	i = surf_lsm_h%i(m)
1150	j = surf_lsm_h%j(m)
1151	local_pf(i,j,nzb+1) = surf_lsm_h%us(m)
1152	ENDDO
1153	DO m = 1, surf_usm_h%ns
1154	i = surf_usm_h%i(m)
1155	j = surf_usm_h%j(m)
1156	local_pf(i,j,nzb+1) = surf_usm_h%us(m)
1157	ENDDO
1158	ELSE
1159	IF ( .NOT. ALLOCATED( us_av ) ) THEN
1160	ALLOCATE( us_av(nysg:nyng,nxlg:nxrg) )
1161	us_av = REAL( fill_value, KIND = wp )
1162	ENDIF
1163	DO i = nxl, nxr
1164	DO j = nys, nyn
1165	local_pf(i,j,nzb+1) = us_av(j,i)
1166	ENDDO
1167	ENDDO
1168	ENDIF
1169	resorted = .TRUE.
1170	two_d = .TRUE.
1171	level_z(nzb+1) = zu(nzb+1)
1172
1173	CASE ( 'v_xy', 'v_xz', 'v_yz' )
1174	flag_nr = 2
1175	IF ( av == 0 ) THEN
1176	to_be_resorted => v
1177	ELSE
1178	IF ( .NOT. ALLOCATED( v_av ) ) THEN
1179	ALLOCATE( v_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1180	v_av = REAL( fill_value, KIND = wp )
1181	ENDIF
1182	to_be_resorted => v_av
1183	ENDIF
1184	IF ( mode == 'xy' ) level_z = zu
1185	!
1186	!-- Substitute the values generated by "mirror" boundary condition
1187	!-- at the bottom boundary by the real surface values.
1188	IF ( do2d(av,if) == 'v_xz' .OR. do2d(av,if) == 'v_yz' ) THEN
1189	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
1190	ENDIF
1191
1192	CASE ( 'thetav_xy', 'thetav_xz', 'thetav_yz' )
1193	IF ( av == 0 ) THEN
1194	to_be_resorted => vpt
1195	ELSE
1196	IF ( .NOT. ALLOCATED( vpt_av ) ) THEN
1197	ALLOCATE( vpt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1198	vpt_av = REAL( fill_value, KIND = wp )
1199	ENDIF
1200	to_be_resorted => vpt_av
1201	ENDIF
1202	IF ( mode == 'xy' ) level_z = zu
1203
1204	CASE ( 'w_xy', 'w_xz', 'w_yz' )
1205	flag_nr = 3
1206	IF ( av == 0 ) THEN
1207	to_be_resorted => w
1208	ELSE
1209	IF ( .NOT. ALLOCATED( w_av ) ) THEN
1210	ALLOCATE( w_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1211	w_av = REAL( fill_value, KIND = wp )
1212	ENDIF
1213	to_be_resorted => w_av
1214	ENDIF
1215	IF ( mode == 'xy' ) level_z = zw
1216
1217	CASE ( 'z0*_xy' ) ! 2d-array
1218	IF ( av == 0 ) THEN
1219	DO m = 1, surf_def_h(0)%ns
1220	i = surf_def_h(0)%i(m)
1221	j = surf_def_h(0)%j(m)
1222	local_pf(i,j,nzb+1) = surf_def_h(0)%z0(m)
1223	ENDDO
1224	DO m = 1, surf_lsm_h%ns
1225	i = surf_lsm_h%i(m)
1226	j = surf_lsm_h%j(m)
1227	local_pf(i,j,nzb+1) = surf_lsm_h%z0(m)
1228	ENDDO
1229	DO m = 1, surf_usm_h%ns
1230	i = surf_usm_h%i(m)
1231	j = surf_usm_h%j(m)
1232	local_pf(i,j,nzb+1) = surf_usm_h%z0(m)
1233	ENDDO
1234	ELSE
1235	IF ( .NOT. ALLOCATED( z0_av ) ) THEN
1236	ALLOCATE( z0_av(nysg:nyng,nxlg:nxrg) )
1237	z0_av = REAL( fill_value, KIND = wp )
1238	ENDIF
1239	DO i = nxl, nxr
1240	DO j = nys, nyn
1241	local_pf(i,j,nzb+1) = z0_av(j,i)
1242	ENDDO
1243	ENDDO
1244	ENDIF
1245	resorted = .TRUE.
1246	two_d = .TRUE.
1247	level_z(nzb+1) = zu(nzb+1)
1248
1249	CASE ( 'z0h*_xy' ) ! 2d-array
1250	IF ( av == 0 ) THEN
1251	DO m = 1, surf_def_h(0)%ns
1252	i = surf_def_h(0)%i(m)
1253	j = surf_def_h(0)%j(m)
1254	local_pf(i,j,nzb+1) = surf_def_h(0)%z0h(m)
1255	ENDDO
1256	DO m = 1, surf_lsm_h%ns
1257	i = surf_lsm_h%i(m)
1258	j = surf_lsm_h%j(m)
1259	local_pf(i,j,nzb+1) = surf_lsm_h%z0h(m)
1260	ENDDO
1261	DO m = 1, surf_usm_h%ns
1262	i = surf_usm_h%i(m)
1263	j = surf_usm_h%j(m)
1264	local_pf(i,j,nzb+1) = surf_usm_h%z0h(m)
1265	ENDDO
1266	ELSE
1267	IF ( .NOT. ALLOCATED( z0h_av ) ) THEN
1268	ALLOCATE( z0h_av(nysg:nyng,nxlg:nxrg) )
1269	z0h_av = REAL( fill_value, KIND = wp )
1270	ENDIF
1271	DO i = nxl, nxr
1272	DO j = nys, nyn
1273	local_pf(i,j,nzb+1) = z0h_av(j,i)
1274	ENDDO
1275	ENDDO
1276	ENDIF
1277	resorted = .TRUE.
1278	two_d = .TRUE.
1279	level_z(nzb+1) = zu(nzb+1)
1280
1281	CASE ( 'z0q*_xy' ) ! 2d-array
1282	IF ( av == 0 ) THEN
1283	DO m = 1, surf_def_h(0)%ns
1284	i = surf_def_h(0)%i(m)
1285	j = surf_def_h(0)%j(m)
1286	local_pf(i,j,nzb+1) = surf_def_h(0)%z0q(m)
1287	ENDDO
1288	DO m = 1, surf_lsm_h%ns
1289	i = surf_lsm_h%i(m)
1290	j = surf_lsm_h%j(m)
1291	local_pf(i,j,nzb+1) = surf_lsm_h%z0q(m)
1292	ENDDO
1293	DO m = 1, surf_usm_h%ns
1294	i = surf_usm_h%i(m)
1295	j = surf_usm_h%j(m)
1296	local_pf(i,j,nzb+1) = surf_usm_h%z0q(m)
1297	ENDDO
1298	ELSE
1299	IF ( .NOT. ALLOCATED( z0q_av ) ) THEN
1300	ALLOCATE( z0q_av(nysg:nyng,nxlg:nxrg) )
1301	z0q_av = REAL( fill_value, KIND = wp )
1302	ENDIF
1303	DO i = nxl, nxr
1304	DO j = nys, nyn
1305	local_pf(i,j,nzb+1) = z0q_av(j,i)
1306	ENDDO
1307	ENDDO
1308	ENDIF
1309	resorted = .TRUE.
1310	two_d = .TRUE.
1311	level_z(nzb+1) = zu(nzb+1)
1312
1313	CASE DEFAULT
1314
1315	!
1316	!-- Quantities of other modules
1317	IF ( .NOT. found .AND. bulk_cloud_model ) THEN
1318	CALL bcm_data_output_2d( av, do2d(av,if), found, grid, mode,&
1319	local_pf, two_d, nzb_do, nzt_do )
1320	ENDIF
1321
1322	IF ( .NOT. found .AND. gust_module_enabled ) THEN
1323	CALL gust_data_output_2d( av, do2d(av,if), found, grid, &
1324	local_pf, two_d, nzb_do, nzt_do )
1325	ENDIF
1326
1327	IF ( .NOT. found .AND. biometeorology &
1328	.AND. mode == 'xy' ) THEN
1329	CALL bio_data_output_2d( av, do2d(av,if), found, grid, &
1330	local_pf, two_d, nzb_do, nzt_do, &
1331	fill_value )
1332	ENDIF
1333
1334	IF ( .NOT. found .AND. land_surface ) THEN
1335	CALL lsm_data_output_2d( av, do2d(av,if), found, grid, mode,&
1336	local_pf, two_d, nzb_do, nzt_do )
1337	ENDIF
1338
1339	IF ( .NOT. found .AND. ocean_mode ) THEN
1340	CALL ocean_data_output_2d( av, do2d(av,if), found, grid, &
1341	mode, local_pf, nzb_do, nzt_do )
1342	ENDIF
1343
1344	IF ( .NOT. found .AND. radiation ) THEN
1345	CALL radiation_data_output_2d( av, do2d(av,if), found, grid,&
1346	mode, local_pf, two_d, &
1347	nzb_do, nzt_do )
1348	ENDIF
1349
1350	IF ( .NOT. found .AND. salsa ) THEN
1351	CALL salsa_data_output_2d( av, do2d(av,if), found, grid, &
1352	mode, local_pf, two_d )
1353	ENDIF
1354
1355	IF ( .NOT. found .AND. uv_exposure ) THEN
1356	CALL uvem_data_output_2d( av, do2d(av,if), found, grid, &
1357	local_pf, two_d, nzb_do, nzt_do )
1358	ENDIF
1359
1360	IF ( .NOT. found .AND. air_chemistry ) THEN
1361	CALL chem_data_output_2d( av, do2d(av,if), found, grid, mode, &
1362	local_pf, two_d, nzb_do, nzt_do, fill_value )
1363	ENDIF
1364	!
1365	!-- User defined quantities
1366	IF ( .NOT. found ) THEN
1367	CALL user_data_output_2d( av, do2d(av,if), found, grid, &
1368	local_pf, two_d, nzb_do, nzt_do )
1369	ENDIF
1370
1371	resorted = .TRUE.
1372
1373	IF ( grid == 'zu' ) THEN
1374	IF ( mode == 'xy' ) level_z = zu
1375	ELSEIF ( grid == 'zw' ) THEN
1376	IF ( mode == 'xy' ) level_z = zw
1377	ELSEIF ( grid == 'zu1' ) THEN
1378	IF ( mode == 'xy' ) level_z(nzb+1) = zu(nzb+1)
1379	ELSEIF ( grid == 'zs' ) THEN
1380	IF ( mode == 'xy' ) level_z = zs
1381	ENDIF
1382
1383	IF ( .NOT. found ) THEN
1384	message_string = 'no output provided for: ' // &
1385	TRIM( do2d(av,if) )
1386	CALL message( 'data_output_2d', 'PA0181', 0, 0, 0, 6, 0 )
1387	ENDIF
1388
1389	END SELECT
1390
1391	!
1392	!-- Resort the array to be output, if not done above. Flag topography
1393	!-- grid points with fill values, using the corresponding maksing flag.
1394	IF ( .NOT. resorted ) THEN
1395	DO i = nxl, nxr
1396	DO j = nys, nyn
1397	DO k = nzb_do, nzt_do
1398	local_pf(i,j,k) = MERGE( to_be_resorted(k,j,i), &
1399	REAL( fill_value, KIND = wp ), &
1400	BTEST( wall_flags_0(k,j,i), &
1401	flag_nr ) )
1402	ENDDO
1403	ENDDO
1404	ENDDO
1405	ENDIF
1406
1407	!
1408	!-- Output of the individual cross-sections, depending on the cross-
1409	!-- section mode chosen.
1410	is = 1
1411	loop1: DO WHILE ( section(is,s_ind) /= -9999 .OR. two_d )
1412
1413	SELECT CASE ( mode )
1414
1415	CASE ( 'xy' )
1416	!
1417	!-- Determine the cross section index
1418	IF ( two_d ) THEN
1419	layer_xy = nzb+1
1420	ELSE
1421	layer_xy = section(is,s_ind)
1422	ENDIF
1423
1424	!
1425	!-- Exit the loop for layers beyond the data output domain
1426	!-- (used for soil model)
1427	IF ( layer_xy > nzt_do ) THEN
1428	EXIT loop1
1429	ENDIF
1430
1431	!
1432	!-- Update the netCDF xy cross section time axis.
1433	!-- In case of parallel output, this is only done by PE0
1434	!-- to increase the performance.
1435	IF ( simulated_time /= do2d_xy_last_time(av) ) THEN
1436	do2d_xy_time_count(av) = do2d_xy_time_count(av) + 1
1437	do2d_xy_last_time(av) = simulated_time
1438	IF ( myid == 0 ) THEN
1439	IF ( .NOT. data_output_2d_on_each_pe &
1440	.OR. netcdf_data_format > 4 ) &
1441	THEN
1442	#if defined( __netcdf )
1443	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1444	id_var_time_xy(av), &
1445	(/ time_since_reference_point /), &
1446	start = (/ do2d_xy_time_count(av) /), &
1447	count = (/ 1 /) )
1448	CALL netcdf_handle_error( 'data_output_2d', 53 )
1449	#endif
1450	ENDIF
1451	ENDIF
1452	ENDIF
1453	!
1454	!-- If required, carry out averaging along z
1455	IF ( section(is,s_ind) == -1 .AND. .NOT. two_d ) THEN
1456
1457	local_2d = 0.0_wp
1458	!
1459	!-- Carry out the averaging (all data are on the PE)
1460	DO k = nzb_do, nzt_do
1461	DO j = nys, nyn
1462	DO i = nxl, nxr
1463	local_2d(i,j) = local_2d(i,j) + local_pf(i,j,k)
1464	ENDDO
1465	ENDDO
1466	ENDDO
1467
1468	local_2d = local_2d / ( nzt_do - nzb_do + 1.0_wp)
1469
1470	ELSE
1471	!
1472	!-- Just store the respective section on the local array
1473	local_2d = local_pf(:,:,layer_xy)
1474
1475	ENDIF
1476
1477	#if defined( __parallel )
1478	IF ( netcdf_data_format > 4 ) THEN
1479	!
1480	!-- Parallel output in netCDF4/HDF5 format.
1481	IF ( two_d ) THEN
1482	iis = 1
1483	ELSE
1484	iis = is
1485	ENDIF
1486
1487	#if defined( __netcdf )
1488	!
1489	!-- For parallel output, all cross sections are first stored
1490	!-- here on a local array and will be written to the output
1491	!-- file afterwards to increase the performance.
1492	DO i = nxl, nxr
1493	DO j = nys, nyn
1494	local_2d_sections(i,j,iis) = local_2d(i,j)
1495	ENDDO
1496	ENDDO
1497	#endif
1498	ELSE
1499
1500	IF ( data_output_2d_on_each_pe ) THEN
1501	!
1502	!-- Output of partial arrays on each PE
1503	#if defined( __netcdf )
1504	IF ( myid == 0 ) THEN
1505	WRITE ( 21 ) time_since_reference_point, &
1506	do2d_xy_time_count(av), av
1507	ENDIF
1508	#endif
1509	DO i = 0, io_blocks-1
1510	IF ( i == io_group ) THEN
1511	WRITE ( 21 ) nxl, nxr, nys, nyn, nys, nyn
1512	WRITE ( 21 ) local_2d
1513	ENDIF
1514	#if defined( __parallel )
1515	CALL MPI_BARRIER( comm2d, ierr )
1516	#endif
1517	ENDDO
1518
1519	ELSE
1520	!
1521	!-- PE0 receives partial arrays from all processors and
1522	!-- then outputs them. Here a barrier has to be set,
1523	!-- because otherwise "-MPI- FATAL: Remote protocol queue
1524	!-- full" may occur.
1525	CALL MPI_BARRIER( comm2d, ierr )
1526
1527	ngp = ( nxr-nxl+1 ) * ( nyn-nys+1 )
1528	IF ( myid == 0 ) THEN
1529	!
1530	!-- Local array can be relocated directly.
1531	total_2d(nxl:nxr,nys:nyn) = local_2d
1532	!
1533	!-- Receive data from all other PEs.
1534	DO n = 1, numprocs-1
1535	!
1536	!-- Receive index limits first, then array.
1537	!-- Index limits are received in arbitrary order from
1538	!-- the PEs.
1539	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1540	MPI_ANY_SOURCE, 0, comm2d, &
1541	status, ierr )
1542	sender = status(MPI_SOURCE)
1543	DEALLOCATE( local_2d )
1544	ALLOCATE( local_2d(ind(1):ind(2),ind(3):ind(4)) )
1545	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1546	MPI_REAL, sender, 1, comm2d, &
1547	status, ierr )
1548	total_2d(ind(1):ind(2),ind(3):ind(4)) = local_2d
1549	ENDDO
1550	!
1551	!-- Relocate the local array for the next loop increment
1552	DEALLOCATE( local_2d )
1553	ALLOCATE( local_2d(nxl:nxr,nys:nyn) )
1554
1555	#if defined( __netcdf )
1556	IF ( two_d ) THEN
1557	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1558	id_var_do2d(av,if), &
1559	total_2d(0:nx,0:ny), &
1560	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
1561	count = (/ nx+1, ny+1, 1, 1 /) )
1562	ELSE
1563	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1564	id_var_do2d(av,if), &
1565	total_2d(0:nx,0:ny), &
1566	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
1567	count = (/ nx+1, ny+1, 1, 1 /) )
1568	ENDIF
1569	CALL netcdf_handle_error( 'data_output_2d', 54 )
1570	#endif
1571
1572	ELSE
1573	!
1574	!-- First send the local index limits to PE0
1575	ind(1) = nxl; ind(2) = nxr
1576	ind(3) = nys; ind(4) = nyn
1577	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1578	comm2d, ierr )
1579	!
1580	!-- Send data to PE0
1581	CALL MPI_SEND( local_2d(nxl,nys), ngp, &
1582	MPI_REAL, 0, 1, comm2d, ierr )
1583	ENDIF
1584	!
1585	!-- A barrier has to be set, because otherwise some PEs may
1586	!-- proceed too fast so that PE0 may receive wrong data on
1587	!-- tag 0
1588	CALL MPI_BARRIER( comm2d, ierr )
1589	ENDIF
1590
1591	ENDIF
1592	#else
1593	#if defined( __netcdf )
1594	IF ( two_d ) THEN
1595	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1596	id_var_do2d(av,if), &
1597	local_2d(nxl:nxr,nys:nyn), &
1598	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
1599	count = (/ nx+1, ny+1, 1, 1 /) )
1600	ELSE
1601	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1602	id_var_do2d(av,if), &
1603	local_2d(nxl:nxr,nys:nyn), &
1604	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
1605	count = (/ nx+1, ny+1, 1, 1 /) )
1606	ENDIF
1607	CALL netcdf_handle_error( 'data_output_2d', 447 )
1608	#endif
1609	#endif
1610
1611	!
1612	!-- For 2D-arrays (e.g. u*) only one cross-section is available.
1613	!-- Hence exit loop of output levels.
1614	IF ( two_d ) THEN
1615	IF ( netcdf_data_format < 5 ) two_d = .FALSE.
1616	EXIT loop1
1617	ENDIF
1618
1619	CASE ( 'xz' )
1620	!
1621	!-- Update the netCDF xz cross section time axis.
1622	!-- In case of parallel output, this is only done by PE0
1623	!-- to increase the performance.
1624	IF ( simulated_time /= do2d_xz_last_time(av) ) THEN
1625	do2d_xz_time_count(av) = do2d_xz_time_count(av) + 1
1626	do2d_xz_last_time(av) = simulated_time
1627	IF ( myid == 0 ) THEN
1628	IF ( .NOT. data_output_2d_on_each_pe &
1629	.OR. netcdf_data_format > 4 ) &
1630	THEN
1631	#if defined( __netcdf )
1632	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1633	id_var_time_xz(av), &
1634	(/ time_since_reference_point /), &
1635	start = (/ do2d_xz_time_count(av) /), &
1636	count = (/ 1 /) )
1637	CALL netcdf_handle_error( 'data_output_2d', 56 )
1638	#endif
1639	ENDIF
1640	ENDIF
1641	ENDIF
1642
1643	!
1644	!-- If required, carry out averaging along y
1645	IF ( section(is,s_ind) == -1 ) THEN
1646
1647	ALLOCATE( local_2d_l(nxl:nxr,nzb_do:nzt_do) )
1648	local_2d_l = 0.0_wp
1649	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
1650	!
1651	!-- First local averaging on the PE
1652	DO k = nzb_do, nzt_do
1653	DO j = nys, nyn
1654	DO i = nxl, nxr
1655	local_2d_l(i,k) = local_2d_l(i,k) + &
1656	local_pf(i,j,k)
1657	ENDDO
1658	ENDDO
1659	ENDDO
1660	#if defined( __parallel )
1661	!
1662	!-- Now do the averaging over all PEs along y
1663	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1664	CALL MPI_ALLREDUCE( local_2d_l(nxl,nzb_do), &
1665	local_2d(nxl,nzb_do), ngp, MPI_REAL, &
1666	MPI_SUM, comm1dy, ierr )
1667	#else
1668	local_2d = local_2d_l
1669	#endif
1670	local_2d = local_2d / ( ny + 1.0_wp )
1671
1672	DEALLOCATE( local_2d_l )
1673
1674	ELSE
1675	!
1676	!-- Just store the respective section on the local array
1677	!-- (but only if it is available on this PE!)
1678	IF ( section(is,s_ind) >= nys .AND. section(is,s_ind) <= nyn ) &
1679	THEN
1680	local_2d = local_pf(:,section(is,s_ind),nzb_do:nzt_do)
1681	ENDIF
1682
1683	ENDIF
1684
1685	#if defined( __parallel )
1686	IF ( netcdf_data_format > 4 ) THEN
1687	!
1688	!-- Output in netCDF4/HDF5 format.
1689	!-- Output only on those PEs where the respective cross
1690	!-- sections reside. Cross sections averaged along y are
1691	!-- output on the respective first PE along y (myidy=0).
1692	IF ( ( section(is,s_ind) >= nys .AND. &
1693	section(is,s_ind) <= nyn ) .OR. &
1694	( section(is,s_ind) == -1 .AND. myidy == 0 ) ) THEN
1695	#if defined( __netcdf )
1696	!
1697	!-- For parallel output, all cross sections are first
1698	!-- stored here on a local array and will be written to the
1699	!-- output file afterwards to increase the performance.
1700	DO i = nxl, nxr
1701	DO k = nzb_do, nzt_do
1702	local_2d_sections_l(i,is,k) = local_2d(i,k)
1703	ENDDO
1704	ENDDO
1705	#endif
1706	ENDIF
1707
1708	ELSE
1709
1710	IF ( data_output_2d_on_each_pe ) THEN
1711	!
1712	!-- Output of partial arrays on each PE. If the cross
1713	!-- section does not reside on the PE, output special
1714	!-- index values.
1715	#if defined( __netcdf )
1716	IF ( myid == 0 ) THEN
1717	WRITE ( 22 ) time_since_reference_point, &
1718	do2d_xz_time_count(av), av
1719	ENDIF
1720	#endif
1721	DO i = 0, io_blocks-1
1722	IF ( i == io_group ) THEN
1723	IF ( ( section(is,s_ind) >= nys .AND. &
1724	section(is,s_ind) <= nyn ) .OR. &
1725	( section(is,s_ind) == -1 .AND. &
1726	nys-1 == -1 ) ) &
1727	THEN
1728	WRITE (22) nxl, nxr, nzb_do, nzt_do, nzb, nzt+1
1729	WRITE (22) local_2d
1730	ELSE
1731	WRITE (22) -1, -1, -1, -1, -1, -1
1732	ENDIF
1733	ENDIF
1734	#if defined( __parallel )
1735	CALL MPI_BARRIER( comm2d, ierr )
1736	#endif
1737	ENDDO
1738
1739	ELSE
1740	!
1741	!-- PE0 receives partial arrays from all processors of the
1742	!-- respective cross section and outputs them. Here a
1743	!-- barrier has to be set, because otherwise
1744	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
1745	CALL MPI_BARRIER( comm2d, ierr )
1746
1747	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
1748	IF ( myid == 0 ) THEN
1749	!
1750	!-- Local array can be relocated directly.
1751	IF ( ( section(is,s_ind) >= nys .AND. &
1752	section(is,s_ind) <= nyn ) .OR. &
1753	( section(is,s_ind) == -1 .AND. &
1754	nys-1 == -1 ) ) THEN
1755	total_2d(nxl:nxr,nzb_do:nzt_do) = local_2d
1756	ENDIF
1757	!
1758	!-- Receive data from all other PEs.
1759	DO n = 1, numprocs-1
1760	!
1761	!-- Receive index limits first, then array.
1762	!-- Index limits are received in arbitrary order from
1763	!-- the PEs.
1764	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1765	MPI_ANY_SOURCE, 0, comm2d, &
1766	status, ierr )
1767	!
1768	!-- Not all PEs have data for XZ-cross-section.
1769	IF ( ind(1) /= -9999 ) THEN
1770	sender = status(MPI_SOURCE)
1771	DEALLOCATE( local_2d )
1772	ALLOCATE( local_2d(ind(1):ind(2), &
1773	ind(3):ind(4)) )
1774	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1775	MPI_REAL, sender, 1, comm2d, &
1776	status, ierr )
1777	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
1778	local_2d
1779	ENDIF
1780	ENDDO
1781	!
1782	!-- Relocate the local array for the next loop increment
1783	DEALLOCATE( local_2d )
1784	ALLOCATE( local_2d(nxl:nxr,nzb_do:nzt_do) )
1785
1786	#if defined( __netcdf )
1787	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1788	id_var_do2d(av,if), &
1789	total_2d(0:nx,nzb_do:nzt_do), &
1790	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
1791	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
1792	CALL netcdf_handle_error( 'data_output_2d', 58 )
1793	#endif
1794
1795	ELSE
1796	!
1797	!-- If the cross section resides on the PE, send the
1798	!-- local index limits, otherwise send -9999 to PE0.
1799	IF ( ( section(is,s_ind) >= nys .AND. &
1800	section(is,s_ind) <= nyn ) .OR. &
1801	( section(is,s_ind) == -1 .AND. nys-1 == -1 ) ) &
1802	THEN
1803	ind(1) = nxl; ind(2) = nxr
1804	ind(3) = nzb_do; ind(4) = nzt_do
1805	ELSE
1806	ind(1) = -9999; ind(2) = -9999
1807	ind(3) = -9999; ind(4) = -9999
1808	ENDIF
1809	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1810	comm2d, ierr )
1811	!
1812	!-- If applicable, send data to PE0.
1813	IF ( ind(1) /= -9999 ) THEN
1814	CALL MPI_SEND( local_2d(nxl,nzb_do), ngp, &
1815	MPI_REAL, 0, 1, comm2d, ierr )
1816	ENDIF
1817	ENDIF
1818	!
1819	!-- A barrier has to be set, because otherwise some PEs may
1820	!-- proceed too fast so that PE0 may receive wrong data on
1821	!-- tag 0
1822	CALL MPI_BARRIER( comm2d, ierr )
1823	ENDIF
1824
1825	ENDIF
1826	#else
1827	#if defined( __netcdf )
1828	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1829	id_var_do2d(av,if), &
1830	local_2d(nxl:nxr,nzb_do:nzt_do), &
1831	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
1832	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
1833	CALL netcdf_handle_error( 'data_output_2d', 451 )
1834	#endif
1835	#endif
1836
1837	CASE ( 'yz' )
1838	!
1839	!-- Update the netCDF yz cross section time axis.
1840	!-- In case of parallel output, this is only done by PE0
1841	!-- to increase the performance.
1842	IF ( simulated_time /= do2d_yz_last_time(av) ) THEN
1843	do2d_yz_time_count(av) = do2d_yz_time_count(av) + 1
1844	do2d_yz_last_time(av) = simulated_time
1845	IF ( myid == 0 ) THEN
1846	IF ( .NOT. data_output_2d_on_each_pe &
1847	.OR. netcdf_data_format > 4 ) &
1848	THEN
1849	#if defined( __netcdf )
1850	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
1851	id_var_time_yz(av), &
1852	(/ time_since_reference_point /), &
1853	start = (/ do2d_yz_time_count(av) /), &
1854	count = (/ 1 /) )
1855	CALL netcdf_handle_error( 'data_output_2d', 59 )
1856	#endif
1857	ENDIF
1858	ENDIF
1859	ENDIF
1860
1861	!
1862	!-- If required, carry out averaging along x
1863	IF ( section(is,s_ind) == -1 ) THEN
1864
1865	ALLOCATE( local_2d_l(nys:nyn,nzb_do:nzt_do) )
1866	local_2d_l = 0.0_wp
1867	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
1868	!
1869	!-- First local averaging on the PE
1870	DO k = nzb_do, nzt_do
1871	DO j = nys, nyn
1872	DO i = nxl, nxr
1873	local_2d_l(j,k) = local_2d_l(j,k) + &
1874	local_pf(i,j,k)
1875	ENDDO
1876	ENDDO
1877	ENDDO
1878	#if defined( __parallel )
1879	!
1880	!-- Now do the averaging over all PEs along x
1881	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1882	CALL MPI_ALLREDUCE( local_2d_l(nys,nzb_do), &
1883	local_2d(nys,nzb_do), ngp, MPI_REAL, &
1884	MPI_SUM, comm1dx, ierr )
1885	#else
1886	local_2d = local_2d_l
1887	#endif
1888	local_2d = local_2d / ( nx + 1.0_wp )
1889
1890	DEALLOCATE( local_2d_l )
1891
1892	ELSE
1893	!
1894	!-- Just store the respective section on the local array
1895	!-- (but only if it is available on this PE!)
1896	IF ( section(is,s_ind) >= nxl .AND. section(is,s_ind) <= nxr ) &
1897	THEN
1898	local_2d = local_pf(section(is,s_ind),:,nzb_do:nzt_do)
1899	ENDIF
1900
1901	ENDIF
1902
1903	#if defined( __parallel )
1904	IF ( netcdf_data_format > 4 ) THEN
1905	!
1906	!-- Output in netCDF4/HDF5 format.
1907	!-- Output only on those PEs where the respective cross
1908	!-- sections reside. Cross sections averaged along x are
1909	!-- output on the respective first PE along x (myidx=0).
1910	IF ( ( section(is,s_ind) >= nxl .AND. &
1911	section(is,s_ind) <= nxr ) .OR. &
1912	( section(is,s_ind) == -1 .AND. myidx == 0 ) ) THEN
1913	#if defined( __netcdf )
1914	!
1915	!-- For parallel output, all cross sections are first
1916	!-- stored here on a local array and will be written to the
1917	!-- output file afterwards to increase the performance.
1918	DO j = nys, nyn
1919	DO k = nzb_do, nzt_do
1920	local_2d_sections_l(is,j,k) = local_2d(j,k)
1921	ENDDO
1922	ENDDO
1923	#endif
1924	ENDIF
1925
1926	ELSE
1927
1928	IF ( data_output_2d_on_each_pe ) THEN
1929	!
1930	!-- Output of partial arrays on each PE. If the cross
1931	!-- section does not reside on the PE, output special
1932	!-- index values.
1933	#if defined( __netcdf )
1934	IF ( myid == 0 ) THEN
1935	WRITE ( 23 ) time_since_reference_point, &
1936	do2d_yz_time_count(av), av
1937	ENDIF
1938	#endif
1939	DO i = 0, io_blocks-1
1940	IF ( i == io_group ) THEN
1941	IF ( ( section(is,s_ind) >= nxl .AND. &
1942	section(is,s_ind) <= nxr ) .OR. &
1943	( section(is,s_ind) == -1 .AND. &
1944	nxl-1 == -1 ) ) &
1945	THEN
1946	WRITE (23) nys, nyn, nzb_do, nzt_do, nzb, nzt+1
1947	WRITE (23) local_2d
1948	ELSE
1949	WRITE (23) -1, -1, -1, -1, -1, -1
1950	ENDIF
1951	ENDIF
1952	#if defined( __parallel )
1953	CALL MPI_BARRIER( comm2d, ierr )
1954	#endif
1955	ENDDO
1956
1957	ELSE
1958	!
1959	!-- PE0 receives partial arrays from all processors of the
1960	!-- respective cross section and outputs them. Here a
1961	!-- barrier has to be set, because otherwise
1962	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
1963	CALL MPI_BARRIER( comm2d, ierr )
1964
1965	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
1966	IF ( myid == 0 ) THEN
1967	!
1968	!-- Local array can be relocated directly.
1969	IF ( ( section(is,s_ind) >= nxl .AND. &
1970	section(is,s_ind) <= nxr ) .OR. &
1971	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
1972	THEN
1973	total_2d(nys:nyn,nzb_do:nzt_do) = local_2d
1974	ENDIF
1975	!
1976	!-- Receive data from all other PEs.
1977	DO n = 1, numprocs-1
1978	!
1979	!-- Receive index limits first, then array.
1980	!-- Index limits are received in arbitrary order from
1981	!-- the PEs.
1982	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1983	MPI_ANY_SOURCE, 0, comm2d, &
1984	status, ierr )
1985	!
1986	!-- Not all PEs have data for YZ-cross-section.
1987	IF ( ind(1) /= -9999 ) THEN
1988	sender = status(MPI_SOURCE)
1989	DEALLOCATE( local_2d )
1990	ALLOCATE( local_2d(ind(1):ind(2), &
1991	ind(3):ind(4)) )
1992	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1993	MPI_REAL, sender, 1, comm2d, &
1994	status, ierr )
1995	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
1996	local_2d
1997	ENDIF
1998	ENDDO
1999	!
2000	!-- Relocate the local array for the next loop increment
2001	DEALLOCATE( local_2d )
2002	ALLOCATE( local_2d(nys:nyn,nzb_do:nzt_do) )
2003
2004	#if defined( __netcdf )
2005	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2006	id_var_do2d(av,if), &
2007	total_2d(0:ny,nzb_do:nzt_do), &
2008	start = (/ is, 1, 1, do2d_yz_time_count(av) /), &
2009	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
2010	CALL netcdf_handle_error( 'data_output_2d', 61 )
2011	#endif
2012
2013	ELSE
2014	!
2015	!-- If the cross section resides on the PE, send the
2016	!-- local index limits, otherwise send -9999 to PE0.
2017	IF ( ( section(is,s_ind) >= nxl .AND. &
2018	section(is,s_ind) <= nxr ) .OR. &
2019	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
2020	THEN
2021	ind(1) = nys; ind(2) = nyn
2022	ind(3) = nzb_do; ind(4) = nzt_do
2023	ELSE
2024	ind(1) = -9999; ind(2) = -9999
2025	ind(3) = -9999; ind(4) = -9999
2026	ENDIF
2027	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
2028	comm2d, ierr )
2029	!
2030	!-- If applicable, send data to PE0.
2031	IF ( ind(1) /= -9999 ) THEN
2032	CALL MPI_SEND( local_2d(nys,nzb_do), ngp, &
2033	MPI_REAL, 0, 1, comm2d, ierr )
2034	ENDIF
2035	ENDIF
2036	!
2037	!-- A barrier has to be set, because otherwise some PEs may
2038	!-- proceed too fast so that PE0 may receive wrong data on
2039	!-- tag 0
2040	CALL MPI_BARRIER( comm2d, ierr )
2041	ENDIF
2042
2043	ENDIF
2044	#else
2045	#if defined( __netcdf )
2046	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2047	id_var_do2d(av,if), &
2048	local_2d(nys:nyn,nzb_do:nzt_do), &
2049	start = (/ is, 1, 1, do2d_xz_time_count(av) /), &
2050	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
2051	CALL netcdf_handle_error( 'data_output_2d', 452 )
2052	#endif
2053	#endif
2054
2055	END SELECT
2056
2057	is = is + 1
2058	ENDDO loop1
2059
2060	!
2061	!-- For parallel output, all data were collected before on a local array
2062	!-- and are written now to the netcdf file. This must be done to increase
2063	!-- the performance of the parallel output.
2064	#if defined( __netcdf )
2065	IF ( netcdf_data_format > 4 ) THEN
2066
2067	SELECT CASE ( mode )
2068
2069	CASE ( 'xy' )
2070	IF ( two_d ) THEN
2071	nis = 1
2072	two_d = .FALSE.
2073	ELSE
2074	nis = ns
2075	ENDIF
2076	!
2077	!-- Do not output redundant ghost point data except for the
2078	!-- boundaries of the total domain.
2079	! IF ( nxr == nx .AND. nyn /= ny ) THEN
2080	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2081	! id_var_do2d(av,if), &
2082	! local_2d_sections(nxl:nxr+1, &
2083	! nys:nyn,1:nis), &
2084	! start = (/ nxl+1, nys+1, 1, &
2085	! do2d_xy_time_count(av) /), &
2086	! count = (/ nxr-nxl+2, &
2087	! nyn-nys+1, nis, 1 &
2088	! /) )
2089	! ELSEIF ( nxr /= nx .AND. nyn == ny ) THEN
2090	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2091	! id_var_do2d(av,if), &
2092	! local_2d_sections(nxl:nxr, &
2093	! nys:nyn+1,1:nis), &
2094	! start = (/ nxl+1, nys+1, 1, &
2095	! do2d_xy_time_count(av) /), &
2096	! count = (/ nxr-nxl+1, &
2097	! nyn-nys+2, nis, 1 &
2098	! /) )
2099	! ELSEIF ( nxr == nx .AND. nyn == ny ) THEN
2100	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2101	! id_var_do2d(av,if), &
2102	! local_2d_sections(nxl:nxr+1, &
2103	! nys:nyn+1,1:nis), &
2104	! start = (/ nxl+1, nys+1, 1, &
2105	! do2d_xy_time_count(av) /), &
2106	! count = (/ nxr-nxl+2, &
2107	! nyn-nys+2, nis, 1 &
2108	! /) )
2109	! ELSE
2110	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2111	id_var_do2d(av,if), &
2112	local_2d_sections(nxl:nxr, &
2113	nys:nyn,1:nis), &
2114	start = (/ nxl+1, nys+1, 1, &
2115	do2d_xy_time_count(av) /), &
2116	count = (/ nxr-nxl+1, &
2117	nyn-nys+1, nis, 1 &
2118	/) )
2119	! ENDIF
2120
2121	CALL netcdf_handle_error( 'data_output_2d', 55 )
2122
2123	CASE ( 'xz' )
2124	!
2125	!-- First, all PEs get the information of all cross-sections.
2126	!-- Then the data are written to the output file by all PEs
2127	!-- while NF90_COLLECTIVE is set in subroutine
2128	!-- define_netcdf_header. Although redundant information are
2129	!-- written to the output file in that case, the performance
2130	!-- is significantly better compared to the case where only
2131	!-- the first row of PEs in x-direction (myidx = 0) is given
2132	!-- the output while NF90_INDEPENDENT is set.
2133	IF ( npey /= 1 ) THEN
2134
2135	#if defined( __parallel )
2136	!
2137	!-- Distribute data over all PEs along y
2138	ngp = ( nxr-nxl+1 ) * ( nzt_do-nzb_do+1 ) * ns
2139	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
2140	CALL MPI_ALLREDUCE( local_2d_sections_l(nxl,1,nzb_do), &
2141	local_2d_sections(nxl,1,nzb_do), &
2142	ngp, MPI_REAL, MPI_SUM, comm1dy, &
2143	ierr )
2144	#else
2145	local_2d_sections = local_2d_sections_l
2146	#endif
2147	ENDIF
2148	!
2149	!-- Do not output redundant ghost point data except for the
2150	!-- boundaries of the total domain.
2151	! IF ( nxr == nx ) THEN
2152	! nc_stat = NF90_PUT_VAR( id_set_xz(av), &
2153	! id_var_do2d(av,if), &
2154	! local_2d_sections(nxl:nxr+1,1:ns, &
2155	! nzb_do:nzt_do), &
2156	! start = (/ nxl+1, 1, 1, &
2157	! do2d_xz_time_count(av) /), &
2158	! count = (/ nxr-nxl+2, ns, nzt_do-nzb_do+1, &
2159	! 1 /) )
2160	! ELSE
2161	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
2162	id_var_do2d(av,if), &
2163	local_2d_sections(nxl:nxr,1:ns, &
2164	nzb_do:nzt_do), &
2165	start = (/ nxl+1, 1, 1, &
2166	do2d_xz_time_count(av) /), &
2167	count = (/ nxr-nxl+1, ns, nzt_do-nzb_do+1, &
2168	1 /) )
2169	! ENDIF
2170
2171	CALL netcdf_handle_error( 'data_output_2d', 57 )
2172
2173	CASE ( 'yz' )
2174	!
2175	!-- First, all PEs get the information of all cross-sections.
2176	!-- Then the data are written to the output file by all PEs
2177	!-- while NF90_COLLECTIVE is set in subroutine
2178	!-- define_netcdf_header. Although redundant information are
2179	!-- written to the output file in that case, the performance
2180	!-- is significantly better compared to the case where only
2181	!-- the first row of PEs in y-direction (myidy = 0) is given
2182	!-- the output while NF90_INDEPENDENT is set.
2183	IF ( npex /= 1 ) THEN
2184
2185	#if defined( __parallel )
2186	!
2187	!-- Distribute data over all PEs along x
2188	ngp = ( nyn-nys+1 ) * ( nzt-nzb + 2 ) * ns
2189	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
2190	CALL MPI_ALLREDUCE( local_2d_sections_l(1,nys,nzb_do), &
2191	local_2d_sections(1,nys,nzb_do), &
2192	ngp, MPI_REAL, MPI_SUM, comm1dx, &
2193	ierr )
2194	#else
2195	local_2d_sections = local_2d_sections_l
2196	#endif
2197	ENDIF
2198	!
2199	!-- Do not output redundant ghost point data except for the
2200	!-- boundaries of the total domain.
2201	! IF ( nyn == ny ) THEN
2202	! nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2203	! id_var_do2d(av,if), &
2204	! local_2d_sections(1:ns, &
2205	! nys:nyn+1,nzb_do:nzt_do), &
2206	! start = (/ 1, nys+1, 1, &
2207	! do2d_yz_time_count(av) /), &
2208	! count = (/ ns, nyn-nys+2, &
2209	! nzt_do-nzb_do+1, 1 /) )
2210	! ELSE
2211	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2212	id_var_do2d(av,if), &
2213	local_2d_sections(1:ns,nys:nyn, &
2214	nzb_do:nzt_do), &
2215	start = (/ 1, nys+1, 1, &
2216	do2d_yz_time_count(av) /), &
2217	count = (/ ns, nyn-nys+1, &
2218	nzt_do-nzb_do+1, 1 /) )
2219	! ENDIF
2220
2221	CALL netcdf_handle_error( 'data_output_2d', 60 )
2222
2223	CASE DEFAULT
2224	message_string = 'unknown cross-section: ' // TRIM( mode )
2225	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
2226
2227	END SELECT
2228
2229	ENDIF
2230	#endif
2231	ENDIF
2232
2233	if = if + 1
2234	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
2235	do2d_mode = do2d(av,if)(l-1:l)
2236
2237	ENDDO
2238
2239	!
2240	!-- Deallocate temporary arrays.
2241	IF ( ALLOCATED( level_z ) ) DEALLOCATE( level_z )
2242	IF ( netcdf_data_format > 4 ) THEN
2243	DEALLOCATE( local_pf, local_2d, local_2d_sections )
2244	IF( mode == 'xz' .OR. mode == 'yz' ) DEALLOCATE( local_2d_sections_l )
2245	ENDIF
2246	#if defined( __parallel )
2247	IF ( .NOT. data_output_2d_on_each_pe .AND. myid == 0 ) THEN
2248	DEALLOCATE( total_2d )
2249	ENDIF
2250	#endif
2251
2252	!
2253	!-- Close plot output file.
2254	file_id = 20 + s_ind
2255
2256	IF ( data_output_2d_on_each_pe ) THEN
2257	DO i = 0, io_blocks-1
2258	IF ( i == io_group ) THEN
2259	CALL close_file( file_id )
2260	ENDIF
2261	#if defined( __parallel )
2262	CALL MPI_BARRIER( comm2d, ierr )
2263	#endif
2264	ENDDO
2265	ELSE
2266	IF ( myid == 0 ) CALL close_file( file_id )
2267	ENDIF
2268
2269	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
2270
2271	END SUBROUTINE data_output_2d

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |