Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 3506

Last change on this file since 3506 was 3467, checked in by suehring, 6 years ago
Branch salsa @3446 re-integrated into trunk
Property svn:keywords set to `Id`
File size: 95.6 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 3467 2018-10-30 19:05:21Z knoop $
27	! Implementation of a new aerosol module salsa.
28	!
29	! 3448 2018-10-29 18:14:31Z kanani
30	! Add biometeorology
31	!
32	! 3421 2018-10-24 18:39:32Z gronemeier
33	! Renamed output variables
34	!
35	! 3419 2018-10-24 17:27:31Z gronemeier
36	! minor formatting (kanani)
37	! chem_data_output_2d subroutine added (basit)
38	!
39	! 3294 2018-10-01 02:37:10Z raasch
40	! changes concerning modularization of ocean option
41	!
42	! 3274 2018-09-24 15:42:55Z knoop
43	! Modularization of all bulk cloud physics code components
44	!
45	! 3241 2018-09-12 15:02:00Z raasch
46	! unused variables removed
47	!
48	! 3176 2018-07-26 17:12:48Z suehring
49	! Remove output of latent heat flux at urban-surfaces and set fill values
50	! instead
51	!
52	! 3052 2018-05-31 06:11:20Z raasch
53	! Do not open FORTRAN binary files in case of parallel netCDF I/O
54	!
55	! 3049 2018-05-29 13:52:36Z Giersch
56	! Error messages revised
57	!
58	! 3045 2018-05-28 07:55:41Z Giersch
59	! Error messages revised
60	!
61	! 3014 2018-05-09 08:42:38Z maronga
62	! Added nzb_do and nzt_do for some modules for 2d output
63	!
64	! 3004 2018-04-27 12:33:25Z Giersch
65	! precipitation_rate removed, case prr*_xy removed, to_be_resorted have to point
66	! to ql_vp_av and not to ql_vp, allocation checks implemented (averaged data
67	! will be assigned to fill values if no allocation happened so far)
68	!
69	! 2963 2018-04-12 14:47:44Z suehring
70	! Introduce index for vegetation/wall, pavement/green-wall and water/window
71	! surfaces, for clearer access of surface fraction, albedo, emissivity, etc. .
72	!
73	! 2817 2018-02-19 16:32:21Z knoop
74	! Preliminary gust module interface implemented
75	!
76	! 2805 2018-02-14 17:00:09Z suehring
77	! Consider also default-type surfaces for surface temperature output.
78	!
79	! 2797 2018-02-08 13:24:35Z suehring
80	! Enable output of ground-heat flux also at urban surfaces.
81	!
82	! 2743 2018-01-12 16:03:39Z suehring
83	! In case of natural- and urban-type surfaces output surfaces fluxes in W/m2.
84	!
85	! 2742 2018-01-12 14:59:47Z suehring
86	! Enable output of surface temperature
87	!
88	! 2735 2018-01-11 12:01:27Z suehring
89	! output of r_a moved from land-surface to consider also urban-type surfaces
90	!
91	! 2718 2018-01-02 08:49:38Z maronga
92	! Corrected "Former revisions" section
93	!
94	! 2696 2017-12-14 17:12:51Z kanani
95	! Change in file header (GPL part)
96	! Implementation of uv exposure model (FK)
97	! Implementation of turbulence_closure_mod (TG)
98	! Set fill values at topography grid points or e.g. non-natural-type surface
99	! in case of LSM output (MS)
100	!
101	! 2512 2017-10-04 08:26:59Z raasch
102	! upper bounds of cross section output changed from nx+1,ny+1 to nx,ny
103	! no output of ghost layer data
104	!
105	! 2292 2017-06-20 09:51:42Z schwenkel
106	! Implementation of new microphysic scheme: cloud_scheme = 'morrison'
107	! includes two more prognostic equations for cloud drop concentration (nc)
108	! and cloud water content (qc).
109	!
110	! 2277 2017-06-12 10:47:51Z kanani
111	! Removed unused variables do2d_xy_n, do2d_xz_n, do2d_yz_n
112	!
113	! 2233 2017-05-30 18:08:54Z suehring
114	!
115	! 2232 2017-05-30 17:47:52Z suehring
116	! Adjustments to new surface concept
117	!
118	!
119	! 2190 2017-03-21 12:16:43Z raasch
120	! bugfix for r2031: string rho replaced by rho_ocean
121	!
122	! 2031 2016-10-21 15:11:58Z knoop
123	! renamed variable rho to rho_ocean and rho_av to rho_ocean_av
124	!
125	! 2000 2016-08-20 18:09:15Z knoop
126	! Forced header and separation lines into 80 columns
127	!
128	! 1980 2016-07-29 15:51:57Z suehring
129	! Bugfix, in order to steer user-defined output, setting flag found explicitly
130	! to .F.
131	!
132	! 1976 2016-07-27 13:28:04Z maronga
133	! Output of radiation quantities is now done directly in the respective module
134	!
135	! 1972 2016-07-26 07:52:02Z maronga
136	! Output of land surface quantities is now done directly in the respective
137	! module
138	!
139	! 1960 2016-07-12 16:34:24Z suehring
140	! Scalar surface flux added
141	! Rename INTEGER variable s into s_ind, as s is already assigned to scalar
142	!
143	! 1849 2016-04-08 11:33:18Z hoffmann
144	! precipitation_amount, precipitation_rate, prr moved to arrays_3d
145	!
146	! 1822 2016-04-07 07:49:42Z hoffmann
147	! Output of bulk cloud physics simplified.
148	!
149	! 1788 2016-03-10 11:01:04Z maronga
150	! Added output of z0q
151	!
152	! 1783 2016-03-06 18:36:17Z raasch
153	! name change of netcdf routines and module + related changes
154	!
155	! 1745 2016-02-05 13:06:51Z gronemeier
156	! Bugfix: test if time axis limit exceeds moved to point after call of check_open
157	!
158	! 1703 2015-11-02 12:38:44Z raasch
159	! bugfix for output of single (*) xy-sections in case of parallel netcdf I/O
160	!
161	! 1701 2015-11-02 07:43:04Z maronga
162	! Bugfix in output of RRTGM data
163	!
164	! 1691 2015-10-26 16:17:44Z maronga
165	! Added output of Obukhov length (ol) and radiative heating rates for RRTMG.
166	! Formatting corrections.
167	!
168	! 1682 2015-10-07 23:56:08Z knoop
169	! Code annotations made doxygen readable
170	!
171	! 1585 2015-04-30 07:05:52Z maronga
172	! Added support for RRTMG
173	!
174	! 1555 2015-03-04 17:44:27Z maronga
175	! Added output of r_a and r_s
176	!
177	! 1551 2015-03-03 14:18:16Z maronga
178	! Added suppport for land surface model and radiation model output. In the course
179	! of this action, the limits for vertical loops have been changed (from nzb and
180	! nzt+1 to nzb_do and nzt_do, respectively in order to allow soil model output).
181	! Moreover, a new vertical grid zs was introduced.
182	!
183	! 1359 2014-04-11 17:15:14Z hoffmann
184	! New particle structure integrated.
185	!
186	! 1353 2014-04-08 15:21:23Z heinze
187	! REAL constants provided with KIND-attribute
188	!
189	! 1327 2014-03-21 11:00:16Z raasch
190	! parts concerning iso2d output removed,
191	! -netcdf output queries
192	!
193	! 1320 2014-03-20 08:40:49Z raasch
194	! ONLY-attribute added to USE-statements,
195	! kind-parameters added to all INTEGER and REAL declaration statements,
196	! kinds are defined in new module kinds,
197	! revision history before 2012 removed,
198	! comment fields (!:) to be used for variable explanations added to
199	! all variable declaration statements
200	!
201	! 1318 2014-03-17 13:35:16Z raasch
202	! barrier argument removed from cpu_log.
203	! module interfaces removed
204	!
205	! 1311 2014-03-14 12:13:39Z heinze
206	! bugfix: close #if defined( __netcdf )
207	!
208	! 1308 2014-03-13 14:58:42Z fricke
209	! +local_2d_sections, local_2d_sections_l, ns
210	! Check, if the limit of the time dimension is exceeded for parallel output
211	! To increase the performance for parallel output, the following is done:
212	! - Update of time axis is only done by PE0
213	! - Cross sections are first stored on a local array and are written
214	! collectively to the output file by all PEs.
215	!
216	! 1115 2013-03-26 18:16:16Z hoffmann
217	! ql is calculated by calc_liquid_water_content
218	!
219	! 1076 2012-12-05 08:30:18Z hoffmann
220	! Bugfix in output of ql
221	!
222	! 1065 2012-11-22 17:42:36Z hoffmann
223	! Bugfix: Output of cross sections of ql
224	!
225	! 1053 2012-11-13 17:11:03Z hoffmann
226	! +qr, nr, qc and cross sections
227	!
228	! 1036 2012-10-22 13:43:42Z raasch
229	! code put under GPL (PALM 3.9)
230	!
231	! 1031 2012-10-19 14:35:30Z raasch
232	! netCDF4 without parallel file support implemented
233	!
234	! 1007 2012-09-19 14:30:36Z franke
235	! Bugfix: missing calculation of ql_vp added
236	!
237	! 978 2012-08-09 08:28:32Z fricke
238	! +z0h
239	!
240	! Revision 1.1 1997/08/11 06:24:09 raasch
241	! Initial revision
242	!
243	!
244	! Description:
245	! ------------
246	!> Data output of cross-sections in netCDF format or binary format
247	!> to be later converted to NetCDF by helper routine combine_plot_fields.
248	!> Attention: The position of the sectional planes is still not always computed
249	!> --------- correctly. (zu is used always)!
250	!------------------------------------------------------------------------------!
251	SUBROUTINE data_output_2d( mode, av )
252
253
254	USE arrays_3d, &
255	ONLY: dzw, e, heatflux_output_conversion, nc, nr, p, pt, &
256	precipitation_amount, prr, q, qc, ql, ql_c, ql_v, qr, s, tend, &
257	u, v, vpt, w, zu, zw, waterflux_output_conversion, hyrho, d_exner
258
259	USE averaging
260
261	USE basic_constants_and_equations_mod, &
262	ONLY: c_p, lv_d_cp, l_v
263
264	USE biometeorology_mod, &
265	ONLY: biom_data_output_2d
266
267	USE bulk_cloud_model_mod, &
268	ONLY: bulk_cloud_model, bcm_data_output_2d
269
270	USE chemistry_model_mod, &
271	ONLY: chem_data_output_2d
272
273	USE control_parameters, &
274	ONLY: air_chemistry, biometeorology, data_output_2d_on_each_pe, &
275	data_output_xy, data_output_xz, data_output_yz, do2d, &
276	do2d_xy_last_time, do2d_xy_time_count, &
277	do2d_xz_last_time, do2d_xz_time_count, &
278	do2d_yz_last_time, do2d_yz_time_count, &
279	ibc_uv_b, io_blocks, io_group, land_surface, message_string, &
280	ntdim_2d_xy, ntdim_2d_xz, ntdim_2d_yz, &
281	ocean_mode, psolver, section, simulated_time, &
282	time_since_reference_point, uv_exposure
283
284	USE cpulog, &
285	ONLY: cpu_log, log_point
286
287	USE gust_mod, &
288	ONLY: gust_data_output_2d, gust_module_enabled
289
290	USE indices, &
291	ONLY: nbgp, nx, nxl, nxlg, nxr, nxrg, ny, nyn, nyng, nys, nysg, &
292	nzb, nzt, wall_flags_0
293
294	USE kinds
295
296	USE land_surface_model_mod, &
297	ONLY: lsm_data_output_2d, zs
298
299	#if defined( __netcdf )
300	USE NETCDF
301	#endif
302
303	USE netcdf_interface, &
304	ONLY: fill_value, id_set_xy, id_set_xz, id_set_yz, id_var_do2d, &
305	id_var_time_xy, id_var_time_xz, id_var_time_yz, nc_stat, &
306	netcdf_data_format, netcdf_handle_error
307
308	USE ocean_mod, &
309	ONLY: ocean_data_output_2d
310
311	USE particle_attributes, &
312	ONLY: grid_particles, number_of_particles, particle_advection_start, &
313	particles, prt_count
314
315	USE pegrid
316
317	USE radiation_model_mod, &
318	ONLY: radiation, radiation_data_output_2d
319
320	USE salsa_mod, &
321	ONLY: salsa, salsa_data_output_2d
322
323	USE surface_mod, &
324	ONLY: ind_pav_green, ind_veg_wall, ind_wat_win, surf_def_h, &
325	surf_lsm_h, surf_usm_h
326
327	USE turbulence_closure_mod, &
328	ONLY: tcm_data_output_2d
329
330	USE uv_exposure_model_mod, &
331	ONLY: uvem_data_output_2d
332
333
334	IMPLICIT NONE
335
336	CHARACTER (LEN=2) :: do2d_mode !<
337	CHARACTER (LEN=2) :: mode !<
338	CHARACTER (LEN=4) :: grid !<
339	CHARACTER (LEN=50) :: rtext !<
340
341	INTEGER(iwp) :: av !<
342	INTEGER(iwp) :: ngp !<
343	INTEGER(iwp) :: file_id !<
344	INTEGER(iwp) :: flag_nr !< number of masking flag
345	INTEGER(iwp) :: i !<
346	INTEGER(iwp) :: if !<
347	INTEGER(iwp) :: is !<
348	INTEGER(iwp) :: iis !<
349	INTEGER(iwp) :: j !<
350	INTEGER(iwp) :: k !<
351	INTEGER(iwp) :: l !<
352	INTEGER(iwp) :: layer_xy !<
353	INTEGER(iwp) :: m !<
354	INTEGER(iwp) :: n !<
355	INTEGER(iwp) :: nis !<
356	INTEGER(iwp) :: ns !<
357	INTEGER(iwp) :: nzb_do !< lower limit of the data field (usually nzb)
358	INTEGER(iwp) :: nzt_do !< upper limit of the data field (usually nzt+1)
359	INTEGER(iwp) :: s_ind !<
360	INTEGER(iwp) :: sender !<
361	INTEGER(iwp) :: ind(4) !<
362
363	LOGICAL :: found !<
364	LOGICAL :: resorted !<
365	LOGICAL :: two_d !<
366
367	REAL(wp) :: mean_r !<
368	REAL(wp) :: s_r2 !<
369	REAL(wp) :: s_r3 !<
370
371	REAL(wp), DIMENSION(:), ALLOCATABLE :: level_z !<
372	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d !<
373	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d_l !<
374
375	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !<
376	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections !<
377	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections_l !<
378
379	#if defined( __parallel )
380	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: total_2d !<
381	#endif
382	REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !<
383
384	NAMELIST /LOCAL/ rtext
385
386	!
387	!-- Immediate return, if no output is requested (no respective sections
388	!-- found in parameter data_output)
389	IF ( mode == 'xy' .AND. .NOT. data_output_xy(av) ) RETURN
390	IF ( mode == 'xz' .AND. .NOT. data_output_xz(av) ) RETURN
391	IF ( mode == 'yz' .AND. .NOT. data_output_yz(av) ) RETURN
392
393	CALL cpu_log (log_point(3),'data_output_2d','start')
394
395	two_d = .FALSE. ! local variable to distinguish between output of pure 2D
396	! arrays and cross-sections of 3D arrays.
397
398	!
399	!-- Depending on the orientation of the cross-section, the respective output
400	!-- files have to be opened.
401	SELECT CASE ( mode )
402
403	CASE ( 'xy' )
404	s_ind = 1
405	ALLOCATE( level_z(nzb:nzt+1), local_2d(nxl:nxr,nys:nyn) )
406
407	IF ( netcdf_data_format > 4 ) THEN
408	ns = 1
409	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
410	ns = ns + 1
411	ENDDO
412	ns = ns - 1
413	ALLOCATE( local_2d_sections(nxl:nxr,nys:nyn,1:ns) )
414	local_2d_sections = 0.0_wp
415	ENDIF
416
417	!
418	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
419	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
420	CALL check_open( 101+av*10 )
421	ENDIF
422	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
423	CALL check_open( 21 )
424	ELSE
425	IF ( myid == 0 ) THEN
426	#if defined( __parallel )
427	ALLOCATE( total_2d(0:nx,0:ny) )
428	#endif
429	ENDIF
430	ENDIF
431
432	CASE ( 'xz' )
433	s_ind = 2
434	ALLOCATE( local_2d(nxl:nxr,nzb:nzt+1) )
435
436	IF ( netcdf_data_format > 4 ) THEN
437	ns = 1
438	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
439	ns = ns + 1
440	ENDDO
441	ns = ns - 1
442	ALLOCATE( local_2d_sections(nxl:nxr,1:ns,nzb:nzt+1) )
443	ALLOCATE( local_2d_sections_l(nxl:nxr,1:ns,nzb:nzt+1) )
444	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
445	ENDIF
446
447	!
448	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
449	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
450	CALL check_open( 102+av*10 )
451	ENDIF
452
453	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
454	CALL check_open( 22 )
455	ELSE
456	IF ( myid == 0 ) THEN
457	#if defined( __parallel )
458	ALLOCATE( total_2d(0:nx,nzb:nzt+1) )
459	#endif
460	ENDIF
461	ENDIF
462
463	CASE ( 'yz' )
464	s_ind = 3
465	ALLOCATE( local_2d(nys:nyn,nzb:nzt+1) )
466
467	IF ( netcdf_data_format > 4 ) THEN
468	ns = 1
469	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
470	ns = ns + 1
471	ENDDO
472	ns = ns - 1
473	ALLOCATE( local_2d_sections(1:ns,nys:nyn,nzb:nzt+1) )
474	ALLOCATE( local_2d_sections_l(1:ns,nys:nyn,nzb:nzt+1) )
475	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
476	ENDIF
477
478	!
479	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
480	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
481	CALL check_open( 103+av*10 )
482	ENDIF
483
484	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
485	CALL check_open( 23 )
486	ELSE
487	IF ( myid == 0 ) THEN
488	#if defined( __parallel )
489	ALLOCATE( total_2d(0:ny,nzb:nzt+1) )
490	#endif
491	ENDIF
492	ENDIF
493
494	CASE DEFAULT
495	message_string = 'unknown cross-section: ' // TRIM( mode )
496	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
497
498	END SELECT
499
500	!
501	!-- For parallel netcdf output the time axis must be limited. Return, if this
502	!-- limit is exceeded. This could be the case, if the simulated time exceeds
503	!-- the given end time by the length of the given output interval.
504	IF ( netcdf_data_format > 4 ) THEN
505	IF ( mode == 'xy' .AND. do2d_xy_time_count(av) + 1 > &
506	ntdim_2d_xy(av) ) THEN
507	WRITE ( message_string, * ) 'Output of xy cross-sections is not ', &
508	'given at t=', simulated_time, '&because the', &
509	' maximum number of output time levels is exceeded.'
510	CALL message( 'data_output_2d', 'PA0384', 0, 1, 0, 6, 0 )
511	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
512	RETURN
513	ENDIF
514	IF ( mode == 'xz' .AND. do2d_xz_time_count(av) + 1 > &
515	ntdim_2d_xz(av) ) THEN
516	WRITE ( message_string, * ) 'Output of xz cross-sections is not ', &
517	'given at t=', simulated_time, '&because the', &
518	' maximum number of output time levels is exceeded.'
519	CALL message( 'data_output_2d', 'PA0385', 0, 1, 0, 6, 0 )
520	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
521	RETURN
522	ENDIF
523	IF ( mode == 'yz' .AND. do2d_yz_time_count(av) + 1 > &
524	ntdim_2d_yz(av) ) THEN
525	WRITE ( message_string, * ) 'Output of yz cross-sections is not ', &
526	'given at t=', simulated_time, '&because the', &
527	' maximum number of output time levels is exceeded.'
528	CALL message( 'data_output_2d', 'PA0386', 0, 1, 0, 6, 0 )
529	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
530	RETURN
531	ENDIF
532	ENDIF
533
534	!
535	!-- Allocate a temporary array for resorting (kji -> ijk).
536	ALLOCATE( local_pf(nxl:nxr,nys:nyn,nzb:nzt+1) )
537	local_pf = 0.0
538
539	!
540	!-- Loop of all variables to be written.
541	!-- Output dimensions chosen
542	if = 1
543	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
544	do2d_mode = do2d(av,if)(l-1:l)
545
546	DO WHILE ( do2d(av,if)(1:1) /= ' ' )
547
548	IF ( do2d_mode == mode ) THEN
549	!
550	!-- Set flag to steer output of radiation, land-surface, or user-defined
551	!-- quantities
552	found = .FALSE.
553
554	nzb_do = nzb
555	nzt_do = nzt+1
556	!
557	!-- Before each output, set array local_pf to fill value
558	local_pf = fill_value
559	!
560	!-- Set masking flag for topography for not resorted arrays
561	flag_nr = 0
562
563	!
564	!-- Store the array chosen on the temporary array.
565	resorted = .FALSE.
566	SELECT CASE ( TRIM( do2d(av,if) ) )
567	CASE ( 'e_xy', 'e_xz', 'e_yz' )
568	IF ( av == 0 ) THEN
569	to_be_resorted => e
570	ELSE
571	IF ( .NOT. ALLOCATED( e_av ) ) THEN
572	ALLOCATE( e_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
573	e_av = REAL( fill_value, KIND = wp )
574	ENDIF
575	to_be_resorted => e_av
576	ENDIF
577	IF ( mode == 'xy' ) level_z = zu
578
579	CASE ( 'thetal_xy', 'thetal_xz', 'thetal_yz' )
580	IF ( av == 0 ) THEN
581	to_be_resorted => pt
582	ELSE
583	IF ( .NOT. ALLOCATED( lpt_av ) ) THEN
584	ALLOCATE( lpt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
585	lpt_av = REAL( fill_value, KIND = wp )
586	ENDIF
587	to_be_resorted => lpt_av
588	ENDIF
589	IF ( mode == 'xy' ) level_z = zu
590
591	CASE ( 'lwp*_xy' ) ! 2d-array
592	IF ( av == 0 ) THEN
593	DO i = nxl, nxr
594	DO j = nys, nyn
595	local_pf(i,j,nzb+1) = SUM( ql(nzb:nzt,j,i) * &
596	dzw(1:nzt+1) )
597	ENDDO
598	ENDDO
599	ELSE
600	IF ( .NOT. ALLOCATED( lwp_av ) ) THEN
601	ALLOCATE( lwp_av(nysg:nyng,nxlg:nxrg) )
602	lwp_av = REAL( fill_value, KIND = wp )
603	ENDIF
604	DO i = nxl, nxr
605	DO j = nys, nyn
606	local_pf(i,j,nzb+1) = lwp_av(j,i)
607	ENDDO
608	ENDDO
609	ENDIF
610	resorted = .TRUE.
611	two_d = .TRUE.
612	level_z(nzb+1) = zu(nzb+1)
613
614	CASE ( 'ghf*_xy' ) ! 2d-array
615	IF ( av == 0 ) THEN
616	DO m = 1, surf_lsm_h%ns
617	i = surf_lsm_h%i(m)
618	j = surf_lsm_h%j(m)
619	local_pf(i,j,nzb+1) = surf_lsm_h%ghf(m)
620	ENDDO
621	DO m = 1, surf_usm_h%ns
622	i = surf_usm_h%i(m)
623	j = surf_usm_h%j(m)
624	local_pf(i,j,nzb+1) = surf_usm_h%frac(ind_veg_wall,m) * &
625	surf_usm_h%wghf_eb(m) + &
626	surf_usm_h%frac(ind_pav_green,m) * &
627	surf_usm_h%wghf_eb_green(m) + &
628	surf_usm_h%frac(ind_wat_win,m) * &
629	surf_usm_h%wghf_eb_window(m)
630	ENDDO
631	ELSE
632	IF ( .NOT. ALLOCATED( ghf_av ) ) THEN
633	ALLOCATE( ghf_av(nysg:nyng,nxlg:nxrg) )
634	ghf_av = REAL( fill_value, KIND = wp )
635	ENDIF
636	DO i = nxl, nxr
637	DO j = nys, nyn
638	local_pf(i,j,nzb+1) = ghf_av(j,i)
639	ENDDO
640	ENDDO
641	ENDIF
642
643	resorted = .TRUE.
644	two_d = .TRUE.
645	level_z(nzb+1) = zu(nzb+1)
646
647	CASE ( 'ol*_xy' ) ! 2d-array
648	IF ( av == 0 ) THEN
649	DO m = 1, surf_def_h(0)%ns
650	i = surf_def_h(0)%i(m)
651	j = surf_def_h(0)%j(m)
652	local_pf(i,j,nzb+1) = surf_def_h(0)%ol(m)
653	ENDDO
654	DO m = 1, surf_lsm_h%ns
655	i = surf_lsm_h%i(m)
656	j = surf_lsm_h%j(m)
657	local_pf(i,j,nzb+1) = surf_lsm_h%ol(m)
658	ENDDO
659	DO m = 1, surf_usm_h%ns
660	i = surf_usm_h%i(m)
661	j = surf_usm_h%j(m)
662	local_pf(i,j,nzb+1) = surf_usm_h%ol(m)
663	ENDDO
664	ELSE
665	IF ( .NOT. ALLOCATED( ol_av ) ) THEN
666	ALLOCATE( ol_av(nysg:nyng,nxlg:nxrg) )
667	ol_av = REAL( fill_value, KIND = wp )
668	ENDIF
669	DO i = nxl, nxr
670	DO j = nys, nyn
671	local_pf(i,j,nzb+1) = ol_av(j,i)
672	ENDDO
673	ENDDO
674	ENDIF
675	resorted = .TRUE.
676	two_d = .TRUE.
677	level_z(nzb+1) = zu(nzb+1)
678
679	CASE ( 'p_xy', 'p_xz', 'p_yz' )
680	IF ( av == 0 ) THEN
681	IF ( psolver /= 'sor' ) CALL exchange_horiz( p, nbgp )
682	to_be_resorted => p
683	ELSE
684	IF ( .NOT. ALLOCATED( p_av ) ) THEN
685	ALLOCATE( p_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
686	p_av = REAL( fill_value, KIND = wp )
687	ENDIF
688	IF ( psolver /= 'sor' ) CALL exchange_horiz( p_av, nbgp )
689	to_be_resorted => p_av
690	ENDIF
691	IF ( mode == 'xy' ) level_z = zu
692
693	CASE ( 'pc_xy', 'pc_xz', 'pc_yz' ) ! particle concentration
694	IF ( av == 0 ) THEN
695	IF ( simulated_time >= particle_advection_start ) THEN
696	tend = prt_count
697	! CALL exchange_horiz( tend, nbgp )
698	ELSE
699	tend = 0.0_wp
700	ENDIF
701	DO i = nxl, nxr
702	DO j = nys, nyn
703	DO k = nzb, nzt+1
704	local_pf(i,j,k) = tend(k,j,i)
705	ENDDO
706	ENDDO
707	ENDDO
708	resorted = .TRUE.
709	ELSE
710	IF ( .NOT. ALLOCATED( pc_av ) ) THEN
711	ALLOCATE( pc_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
712	pc_av = REAL( fill_value, KIND = wp )
713	ENDIF
714	! CALL exchange_horiz( pc_av, nbgp )
715	to_be_resorted => pc_av
716	ENDIF
717
718	CASE ( 'pr_xy', 'pr_xz', 'pr_yz' ) ! mean particle radius (effective radius)
719	IF ( av == 0 ) THEN
720	IF ( simulated_time >= particle_advection_start ) THEN
721	DO i = nxl, nxr
722	DO j = nys, nyn
723	DO k = nzb, nzt+1
724	number_of_particles = prt_count(k,j,i)
725	IF (number_of_particles <= 0) CYCLE
726	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
727	s_r2 = 0.0_wp
728	s_r3 = 0.0_wp
729	DO n = 1, number_of_particles
730	IF ( particles(n)%particle_mask ) THEN
731	s_r2 = s_r2 + particles(n)%radius*2 &
732	particles(n)%weight_factor
733	s_r3 = s_r3 + particles(n)%radius*3 &
734	particles(n)%weight_factor
735	ENDIF
736	ENDDO
737	IF ( s_r2 > 0.0_wp ) THEN
738	mean_r = s_r3 / s_r2
739	ELSE
740	mean_r = 0.0_wp
741	ENDIF
742	tend(k,j,i) = mean_r
743	ENDDO
744	ENDDO
745	ENDDO
746	! CALL exchange_horiz( tend, nbgp )
747	ELSE
748	tend = 0.0_wp
749	ENDIF
750	DO i = nxl, nxr
751	DO j = nys, nyn
752	DO k = nzb, nzt+1
753	local_pf(i,j,k) = tend(k,j,i)
754	ENDDO
755	ENDDO
756	ENDDO
757	resorted = .TRUE.
758	ELSE
759	IF ( .NOT. ALLOCATED( pr_av ) ) THEN
760	ALLOCATE( pr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
761	pr_av = REAL( fill_value, KIND = wp )
762	ENDIF
763	! CALL exchange_horiz( pr_av, nbgp )
764	to_be_resorted => pr_av
765	ENDIF
766
767	CASE ( 'theta_xy', 'theta_xz', 'theta_yz' )
768	IF ( av == 0 ) THEN
769	IF ( .NOT. bulk_cloud_model ) THEN
770	to_be_resorted => pt
771	ELSE
772	DO i = nxl, nxr
773	DO j = nys, nyn
774	DO k = nzb, nzt+1
775	local_pf(i,j,k) = pt(k,j,i) + lv_d_cp * &
776	d_exner(k) * &
777	ql(k,j,i)
778	ENDDO
779	ENDDO
780	ENDDO
781	resorted = .TRUE.
782	ENDIF
783	ELSE
784	IF ( .NOT. ALLOCATED( pt_av ) ) THEN
785	ALLOCATE( pt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
786	pt_av = REAL( fill_value, KIND = wp )
787	ENDIF
788	to_be_resorted => pt_av
789	ENDIF
790	IF ( mode == 'xy' ) level_z = zu
791
792	CASE ( 'q_xy', 'q_xz', 'q_yz' )
793	IF ( av == 0 ) THEN
794	to_be_resorted => q
795	ELSE
796	IF ( .NOT. ALLOCATED( q_av ) ) THEN
797	ALLOCATE( q_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
798	q_av = REAL( fill_value, KIND = wp )
799	ENDIF
800	to_be_resorted => q_av
801	ENDIF
802	IF ( mode == 'xy' ) level_z = zu
803
804	CASE ( 'ql_xy', 'ql_xz', 'ql_yz' )
805	IF ( av == 0 ) THEN
806	to_be_resorted => ql
807	ELSE
808	IF ( .NOT. ALLOCATED( ql_av ) ) THEN
809	ALLOCATE( ql_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
810	ql_av = REAL( fill_value, KIND = wp )
811	ENDIF
812	to_be_resorted => ql_av
813	ENDIF
814	IF ( mode == 'xy' ) level_z = zu
815
816	CASE ( 'ql_c_xy', 'ql_c_xz', 'ql_c_yz' )
817	IF ( av == 0 ) THEN
818	to_be_resorted => ql_c
819	ELSE
820	IF ( .NOT. ALLOCATED( ql_c_av ) ) THEN
821	ALLOCATE( ql_c_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
822	ql_c_av = REAL( fill_value, KIND = wp )
823	ENDIF
824	to_be_resorted => ql_c_av
825	ENDIF
826	IF ( mode == 'xy' ) level_z = zu
827
828	CASE ( 'ql_v_xy', 'ql_v_xz', 'ql_v_yz' )
829	IF ( av == 0 ) THEN
830	to_be_resorted => ql_v
831	ELSE
832	IF ( .NOT. ALLOCATED( ql_v_av ) ) THEN
833	ALLOCATE( ql_v_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
834	ql_v_av = REAL( fill_value, KIND = wp )
835	ENDIF
836	to_be_resorted => ql_v_av
837	ENDIF
838	IF ( mode == 'xy' ) level_z = zu
839
840	CASE ( 'ql_vp_xy', 'ql_vp_xz', 'ql_vp_yz' )
841	IF ( av == 0 ) THEN
842	IF ( simulated_time >= particle_advection_start ) THEN
843	DO i = nxl, nxr
844	DO j = nys, nyn
845	DO k = nzb, nzt+1
846	number_of_particles = prt_count(k,j,i)
847	IF (number_of_particles <= 0) CYCLE
848	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
849	DO n = 1, number_of_particles
850	IF ( particles(n)%particle_mask ) THEN
851	tend(k,j,i) = tend(k,j,i) + &
852	particles(n)%weight_factor / &
853	prt_count(k,j,i)
854	ENDIF
855	ENDDO
856	ENDDO
857	ENDDO
858	ENDDO
859	! CALL exchange_horiz( tend, nbgp )
860	ELSE
861	tend = 0.0_wp
862	ENDIF
863	DO i = nxl, nxr
864	DO j = nys, nyn
865	DO k = nzb, nzt+1
866	local_pf(i,j,k) = tend(k,j,i)
867	ENDDO
868	ENDDO
869	ENDDO
870	resorted = .TRUE.
871	ELSE
872	IF ( .NOT. ALLOCATED( ql_vp_av ) ) THEN
873	ALLOCATE( ql_vp_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
874	ql_vp_av = REAL( fill_value, KIND = wp )
875	ENDIF
876	! CALL exchange_horiz( ql_vp_av, nbgp )
877	to_be_resorted => ql_vp_av
878	ENDIF
879	IF ( mode == 'xy' ) level_z = zu
880
881	CASE ( 'qsws*_xy' ) ! 2d-array
882	IF ( av == 0 ) THEN
883	local_pf(:,:,nzb+1) = REAL( fill_value, KIND = wp )
884	!
885	!-- In case of default surfaces, clean-up flux by density.
886	!-- In case of land-surfaces, convert fluxes into
887	!-- dynamic units
888	DO m = 1, surf_def_h(0)%ns
889	i = surf_def_h(0)%i(m)
890	j = surf_def_h(0)%j(m)
891	k = surf_def_h(0)%k(m)
892	local_pf(i,j,nzb+1) = surf_def_h(0)%qsws(m) * &
893	waterflux_output_conversion(k)
894	ENDDO
895	DO m = 1, surf_lsm_h%ns
896	i = surf_lsm_h%i(m)
897	j = surf_lsm_h%j(m)
898	k = surf_lsm_h%k(m)
899	local_pf(i,j,nzb+1) = surf_lsm_h%qsws(m) * l_v
900	ENDDO
901	ELSE
902	IF ( .NOT. ALLOCATED( qsws_av ) ) THEN
903	ALLOCATE( qsws_av(nysg:nyng,nxlg:nxrg) )
904	qsws_av = REAL( fill_value, KIND = wp )
905	ENDIF
906	DO i = nxl, nxr
907	DO j = nys, nyn
908	local_pf(i,j,nzb+1) = qsws_av(j,i)
909	ENDDO
910	ENDDO
911	ENDIF
912	resorted = .TRUE.
913	two_d = .TRUE.
914	level_z(nzb+1) = zu(nzb+1)
915
916	CASE ( 'qv_xy', 'qv_xz', 'qv_yz' )
917	IF ( av == 0 ) THEN
918	DO i = nxl, nxr
919	DO j = nys, nyn
920	DO k = nzb, nzt+1
921	local_pf(i,j,k) = q(k,j,i) - ql(k,j,i)
922	ENDDO
923	ENDDO
924	ENDDO
925	resorted = .TRUE.
926	ELSE
927	IF ( .NOT. ALLOCATED( qv_av ) ) THEN
928	ALLOCATE( qv_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
929	qv_av = REAL( fill_value, KIND = wp )
930	ENDIF
931	to_be_resorted => qv_av
932	ENDIF
933	IF ( mode == 'xy' ) level_z = zu
934
935	CASE ( 'r_a*_xy' ) ! 2d-array
936	IF ( av == 0 ) THEN
937	DO m = 1, surf_lsm_h%ns
938	i = surf_lsm_h%i(m)
939	j = surf_lsm_h%j(m)
940	local_pf(i,j,nzb+1) = surf_lsm_h%r_a(m)
941	ENDDO
942
943	DO m = 1, surf_usm_h%ns
944	i = surf_usm_h%i(m)
945	j = surf_usm_h%j(m)
946	local_pf(i,j,nzb+1) = &
947	( surf_usm_h%frac(ind_veg_wall,m) * &
948	surf_usm_h%r_a(m) + &
949	surf_usm_h%frac(ind_pav_green,m) * &
950	surf_usm_h%r_a_green(m) + &
951	surf_usm_h%frac(ind_wat_win,m) * &
952	surf_usm_h%r_a_window(m) )
953	ENDDO
954	ELSE
955	IF ( .NOT. ALLOCATED( r_a_av ) ) THEN
956	ALLOCATE( r_a_av(nysg:nyng,nxlg:nxrg) )
957	r_a_av = REAL( fill_value, KIND = wp )
958	ENDIF
959	DO i = nxl, nxr
960	DO j = nys, nyn
961	local_pf(i,j,nzb+1) = r_a_av(j,i)
962	ENDDO
963	ENDDO
964	ENDIF
965	resorted = .TRUE.
966	two_d = .TRUE.
967	level_z(nzb+1) = zu(nzb+1)
968
969	CASE ( 's_xy', 's_xz', 's_yz' )
970	IF ( av == 0 ) THEN
971	to_be_resorted => s
972	ELSE
973	IF ( .NOT. ALLOCATED( s_av ) ) THEN
974	ALLOCATE( s_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
975	s_av = REAL( fill_value, KIND = wp )
976	ENDIF
977	to_be_resorted => s_av
978	ENDIF
979
980	CASE ( 'shf*_xy' ) ! 2d-array
981	IF ( av == 0 ) THEN
982	!
983	!-- In case of default surfaces, clean-up flux by density.
984	!-- In case of land- and urban-surfaces, convert fluxes into
985	!-- dynamic units.
986	DO m = 1, surf_def_h(0)%ns
987	i = surf_def_h(0)%i(m)
988	j = surf_def_h(0)%j(m)
989	k = surf_def_h(0)%k(m)
990	local_pf(i,j,nzb+1) = surf_def_h(0)%shf(m) * &
991	heatflux_output_conversion(k)
992	ENDDO
993	DO m = 1, surf_lsm_h%ns
994	i = surf_lsm_h%i(m)
995	j = surf_lsm_h%j(m)
996	k = surf_lsm_h%k(m)
997	local_pf(i,j,nzb+1) = surf_lsm_h%shf(m) * c_p
998	ENDDO
999	DO m = 1, surf_usm_h%ns
1000	i = surf_usm_h%i(m)
1001	j = surf_usm_h%j(m)
1002	k = surf_usm_h%k(m)
1003	local_pf(i,j,nzb+1) = surf_usm_h%shf(m) * c_p
1004	ENDDO
1005	ELSE
1006	IF ( .NOT. ALLOCATED( shf_av ) ) THEN
1007	ALLOCATE( shf_av(nysg:nyng,nxlg:nxrg) )
1008	shf_av = REAL( fill_value, KIND = wp )
1009	ENDIF
1010	DO i = nxl, nxr
1011	DO j = nys, nyn
1012	local_pf(i,j,nzb+1) = shf_av(j,i)
1013	ENDDO
1014	ENDDO
1015	ENDIF
1016	resorted = .TRUE.
1017	two_d = .TRUE.
1018	level_z(nzb+1) = zu(nzb+1)
1019
1020	CASE ( 'ssws*_xy' ) ! 2d-array
1021	IF ( av == 0 ) THEN
1022	DO m = 1, surf_def_h(0)%ns
1023	i = surf_def_h(0)%i(m)
1024	j = surf_def_h(0)%j(m)
1025	local_pf(i,j,nzb+1) = surf_def_h(0)%ssws(m)
1026	ENDDO
1027	DO m = 1, surf_lsm_h%ns
1028	i = surf_lsm_h%i(m)
1029	j = surf_lsm_h%j(m)
1030	local_pf(i,j,nzb+1) = surf_lsm_h%ssws(m)
1031	ENDDO
1032	DO m = 1, surf_usm_h%ns
1033	i = surf_usm_h%i(m)
1034	j = surf_usm_h%j(m)
1035	local_pf(i,j,nzb+1) = surf_usm_h%ssws(m)
1036	ENDDO
1037	ELSE
1038	IF ( .NOT. ALLOCATED( ssws_av ) ) THEN
1039	ALLOCATE( ssws_av(nysg:nyng,nxlg:nxrg) )
1040	ssws_av = REAL( fill_value, KIND = wp )
1041	ENDIF
1042	DO i = nxl, nxr
1043	DO j = nys, nyn
1044	local_pf(i,j,nzb+1) = ssws_av(j,i)
1045	ENDDO
1046	ENDDO
1047	ENDIF
1048	resorted = .TRUE.
1049	two_d = .TRUE.
1050	level_z(nzb+1) = zu(nzb+1)
1051
1052	CASE ( 't*_xy' ) ! 2d-array
1053	IF ( av == 0 ) THEN
1054	DO m = 1, surf_def_h(0)%ns
1055	i = surf_def_h(0)%i(m)
1056	j = surf_def_h(0)%j(m)
1057	local_pf(i,j,nzb+1) = surf_def_h(0)%ts(m)
1058	ENDDO
1059	DO m = 1, surf_lsm_h%ns
1060	i = surf_lsm_h%i(m)
1061	j = surf_lsm_h%j(m)
1062	local_pf(i,j,nzb+1) = surf_lsm_h%ts(m)
1063	ENDDO
1064	DO m = 1, surf_usm_h%ns
1065	i = surf_usm_h%i(m)
1066	j = surf_usm_h%j(m)
1067	local_pf(i,j,nzb+1) = surf_usm_h%ts(m)
1068	ENDDO
1069	ELSE
1070	IF ( .NOT. ALLOCATED( ts_av ) ) THEN
1071	ALLOCATE( ts_av(nysg:nyng,nxlg:nxrg) )
1072	ts_av = REAL( fill_value, KIND = wp )
1073	ENDIF
1074	DO i = nxl, nxr
1075	DO j = nys, nyn
1076	local_pf(i,j,nzb+1) = ts_av(j,i)
1077	ENDDO
1078	ENDDO
1079	ENDIF
1080	resorted = .TRUE.
1081	two_d = .TRUE.
1082	level_z(nzb+1) = zu(nzb+1)
1083
1084	CASE ( 'tsurf*_xy' ) ! 2d-array
1085	IF ( av == 0 ) THEN
1086	DO m = 1, surf_def_h(0)%ns
1087	i = surf_def_h(0)%i(m)
1088	j = surf_def_h(0)%j(m)
1089	local_pf(i,j,nzb+1) = surf_def_h(0)%pt_surface(m)
1090	ENDDO
1091
1092	DO m = 1, surf_lsm_h%ns
1093	i = surf_lsm_h%i(m)
1094	j = surf_lsm_h%j(m)
1095	local_pf(i,j,nzb+1) = surf_lsm_h%pt_surface(m)
1096	ENDDO
1097
1098	DO m = 1, surf_usm_h%ns
1099	i = surf_usm_h%i(m)
1100	j = surf_usm_h%j(m)
1101	local_pf(i,j,nzb+1) = surf_usm_h%pt_surface(m)
1102	ENDDO
1103
1104	ELSE
1105	IF ( .NOT. ALLOCATED( tsurf_av ) ) THEN
1106	ALLOCATE( tsurf_av(nysg:nyng,nxlg:nxrg) )
1107	tsurf_av = REAL( fill_value, KIND = wp )
1108	ENDIF
1109	DO i = nxl, nxr
1110	DO j = nys, nyn
1111	local_pf(i,j,nzb+1) = tsurf_av(j,i)
1112	ENDDO
1113	ENDDO
1114	ENDIF
1115	resorted = .TRUE.
1116	two_d = .TRUE.
1117	level_z(nzb+1) = zu(nzb+1)
1118
1119	CASE ( 'u_xy', 'u_xz', 'u_yz' )
1120	flag_nr = 1
1121	IF ( av == 0 ) THEN
1122	to_be_resorted => u
1123	ELSE
1124	IF ( .NOT. ALLOCATED( u_av ) ) THEN
1125	ALLOCATE( u_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1126	u_av = REAL( fill_value, KIND = wp )
1127	ENDIF
1128	to_be_resorted => u_av
1129	ENDIF
1130	IF ( mode == 'xy' ) level_z = zu
1131	!
1132	!-- Substitute the values generated by "mirror" boundary condition
1133	!-- at the bottom boundary by the real surface values.
1134	IF ( do2d(av,if) == 'u_xz' .OR. do2d(av,if) == 'u_yz' ) THEN
1135	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
1136	ENDIF
1137
1138	CASE ( 'us*_xy' ) ! 2d-array
1139	IF ( av == 0 ) THEN
1140	DO m = 1, surf_def_h(0)%ns
1141	i = surf_def_h(0)%i(m)
1142	j = surf_def_h(0)%j(m)
1143	local_pf(i,j,nzb+1) = surf_def_h(0)%us(m)
1144	ENDDO
1145	DO m = 1, surf_lsm_h%ns
1146	i = surf_lsm_h%i(m)
1147	j = surf_lsm_h%j(m)
1148	local_pf(i,j,nzb+1) = surf_lsm_h%us(m)
1149	ENDDO
1150	DO m = 1, surf_usm_h%ns
1151	i = surf_usm_h%i(m)
1152	j = surf_usm_h%j(m)
1153	local_pf(i,j,nzb+1) = surf_usm_h%us(m)
1154	ENDDO
1155	ELSE
1156	IF ( .NOT. ALLOCATED( us_av ) ) THEN
1157	ALLOCATE( us_av(nysg:nyng,nxlg:nxrg) )
1158	us_av = REAL( fill_value, KIND = wp )
1159	ENDIF
1160	DO i = nxl, nxr
1161	DO j = nys, nyn
1162	local_pf(i,j,nzb+1) = us_av(j,i)
1163	ENDDO
1164	ENDDO
1165	ENDIF
1166	resorted = .TRUE.
1167	two_d = .TRUE.
1168	level_z(nzb+1) = zu(nzb+1)
1169
1170	CASE ( 'v_xy', 'v_xz', 'v_yz' )
1171	flag_nr = 2
1172	IF ( av == 0 ) THEN
1173	to_be_resorted => v
1174	ELSE
1175	IF ( .NOT. ALLOCATED( v_av ) ) THEN
1176	ALLOCATE( v_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1177	v_av = REAL( fill_value, KIND = wp )
1178	ENDIF
1179	to_be_resorted => v_av
1180	ENDIF
1181	IF ( mode == 'xy' ) level_z = zu
1182	!
1183	!-- Substitute the values generated by "mirror" boundary condition
1184	!-- at the bottom boundary by the real surface values.
1185	IF ( do2d(av,if) == 'v_xz' .OR. do2d(av,if) == 'v_yz' ) THEN
1186	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
1187	ENDIF
1188
1189	CASE ( 'thetav_xy', 'thetav_xz', 'thetav_yz' )
1190	IF ( av == 0 ) THEN
1191	to_be_resorted => vpt
1192	ELSE
1193	IF ( .NOT. ALLOCATED( vpt_av ) ) THEN
1194	ALLOCATE( vpt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1195	vpt_av = REAL( fill_value, KIND = wp )
1196	ENDIF
1197	to_be_resorted => vpt_av
1198	ENDIF
1199	IF ( mode == 'xy' ) level_z = zu
1200
1201	CASE ( 'w_xy', 'w_xz', 'w_yz' )
1202	flag_nr = 3
1203	IF ( av == 0 ) THEN
1204	to_be_resorted => w
1205	ELSE
1206	IF ( .NOT. ALLOCATED( w_av ) ) THEN
1207	ALLOCATE( w_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1208	w_av = REAL( fill_value, KIND = wp )
1209	ENDIF
1210	to_be_resorted => w_av
1211	ENDIF
1212	IF ( mode == 'xy' ) level_z = zw
1213
1214	CASE ( 'z0*_xy' ) ! 2d-array
1215	IF ( av == 0 ) THEN
1216	DO m = 1, surf_def_h(0)%ns
1217	i = surf_def_h(0)%i(m)
1218	j = surf_def_h(0)%j(m)
1219	local_pf(i,j,nzb+1) = surf_def_h(0)%z0(m)
1220	ENDDO
1221	DO m = 1, surf_lsm_h%ns
1222	i = surf_lsm_h%i(m)
1223	j = surf_lsm_h%j(m)
1224	local_pf(i,j,nzb+1) = surf_lsm_h%z0(m)
1225	ENDDO
1226	DO m = 1, surf_usm_h%ns
1227	i = surf_usm_h%i(m)
1228	j = surf_usm_h%j(m)
1229	local_pf(i,j,nzb+1) = surf_usm_h%z0(m)
1230	ENDDO
1231	ELSE
1232	IF ( .NOT. ALLOCATED( z0_av ) ) THEN
1233	ALLOCATE( z0_av(nysg:nyng,nxlg:nxrg) )
1234	z0_av = REAL( fill_value, KIND = wp )
1235	ENDIF
1236	DO i = nxl, nxr
1237	DO j = nys, nyn
1238	local_pf(i,j,nzb+1) = z0_av(j,i)
1239	ENDDO
1240	ENDDO
1241	ENDIF
1242	resorted = .TRUE.
1243	two_d = .TRUE.
1244	level_z(nzb+1) = zu(nzb+1)
1245
1246	CASE ( 'z0h*_xy' ) ! 2d-array
1247	IF ( av == 0 ) THEN
1248	DO m = 1, surf_def_h(0)%ns
1249	i = surf_def_h(0)%i(m)
1250	j = surf_def_h(0)%j(m)
1251	local_pf(i,j,nzb+1) = surf_def_h(0)%z0h(m)
1252	ENDDO
1253	DO m = 1, surf_lsm_h%ns
1254	i = surf_lsm_h%i(m)
1255	j = surf_lsm_h%j(m)
1256	local_pf(i,j,nzb+1) = surf_lsm_h%z0h(m)
1257	ENDDO
1258	DO m = 1, surf_usm_h%ns
1259	i = surf_usm_h%i(m)
1260	j = surf_usm_h%j(m)
1261	local_pf(i,j,nzb+1) = surf_usm_h%z0h(m)
1262	ENDDO
1263	ELSE
1264	IF ( .NOT. ALLOCATED( z0h_av ) ) THEN
1265	ALLOCATE( z0h_av(nysg:nyng,nxlg:nxrg) )
1266	z0h_av = REAL( fill_value, KIND = wp )
1267	ENDIF
1268	DO i = nxl, nxr
1269	DO j = nys, nyn
1270	local_pf(i,j,nzb+1) = z0h_av(j,i)
1271	ENDDO
1272	ENDDO
1273	ENDIF
1274	resorted = .TRUE.
1275	two_d = .TRUE.
1276	level_z(nzb+1) = zu(nzb+1)
1277
1278	CASE ( 'z0q*_xy' ) ! 2d-array
1279	IF ( av == 0 ) THEN
1280	DO m = 1, surf_def_h(0)%ns
1281	i = surf_def_h(0)%i(m)
1282	j = surf_def_h(0)%j(m)
1283	local_pf(i,j,nzb+1) = surf_def_h(0)%z0q(m)
1284	ENDDO
1285	DO m = 1, surf_lsm_h%ns
1286	i = surf_lsm_h%i(m)
1287	j = surf_lsm_h%j(m)
1288	local_pf(i,j,nzb+1) = surf_lsm_h%z0q(m)
1289	ENDDO
1290	DO m = 1, surf_usm_h%ns
1291	i = surf_usm_h%i(m)
1292	j = surf_usm_h%j(m)
1293	local_pf(i,j,nzb+1) = surf_usm_h%z0q(m)
1294	ENDDO
1295	ELSE
1296	IF ( .NOT. ALLOCATED( z0q_av ) ) THEN
1297	ALLOCATE( z0q_av(nysg:nyng,nxlg:nxrg) )
1298	z0q_av = REAL( fill_value, KIND = wp )
1299	ENDIF
1300	DO i = nxl, nxr
1301	DO j = nys, nyn
1302	local_pf(i,j,nzb+1) = z0q_av(j,i)
1303	ENDDO
1304	ENDDO
1305	ENDIF
1306	resorted = .TRUE.
1307	two_d = .TRUE.
1308	level_z(nzb+1) = zu(nzb+1)
1309
1310	CASE DEFAULT
1311
1312	!
1313	!-- Quantities of other modules
1314	IF ( .NOT. found .AND. bulk_cloud_model ) THEN
1315	CALL bcm_data_output_2d( av, do2d(av,if), found, grid, mode,&
1316	local_pf, two_d, nzb_do, nzt_do )
1317	ENDIF
1318
1319	IF ( .NOT. found .AND. gust_module_enabled ) THEN
1320	CALL gust_data_output_2d( av, do2d(av,if), found, grid, &
1321	local_pf, two_d, nzb_do, nzt_do )
1322	ENDIF
1323
1324	IF ( .NOT. found .AND. biometeorology &
1325	.AND. mode == 'xy' ) THEN
1326	CALL biom_data_output_2d( av, do2d(av,if), found, grid, &
1327	local_pf, two_d, nzb_do, nzt_do, &
1328	fill_value )
1329	ENDIF
1330
1331	IF ( .NOT. found .AND. land_surface ) THEN
1332	CALL lsm_data_output_2d( av, do2d(av,if), found, grid, mode,&
1333	local_pf, two_d, nzb_do, nzt_do )
1334	ENDIF
1335
1336	IF ( .NOT. found .AND. ocean_mode ) THEN
1337	CALL ocean_data_output_2d( av, do2d(av,if), found, grid, &
1338	mode, local_pf, nzb_do, nzt_do )
1339	ENDIF
1340
1341	IF ( .NOT. found .AND. radiation ) THEN
1342	CALL radiation_data_output_2d( av, do2d(av,if), found, grid,&
1343	mode, local_pf, two_d, &
1344	nzb_do, nzt_do )
1345	ENDIF
1346
1347	IF ( .NOT. found .AND. salsa ) THEN
1348	CALL salsa_data_output_2d( av, do2d(av,if), found, grid, &
1349	mode, local_pf, two_d )
1350	ENDIF
1351
1352	IF ( .NOT. found .AND. uv_exposure ) THEN
1353	CALL uvem_data_output_2d( av, do2d(av,if), found, grid, &
1354	local_pf, two_d, nzb_do, nzt_do )
1355	ENDIF
1356
1357	IF ( .NOT. found .AND. air_chemistry ) THEN
1358	CALL chem_data_output_2d( av, do2d(av,if), found, grid, mode, &
1359	local_pf, two_d, nzb_do, nzt_do, fill_value )
1360	ENDIF
1361	!
1362	!-- User defined quantities
1363	IF ( .NOT. found ) THEN
1364	CALL user_data_output_2d( av, do2d(av,if), found, grid, &
1365	local_pf, two_d, nzb_do, nzt_do )
1366	ENDIF
1367
1368	resorted = .TRUE.
1369
1370	IF ( grid == 'zu' ) THEN
1371	IF ( mode == 'xy' ) level_z = zu
1372	ELSEIF ( grid == 'zw' ) THEN
1373	IF ( mode == 'xy' ) level_z = zw
1374	ELSEIF ( grid == 'zu1' ) THEN
1375	IF ( mode == 'xy' ) level_z(nzb+1) = zu(nzb+1)
1376	ELSEIF ( grid == 'zs' ) THEN
1377	IF ( mode == 'xy' ) level_z = zs
1378	ENDIF
1379
1380	IF ( .NOT. found ) THEN
1381	message_string = 'no output provided for: ' // &
1382	TRIM( do2d(av,if) )
1383	CALL message( 'data_output_2d', 'PA0181', 0, 0, 0, 6, 0 )
1384	ENDIF
1385
1386	END SELECT
1387
1388	!
1389	!-- Resort the array to be output, if not done above. Flag topography
1390	!-- grid points with fill values, using the corresponding maksing flag.
1391	IF ( .NOT. resorted ) THEN
1392	DO i = nxl, nxr
1393	DO j = nys, nyn
1394	DO k = nzb_do, nzt_do
1395	local_pf(i,j,k) = MERGE( to_be_resorted(k,j,i), &
1396	REAL( fill_value, KIND = wp ), &
1397	BTEST( wall_flags_0(k,j,i), &
1398	flag_nr ) )
1399	ENDDO
1400	ENDDO
1401	ENDDO
1402	ENDIF
1403
1404	!
1405	!-- Output of the individual cross-sections, depending on the cross-
1406	!-- section mode chosen.
1407	is = 1
1408	loop1: DO WHILE ( section(is,s_ind) /= -9999 .OR. two_d )
1409
1410	SELECT CASE ( mode )
1411
1412	CASE ( 'xy' )
1413	!
1414	!-- Determine the cross section index
1415	IF ( two_d ) THEN
1416	layer_xy = nzb+1
1417	ELSE
1418	layer_xy = section(is,s_ind)
1419	ENDIF
1420
1421	!
1422	!-- Exit the loop for layers beyond the data output domain
1423	!-- (used for soil model)
1424	IF ( layer_xy > nzt_do ) THEN
1425	EXIT loop1
1426	ENDIF
1427
1428	!
1429	!-- Update the netCDF xy cross section time axis.
1430	!-- In case of parallel output, this is only done by PE0
1431	!-- to increase the performance.
1432	IF ( simulated_time /= do2d_xy_last_time(av) ) THEN
1433	do2d_xy_time_count(av) = do2d_xy_time_count(av) + 1
1434	do2d_xy_last_time(av) = simulated_time
1435	IF ( myid == 0 ) THEN
1436	IF ( .NOT. data_output_2d_on_each_pe &
1437	.OR. netcdf_data_format > 4 ) &
1438	THEN
1439	#if defined( __netcdf )
1440	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1441	id_var_time_xy(av), &
1442	(/ time_since_reference_point /), &
1443	start = (/ do2d_xy_time_count(av) /), &
1444	count = (/ 1 /) )
1445	CALL netcdf_handle_error( 'data_output_2d', 53 )
1446	#endif
1447	ENDIF
1448	ENDIF
1449	ENDIF
1450	!
1451	!-- If required, carry out averaging along z
1452	IF ( section(is,s_ind) == -1 .AND. .NOT. two_d ) THEN
1453
1454	local_2d = 0.0_wp
1455	!
1456	!-- Carry out the averaging (all data are on the PE)
1457	DO k = nzb_do, nzt_do
1458	DO j = nys, nyn
1459	DO i = nxl, nxr
1460	local_2d(i,j) = local_2d(i,j) + local_pf(i,j,k)
1461	ENDDO
1462	ENDDO
1463	ENDDO
1464
1465	local_2d = local_2d / ( nzt_do - nzb_do + 1.0_wp)
1466
1467	ELSE
1468	!
1469	!-- Just store the respective section on the local array
1470	local_2d = local_pf(:,:,layer_xy)
1471
1472	ENDIF
1473
1474	#if defined( __parallel )
1475	IF ( netcdf_data_format > 4 ) THEN
1476	!
1477	!-- Parallel output in netCDF4/HDF5 format.
1478	IF ( two_d ) THEN
1479	iis = 1
1480	ELSE
1481	iis = is
1482	ENDIF
1483
1484	#if defined( __netcdf )
1485	!
1486	!-- For parallel output, all cross sections are first stored
1487	!-- here on a local array and will be written to the output
1488	!-- file afterwards to increase the performance.
1489	DO i = nxl, nxr
1490	DO j = nys, nyn
1491	local_2d_sections(i,j,iis) = local_2d(i,j)
1492	ENDDO
1493	ENDDO
1494	#endif
1495	ELSE
1496
1497	IF ( data_output_2d_on_each_pe ) THEN
1498	!
1499	!-- Output of partial arrays on each PE
1500	#if defined( __netcdf )
1501	IF ( myid == 0 ) THEN
1502	WRITE ( 21 ) time_since_reference_point, &
1503	do2d_xy_time_count(av), av
1504	ENDIF
1505	#endif
1506	DO i = 0, io_blocks-1
1507	IF ( i == io_group ) THEN
1508	WRITE ( 21 ) nxl, nxr, nys, nyn, nys, nyn
1509	WRITE ( 21 ) local_2d
1510	ENDIF
1511	#if defined( __parallel )
1512	CALL MPI_BARRIER( comm2d, ierr )
1513	#endif
1514	ENDDO
1515
1516	ELSE
1517	!
1518	!-- PE0 receives partial arrays from all processors and
1519	!-- then outputs them. Here a barrier has to be set,
1520	!-- because otherwise "-MPI- FATAL: Remote protocol queue
1521	!-- full" may occur.
1522	CALL MPI_BARRIER( comm2d, ierr )
1523
1524	ngp = ( nxr-nxl+1 ) * ( nyn-nys+1 )
1525	IF ( myid == 0 ) THEN
1526	!
1527	!-- Local array can be relocated directly.
1528	total_2d(nxl:nxr,nys:nyn) = local_2d
1529	!
1530	!-- Receive data from all other PEs.
1531	DO n = 1, numprocs-1
1532	!
1533	!-- Receive index limits first, then array.
1534	!-- Index limits are received in arbitrary order from
1535	!-- the PEs.
1536	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1537	MPI_ANY_SOURCE, 0, comm2d, &
1538	status, ierr )
1539	sender = status(MPI_SOURCE)
1540	DEALLOCATE( local_2d )
1541	ALLOCATE( local_2d(ind(1):ind(2),ind(3):ind(4)) )
1542	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1543	MPI_REAL, sender, 1, comm2d, &
1544	status, ierr )
1545	total_2d(ind(1):ind(2),ind(3):ind(4)) = local_2d
1546	ENDDO
1547	!
1548	!-- Relocate the local array for the next loop increment
1549	DEALLOCATE( local_2d )
1550	ALLOCATE( local_2d(nxl:nxr,nys:nyn) )
1551
1552	#if defined( __netcdf )
1553	IF ( two_d ) THEN
1554	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1555	id_var_do2d(av,if), &
1556	total_2d(0:nx,0:ny), &
1557	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
1558	count = (/ nx+1, ny+1, 1, 1 /) )
1559	ELSE
1560	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1561	id_var_do2d(av,if), &
1562	total_2d(0:nx,0:ny), &
1563	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
1564	count = (/ nx+1, ny+1, 1, 1 /) )
1565	ENDIF
1566	CALL netcdf_handle_error( 'data_output_2d', 54 )
1567	#endif
1568
1569	ELSE
1570	!
1571	!-- First send the local index limits to PE0
1572	ind(1) = nxl; ind(2) = nxr
1573	ind(3) = nys; ind(4) = nyn
1574	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1575	comm2d, ierr )
1576	!
1577	!-- Send data to PE0
1578	CALL MPI_SEND( local_2d(nxl,nys), ngp, &
1579	MPI_REAL, 0, 1, comm2d, ierr )
1580	ENDIF
1581	!
1582	!-- A barrier has to be set, because otherwise some PEs may
1583	!-- proceed too fast so that PE0 may receive wrong data on
1584	!-- tag 0
1585	CALL MPI_BARRIER( comm2d, ierr )
1586	ENDIF
1587
1588	ENDIF
1589	#else
1590	#if defined( __netcdf )
1591	IF ( two_d ) THEN
1592	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1593	id_var_do2d(av,if), &
1594	local_2d(nxl:nxr,nys:nyn), &
1595	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
1596	count = (/ nx+1, ny+1, 1, 1 /) )
1597	ELSE
1598	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1599	id_var_do2d(av,if), &
1600	local_2d(nxl:nxr,nys:nyn), &
1601	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
1602	count = (/ nx+1, ny+1, 1, 1 /) )
1603	ENDIF
1604	CALL netcdf_handle_error( 'data_output_2d', 447 )
1605	#endif
1606	#endif
1607
1608	!
1609	!-- For 2D-arrays (e.g. u*) only one cross-section is available.
1610	!-- Hence exit loop of output levels.
1611	IF ( two_d ) THEN
1612	IF ( netcdf_data_format < 5 ) two_d = .FALSE.
1613	EXIT loop1
1614	ENDIF
1615
1616	CASE ( 'xz' )
1617	!
1618	!-- Update the netCDF xz cross section time axis.
1619	!-- In case of parallel output, this is only done by PE0
1620	!-- to increase the performance.
1621	IF ( simulated_time /= do2d_xz_last_time(av) ) THEN
1622	do2d_xz_time_count(av) = do2d_xz_time_count(av) + 1
1623	do2d_xz_last_time(av) = simulated_time
1624	IF ( myid == 0 ) THEN
1625	IF ( .NOT. data_output_2d_on_each_pe &
1626	.OR. netcdf_data_format > 4 ) &
1627	THEN
1628	#if defined( __netcdf )
1629	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1630	id_var_time_xz(av), &
1631	(/ time_since_reference_point /), &
1632	start = (/ do2d_xz_time_count(av) /), &
1633	count = (/ 1 /) )
1634	CALL netcdf_handle_error( 'data_output_2d', 56 )
1635	#endif
1636	ENDIF
1637	ENDIF
1638	ENDIF
1639
1640	!
1641	!-- If required, carry out averaging along y
1642	IF ( section(is,s_ind) == -1 ) THEN
1643
1644	ALLOCATE( local_2d_l(nxl:nxr,nzb_do:nzt_do) )
1645	local_2d_l = 0.0_wp
1646	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
1647	!
1648	!-- First local averaging on the PE
1649	DO k = nzb_do, nzt_do
1650	DO j = nys, nyn
1651	DO i = nxl, nxr
1652	local_2d_l(i,k) = local_2d_l(i,k) + &
1653	local_pf(i,j,k)
1654	ENDDO
1655	ENDDO
1656	ENDDO
1657	#if defined( __parallel )
1658	!
1659	!-- Now do the averaging over all PEs along y
1660	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1661	CALL MPI_ALLREDUCE( local_2d_l(nxl,nzb_do), &
1662	local_2d(nxl,nzb_do), ngp, MPI_REAL, &
1663	MPI_SUM, comm1dy, ierr )
1664	#else
1665	local_2d = local_2d_l
1666	#endif
1667	local_2d = local_2d / ( ny + 1.0_wp )
1668
1669	DEALLOCATE( local_2d_l )
1670
1671	ELSE
1672	!
1673	!-- Just store the respective section on the local array
1674	!-- (but only if it is available on this PE!)
1675	IF ( section(is,s_ind) >= nys .AND. section(is,s_ind) <= nyn ) &
1676	THEN
1677	local_2d = local_pf(:,section(is,s_ind),nzb_do:nzt_do)
1678	ENDIF
1679
1680	ENDIF
1681
1682	#if defined( __parallel )
1683	IF ( netcdf_data_format > 4 ) THEN
1684	!
1685	!-- Output in netCDF4/HDF5 format.
1686	!-- Output only on those PEs where the respective cross
1687	!-- sections reside. Cross sections averaged along y are
1688	!-- output on the respective first PE along y (myidy=0).
1689	IF ( ( section(is,s_ind) >= nys .AND. &
1690	section(is,s_ind) <= nyn ) .OR. &
1691	( section(is,s_ind) == -1 .AND. myidy == 0 ) ) THEN
1692	#if defined( __netcdf )
1693	!
1694	!-- For parallel output, all cross sections are first
1695	!-- stored here on a local array and will be written to the
1696	!-- output file afterwards to increase the performance.
1697	DO i = nxl, nxr
1698	DO k = nzb_do, nzt_do
1699	local_2d_sections_l(i,is,k) = local_2d(i,k)
1700	ENDDO
1701	ENDDO
1702	#endif
1703	ENDIF
1704
1705	ELSE
1706
1707	IF ( data_output_2d_on_each_pe ) THEN
1708	!
1709	!-- Output of partial arrays on each PE. If the cross
1710	!-- section does not reside on the PE, output special
1711	!-- index values.
1712	#if defined( __netcdf )
1713	IF ( myid == 0 ) THEN
1714	WRITE ( 22 ) time_since_reference_point, &
1715	do2d_xz_time_count(av), av
1716	ENDIF
1717	#endif
1718	DO i = 0, io_blocks-1
1719	IF ( i == io_group ) THEN
1720	IF ( ( section(is,s_ind) >= nys .AND. &
1721	section(is,s_ind) <= nyn ) .OR. &
1722	( section(is,s_ind) == -1 .AND. &
1723	nys-1 == -1 ) ) &
1724	THEN
1725	WRITE (22) nxl, nxr, nzb_do, nzt_do, nzb, nzt+1
1726	WRITE (22) local_2d
1727	ELSE
1728	WRITE (22) -1, -1, -1, -1, -1, -1
1729	ENDIF
1730	ENDIF
1731	#if defined( __parallel )
1732	CALL MPI_BARRIER( comm2d, ierr )
1733	#endif
1734	ENDDO
1735
1736	ELSE
1737	!
1738	!-- PE0 receives partial arrays from all processors of the
1739	!-- respective cross section and outputs them. Here a
1740	!-- barrier has to be set, because otherwise
1741	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
1742	CALL MPI_BARRIER( comm2d, ierr )
1743
1744	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
1745	IF ( myid == 0 ) THEN
1746	!
1747	!-- Local array can be relocated directly.
1748	IF ( ( section(is,s_ind) >= nys .AND. &
1749	section(is,s_ind) <= nyn ) .OR. &
1750	( section(is,s_ind) == -1 .AND. &
1751	nys-1 == -1 ) ) THEN
1752	total_2d(nxl:nxr,nzb_do:nzt_do) = local_2d
1753	ENDIF
1754	!
1755	!-- Receive data from all other PEs.
1756	DO n = 1, numprocs-1
1757	!
1758	!-- Receive index limits first, then array.
1759	!-- Index limits are received in arbitrary order from
1760	!-- the PEs.
1761	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1762	MPI_ANY_SOURCE, 0, comm2d, &
1763	status, ierr )
1764	!
1765	!-- Not all PEs have data for XZ-cross-section.
1766	IF ( ind(1) /= -9999 ) THEN
1767	sender = status(MPI_SOURCE)
1768	DEALLOCATE( local_2d )
1769	ALLOCATE( local_2d(ind(1):ind(2), &
1770	ind(3):ind(4)) )
1771	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1772	MPI_REAL, sender, 1, comm2d, &
1773	status, ierr )
1774	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
1775	local_2d
1776	ENDIF
1777	ENDDO
1778	!
1779	!-- Relocate the local array for the next loop increment
1780	DEALLOCATE( local_2d )
1781	ALLOCATE( local_2d(nxl:nxr,nzb_do:nzt_do) )
1782
1783	#if defined( __netcdf )
1784	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1785	id_var_do2d(av,if), &
1786	total_2d(0:nx,nzb_do:nzt_do), &
1787	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
1788	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
1789	CALL netcdf_handle_error( 'data_output_2d', 58 )
1790	#endif
1791
1792	ELSE
1793	!
1794	!-- If the cross section resides on the PE, send the
1795	!-- local index limits, otherwise send -9999 to PE0.
1796	IF ( ( section(is,s_ind) >= nys .AND. &
1797	section(is,s_ind) <= nyn ) .OR. &
1798	( section(is,s_ind) == -1 .AND. nys-1 == -1 ) ) &
1799	THEN
1800	ind(1) = nxl; ind(2) = nxr
1801	ind(3) = nzb_do; ind(4) = nzt_do
1802	ELSE
1803	ind(1) = -9999; ind(2) = -9999
1804	ind(3) = -9999; ind(4) = -9999
1805	ENDIF
1806	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1807	comm2d, ierr )
1808	!
1809	!-- If applicable, send data to PE0.
1810	IF ( ind(1) /= -9999 ) THEN
1811	CALL MPI_SEND( local_2d(nxl,nzb_do), ngp, &
1812	MPI_REAL, 0, 1, comm2d, ierr )
1813	ENDIF
1814	ENDIF
1815	!
1816	!-- A barrier has to be set, because otherwise some PEs may
1817	!-- proceed too fast so that PE0 may receive wrong data on
1818	!-- tag 0
1819	CALL MPI_BARRIER( comm2d, ierr )
1820	ENDIF
1821
1822	ENDIF
1823	#else
1824	#if defined( __netcdf )
1825	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1826	id_var_do2d(av,if), &
1827	local_2d(nxl:nxr,nzb_do:nzt_do), &
1828	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
1829	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
1830	CALL netcdf_handle_error( 'data_output_2d', 451 )
1831	#endif
1832	#endif
1833
1834	CASE ( 'yz' )
1835	!
1836	!-- Update the netCDF yz cross section time axis.
1837	!-- In case of parallel output, this is only done by PE0
1838	!-- to increase the performance.
1839	IF ( simulated_time /= do2d_yz_last_time(av) ) THEN
1840	do2d_yz_time_count(av) = do2d_yz_time_count(av) + 1
1841	do2d_yz_last_time(av) = simulated_time
1842	IF ( myid == 0 ) THEN
1843	IF ( .NOT. data_output_2d_on_each_pe &
1844	.OR. netcdf_data_format > 4 ) &
1845	THEN
1846	#if defined( __netcdf )
1847	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
1848	id_var_time_yz(av), &
1849	(/ time_since_reference_point /), &
1850	start = (/ do2d_yz_time_count(av) /), &
1851	count = (/ 1 /) )
1852	CALL netcdf_handle_error( 'data_output_2d', 59 )
1853	#endif
1854	ENDIF
1855	ENDIF
1856	ENDIF
1857
1858	!
1859	!-- If required, carry out averaging along x
1860	IF ( section(is,s_ind) == -1 ) THEN
1861
1862	ALLOCATE( local_2d_l(nys:nyn,nzb_do:nzt_do) )
1863	local_2d_l = 0.0_wp
1864	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
1865	!
1866	!-- First local averaging on the PE
1867	DO k = nzb_do, nzt_do
1868	DO j = nys, nyn
1869	DO i = nxl, nxr
1870	local_2d_l(j,k) = local_2d_l(j,k) + &
1871	local_pf(i,j,k)
1872	ENDDO
1873	ENDDO
1874	ENDDO
1875	#if defined( __parallel )
1876	!
1877	!-- Now do the averaging over all PEs along x
1878	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1879	CALL MPI_ALLREDUCE( local_2d_l(nys,nzb_do), &
1880	local_2d(nys,nzb_do), ngp, MPI_REAL, &
1881	MPI_SUM, comm1dx, ierr )
1882	#else
1883	local_2d = local_2d_l
1884	#endif
1885	local_2d = local_2d / ( nx + 1.0_wp )
1886
1887	DEALLOCATE( local_2d_l )
1888
1889	ELSE
1890	!
1891	!-- Just store the respective section on the local array
1892	!-- (but only if it is available on this PE!)
1893	IF ( section(is,s_ind) >= nxl .AND. section(is,s_ind) <= nxr ) &
1894	THEN
1895	local_2d = local_pf(section(is,s_ind),:,nzb_do:nzt_do)
1896	ENDIF
1897
1898	ENDIF
1899
1900	#if defined( __parallel )
1901	IF ( netcdf_data_format > 4 ) THEN
1902	!
1903	!-- Output in netCDF4/HDF5 format.
1904	!-- Output only on those PEs where the respective cross
1905	!-- sections reside. Cross sections averaged along x are
1906	!-- output on the respective first PE along x (myidx=0).
1907	IF ( ( section(is,s_ind) >= nxl .AND. &
1908	section(is,s_ind) <= nxr ) .OR. &
1909	( section(is,s_ind) == -1 .AND. myidx == 0 ) ) THEN
1910	#if defined( __netcdf )
1911	!
1912	!-- For parallel output, all cross sections are first
1913	!-- stored here on a local array and will be written to the
1914	!-- output file afterwards to increase the performance.
1915	DO j = nys, nyn
1916	DO k = nzb_do, nzt_do
1917	local_2d_sections_l(is,j,k) = local_2d(j,k)
1918	ENDDO
1919	ENDDO
1920	#endif
1921	ENDIF
1922
1923	ELSE
1924
1925	IF ( data_output_2d_on_each_pe ) THEN
1926	!
1927	!-- Output of partial arrays on each PE. If the cross
1928	!-- section does not reside on the PE, output special
1929	!-- index values.
1930	#if defined( __netcdf )
1931	IF ( myid == 0 ) THEN
1932	WRITE ( 23 ) time_since_reference_point, &
1933	do2d_yz_time_count(av), av
1934	ENDIF
1935	#endif
1936	DO i = 0, io_blocks-1
1937	IF ( i == io_group ) THEN
1938	IF ( ( section(is,s_ind) >= nxl .AND. &
1939	section(is,s_ind) <= nxr ) .OR. &
1940	( section(is,s_ind) == -1 .AND. &
1941	nxl-1 == -1 ) ) &
1942	THEN
1943	WRITE (23) nys, nyn, nzb_do, nzt_do, nzb, nzt+1
1944	WRITE (23) local_2d
1945	ELSE
1946	WRITE (23) -1, -1, -1, -1, -1, -1
1947	ENDIF
1948	ENDIF
1949	#if defined( __parallel )
1950	CALL MPI_BARRIER( comm2d, ierr )
1951	#endif
1952	ENDDO
1953
1954	ELSE
1955	!
1956	!-- PE0 receives partial arrays from all processors of the
1957	!-- respective cross section and outputs them. Here a
1958	!-- barrier has to be set, because otherwise
1959	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
1960	CALL MPI_BARRIER( comm2d, ierr )
1961
1962	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
1963	IF ( myid == 0 ) THEN
1964	!
1965	!-- Local array can be relocated directly.
1966	IF ( ( section(is,s_ind) >= nxl .AND. &
1967	section(is,s_ind) <= nxr ) .OR. &
1968	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
1969	THEN
1970	total_2d(nys:nyn,nzb_do:nzt_do) = local_2d
1971	ENDIF
1972	!
1973	!-- Receive data from all other PEs.
1974	DO n = 1, numprocs-1
1975	!
1976	!-- Receive index limits first, then array.
1977	!-- Index limits are received in arbitrary order from
1978	!-- the PEs.
1979	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1980	MPI_ANY_SOURCE, 0, comm2d, &
1981	status, ierr )
1982	!
1983	!-- Not all PEs have data for YZ-cross-section.
1984	IF ( ind(1) /= -9999 ) THEN
1985	sender = status(MPI_SOURCE)
1986	DEALLOCATE( local_2d )
1987	ALLOCATE( local_2d(ind(1):ind(2), &
1988	ind(3):ind(4)) )
1989	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1990	MPI_REAL, sender, 1, comm2d, &
1991	status, ierr )
1992	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
1993	local_2d
1994	ENDIF
1995	ENDDO
1996	!
1997	!-- Relocate the local array for the next loop increment
1998	DEALLOCATE( local_2d )
1999	ALLOCATE( local_2d(nys:nyn,nzb_do:nzt_do) )
2000
2001	#if defined( __netcdf )
2002	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2003	id_var_do2d(av,if), &
2004	total_2d(0:ny,nzb_do:nzt_do), &
2005	start = (/ is, 1, 1, do2d_yz_time_count(av) /), &
2006	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
2007	CALL netcdf_handle_error( 'data_output_2d', 61 )
2008	#endif
2009
2010	ELSE
2011	!
2012	!-- If the cross section resides on the PE, send the
2013	!-- local index limits, otherwise send -9999 to PE0.
2014	IF ( ( section(is,s_ind) >= nxl .AND. &
2015	section(is,s_ind) <= nxr ) .OR. &
2016	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
2017	THEN
2018	ind(1) = nys; ind(2) = nyn
2019	ind(3) = nzb_do; ind(4) = nzt_do
2020	ELSE
2021	ind(1) = -9999; ind(2) = -9999
2022	ind(3) = -9999; ind(4) = -9999
2023	ENDIF
2024	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
2025	comm2d, ierr )
2026	!
2027	!-- If applicable, send data to PE0.
2028	IF ( ind(1) /= -9999 ) THEN
2029	CALL MPI_SEND( local_2d(nys,nzb_do), ngp, &
2030	MPI_REAL, 0, 1, comm2d, ierr )
2031	ENDIF
2032	ENDIF
2033	!
2034	!-- A barrier has to be set, because otherwise some PEs may
2035	!-- proceed too fast so that PE0 may receive wrong data on
2036	!-- tag 0
2037	CALL MPI_BARRIER( comm2d, ierr )
2038	ENDIF
2039
2040	ENDIF
2041	#else
2042	#if defined( __netcdf )
2043	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2044	id_var_do2d(av,if), &
2045	local_2d(nys:nyn,nzb_do:nzt_do), &
2046	start = (/ is, 1, 1, do2d_xz_time_count(av) /), &
2047	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
2048	CALL netcdf_handle_error( 'data_output_2d', 452 )
2049	#endif
2050	#endif
2051
2052	END SELECT
2053
2054	is = is + 1
2055	ENDDO loop1
2056
2057	!
2058	!-- For parallel output, all data were collected before on a local array
2059	!-- and are written now to the netcdf file. This must be done to increase
2060	!-- the performance of the parallel output.
2061	#if defined( __netcdf )
2062	IF ( netcdf_data_format > 4 ) THEN
2063
2064	SELECT CASE ( mode )
2065
2066	CASE ( 'xy' )
2067	IF ( two_d ) THEN
2068	nis = 1
2069	two_d = .FALSE.
2070	ELSE
2071	nis = ns
2072	ENDIF
2073	!
2074	!-- Do not output redundant ghost point data except for the
2075	!-- boundaries of the total domain.
2076	! IF ( nxr == nx .AND. nyn /= ny ) THEN
2077	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2078	! id_var_do2d(av,if), &
2079	! local_2d_sections(nxl:nxr+1, &
2080	! nys:nyn,1:nis), &
2081	! start = (/ nxl+1, nys+1, 1, &
2082	! do2d_xy_time_count(av) /), &
2083	! count = (/ nxr-nxl+2, &
2084	! nyn-nys+1, nis, 1 &
2085	! /) )
2086	! ELSEIF ( nxr /= nx .AND. nyn == ny ) THEN
2087	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2088	! id_var_do2d(av,if), &
2089	! local_2d_sections(nxl:nxr, &
2090	! nys:nyn+1,1:nis), &
2091	! start = (/ nxl+1, nys+1, 1, &
2092	! do2d_xy_time_count(av) /), &
2093	! count = (/ nxr-nxl+1, &
2094	! nyn-nys+2, nis, 1 &
2095	! /) )
2096	! ELSEIF ( nxr == nx .AND. nyn == ny ) THEN
2097	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2098	! id_var_do2d(av,if), &
2099	! local_2d_sections(nxl:nxr+1, &
2100	! nys:nyn+1,1:nis), &
2101	! start = (/ nxl+1, nys+1, 1, &
2102	! do2d_xy_time_count(av) /), &
2103	! count = (/ nxr-nxl+2, &
2104	! nyn-nys+2, nis, 1 &
2105	! /) )
2106	! ELSE
2107	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2108	id_var_do2d(av,if), &
2109	local_2d_sections(nxl:nxr, &
2110	nys:nyn,1:nis), &
2111	start = (/ nxl+1, nys+1, 1, &
2112	do2d_xy_time_count(av) /), &
2113	count = (/ nxr-nxl+1, &
2114	nyn-nys+1, nis, 1 &
2115	/) )
2116	! ENDIF
2117
2118	CALL netcdf_handle_error( 'data_output_2d', 55 )
2119
2120	CASE ( 'xz' )
2121	!
2122	!-- First, all PEs get the information of all cross-sections.
2123	!-- Then the data are written to the output file by all PEs
2124	!-- while NF90_COLLECTIVE is set in subroutine
2125	!-- define_netcdf_header. Although redundant information are
2126	!-- written to the output file in that case, the performance
2127	!-- is significantly better compared to the case where only
2128	!-- the first row of PEs in x-direction (myidx = 0) is given
2129	!-- the output while NF90_INDEPENDENT is set.
2130	IF ( npey /= 1 ) THEN
2131
2132	#if defined( __parallel )
2133	!
2134	!-- Distribute data over all PEs along y
2135	ngp = ( nxr-nxl+1 ) * ( nzt_do-nzb_do+1 ) * ns
2136	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
2137	CALL MPI_ALLREDUCE( local_2d_sections_l(nxl,1,nzb_do), &
2138	local_2d_sections(nxl,1,nzb_do), &
2139	ngp, MPI_REAL, MPI_SUM, comm1dy, &
2140	ierr )
2141	#else
2142	local_2d_sections = local_2d_sections_l
2143	#endif
2144	ENDIF
2145	!
2146	!-- Do not output redundant ghost point data except for the
2147	!-- boundaries of the total domain.
2148	! IF ( nxr == nx ) THEN
2149	! nc_stat = NF90_PUT_VAR( id_set_xz(av), &
2150	! id_var_do2d(av,if), &
2151	! local_2d_sections(nxl:nxr+1,1:ns, &
2152	! nzb_do:nzt_do), &
2153	! start = (/ nxl+1, 1, 1, &
2154	! do2d_xz_time_count(av) /), &
2155	! count = (/ nxr-nxl+2, ns, nzt_do-nzb_do+1, &
2156	! 1 /) )
2157	! ELSE
2158	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
2159	id_var_do2d(av,if), &
2160	local_2d_sections(nxl:nxr,1:ns, &
2161	nzb_do:nzt_do), &
2162	start = (/ nxl+1, 1, 1, &
2163	do2d_xz_time_count(av) /), &
2164	count = (/ nxr-nxl+1, ns, nzt_do-nzb_do+1, &
2165	1 /) )
2166	! ENDIF
2167
2168	CALL netcdf_handle_error( 'data_output_2d', 57 )
2169
2170	CASE ( 'yz' )
2171	!
2172	!-- First, all PEs get the information of all cross-sections.
2173	!-- Then the data are written to the output file by all PEs
2174	!-- while NF90_COLLECTIVE is set in subroutine
2175	!-- define_netcdf_header. Although redundant information are
2176	!-- written to the output file in that case, the performance
2177	!-- is significantly better compared to the case where only
2178	!-- the first row of PEs in y-direction (myidy = 0) is given
2179	!-- the output while NF90_INDEPENDENT is set.
2180	IF ( npex /= 1 ) THEN
2181
2182	#if defined( __parallel )
2183	!
2184	!-- Distribute data over all PEs along x
2185	ngp = ( nyn-nys+1 ) * ( nzt-nzb + 2 ) * ns
2186	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
2187	CALL MPI_ALLREDUCE( local_2d_sections_l(1,nys,nzb_do), &
2188	local_2d_sections(1,nys,nzb_do), &
2189	ngp, MPI_REAL, MPI_SUM, comm1dx, &
2190	ierr )
2191	#else
2192	local_2d_sections = local_2d_sections_l
2193	#endif
2194	ENDIF
2195	!
2196	!-- Do not output redundant ghost point data except for the
2197	!-- boundaries of the total domain.
2198	! IF ( nyn == ny ) THEN
2199	! nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2200	! id_var_do2d(av,if), &
2201	! local_2d_sections(1:ns, &
2202	! nys:nyn+1,nzb_do:nzt_do), &
2203	! start = (/ 1, nys+1, 1, &
2204	! do2d_yz_time_count(av) /), &
2205	! count = (/ ns, nyn-nys+2, &
2206	! nzt_do-nzb_do+1, 1 /) )
2207	! ELSE
2208	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2209	id_var_do2d(av,if), &
2210	local_2d_sections(1:ns,nys:nyn, &
2211	nzb_do:nzt_do), &
2212	start = (/ 1, nys+1, 1, &
2213	do2d_yz_time_count(av) /), &
2214	count = (/ ns, nyn-nys+1, &
2215	nzt_do-nzb_do+1, 1 /) )
2216	! ENDIF
2217
2218	CALL netcdf_handle_error( 'data_output_2d', 60 )
2219
2220	CASE DEFAULT
2221	message_string = 'unknown cross-section: ' // TRIM( mode )
2222	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
2223
2224	END SELECT
2225
2226	ENDIF
2227	#endif
2228	ENDIF
2229
2230	if = if + 1
2231	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
2232	do2d_mode = do2d(av,if)(l-1:l)
2233
2234	ENDDO
2235
2236	!
2237	!-- Deallocate temporary arrays.
2238	IF ( ALLOCATED( level_z ) ) DEALLOCATE( level_z )
2239	IF ( netcdf_data_format > 4 ) THEN
2240	DEALLOCATE( local_pf, local_2d, local_2d_sections )
2241	IF( mode == 'xz' .OR. mode == 'yz' ) DEALLOCATE( local_2d_sections_l )
2242	ENDIF
2243	#if defined( __parallel )
2244	IF ( .NOT. data_output_2d_on_each_pe .AND. myid == 0 ) THEN
2245	DEALLOCATE( total_2d )
2246	ENDIF
2247	#endif
2248
2249	!
2250	!-- Close plot output file.
2251	file_id = 20 + s_ind
2252
2253	IF ( data_output_2d_on_each_pe ) THEN
2254	DO i = 0, io_blocks-1
2255	IF ( i == io_group ) THEN
2256	CALL close_file( file_id )
2257	ENDIF
2258	#if defined( __parallel )
2259	CALL MPI_BARRIER( comm2d, ierr )
2260	#endif
2261	ENDDO
2262	ELSE
2263	IF ( myid == 0 ) CALL close_file( file_id )
2264	ENDIF
2265
2266	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
2267
2268	END SUBROUTINE data_output_2d

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |