Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 4553

Last change on this file since 4553 was 4518, checked in by suehring, 5 years ago
Diagnostic output: Define arrays over ghost points in order to allow for standard mpi-io treatment. By this modularization of restart-data input is possible with the module interface. Move input of restart data to doq_rrd_local. Enable mpi-io for restart data. Bugfix: add missing restart input of wtheta_av, wq_av, wu_av, and wv_av.
Property svn:keywords set to `Id`
File size: 91.8 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 1997-2020 Leibniz Universitaet Hannover
18	!------------------------------------------------------------------------------!
19	!
20	! Current revisions:
21	! ------------------
22	!
23	!
24	! Former revisions:
25	! -----------------
26	! $Id: data_output_2d.f90 4518 2020-05-04 15:44:28Z eckhard $
27	! remove double index
28	!
29	! 4514 2020-04-30 16:29:59Z suehring
30	! Enable output of qsurf and ssurf
31	!
32	! 4500 2020-04-17 10:12:45Z suehring
33	! Unify output conversion of sensible and latent heat flux
34	!
35	! 4457 2020-03-11 14:20:43Z raasch
36	! use statement for exchange horiz added
37	!
38	! 4444 2020-03-05 15:59:50Z raasch
39	! bugfix: cpp-directives for serial mode added
40	!
41	! 4442 2020-03-04 19:21:13Z suehring
42	! Change order of dimension in surface array %frac to allow for better
43	! vectorization.
44	!
45	! 4441 2020-03-04 19:20:35Z suehring
46	! Introduction of wall_flags_total_0, which currently sets bits based on static
47	! topography information used in wall_flags_static_0
48	!
49	! 4331 2019-12-10 18:25:02Z suehring
50	! Move 2-m potential temperature output to diagnostic_output_quantities
51	!
52	! 4329 2019-12-10 15:46:36Z motisi
53	! Renamed wall_flags_0 to wall_flags_static_0
54	!
55	! 4182 2019-08-22 15:20:23Z scharf
56	! Corrected "Former revisions" section
57	!
58	! 4048 2019-06-21 21:00:21Z knoop
59	! Removed turbulence_closure_mod dependency
60	!
61	! 4039 2019-06-18 10:32:41Z suehring
62	! modularize diagnostic output
63	!
64	! 3994 2019-05-22 18:08:09Z suehring
65	! output of turbulence intensity added
66	!
67	! 3987 2019-05-22 09:52:13Z kanani
68	! Introduce alternative switch for debug output during timestepping
69	!
70	! 3943 2019-05-02 09:50:41Z maronga
71	! Added output of qsws for green roofs.
72	!
73	! 3885 2019-04-11 11:29:34Z kanani
74	! Changes related to global restructuring of location messages and introduction
75	! of additional debug messages
76	!
77	! 3766 2019-02-26 16:23:41Z raasch
78	! unused variables removed
79	!
80	! 3655 2019-01-07 16:51:22Z knoop
81	! Bugfix: use time_since_reference_point instead of simulated_time (relevant
82	! when using wall/soil spinup)
83	!
84	! Revision 1.1 1997/08/11 06:24:09 raasch
85	! Initial revision
86	!
87	!
88	! Description:
89	! ------------
90	!> Data output of cross-sections in netCDF format or binary format
91	!> to be later converted to NetCDF by helper routine combine_plot_fields.
92	!> Attention: The position of the sectional planes is still not always computed
93	!> --------- correctly. (zu is used always)!
94	!------------------------------------------------------------------------------!
95	SUBROUTINE data_output_2d( mode, av )
96
97
98	USE arrays_3d, &
99	ONLY: dzw, d_exner, e, heatflux_output_conversion, p, pt, q, ql, ql_c, ql_v, s, tend, u, &
100	v, vpt, w, waterflux_output_conversion, zu, zw
101
102	USE averaging
103
104	USE basic_constants_and_equations_mod, &
105	ONLY: lv_d_cp
106
107	USE bulk_cloud_model_mod, &
108	ONLY: bulk_cloud_model
109
110	USE control_parameters, &
111	ONLY: data_output_2d_on_each_pe, &
112	data_output_xy, data_output_xz, data_output_yz, &
113	debug_output_timestep, &
114	do2d, &
115	do2d_xy_last_time, do2d_xy_time_count, &
116	do2d_xz_last_time, do2d_xz_time_count, &
117	do2d_yz_last_time, do2d_yz_time_count, &
118	ibc_uv_b, io_blocks, io_group, message_string, &
119	ntdim_2d_xy, ntdim_2d_xz, ntdim_2d_yz, &
120	psolver, section, &
121	time_since_reference_point
122
123	USE cpulog, &
124	ONLY: cpu_log, log_point
125
126	USE exchange_horiz_mod, &
127	ONLY: exchange_horiz
128
129	USE indices, &
130	ONLY: nbgp, nx, nxl, nxlg, nxr, nxrg, ny, nyn, nyng, nys, nysg, &
131	nzb, nzt, &
132	topo_top_ind, &
133	wall_flags_total_0
134
135	USE kinds
136
137	USE land_surface_model_mod, &
138	ONLY: zs
139
140	USE module_interface, &
141	ONLY: module_interface_data_output_2d
142
143	#if defined( __netcdf )
144	USE NETCDF
145	#endif
146
147	USE netcdf_interface, &
148	ONLY: fill_value, id_set_xy, id_set_xz, id_set_yz, id_var_do2d, &
149	id_var_time_xy, id_var_time_xz, id_var_time_yz, nc_stat, &
150	netcdf_data_format, netcdf_handle_error
151
152	USE particle_attributes, &
153	ONLY: grid_particles, number_of_particles, particle_advection_start, &
154	particles, prt_count
155
156	USE pegrid
157
158	USE surface_mod, &
159	ONLY: ind_pav_green, ind_veg_wall, ind_wat_win, surf_def_h, &
160	surf_lsm_h, surf_usm_h
161
162
163	IMPLICIT NONE
164
165	CHARACTER (LEN=2) :: do2d_mode !< output mode of variable ('xy', 'xz', 'yz')
166	CHARACTER (LEN=2) :: mode !< mode with which the routine is called ('xy', 'xz', 'yz')
167	CHARACTER (LEN=4) :: grid !< string defining the vertical grid
168
169	INTEGER(iwp) :: av !< flag for (non-)average output
170	INTEGER(iwp) :: ngp !< number of grid points of an output slice
171	INTEGER(iwp) :: file_id !< id of output files
172	INTEGER(iwp) :: flag_nr !< number of masking flag
173	INTEGER(iwp) :: i !< loop index
174	INTEGER(iwp) :: is !< slice index
175	INTEGER(iwp) :: ivar !< variable index
176	INTEGER(iwp) :: j !< loop index
177	INTEGER(iwp) :: k !< loop index
178	INTEGER(iwp) :: l !< loop index
179	INTEGER(iwp) :: layer_xy !< vertical index of a xy slice in array 'local_pf'
180	INTEGER(iwp) :: m !< loop index
181	INTEGER(iwp) :: n !< loop index
182	INTEGER(iwp) :: nis !< number of vertical slices to be written via parallel NetCDF output
183	INTEGER(iwp) :: ns !< number of output slices
184	INTEGER(iwp) :: nzb_do !< lower limit of the data field (usually nzb)
185	INTEGER(iwp) :: nzt_do !< upper limit of the data field (usually nzt+1)
186	INTEGER(iwp) :: s_ind !< index of slice types (xy=1, xz=2, yz=3)
187	#if defined( __parallel )
188	INTEGER(iwp) :: iis !< vertical index of a xy slice in array 'local_2d_sections'
189	INTEGER(iwp) :: sender !< PE id of sending PE
190	INTEGER(iwp) :: ind(4) !< index limits (lower/upper bounds) of array 'local_2d'
191	#endif
192
193	LOGICAL :: found !< true if output variable was found
194	LOGICAL :: resorted !< true if variable is resorted
195	LOGICAL :: two_d !< true if variable is only two dimensional
196
197	REAL(wp) :: mean_r !< mean particle radius
198	REAL(wp) :: s_r2 !< sum( particle-radius**2 )
199	REAL(wp) :: s_r3 !< sum( particle-radius**3 )
200
201	REAL(wp), DIMENSION(:), ALLOCATABLE :: level_z !< z levels for output array
202	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d !< local 2-dimensional array containing output values
203	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d_l !< local 2-dimensional array containing output values
204
205	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !< output array
206	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections !< local array containing values at all slices
207	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections_l !< local array containing values at all slices
208
209	#if defined( __parallel )
210	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: total_2d !< same as local_2d
211	#endif
212	REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !< points to array which shall be output
213
214
215	IF ( debug_output_timestep ) CALL debug_message( 'data_output_2d', 'start' )
216	!
217	!-- Immediate return, if no output is requested (no respective sections
218	!-- found in parameter data_output)
219	IF ( mode == 'xy' .AND. .NOT. data_output_xy(av) ) RETURN
220	IF ( mode == 'xz' .AND. .NOT. data_output_xz(av) ) RETURN
221	IF ( mode == 'yz' .AND. .NOT. data_output_yz(av) ) RETURN
222
223	CALL cpu_log (log_point(3),'data_output_2d','start')
224
225	two_d = .FALSE. ! local variable to distinguish between output of pure 2D
226	! arrays and cross-sections of 3D arrays.
227
228	!
229	!-- Depending on the orientation of the cross-section, the respective output
230	!-- files have to be opened.
231	SELECT CASE ( mode )
232
233	CASE ( 'xy' )
234	s_ind = 1
235	ALLOCATE( level_z(nzb:nzt+1), local_2d(nxl:nxr,nys:nyn) )
236
237	IF ( netcdf_data_format > 4 ) THEN
238	ns = 1
239	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
240	ns = ns + 1
241	ENDDO
242	ns = ns - 1
243	ALLOCATE( local_2d_sections(nxl:nxr,nys:nyn,1:ns) )
244	local_2d_sections = 0.0_wp
245	ENDIF
246
247	!
248	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
249	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
250	CALL check_open( 101+av*10 )
251	ENDIF
252	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
253	CALL check_open( 21 )
254	ELSE
255	IF ( myid == 0 ) THEN
256	#if defined( __parallel )
257	ALLOCATE( total_2d(0:nx,0:ny) )
258	#endif
259	ENDIF
260	ENDIF
261
262	CASE ( 'xz' )
263	s_ind = 2
264	ALLOCATE( local_2d(nxl:nxr,nzb:nzt+1) )
265
266	IF ( netcdf_data_format > 4 ) THEN
267	ns = 1
268	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
269	ns = ns + 1
270	ENDDO
271	ns = ns - 1
272	ALLOCATE( local_2d_sections(nxl:nxr,1:ns,nzb:nzt+1) )
273	ALLOCATE( local_2d_sections_l(nxl:nxr,1:ns,nzb:nzt+1) )
274	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
275	ENDIF
276
277	!
278	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
279	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
280	CALL check_open( 102+av*10 )
281	ENDIF
282
283	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
284	CALL check_open( 22 )
285	ELSE
286	IF ( myid == 0 ) THEN
287	#if defined( __parallel )
288	ALLOCATE( total_2d(0:nx,nzb:nzt+1) )
289	#endif
290	ENDIF
291	ENDIF
292
293	CASE ( 'yz' )
294	s_ind = 3
295	ALLOCATE( local_2d(nys:nyn,nzb:nzt+1) )
296
297	IF ( netcdf_data_format > 4 ) THEN
298	ns = 1
299	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
300	ns = ns + 1
301	ENDDO
302	ns = ns - 1
303	ALLOCATE( local_2d_sections(1:ns,nys:nyn,nzb:nzt+1) )
304	ALLOCATE( local_2d_sections_l(1:ns,nys:nyn,nzb:nzt+1) )
305	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
306	ENDIF
307
308	!
309	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
310	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
311	CALL check_open( 103+av*10 )
312	ENDIF
313
314	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
315	CALL check_open( 23 )
316	ELSE
317	IF ( myid == 0 ) THEN
318	#if defined( __parallel )
319	ALLOCATE( total_2d(0:ny,nzb:nzt+1) )
320	#endif
321	ENDIF
322	ENDIF
323
324	CASE DEFAULT
325	message_string = 'unknown cross-section: ' // TRIM( mode )
326	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
327
328	END SELECT
329
330	!
331	!-- For parallel netcdf output the time axis must be limited. Return, if this
332	!-- limit is exceeded. This could be the case, if the simulated time exceeds
333	!-- the given end time by the length of the given output interval.
334	IF ( netcdf_data_format > 4 ) THEN
335	IF ( mode == 'xy' .AND. do2d_xy_time_count(av) + 1 > &
336	ntdim_2d_xy(av) ) THEN
337	WRITE ( message_string, * ) 'Output of xy cross-sections is not ', &
338	'given at t=', time_since_reference_point, 's because the', &
339	' maximum number of output time levels is exceeded.'
340	CALL message( 'data_output_2d', 'PA0384', 0, 1, 0, 6, 0 )
341	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
342	RETURN
343	ENDIF
344	IF ( mode == 'xz' .AND. do2d_xz_time_count(av) + 1 > &
345	ntdim_2d_xz(av) ) THEN
346	WRITE ( message_string, * ) 'Output of xz cross-sections is not ', &
347	'given at t=', time_since_reference_point, 's because the', &
348	' maximum number of output time levels is exceeded.'
349	CALL message( 'data_output_2d', 'PA0385', 0, 1, 0, 6, 0 )
350	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
351	RETURN
352	ENDIF
353	IF ( mode == 'yz' .AND. do2d_yz_time_count(av) + 1 > &
354	ntdim_2d_yz(av) ) THEN
355	WRITE ( message_string, * ) 'Output of yz cross-sections is not ', &
356	'given at t=', time_since_reference_point, 's because the', &
357	' maximum number of output time levels is exceeded.'
358	CALL message( 'data_output_2d', 'PA0386', 0, 1, 0, 6, 0 )
359	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
360	RETURN
361	ENDIF
362	ENDIF
363
364	!
365	!-- Allocate a temporary array for resorting (kji -> ijk).
366	ALLOCATE( local_pf(nxl:nxr,nys:nyn,nzb:nzt+1) )
367	local_pf = 0.0
368
369	!
370	!-- Loop of all variables to be written.
371	!-- Output dimensions chosen
372	ivar = 1
373	l = MAX( 2, LEN_TRIM( do2d(av,ivar) ) )
374	do2d_mode = do2d(av,ivar)(l-1:l)
375
376	DO WHILE ( do2d(av,ivar)(1:1) /= ' ' )
377
378	IF ( do2d_mode == mode ) THEN
379	!
380	!-- Set flag to steer output of radiation, land-surface, or user-defined
381	!-- quantities
382	found = .FALSE.
383
384	nzb_do = nzb
385	nzt_do = nzt+1
386	!
387	!-- Before each output, set array local_pf to fill value
388	local_pf = fill_value
389	!
390	!-- Set masking flag for topography for not resorted arrays
391	flag_nr = 0
392
393	!
394	!-- Store the array chosen on the temporary array.
395	resorted = .FALSE.
396	SELECT CASE ( TRIM( do2d(av,ivar) ) )
397	CASE ( 'e_xy', 'e_xz', 'e_yz' )
398	IF ( av == 0 ) THEN
399	to_be_resorted => e
400	ELSE
401	IF ( .NOT. ALLOCATED( e_av ) ) THEN
402	ALLOCATE( e_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
403	e_av = REAL( fill_value, KIND = wp )
404	ENDIF
405	to_be_resorted => e_av
406	ENDIF
407	IF ( mode == 'xy' ) level_z = zu
408
409	CASE ( 'thetal_xy', 'thetal_xz', 'thetal_yz' )
410	IF ( av == 0 ) THEN
411	to_be_resorted => pt
412	ELSE
413	IF ( .NOT. ALLOCATED( lpt_av ) ) THEN
414	ALLOCATE( lpt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
415	lpt_av = REAL( fill_value, KIND = wp )
416	ENDIF
417	to_be_resorted => lpt_av
418	ENDIF
419	IF ( mode == 'xy' ) level_z = zu
420
421	CASE ( 'lwp*_xy' ) ! 2d-array
422	IF ( av == 0 ) THEN
423	DO i = nxl, nxr
424	DO j = nys, nyn
425	local_pf(i,j,nzb+1) = SUM( ql(nzb:nzt,j,i) * &
426	dzw(1:nzt+1) )
427	ENDDO
428	ENDDO
429	ELSE
430	IF ( .NOT. ALLOCATED( lwp_av ) ) THEN
431	ALLOCATE( lwp_av(nysg:nyng,nxlg:nxrg) )
432	lwp_av = REAL( fill_value, KIND = wp )
433	ENDIF
434	DO i = nxl, nxr
435	DO j = nys, nyn
436	local_pf(i,j,nzb+1) = lwp_av(j,i)
437	ENDDO
438	ENDDO
439	ENDIF
440	resorted = .TRUE.
441	two_d = .TRUE.
442	level_z(nzb+1) = zu(nzb+1)
443
444	CASE ( 'ghf*_xy' ) ! 2d-array
445	IF ( av == 0 ) THEN
446	DO m = 1, surf_lsm_h%ns
447	i = surf_lsm_h%i(m)
448	j = surf_lsm_h%j(m)
449	local_pf(i,j,nzb+1) = surf_lsm_h%ghf(m)
450	ENDDO
451	DO m = 1, surf_usm_h%ns
452	i = surf_usm_h%i(m)
453	j = surf_usm_h%j(m)
454	local_pf(i,j,nzb+1) = surf_usm_h%frac(m,ind_veg_wall) * &
455	surf_usm_h%wghf_eb(m) + &
456	surf_usm_h%frac(m,ind_pav_green) * &
457	surf_usm_h%wghf_eb_green(m) + &
458	surf_usm_h%frac(m,ind_wat_win) * &
459	surf_usm_h%wghf_eb_window(m)
460	ENDDO
461	ELSE
462	IF ( .NOT. ALLOCATED( ghf_av ) ) THEN
463	ALLOCATE( ghf_av(nysg:nyng,nxlg:nxrg) )
464	ghf_av = REAL( fill_value, KIND = wp )
465	ENDIF
466	DO i = nxl, nxr
467	DO j = nys, nyn
468	local_pf(i,j,nzb+1) = ghf_av(j,i)
469	ENDDO
470	ENDDO
471	ENDIF
472
473	resorted = .TRUE.
474	two_d = .TRUE.
475	level_z(nzb+1) = zu(nzb+1)
476
477	CASE ( 'ol*_xy' ) ! 2d-array
478	IF ( av == 0 ) THEN
479	DO m = 1, surf_def_h(0)%ns
480	i = surf_def_h(0)%i(m)
481	j = surf_def_h(0)%j(m)
482	local_pf(i,j,nzb+1) = surf_def_h(0)%ol(m)
483	ENDDO
484	DO m = 1, surf_lsm_h%ns
485	i = surf_lsm_h%i(m)
486	j = surf_lsm_h%j(m)
487	local_pf(i,j,nzb+1) = surf_lsm_h%ol(m)
488	ENDDO
489	DO m = 1, surf_usm_h%ns
490	i = surf_usm_h%i(m)
491	j = surf_usm_h%j(m)
492	local_pf(i,j,nzb+1) = surf_usm_h%ol(m)
493	ENDDO
494	ELSE
495	IF ( .NOT. ALLOCATED( ol_av ) ) THEN
496	ALLOCATE( ol_av(nysg:nyng,nxlg:nxrg) )
497	ol_av = REAL( fill_value, KIND = wp )
498	ENDIF
499	DO i = nxl, nxr
500	DO j = nys, nyn
501	local_pf(i,j,nzb+1) = ol_av(j,i)
502	ENDDO
503	ENDDO
504	ENDIF
505	resorted = .TRUE.
506	two_d = .TRUE.
507	level_z(nzb+1) = zu(nzb+1)
508
509	CASE ( 'p_xy', 'p_xz', 'p_yz' )
510	IF ( av == 0 ) THEN
511	IF ( psolver /= 'sor' ) CALL exchange_horiz( p, nbgp )
512	to_be_resorted => p
513	ELSE
514	IF ( .NOT. ALLOCATED( p_av ) ) THEN
515	ALLOCATE( p_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
516	p_av = REAL( fill_value, KIND = wp )
517	ENDIF
518	IF ( psolver /= 'sor' ) CALL exchange_horiz( p_av, nbgp )
519	to_be_resorted => p_av
520	ENDIF
521	IF ( mode == 'xy' ) level_z = zu
522
523	CASE ( 'pc_xy', 'pc_xz', 'pc_yz' ) ! particle concentration
524	IF ( av == 0 ) THEN
525	IF ( time_since_reference_point >= particle_advection_start ) THEN
526	tend = prt_count
527	! CALL exchange_horiz( tend, nbgp )
528	ELSE
529	tend = 0.0_wp
530	ENDIF
531	DO i = nxl, nxr
532	DO j = nys, nyn
533	DO k = nzb, nzt+1
534	local_pf(i,j,k) = tend(k,j,i)
535	ENDDO
536	ENDDO
537	ENDDO
538	resorted = .TRUE.
539	ELSE
540	IF ( .NOT. ALLOCATED( pc_av ) ) THEN
541	ALLOCATE( pc_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
542	pc_av = REAL( fill_value, KIND = wp )
543	ENDIF
544	! CALL exchange_horiz( pc_av, nbgp )
545	to_be_resorted => pc_av
546	ENDIF
547
548	CASE ( 'pr_xy', 'pr_xz', 'pr_yz' ) ! mean particle radius (effective radius)
549	IF ( av == 0 ) THEN
550	IF ( time_since_reference_point >= particle_advection_start ) THEN
551	DO i = nxl, nxr
552	DO j = nys, nyn
553	DO k = nzb, nzt+1
554	number_of_particles = prt_count(k,j,i)
555	IF (number_of_particles <= 0) CYCLE
556	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
557	s_r2 = 0.0_wp
558	s_r3 = 0.0_wp
559	DO n = 1, number_of_particles
560	IF ( particles(n)%particle_mask ) THEN
561	s_r2 = s_r2 + particles(n)%radius*2 &
562	particles(n)%weight_factor
563	s_r3 = s_r3 + particles(n)%radius*3 &
564	particles(n)%weight_factor
565	ENDIF
566	ENDDO
567	IF ( s_r2 > 0.0_wp ) THEN
568	mean_r = s_r3 / s_r2
569	ELSE
570	mean_r = 0.0_wp
571	ENDIF
572	tend(k,j,i) = mean_r
573	ENDDO
574	ENDDO
575	ENDDO
576	! CALL exchange_horiz( tend, nbgp )
577	ELSE
578	tend = 0.0_wp
579	ENDIF
580	DO i = nxl, nxr
581	DO j = nys, nyn
582	DO k = nzb, nzt+1
583	local_pf(i,j,k) = tend(k,j,i)
584	ENDDO
585	ENDDO
586	ENDDO
587	resorted = .TRUE.
588	ELSE
589	IF ( .NOT. ALLOCATED( pr_av ) ) THEN
590	ALLOCATE( pr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
591	pr_av = REAL( fill_value, KIND = wp )
592	ENDIF
593	! CALL exchange_horiz( pr_av, nbgp )
594	to_be_resorted => pr_av
595	ENDIF
596
597	CASE ( 'theta_xy', 'theta_xz', 'theta_yz' )
598	IF ( av == 0 ) THEN
599	IF ( .NOT. bulk_cloud_model ) THEN
600	to_be_resorted => pt
601	ELSE
602	DO i = nxl, nxr
603	DO j = nys, nyn
604	DO k = nzb, nzt+1
605	local_pf(i,j,k) = pt(k,j,i) + lv_d_cp * &
606	d_exner(k) * &
607	ql(k,j,i)
608	ENDDO
609	ENDDO
610	ENDDO
611	resorted = .TRUE.
612	ENDIF
613	ELSE
614	IF ( .NOT. ALLOCATED( pt_av ) ) THEN
615	ALLOCATE( pt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
616	pt_av = REAL( fill_value, KIND = wp )
617	ENDIF
618	to_be_resorted => pt_av
619	ENDIF
620	IF ( mode == 'xy' ) level_z = zu
621
622	CASE ( 'q_xy', 'q_xz', 'q_yz' )
623	IF ( av == 0 ) THEN
624	to_be_resorted => q
625	ELSE
626	IF ( .NOT. ALLOCATED( q_av ) ) THEN
627	ALLOCATE( q_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
628	q_av = REAL( fill_value, KIND = wp )
629	ENDIF
630	to_be_resorted => q_av
631	ENDIF
632	IF ( mode == 'xy' ) level_z = zu
633
634	CASE ( 'ql_xy', 'ql_xz', 'ql_yz' )
635	IF ( av == 0 ) THEN
636	to_be_resorted => ql
637	ELSE
638	IF ( .NOT. ALLOCATED( ql_av ) ) THEN
639	ALLOCATE( ql_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
640	ql_av = REAL( fill_value, KIND = wp )
641	ENDIF
642	to_be_resorted => ql_av
643	ENDIF
644	IF ( mode == 'xy' ) level_z = zu
645
646	CASE ( 'ql_c_xy', 'ql_c_xz', 'ql_c_yz' )
647	IF ( av == 0 ) THEN
648	to_be_resorted => ql_c
649	ELSE
650	IF ( .NOT. ALLOCATED( ql_c_av ) ) THEN
651	ALLOCATE( ql_c_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
652	ql_c_av = REAL( fill_value, KIND = wp )
653	ENDIF
654	to_be_resorted => ql_c_av
655	ENDIF
656	IF ( mode == 'xy' ) level_z = zu
657
658	CASE ( 'ql_v_xy', 'ql_v_xz', 'ql_v_yz' )
659	IF ( av == 0 ) THEN
660	to_be_resorted => ql_v
661	ELSE
662	IF ( .NOT. ALLOCATED( ql_v_av ) ) THEN
663	ALLOCATE( ql_v_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
664	ql_v_av = REAL( fill_value, KIND = wp )
665	ENDIF
666	to_be_resorted => ql_v_av
667	ENDIF
668	IF ( mode == 'xy' ) level_z = zu
669
670	CASE ( 'ql_vp_xy', 'ql_vp_xz', 'ql_vp_yz' )
671	IF ( av == 0 ) THEN
672	IF ( time_since_reference_point >= particle_advection_start ) THEN
673	DO i = nxl, nxr
674	DO j = nys, nyn
675	DO k = nzb, nzt+1
676	number_of_particles = prt_count(k,j,i)
677	IF (number_of_particles <= 0) CYCLE
678	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
679	DO n = 1, number_of_particles
680	IF ( particles(n)%particle_mask ) THEN
681	tend(k,j,i) = tend(k,j,i) + &
682	particles(n)%weight_factor / &
683	prt_count(k,j,i)
684	ENDIF
685	ENDDO
686	ENDDO
687	ENDDO
688	ENDDO
689	! CALL exchange_horiz( tend, nbgp )
690	ELSE
691	tend = 0.0_wp
692	ENDIF
693	DO i = nxl, nxr
694	DO j = nys, nyn
695	DO k = nzb, nzt+1
696	local_pf(i,j,k) = tend(k,j,i)
697	ENDDO
698	ENDDO
699	ENDDO
700	resorted = .TRUE.
701	ELSE
702	IF ( .NOT. ALLOCATED( ql_vp_av ) ) THEN
703	ALLOCATE( ql_vp_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
704	ql_vp_av = REAL( fill_value, KIND = wp )
705	ENDIF
706	! CALL exchange_horiz( ql_vp_av, nbgp )
707	to_be_resorted => ql_vp_av
708	ENDIF
709	IF ( mode == 'xy' ) level_z = zu
710
711	CASE ( 'qsurf*_xy' ) ! 2d-array
712	IF ( av == 0 ) THEN
713	DO m = 1, surf_def_h(0)%ns
714	i = surf_def_h(0)%i(m)
715	j = surf_def_h(0)%j(m)
716	local_pf(i,j,nzb+1) = surf_def_h(0)%q_surface(m)
717	ENDDO
718
719	DO m = 1, surf_lsm_h%ns
720	i = surf_lsm_h%i(m)
721	j = surf_lsm_h%j(m)
722	local_pf(i,j,nzb+1) = surf_lsm_h%q_surface(m)
723	ENDDO
724
725	DO m = 1, surf_usm_h%ns
726	i = surf_usm_h%i(m)
727	j = surf_usm_h%j(m)
728	local_pf(i,j,nzb+1) = surf_usm_h%q_surface(m)
729	ENDDO
730
731	ELSE
732	IF ( .NOT. ALLOCATED( qsurf_av ) ) THEN
733	ALLOCATE( qsurf_av(nysg:nyng,nxlg:nxrg) )
734	qsurf_av = REAL( fill_value, KIND = wp )
735	ENDIF
736	DO i = nxl, nxr
737	DO j = nys, nyn
738	local_pf(i,j,nzb+1) = qsurf_av(j,i)
739	ENDDO
740	ENDDO
741	ENDIF
742	resorted = .TRUE.
743	two_d = .TRUE.
744	level_z(nzb+1) = zu(nzb+1)
745
746	CASE ( 'qsws*_xy' ) ! 2d-array
747	IF ( av == 0 ) THEN
748	local_pf(:,:,nzb+1) = REAL( fill_value, KIND = wp )
749	!
750	!-- In case of default surfaces, clean-up flux by density.
751	!-- In case of land-surfaces, convert fluxes into
752	!-- dynamic units
753	DO m = 1, surf_def_h(0)%ns
754	i = surf_def_h(0)%i(m)
755	j = surf_def_h(0)%j(m)
756	k = surf_def_h(0)%k(m)
757	local_pf(i,j,nzb+1) = surf_def_h(0)%qsws(m) * &
758	waterflux_output_conversion(k)
759	ENDDO
760	DO m = 1, surf_lsm_h%ns
761	i = surf_lsm_h%i(m)
762	j = surf_lsm_h%j(m)
763	k = surf_lsm_h%k(m)
764	local_pf(i,j,nzb+1) = surf_lsm_h%qsws(m) * waterflux_output_conversion(k)
765	ENDDO
766	DO m = 1, surf_usm_h%ns
767	i = surf_usm_h%i(m)
768	j = surf_usm_h%j(m)
769	k = surf_usm_h%k(m)
770	local_pf(i,j,nzb+1) = surf_usm_h%qsws(m) * waterflux_output_conversion(k)
771	ENDDO
772	ELSE
773	IF ( .NOT. ALLOCATED( qsws_av ) ) THEN
774	ALLOCATE( qsws_av(nysg:nyng,nxlg:nxrg) )
775	qsws_av = REAL( fill_value, KIND = wp )
776	ENDIF
777	DO i = nxl, nxr
778	DO j = nys, nyn
779	local_pf(i,j,nzb+1) = qsws_av(j,i)
780	ENDDO
781	ENDDO
782	ENDIF
783	resorted = .TRUE.
784	two_d = .TRUE.
785	level_z(nzb+1) = zu(nzb+1)
786
787	CASE ( 'qv_xy', 'qv_xz', 'qv_yz' )
788	IF ( av == 0 ) THEN
789	DO i = nxl, nxr
790	DO j = nys, nyn
791	DO k = nzb, nzt+1
792	local_pf(i,j,k) = q(k,j,i) - ql(k,j,i)
793	ENDDO
794	ENDDO
795	ENDDO
796	resorted = .TRUE.
797	ELSE
798	IF ( .NOT. ALLOCATED( qv_av ) ) THEN
799	ALLOCATE( qv_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
800	qv_av = REAL( fill_value, KIND = wp )
801	ENDIF
802	to_be_resorted => qv_av
803	ENDIF
804	IF ( mode == 'xy' ) level_z = zu
805
806	CASE ( 'r_a*_xy' ) ! 2d-array
807	IF ( av == 0 ) THEN
808	DO m = 1, surf_lsm_h%ns
809	i = surf_lsm_h%i(m)
810	j = surf_lsm_h%j(m)
811	local_pf(i,j,nzb+1) = surf_lsm_h%r_a(m)
812	ENDDO
813
814	DO m = 1, surf_usm_h%ns
815	i = surf_usm_h%i(m)
816	j = surf_usm_h%j(m)
817	local_pf(i,j,nzb+1) = &
818	( surf_usm_h%frac(m,ind_veg_wall) * &
819	surf_usm_h%r_a(m) + &
820	surf_usm_h%frac(m,ind_pav_green) * &
821	surf_usm_h%r_a_green(m) + &
822	surf_usm_h%frac(m,ind_wat_win) * &
823	surf_usm_h%r_a_window(m) )
824	ENDDO
825	ELSE
826	IF ( .NOT. ALLOCATED( r_a_av ) ) THEN
827	ALLOCATE( r_a_av(nysg:nyng,nxlg:nxrg) )
828	r_a_av = REAL( fill_value, KIND = wp )
829	ENDIF
830	DO i = nxl, nxr
831	DO j = nys, nyn
832	local_pf(i,j,nzb+1) = r_a_av(j,i)
833	ENDDO
834	ENDDO
835	ENDIF
836	resorted = .TRUE.
837	two_d = .TRUE.
838	level_z(nzb+1) = zu(nzb+1)
839
840	CASE ( 's_xy', 's_xz', 's_yz' )
841	IF ( av == 0 ) THEN
842	to_be_resorted => s
843	ELSE
844	IF ( .NOT. ALLOCATED( s_av ) ) THEN
845	ALLOCATE( s_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
846	s_av = REAL( fill_value, KIND = wp )
847	ENDIF
848	to_be_resorted => s_av
849	ENDIF
850
851	CASE ( 'shf*_xy' ) ! 2d-array
852	IF ( av == 0 ) THEN
853	!
854	!-- In case of default surfaces, clean-up flux by density.
855	!-- In case of land- and urban-surfaces, convert fluxes into
856	!-- dynamic units.
857	DO m = 1, surf_def_h(0)%ns
858	i = surf_def_h(0)%i(m)
859	j = surf_def_h(0)%j(m)
860	k = surf_def_h(0)%k(m)
861	local_pf(i,j,nzb+1) = surf_def_h(0)%shf(m) * &
862	heatflux_output_conversion(k)
863	ENDDO
864	DO m = 1, surf_lsm_h%ns
865	i = surf_lsm_h%i(m)
866	j = surf_lsm_h%j(m)
867	k = surf_lsm_h%k(m)
868	local_pf(i,j,nzb+1) = surf_lsm_h%shf(m) * heatflux_output_conversion(k)
869	ENDDO
870	DO m = 1, surf_usm_h%ns
871	i = surf_usm_h%i(m)
872	j = surf_usm_h%j(m)
873	k = surf_usm_h%k(m)
874	local_pf(i,j,nzb+1) = surf_usm_h%shf(m) * heatflux_output_conversion(k)
875	ENDDO
876	ELSE
877	IF ( .NOT. ALLOCATED( shf_av ) ) THEN
878	ALLOCATE( shf_av(nysg:nyng,nxlg:nxrg) )
879	shf_av = REAL( fill_value, KIND = wp )
880	ENDIF
881	DO i = nxl, nxr
882	DO j = nys, nyn
883	local_pf(i,j,nzb+1) = shf_av(j,i)
884	ENDDO
885	ENDDO
886	ENDIF
887	resorted = .TRUE.
888	two_d = .TRUE.
889	level_z(nzb+1) = zu(nzb+1)
890
891	CASE ( 'ssurf*_xy' ) ! 2d-array
892	IF ( av == 0 ) THEN
893	DO i = nxl, nxr
894	DO j = nys, nyn
895	k = topo_top_ind(j,i,0)
896	local_pf(i,j,nzb+1) = s(k,j,i)
897	ENDDO
898	ENDDO
899	ELSE
900	IF ( .NOT. ALLOCATED( ssurf_av ) ) THEN
901	ALLOCATE( ssurf_av(nysg:nyng,nxlg:nxrg) )
902	ssurf_av = REAL( fill_value, KIND = wp )
903	ENDIF
904	DO i = nxl, nxr
905	DO j = nys, nyn
906	local_pf(i,j,nzb+1) = ssurf_av(j,i)
907	ENDDO
908	ENDDO
909	ENDIF
910	resorted = .TRUE.
911	two_d = .TRUE.
912	level_z(nzb+1) = zu(nzb+1)
913
914	CASE ( 'ssws*_xy' ) ! 2d-array
915	IF ( av == 0 ) THEN
916	DO m = 1, surf_def_h(0)%ns
917	i = surf_def_h(0)%i(m)
918	j = surf_def_h(0)%j(m)
919	local_pf(i,j,nzb+1) = surf_def_h(0)%ssws(m)
920	ENDDO
921	DO m = 1, surf_lsm_h%ns
922	i = surf_lsm_h%i(m)
923	j = surf_lsm_h%j(m)
924	local_pf(i,j,nzb+1) = surf_lsm_h%ssws(m)
925	ENDDO
926	DO m = 1, surf_usm_h%ns
927	i = surf_usm_h%i(m)
928	j = surf_usm_h%j(m)
929	local_pf(i,j,nzb+1) = surf_usm_h%ssws(m)
930	ENDDO
931	ELSE
932	IF ( .NOT. ALLOCATED( ssws_av ) ) THEN
933	ALLOCATE( ssws_av(nysg:nyng,nxlg:nxrg) )
934	ssws_av = REAL( fill_value, KIND = wp )
935	ENDIF
936	DO i = nxl, nxr
937	DO j = nys, nyn
938	local_pf(i,j,nzb+1) = ssws_av(j,i)
939	ENDDO
940	ENDDO
941	ENDIF
942	resorted = .TRUE.
943	two_d = .TRUE.
944	level_z(nzb+1) = zu(nzb+1)
945
946	CASE ( 't*_xy' ) ! 2d-array
947	IF ( av == 0 ) THEN
948	DO m = 1, surf_def_h(0)%ns
949	i = surf_def_h(0)%i(m)
950	j = surf_def_h(0)%j(m)
951	local_pf(i,j,nzb+1) = surf_def_h(0)%ts(m)
952	ENDDO
953	DO m = 1, surf_lsm_h%ns
954	i = surf_lsm_h%i(m)
955	j = surf_lsm_h%j(m)
956	local_pf(i,j,nzb+1) = surf_lsm_h%ts(m)
957	ENDDO
958	DO m = 1, surf_usm_h%ns
959	i = surf_usm_h%i(m)
960	j = surf_usm_h%j(m)
961	local_pf(i,j,nzb+1) = surf_usm_h%ts(m)
962	ENDDO
963	ELSE
964	IF ( .NOT. ALLOCATED( ts_av ) ) THEN
965	ALLOCATE( ts_av(nysg:nyng,nxlg:nxrg) )
966	ts_av = REAL( fill_value, KIND = wp )
967	ENDIF
968	DO i = nxl, nxr
969	DO j = nys, nyn
970	local_pf(i,j,nzb+1) = ts_av(j,i)
971	ENDDO
972	ENDDO
973	ENDIF
974	resorted = .TRUE.
975	two_d = .TRUE.
976	level_z(nzb+1) = zu(nzb+1)
977
978	CASE ( 'tsurf*_xy' ) ! 2d-array
979	IF ( av == 0 ) THEN
980	DO m = 1, surf_def_h(0)%ns
981	i = surf_def_h(0)%i(m)
982	j = surf_def_h(0)%j(m)
983	local_pf(i,j,nzb+1) = surf_def_h(0)%pt_surface(m)
984	ENDDO
985
986	DO m = 1, surf_lsm_h%ns
987	i = surf_lsm_h%i(m)
988	j = surf_lsm_h%j(m)
989	local_pf(i,j,nzb+1) = surf_lsm_h%pt_surface(m)
990	ENDDO
991
992	DO m = 1, surf_usm_h%ns
993	i = surf_usm_h%i(m)
994	j = surf_usm_h%j(m)
995	local_pf(i,j,nzb+1) = surf_usm_h%pt_surface(m)
996	ENDDO
997
998	ELSE
999	IF ( .NOT. ALLOCATED( tsurf_av ) ) THEN
1000	ALLOCATE( tsurf_av(nysg:nyng,nxlg:nxrg) )
1001	tsurf_av = REAL( fill_value, KIND = wp )
1002	ENDIF
1003	DO i = nxl, nxr
1004	DO j = nys, nyn
1005	local_pf(i,j,nzb+1) = tsurf_av(j,i)
1006	ENDDO
1007	ENDDO
1008	ENDIF
1009	resorted = .TRUE.
1010	two_d = .TRUE.
1011	level_z(nzb+1) = zu(nzb+1)
1012
1013	CASE ( 'u_xy', 'u_xz', 'u_yz' )
1014	flag_nr = 1
1015	IF ( av == 0 ) THEN
1016	to_be_resorted => u
1017	ELSE
1018	IF ( .NOT. ALLOCATED( u_av ) ) THEN
1019	ALLOCATE( u_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1020	u_av = REAL( fill_value, KIND = wp )
1021	ENDIF
1022	to_be_resorted => u_av
1023	ENDIF
1024	IF ( mode == 'xy' ) level_z = zu
1025	!
1026	!-- Substitute the values generated by "mirror" boundary condition
1027	!-- at the bottom boundary by the real surface values.
1028	IF ( do2d(av,ivar) == 'u_xz' .OR. do2d(av,ivar) == 'u_yz' ) THEN
1029	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
1030	ENDIF
1031
1032	CASE ( 'us*_xy' ) ! 2d-array
1033	IF ( av == 0 ) THEN
1034	DO m = 1, surf_def_h(0)%ns
1035	i = surf_def_h(0)%i(m)
1036	j = surf_def_h(0)%j(m)
1037	local_pf(i,j,nzb+1) = surf_def_h(0)%us(m)
1038	ENDDO
1039	DO m = 1, surf_lsm_h%ns
1040	i = surf_lsm_h%i(m)
1041	j = surf_lsm_h%j(m)
1042	local_pf(i,j,nzb+1) = surf_lsm_h%us(m)
1043	ENDDO
1044	DO m = 1, surf_usm_h%ns
1045	i = surf_usm_h%i(m)
1046	j = surf_usm_h%j(m)
1047	local_pf(i,j,nzb+1) = surf_usm_h%us(m)
1048	ENDDO
1049	ELSE
1050	IF ( .NOT. ALLOCATED( us_av ) ) THEN
1051	ALLOCATE( us_av(nysg:nyng,nxlg:nxrg) )
1052	us_av = REAL( fill_value, KIND = wp )
1053	ENDIF
1054	DO i = nxl, nxr
1055	DO j = nys, nyn
1056	local_pf(i,j,nzb+1) = us_av(j,i)
1057	ENDDO
1058	ENDDO
1059	ENDIF
1060	resorted = .TRUE.
1061	two_d = .TRUE.
1062	level_z(nzb+1) = zu(nzb+1)
1063
1064	CASE ( 'v_xy', 'v_xz', 'v_yz' )
1065	flag_nr = 2
1066	IF ( av == 0 ) THEN
1067	to_be_resorted => v
1068	ELSE
1069	IF ( .NOT. ALLOCATED( v_av ) ) THEN
1070	ALLOCATE( v_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1071	v_av = REAL( fill_value, KIND = wp )
1072	ENDIF
1073	to_be_resorted => v_av
1074	ENDIF
1075	IF ( mode == 'xy' ) level_z = zu
1076	!
1077	!-- Substitute the values generated by "mirror" boundary condition
1078	!-- at the bottom boundary by the real surface values.
1079	IF ( do2d(av,ivar) == 'v_xz' .OR. do2d(av,ivar) == 'v_yz' ) THEN
1080	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
1081	ENDIF
1082
1083	CASE ( 'thetav_xy', 'thetav_xz', 'thetav_yz' )
1084	IF ( av == 0 ) THEN
1085	to_be_resorted => vpt
1086	ELSE
1087	IF ( .NOT. ALLOCATED( vpt_av ) ) THEN
1088	ALLOCATE( vpt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1089	vpt_av = REAL( fill_value, KIND = wp )
1090	ENDIF
1091	to_be_resorted => vpt_av
1092	ENDIF
1093	IF ( mode == 'xy' ) level_z = zu
1094
1095	CASE ( 'w_xy', 'w_xz', 'w_yz' )
1096	flag_nr = 3
1097	IF ( av == 0 ) THEN
1098	to_be_resorted => w
1099	ELSE
1100	IF ( .NOT. ALLOCATED( w_av ) ) THEN
1101	ALLOCATE( w_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1102	w_av = REAL( fill_value, KIND = wp )
1103	ENDIF
1104	to_be_resorted => w_av
1105	ENDIF
1106	IF ( mode == 'xy' ) level_z = zw
1107
1108	CASE ( 'z0*_xy' ) ! 2d-array
1109	IF ( av == 0 ) THEN
1110	DO m = 1, surf_def_h(0)%ns
1111	i = surf_def_h(0)%i(m)
1112	j = surf_def_h(0)%j(m)
1113	local_pf(i,j,nzb+1) = surf_def_h(0)%z0(m)
1114	ENDDO
1115	DO m = 1, surf_lsm_h%ns
1116	i = surf_lsm_h%i(m)
1117	j = surf_lsm_h%j(m)
1118	local_pf(i,j,nzb+1) = surf_lsm_h%z0(m)
1119	ENDDO
1120	DO m = 1, surf_usm_h%ns
1121	i = surf_usm_h%i(m)
1122	j = surf_usm_h%j(m)
1123	local_pf(i,j,nzb+1) = surf_usm_h%z0(m)
1124	ENDDO
1125	ELSE
1126	IF ( .NOT. ALLOCATED( z0_av ) ) THEN
1127	ALLOCATE( z0_av(nysg:nyng,nxlg:nxrg) )
1128	z0_av = REAL( fill_value, KIND = wp )
1129	ENDIF
1130	DO i = nxl, nxr
1131	DO j = nys, nyn
1132	local_pf(i,j,nzb+1) = z0_av(j,i)
1133	ENDDO
1134	ENDDO
1135	ENDIF
1136	resorted = .TRUE.
1137	two_d = .TRUE.
1138	level_z(nzb+1) = zu(nzb+1)
1139
1140	CASE ( 'z0h*_xy' ) ! 2d-array
1141	IF ( av == 0 ) THEN
1142	DO m = 1, surf_def_h(0)%ns
1143	i = surf_def_h(0)%i(m)
1144	j = surf_def_h(0)%j(m)
1145	local_pf(i,j,nzb+1) = surf_def_h(0)%z0h(m)
1146	ENDDO
1147	DO m = 1, surf_lsm_h%ns
1148	i = surf_lsm_h%i(m)
1149	j = surf_lsm_h%j(m)
1150	local_pf(i,j,nzb+1) = surf_lsm_h%z0h(m)
1151	ENDDO
1152	DO m = 1, surf_usm_h%ns
1153	i = surf_usm_h%i(m)
1154	j = surf_usm_h%j(m)
1155	local_pf(i,j,nzb+1) = surf_usm_h%z0h(m)
1156	ENDDO
1157	ELSE
1158	IF ( .NOT. ALLOCATED( z0h_av ) ) THEN
1159	ALLOCATE( z0h_av(nysg:nyng,nxlg:nxrg) )
1160	z0h_av = REAL( fill_value, KIND = wp )
1161	ENDIF
1162	DO i = nxl, nxr
1163	DO j = nys, nyn
1164	local_pf(i,j,nzb+1) = z0h_av(j,i)
1165	ENDDO
1166	ENDDO
1167	ENDIF
1168	resorted = .TRUE.
1169	two_d = .TRUE.
1170	level_z(nzb+1) = zu(nzb+1)
1171
1172	CASE ( 'z0q*_xy' ) ! 2d-array
1173	IF ( av == 0 ) THEN
1174	DO m = 1, surf_def_h(0)%ns
1175	i = surf_def_h(0)%i(m)
1176	j = surf_def_h(0)%j(m)
1177	local_pf(i,j,nzb+1) = surf_def_h(0)%z0q(m)
1178	ENDDO
1179	DO m = 1, surf_lsm_h%ns
1180	i = surf_lsm_h%i(m)
1181	j = surf_lsm_h%j(m)
1182	local_pf(i,j,nzb+1) = surf_lsm_h%z0q(m)
1183	ENDDO
1184	DO m = 1, surf_usm_h%ns
1185	i = surf_usm_h%i(m)
1186	j = surf_usm_h%j(m)
1187	local_pf(i,j,nzb+1) = surf_usm_h%z0q(m)
1188	ENDDO
1189	ELSE
1190	IF ( .NOT. ALLOCATED( z0q_av ) ) THEN
1191	ALLOCATE( z0q_av(nysg:nyng,nxlg:nxrg) )
1192	z0q_av = REAL( fill_value, KIND = wp )
1193	ENDIF
1194	DO i = nxl, nxr
1195	DO j = nys, nyn
1196	local_pf(i,j,nzb+1) = z0q_av(j,i)
1197	ENDDO
1198	ENDDO
1199	ENDIF
1200	resorted = .TRUE.
1201	two_d = .TRUE.
1202	level_z(nzb+1) = zu(nzb+1)
1203
1204	CASE DEFAULT
1205
1206	!
1207	!-- Quantities of other modules
1208	IF ( .NOT. found ) THEN
1209	CALL module_interface_data_output_2d( &
1210	av, do2d(av,ivar), found, grid, mode, &
1211	local_pf, two_d, nzb_do, nzt_do, &
1212	fill_value &
1213	)
1214	ENDIF
1215
1216	resorted = .TRUE.
1217
1218	IF ( grid == 'zu' ) THEN
1219	IF ( mode == 'xy' ) level_z = zu
1220	ELSEIF ( grid == 'zw' ) THEN
1221	IF ( mode == 'xy' ) level_z = zw
1222	ELSEIF ( grid == 'zu1' ) THEN
1223	IF ( mode == 'xy' ) level_z(nzb+1) = zu(nzb+1)
1224	ELSEIF ( grid == 'zs' ) THEN
1225	IF ( mode == 'xy' ) level_z = zs
1226	ENDIF
1227
1228	IF ( .NOT. found ) THEN
1229	message_string = 'no output provided for: ' // &
1230	TRIM( do2d(av,ivar) )
1231	CALL message( 'data_output_2d', 'PA0181', 0, 0, 0, 6, 0 )
1232	ENDIF
1233
1234	END SELECT
1235
1236	!
1237	!-- Resort the array to be output, if not done above. Flag topography
1238	!-- grid points with fill values, using the corresponding maksing flag.
1239	IF ( .NOT. resorted ) THEN
1240	DO i = nxl, nxr
1241	DO j = nys, nyn
1242	DO k = nzb_do, nzt_do
1243	local_pf(i,j,k) = MERGE( to_be_resorted(k,j,i), &
1244	REAL( fill_value, KIND = wp ), &
1245	BTEST( wall_flags_total_0(k,j,i), &
1246	flag_nr ) )
1247	ENDDO
1248	ENDDO
1249	ENDDO
1250	ENDIF
1251
1252	!
1253	!-- Output of the individual cross-sections, depending on the cross-
1254	!-- section mode chosen.
1255	is = 1
1256	loop1: DO WHILE ( section(is,s_ind) /= -9999 .OR. two_d )
1257
1258	SELECT CASE ( mode )
1259
1260	CASE ( 'xy' )
1261	!
1262	!-- Determine the cross section index
1263	IF ( two_d ) THEN
1264	layer_xy = nzb+1
1265	ELSE
1266	layer_xy = section(is,s_ind)
1267	ENDIF
1268
1269	!
1270	!-- Exit the loop for layers beyond the data output domain
1271	!-- (used for soil model)
1272	IF ( layer_xy > nzt_do ) THEN
1273	EXIT loop1
1274	ENDIF
1275
1276	!
1277	!-- Update the netCDF xy cross section time axis.
1278	!-- In case of parallel output, this is only done by PE0
1279	!-- to increase the performance.
1280	IF ( time_since_reference_point /= do2d_xy_last_time(av) ) THEN
1281	do2d_xy_time_count(av) = do2d_xy_time_count(av) + 1
1282	do2d_xy_last_time(av) = time_since_reference_point
1283	IF ( myid == 0 ) THEN
1284	IF ( .NOT. data_output_2d_on_each_pe &
1285	.OR. netcdf_data_format > 4 ) &
1286	THEN
1287	#if defined( __netcdf )
1288	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1289	id_var_time_xy(av), &
1290	(/ time_since_reference_point /), &
1291	start = (/ do2d_xy_time_count(av) /), &
1292	count = (/ 1 /) )
1293	CALL netcdf_handle_error( 'data_output_2d', 53 )
1294	#endif
1295	ENDIF
1296	ENDIF
1297	ENDIF
1298	!
1299	!-- If required, carry out averaging along z
1300	IF ( section(is,s_ind) == -1 .AND. .NOT. two_d ) THEN
1301
1302	local_2d = 0.0_wp
1303	!
1304	!-- Carry out the averaging (all data are on the PE)
1305	DO k = nzb_do, nzt_do
1306	DO j = nys, nyn
1307	DO i = nxl, nxr
1308	local_2d(i,j) = local_2d(i,j) + local_pf(i,j,k)
1309	ENDDO
1310	ENDDO
1311	ENDDO
1312
1313	local_2d = local_2d / ( nzt_do - nzb_do + 1.0_wp)
1314
1315	ELSE
1316	!
1317	!-- Just store the respective section on the local array
1318	local_2d = local_pf(:,:,layer_xy)
1319
1320	ENDIF
1321
1322	#if defined( __parallel )
1323	IF ( netcdf_data_format > 4 ) THEN
1324	!
1325	!-- Parallel output in netCDF4/HDF5 format.
1326	IF ( two_d ) THEN
1327	iis = 1
1328	ELSE
1329	iis = is
1330	ENDIF
1331
1332	#if defined( __netcdf )
1333	!
1334	!-- For parallel output, all cross sections are first stored
1335	!-- here on a local array and will be written to the output
1336	!-- file afterwards to increase the performance.
1337	DO i = nxl, nxr
1338	DO j = nys, nyn
1339	local_2d_sections(i,j,iis) = local_2d(i,j)
1340	ENDDO
1341	ENDDO
1342	#endif
1343	ELSE
1344
1345	IF ( data_output_2d_on_each_pe ) THEN
1346	!
1347	!-- Output of partial arrays on each PE
1348	#if defined( __netcdf )
1349	IF ( myid == 0 ) THEN
1350	WRITE ( 21 ) time_since_reference_point, &
1351	do2d_xy_time_count(av), av
1352	ENDIF
1353	#endif
1354	DO i = 0, io_blocks-1
1355	IF ( i == io_group ) THEN
1356	WRITE ( 21 ) nxl, nxr, nys, nyn, nys, nyn
1357	WRITE ( 21 ) local_2d
1358	ENDIF
1359	#if defined( __parallel )
1360	CALL MPI_BARRIER( comm2d, ierr )
1361	#endif
1362	ENDDO
1363
1364	ELSE
1365	!
1366	!-- PE0 receives partial arrays from all processors and
1367	!-- then outputs them. Here a barrier has to be set,
1368	!-- because otherwise "-MPI- FATAL: Remote protocol queue
1369	!-- full" may occur.
1370	CALL MPI_BARRIER( comm2d, ierr )
1371
1372	ngp = ( nxr-nxl+1 ) * ( nyn-nys+1 )
1373	IF ( myid == 0 ) THEN
1374	!
1375	!-- Local array can be relocated directly.
1376	total_2d(nxl:nxr,nys:nyn) = local_2d
1377	!
1378	!-- Receive data from all other PEs.
1379	DO n = 1, numprocs-1
1380	!
1381	!-- Receive index limits first, then array.
1382	!-- Index limits are received in arbitrary order from
1383	!-- the PEs.
1384	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1385	MPI_ANY_SOURCE, 0, comm2d, &
1386	status, ierr )
1387	sender = status(MPI_SOURCE)
1388	DEALLOCATE( local_2d )
1389	ALLOCATE( local_2d(ind(1):ind(2),ind(3):ind(4)) )
1390	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1391	MPI_REAL, sender, 1, comm2d, &
1392	status, ierr )
1393	total_2d(ind(1):ind(2),ind(3):ind(4)) = local_2d
1394	ENDDO
1395	!
1396	!-- Relocate the local array for the next loop increment
1397	DEALLOCATE( local_2d )
1398	ALLOCATE( local_2d(nxl:nxr,nys:nyn) )
1399
1400	#if defined( __netcdf )
1401	IF ( two_d ) THEN
1402	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1403	id_var_do2d(av,ivar), &
1404	total_2d(0:nx,0:ny), &
1405	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
1406	count = (/ nx+1, ny+1, 1, 1 /) )
1407	ELSE
1408	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1409	id_var_do2d(av,ivar), &
1410	total_2d(0:nx,0:ny), &
1411	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
1412	count = (/ nx+1, ny+1, 1, 1 /) )
1413	ENDIF
1414	CALL netcdf_handle_error( 'data_output_2d', 54 )
1415	#endif
1416
1417	ELSE
1418	!
1419	!-- First send the local index limits to PE0
1420	ind(1) = nxl; ind(2) = nxr
1421	ind(3) = nys; ind(4) = nyn
1422	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1423	comm2d, ierr )
1424	!
1425	!-- Send data to PE0
1426	CALL MPI_SEND( local_2d(nxl,nys), ngp, &
1427	MPI_REAL, 0, 1, comm2d, ierr )
1428	ENDIF
1429	!
1430	!-- A barrier has to be set, because otherwise some PEs may
1431	!-- proceed too fast so that PE0 may receive wrong data on
1432	!-- tag 0
1433	CALL MPI_BARRIER( comm2d, ierr )
1434	ENDIF
1435
1436	ENDIF
1437	#else
1438	#if defined( __netcdf )
1439	IF ( two_d ) THEN
1440	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1441	id_var_do2d(av,ivar), &
1442	local_2d(nxl:nxr,nys:nyn), &
1443	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
1444	count = (/ nx+1, ny+1, 1, 1 /) )
1445	ELSE
1446	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1447	id_var_do2d(av,ivar), &
1448	local_2d(nxl:nxr,nys:nyn), &
1449	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
1450	count = (/ nx+1, ny+1, 1, 1 /) )
1451	ENDIF
1452	CALL netcdf_handle_error( 'data_output_2d', 447 )
1453	#endif
1454	#endif
1455
1456	!
1457	!-- For 2D-arrays (e.g. u*) only one cross-section is available.
1458	!-- Hence exit loop of output levels.
1459	IF ( two_d ) THEN
1460	IF ( netcdf_data_format < 5 ) two_d = .FALSE.
1461	EXIT loop1
1462	ENDIF
1463
1464	CASE ( 'xz' )
1465	!
1466	!-- Update the netCDF xz cross section time axis.
1467	!-- In case of parallel output, this is only done by PE0
1468	!-- to increase the performance.
1469	IF ( time_since_reference_point /= do2d_xz_last_time(av) ) THEN
1470	do2d_xz_time_count(av) = do2d_xz_time_count(av) + 1
1471	do2d_xz_last_time(av) = time_since_reference_point
1472	IF ( myid == 0 ) THEN
1473	IF ( .NOT. data_output_2d_on_each_pe &
1474	.OR. netcdf_data_format > 4 ) &
1475	THEN
1476	#if defined( __netcdf )
1477	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1478	id_var_time_xz(av), &
1479	(/ time_since_reference_point /), &
1480	start = (/ do2d_xz_time_count(av) /), &
1481	count = (/ 1 /) )
1482	CALL netcdf_handle_error( 'data_output_2d', 56 )
1483	#endif
1484	ENDIF
1485	ENDIF
1486	ENDIF
1487
1488	!
1489	!-- If required, carry out averaging along y
1490	IF ( section(is,s_ind) == -1 ) THEN
1491
1492	ALLOCATE( local_2d_l(nxl:nxr,nzb_do:nzt_do) )
1493	local_2d_l = 0.0_wp
1494	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
1495	!
1496	!-- First local averaging on the PE
1497	DO k = nzb_do, nzt_do
1498	DO j = nys, nyn
1499	DO i = nxl, nxr
1500	local_2d_l(i,k) = local_2d_l(i,k) + &
1501	local_pf(i,j,k)
1502	ENDDO
1503	ENDDO
1504	ENDDO
1505	#if defined( __parallel )
1506	!
1507	!-- Now do the averaging over all PEs along y
1508	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1509	CALL MPI_ALLREDUCE( local_2d_l(nxl,nzb_do), &
1510	local_2d(nxl,nzb_do), ngp, MPI_REAL, &
1511	MPI_SUM, comm1dy, ierr )
1512	#else
1513	local_2d = local_2d_l
1514	#endif
1515	local_2d = local_2d / ( ny + 1.0_wp )
1516
1517	DEALLOCATE( local_2d_l )
1518
1519	ELSE
1520	!
1521	!-- Just store the respective section on the local array
1522	!-- (but only if it is available on this PE!)
1523	IF ( section(is,s_ind) >= nys .AND. section(is,s_ind) <= nyn ) &
1524	THEN
1525	local_2d = local_pf(:,section(is,s_ind),nzb_do:nzt_do)
1526	ENDIF
1527
1528	ENDIF
1529
1530	#if defined( __parallel )
1531	IF ( netcdf_data_format > 4 ) THEN
1532	!
1533	!-- Output in netCDF4/HDF5 format.
1534	!-- Output only on those PEs where the respective cross
1535	!-- sections reside. Cross sections averaged along y are
1536	!-- output on the respective first PE along y (myidy=0).
1537	IF ( ( section(is,s_ind) >= nys .AND. &
1538	section(is,s_ind) <= nyn ) .OR. &
1539	( section(is,s_ind) == -1 .AND. myidy == 0 ) ) THEN
1540	#if defined( __netcdf )
1541	!
1542	!-- For parallel output, all cross sections are first
1543	!-- stored here on a local array and will be written to the
1544	!-- output file afterwards to increase the performance.
1545	DO i = nxl, nxr
1546	DO k = nzb_do, nzt_do
1547	local_2d_sections_l(i,is,k) = local_2d(i,k)
1548	ENDDO
1549	ENDDO
1550	#endif
1551	ENDIF
1552
1553	ELSE
1554
1555	IF ( data_output_2d_on_each_pe ) THEN
1556	!
1557	!-- Output of partial arrays on each PE. If the cross
1558	!-- section does not reside on the PE, output special
1559	!-- index values.
1560	#if defined( __netcdf )
1561	IF ( myid == 0 ) THEN
1562	WRITE ( 22 ) time_since_reference_point, &
1563	do2d_xz_time_count(av), av
1564	ENDIF
1565	#endif
1566	DO i = 0, io_blocks-1
1567	IF ( i == io_group ) THEN
1568	IF ( ( section(is,s_ind) >= nys .AND. &
1569	section(is,s_ind) <= nyn ) .OR. &
1570	( section(is,s_ind) == -1 .AND. &
1571	nys-1 == -1 ) ) &
1572	THEN
1573	WRITE (22) nxl, nxr, nzb_do, nzt_do, nzb, nzt+1
1574	WRITE (22) local_2d
1575	ELSE
1576	WRITE (22) -1, -1, -1, -1, -1, -1
1577	ENDIF
1578	ENDIF
1579	#if defined( __parallel )
1580	CALL MPI_BARRIER( comm2d, ierr )
1581	#endif
1582	ENDDO
1583
1584	ELSE
1585	!
1586	!-- PE0 receives partial arrays from all processors of the
1587	!-- respective cross section and outputs them. Here a
1588	!-- barrier has to be set, because otherwise
1589	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
1590	CALL MPI_BARRIER( comm2d, ierr )
1591
1592	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
1593	IF ( myid == 0 ) THEN
1594	!
1595	!-- Local array can be relocated directly.
1596	IF ( ( section(is,s_ind) >= nys .AND. &
1597	section(is,s_ind) <= nyn ) .OR. &
1598	( section(is,s_ind) == -1 .AND. &
1599	nys-1 == -1 ) ) THEN
1600	total_2d(nxl:nxr,nzb_do:nzt_do) = local_2d
1601	ENDIF
1602	!
1603	!-- Receive data from all other PEs.
1604	DO n = 1, numprocs-1
1605	!
1606	!-- Receive index limits first, then array.
1607	!-- Index limits are received in arbitrary order from
1608	!-- the PEs.
1609	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1610	MPI_ANY_SOURCE, 0, comm2d, &
1611	status, ierr )
1612	!
1613	!-- Not all PEs have data for XZ-cross-section.
1614	IF ( ind(1) /= -9999 ) THEN
1615	sender = status(MPI_SOURCE)
1616	DEALLOCATE( local_2d )
1617	ALLOCATE( local_2d(ind(1):ind(2), &
1618	ind(3):ind(4)) )
1619	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1620	MPI_REAL, sender, 1, comm2d, &
1621	status, ierr )
1622	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
1623	local_2d
1624	ENDIF
1625	ENDDO
1626	!
1627	!-- Relocate the local array for the next loop increment
1628	DEALLOCATE( local_2d )
1629	ALLOCATE( local_2d(nxl:nxr,nzb_do:nzt_do) )
1630
1631	#if defined( __netcdf )
1632	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1633	id_var_do2d(av,ivar), &
1634	total_2d(0:nx,nzb_do:nzt_do), &
1635	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
1636	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
1637	CALL netcdf_handle_error( 'data_output_2d', 58 )
1638	#endif
1639
1640	ELSE
1641	!
1642	!-- If the cross section resides on the PE, send the
1643	!-- local index limits, otherwise send -9999 to PE0.
1644	IF ( ( section(is,s_ind) >= nys .AND. &
1645	section(is,s_ind) <= nyn ) .OR. &
1646	( section(is,s_ind) == -1 .AND. nys-1 == -1 ) ) &
1647	THEN
1648	ind(1) = nxl; ind(2) = nxr
1649	ind(3) = nzb_do; ind(4) = nzt_do
1650	ELSE
1651	ind(1) = -9999; ind(2) = -9999
1652	ind(3) = -9999; ind(4) = -9999
1653	ENDIF
1654	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1655	comm2d, ierr )
1656	!
1657	!-- If applicable, send data to PE0.
1658	IF ( ind(1) /= -9999 ) THEN
1659	CALL MPI_SEND( local_2d(nxl,nzb_do), ngp, &
1660	MPI_REAL, 0, 1, comm2d, ierr )
1661	ENDIF
1662	ENDIF
1663	!
1664	!-- A barrier has to be set, because otherwise some PEs may
1665	!-- proceed too fast so that PE0 may receive wrong data on
1666	!-- tag 0
1667	CALL MPI_BARRIER( comm2d, ierr )
1668	ENDIF
1669
1670	ENDIF
1671	#else
1672	#if defined( __netcdf )
1673	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1674	id_var_do2d(av,ivar), &
1675	local_2d(nxl:nxr,nzb_do:nzt_do), &
1676	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
1677	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
1678	CALL netcdf_handle_error( 'data_output_2d', 451 )
1679	#endif
1680	#endif
1681
1682	CASE ( 'yz' )
1683	!
1684	!-- Update the netCDF yz cross section time axis.
1685	!-- In case of parallel output, this is only done by PE0
1686	!-- to increase the performance.
1687	IF ( time_since_reference_point /= do2d_yz_last_time(av) ) THEN
1688	do2d_yz_time_count(av) = do2d_yz_time_count(av) + 1
1689	do2d_yz_last_time(av) = time_since_reference_point
1690	IF ( myid == 0 ) THEN
1691	IF ( .NOT. data_output_2d_on_each_pe &
1692	.OR. netcdf_data_format > 4 ) &
1693	THEN
1694	#if defined( __netcdf )
1695	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
1696	id_var_time_yz(av), &
1697	(/ time_since_reference_point /), &
1698	start = (/ do2d_yz_time_count(av) /), &
1699	count = (/ 1 /) )
1700	CALL netcdf_handle_error( 'data_output_2d', 59 )
1701	#endif
1702	ENDIF
1703	ENDIF
1704	ENDIF
1705
1706	!
1707	!-- If required, carry out averaging along x
1708	IF ( section(is,s_ind) == -1 ) THEN
1709
1710	ALLOCATE( local_2d_l(nys:nyn,nzb_do:nzt_do) )
1711	local_2d_l = 0.0_wp
1712	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
1713	!
1714	!-- First local averaging on the PE
1715	DO k = nzb_do, nzt_do
1716	DO j = nys, nyn
1717	DO i = nxl, nxr
1718	local_2d_l(j,k) = local_2d_l(j,k) + &
1719	local_pf(i,j,k)
1720	ENDDO
1721	ENDDO
1722	ENDDO
1723	#if defined( __parallel )
1724	!
1725	!-- Now do the averaging over all PEs along x
1726	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1727	CALL MPI_ALLREDUCE( local_2d_l(nys,nzb_do), &
1728	local_2d(nys,nzb_do), ngp, MPI_REAL, &
1729	MPI_SUM, comm1dx, ierr )
1730	#else
1731	local_2d = local_2d_l
1732	#endif
1733	local_2d = local_2d / ( nx + 1.0_wp )
1734
1735	DEALLOCATE( local_2d_l )
1736
1737	ELSE
1738	!
1739	!-- Just store the respective section on the local array
1740	!-- (but only if it is available on this PE!)
1741	IF ( section(is,s_ind) >= nxl .AND. section(is,s_ind) <= nxr ) &
1742	THEN
1743	local_2d = local_pf(section(is,s_ind),:,nzb_do:nzt_do)
1744	ENDIF
1745
1746	ENDIF
1747
1748	#if defined( __parallel )
1749	IF ( netcdf_data_format > 4 ) THEN
1750	!
1751	!-- Output in netCDF4/HDF5 format.
1752	!-- Output only on those PEs where the respective cross
1753	!-- sections reside. Cross sections averaged along x are
1754	!-- output on the respective first PE along x (myidx=0).
1755	IF ( ( section(is,s_ind) >= nxl .AND. &
1756	section(is,s_ind) <= nxr ) .OR. &
1757	( section(is,s_ind) == -1 .AND. myidx == 0 ) ) THEN
1758	#if defined( __netcdf )
1759	!
1760	!-- For parallel output, all cross sections are first
1761	!-- stored here on a local array and will be written to the
1762	!-- output file afterwards to increase the performance.
1763	DO j = nys, nyn
1764	DO k = nzb_do, nzt_do
1765	local_2d_sections_l(is,j,k) = local_2d(j,k)
1766	ENDDO
1767	ENDDO
1768	#endif
1769	ENDIF
1770
1771	ELSE
1772
1773	IF ( data_output_2d_on_each_pe ) THEN
1774	!
1775	!-- Output of partial arrays on each PE. If the cross
1776	!-- section does not reside on the PE, output special
1777	!-- index values.
1778	#if defined( __netcdf )
1779	IF ( myid == 0 ) THEN
1780	WRITE ( 23 ) time_since_reference_point, &
1781	do2d_yz_time_count(av), av
1782	ENDIF
1783	#endif
1784	DO i = 0, io_blocks-1
1785	IF ( i == io_group ) THEN
1786	IF ( ( section(is,s_ind) >= nxl .AND. &
1787	section(is,s_ind) <= nxr ) .OR. &
1788	( section(is,s_ind) == -1 .AND. &
1789	nxl-1 == -1 ) ) &
1790	THEN
1791	WRITE (23) nys, nyn, nzb_do, nzt_do, nzb, nzt+1
1792	WRITE (23) local_2d
1793	ELSE
1794	WRITE (23) -1, -1, -1, -1, -1, -1
1795	ENDIF
1796	ENDIF
1797	#if defined( __parallel )
1798	CALL MPI_BARRIER( comm2d, ierr )
1799	#endif
1800	ENDDO
1801
1802	ELSE
1803	!
1804	!-- PE0 receives partial arrays from all processors of the
1805	!-- respective cross section and outputs them. Here a
1806	!-- barrier has to be set, because otherwise
1807	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
1808	CALL MPI_BARRIER( comm2d, ierr )
1809
1810	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
1811	IF ( myid == 0 ) THEN
1812	!
1813	!-- Local array can be relocated directly.
1814	IF ( ( section(is,s_ind) >= nxl .AND. &
1815	section(is,s_ind) <= nxr ) .OR. &
1816	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
1817	THEN
1818	total_2d(nys:nyn,nzb_do:nzt_do) = local_2d
1819	ENDIF
1820	!
1821	!-- Receive data from all other PEs.
1822	DO n = 1, numprocs-1
1823	!
1824	!-- Receive index limits first, then array.
1825	!-- Index limits are received in arbitrary order from
1826	!-- the PEs.
1827	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1828	MPI_ANY_SOURCE, 0, comm2d, &
1829	status, ierr )
1830	!
1831	!-- Not all PEs have data for YZ-cross-section.
1832	IF ( ind(1) /= -9999 ) THEN
1833	sender = status(MPI_SOURCE)
1834	DEALLOCATE( local_2d )
1835	ALLOCATE( local_2d(ind(1):ind(2), &
1836	ind(3):ind(4)) )
1837	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1838	MPI_REAL, sender, 1, comm2d, &
1839	status, ierr )
1840	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
1841	local_2d
1842	ENDIF
1843	ENDDO
1844	!
1845	!-- Relocate the local array for the next loop increment
1846	DEALLOCATE( local_2d )
1847	ALLOCATE( local_2d(nys:nyn,nzb_do:nzt_do) )
1848
1849	#if defined( __netcdf )
1850	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
1851	id_var_do2d(av,ivar), &
1852	total_2d(0:ny,nzb_do:nzt_do), &
1853	start = (/ is, 1, 1, do2d_yz_time_count(av) /), &
1854	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
1855	CALL netcdf_handle_error( 'data_output_2d', 61 )
1856	#endif
1857
1858	ELSE
1859	!
1860	!-- If the cross section resides on the PE, send the
1861	!-- local index limits, otherwise send -9999 to PE0.
1862	IF ( ( section(is,s_ind) >= nxl .AND. &
1863	section(is,s_ind) <= nxr ) .OR. &
1864	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
1865	THEN
1866	ind(1) = nys; ind(2) = nyn
1867	ind(3) = nzb_do; ind(4) = nzt_do
1868	ELSE
1869	ind(1) = -9999; ind(2) = -9999
1870	ind(3) = -9999; ind(4) = -9999
1871	ENDIF
1872	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1873	comm2d, ierr )
1874	!
1875	!-- If applicable, send data to PE0.
1876	IF ( ind(1) /= -9999 ) THEN
1877	CALL MPI_SEND( local_2d(nys,nzb_do), ngp, &
1878	MPI_REAL, 0, 1, comm2d, ierr )
1879	ENDIF
1880	ENDIF
1881	!
1882	!-- A barrier has to be set, because otherwise some PEs may
1883	!-- proceed too fast so that PE0 may receive wrong data on
1884	!-- tag 0
1885	CALL MPI_BARRIER( comm2d, ierr )
1886	ENDIF
1887
1888	ENDIF
1889	#else
1890	#if defined( __netcdf )
1891	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
1892	id_var_do2d(av,ivar), &
1893	local_2d(nys:nyn,nzb_do:nzt_do), &
1894	start = (/ is, 1, 1, do2d_xz_time_count(av) /), &
1895	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
1896	CALL netcdf_handle_error( 'data_output_2d', 452 )
1897	#endif
1898	#endif
1899
1900	END SELECT
1901
1902	is = is + 1
1903	ENDDO loop1
1904
1905	!
1906	!-- For parallel output, all data were collected before on a local array
1907	!-- and are written now to the netcdf file. This must be done to increase
1908	!-- the performance of the parallel output.
1909	#if defined( __netcdf )
1910	IF ( netcdf_data_format > 4 ) THEN
1911
1912	SELECT CASE ( mode )
1913
1914	CASE ( 'xy' )
1915	IF ( two_d ) THEN
1916	nis = 1
1917	two_d = .FALSE.
1918	ELSE
1919	nis = ns
1920	ENDIF
1921	!
1922	!-- Do not output redundant ghost point data except for the
1923	!-- boundaries of the total domain.
1924	! IF ( nxr == nx .AND. nyn /= ny ) THEN
1925	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1926	! id_var_do2d(av,ivar), &
1927	! local_2d_sections(nxl:nxr+1, &
1928	! nys:nyn,1:nis), &
1929	! start = (/ nxl+1, nys+1, 1, &
1930	! do2d_xy_time_count(av) /), &
1931	! count = (/ nxr-nxl+2, &
1932	! nyn-nys+1, nis, 1 &
1933	! /) )
1934	! ELSEIF ( nxr /= nx .AND. nyn == ny ) THEN
1935	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1936	! id_var_do2d(av,ivar), &
1937	! local_2d_sections(nxl:nxr, &
1938	! nys:nyn+1,1:nis), &
1939	! start = (/ nxl+1, nys+1, 1, &
1940	! do2d_xy_time_count(av) /), &
1941	! count = (/ nxr-nxl+1, &
1942	! nyn-nys+2, nis, 1 &
1943	! /) )
1944	! ELSEIF ( nxr == nx .AND. nyn == ny ) THEN
1945	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1946	! id_var_do2d(av,ivar), &
1947	! local_2d_sections(nxl:nxr+1, &
1948	! nys:nyn+1,1:nis), &
1949	! start = (/ nxl+1, nys+1, 1, &
1950	! do2d_xy_time_count(av) /), &
1951	! count = (/ nxr-nxl+2, &
1952	! nyn-nys+2, nis, 1 &
1953	! /) )
1954	! ELSE
1955	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1956	id_var_do2d(av,ivar), &
1957	local_2d_sections(nxl:nxr, &
1958	nys:nyn,1:nis), &
1959	start = (/ nxl+1, nys+1, 1, &
1960	do2d_xy_time_count(av) /), &
1961	count = (/ nxr-nxl+1, &
1962	nyn-nys+1, nis, 1 &
1963	/) )
1964	! ENDIF
1965
1966	CALL netcdf_handle_error( 'data_output_2d', 55 )
1967
1968	CASE ( 'xz' )
1969	!
1970	!-- First, all PEs get the information of all cross-sections.
1971	!-- Then the data are written to the output file by all PEs
1972	!-- while NF90_COLLECTIVE is set in subroutine
1973	!-- define_netcdf_header. Although redundant information are
1974	!-- written to the output file in that case, the performance
1975	!-- is significantly better compared to the case where only
1976	!-- the first row of PEs in x-direction (myidx = 0) is given
1977	!-- the output while NF90_INDEPENDENT is set.
1978	IF ( npey /= 1 ) THEN
1979
1980	#if defined( __parallel )
1981	!
1982	!-- Distribute data over all PEs along y
1983	ngp = ( nxr-nxl+1 ) * ( nzt_do-nzb_do+1 ) * ns
1984	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1985	CALL MPI_ALLREDUCE( local_2d_sections_l(nxl,1,nzb_do), &
1986	local_2d_sections(nxl,1,nzb_do), &
1987	ngp, MPI_REAL, MPI_SUM, comm1dy, &
1988	ierr )
1989	#else
1990	local_2d_sections = local_2d_sections_l
1991	#endif
1992	ENDIF
1993	!
1994	!-- Do not output redundant ghost point data except for the
1995	!-- boundaries of the total domain.
1996	! IF ( nxr == nx ) THEN
1997	! nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1998	! id_var_do2d(av,ivar), &
1999	! local_2d_sections(nxl:nxr+1,1:ns, &
2000	! nzb_do:nzt_do), &
2001	! start = (/ nxl+1, 1, 1, &
2002	! do2d_xz_time_count(av) /), &
2003	! count = (/ nxr-nxl+2, ns, nzt_do-nzb_do+1, &
2004	! 1 /) )
2005	! ELSE
2006	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
2007	id_var_do2d(av,ivar), &
2008	local_2d_sections(nxl:nxr,1:ns, &
2009	nzb_do:nzt_do), &
2010	start = (/ nxl+1, 1, 1, &
2011	do2d_xz_time_count(av) /), &
2012	count = (/ nxr-nxl+1, ns, nzt_do-nzb_do+1, &
2013	1 /) )
2014	! ENDIF
2015
2016	CALL netcdf_handle_error( 'data_output_2d', 57 )
2017
2018	CASE ( 'yz' )
2019	!
2020	!-- First, all PEs get the information of all cross-sections.
2021	!-- Then the data are written to the output file by all PEs
2022	!-- while NF90_COLLECTIVE is set in subroutine
2023	!-- define_netcdf_header. Although redundant information are
2024	!-- written to the output file in that case, the performance
2025	!-- is significantly better compared to the case where only
2026	!-- the first row of PEs in y-direction (myidy = 0) is given
2027	!-- the output while NF90_INDEPENDENT is set.
2028	IF ( npex /= 1 ) THEN
2029
2030	#if defined( __parallel )
2031	!
2032	!-- Distribute data over all PEs along x
2033	ngp = ( nyn-nys+1 ) * ( nzt-nzb + 2 ) * ns
2034	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
2035	CALL MPI_ALLREDUCE( local_2d_sections_l(1,nys,nzb_do), &
2036	local_2d_sections(1,nys,nzb_do), &
2037	ngp, MPI_REAL, MPI_SUM, comm1dx, &
2038	ierr )
2039	#else
2040	local_2d_sections = local_2d_sections_l
2041	#endif
2042	ENDIF
2043	!
2044	!-- Do not output redundant ghost point data except for the
2045	!-- boundaries of the total domain.
2046	! IF ( nyn == ny ) THEN
2047	! nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2048	! id_var_do2d(av,ivar), &
2049	! local_2d_sections(1:ns, &
2050	! nys:nyn+1,nzb_do:nzt_do), &
2051	! start = (/ 1, nys+1, 1, &
2052	! do2d_yz_time_count(av) /), &
2053	! count = (/ ns, nyn-nys+2, &
2054	! nzt_do-nzb_do+1, 1 /) )
2055	! ELSE
2056	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2057	id_var_do2d(av,ivar), &
2058	local_2d_sections(1:ns,nys:nyn, &
2059	nzb_do:nzt_do), &
2060	start = (/ 1, nys+1, 1, &
2061	do2d_yz_time_count(av) /), &
2062	count = (/ ns, nyn-nys+1, &
2063	nzt_do-nzb_do+1, 1 /) )
2064	! ENDIF
2065
2066	CALL netcdf_handle_error( 'data_output_2d', 60 )
2067
2068	CASE DEFAULT
2069	message_string = 'unknown cross-section: ' // TRIM( mode )
2070	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
2071
2072	END SELECT
2073
2074	ENDIF
2075	#endif
2076	ENDIF
2077
2078	ivar = ivar + 1
2079	l = MAX( 2, LEN_TRIM( do2d(av,ivar) ) )
2080	do2d_mode = do2d(av,ivar)(l-1:l)
2081
2082	ENDDO
2083
2084	!
2085	!-- Deallocate temporary arrays.
2086	IF ( ALLOCATED( level_z ) ) DEALLOCATE( level_z )
2087	IF ( netcdf_data_format > 4 ) THEN
2088	DEALLOCATE( local_pf, local_2d, local_2d_sections )
2089	IF( mode == 'xz' .OR. mode == 'yz' ) DEALLOCATE( local_2d_sections_l )
2090	ENDIF
2091	#if defined( __parallel )
2092	IF ( .NOT. data_output_2d_on_each_pe .AND. myid == 0 ) THEN
2093	DEALLOCATE( total_2d )
2094	ENDIF
2095	#endif
2096
2097	!
2098	!-- Close plot output file.
2099	file_id = 20 + s_ind
2100
2101	IF ( data_output_2d_on_each_pe ) THEN
2102	DO i = 0, io_blocks-1
2103	IF ( i == io_group ) THEN
2104	CALL close_file( file_id )
2105	ENDIF
2106	#if defined( __parallel )
2107	CALL MPI_BARRIER( comm2d, ierr )
2108	#endif
2109	ENDDO
2110	ELSE
2111	IF ( myid == 0 ) CALL close_file( file_id )
2112	ENDIF
2113
2114	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
2115
2116	IF ( debug_output_timestep ) CALL debug_message( 'data_output_2d', 'end' )
2117
2118
2119	END SUBROUTINE data_output_2d

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |