Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 3574

Last change on this file since 3574 was 3569, checked in by kanani, 6 years ago
Fix for biomet output (ticket:757), merge of uv_exposure into biometeorology_mod
Property svn:keywords set to `Id`
File size: 96.6 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 3569 2018-11-27 17:03:40Z raasch $
27	! Remove fill_value from bio_data_output_2d call
28	! dom_dwd_user, Schrempf:
29	! Clean up of biometeorology calls,
30	! remove uv exposure model code, this is now part of biometeorology_mod.
31	!
32	! 3554 2018-11-22 11:24:52Z gronemeier
33	! - add variable description
34	! - rename variable 'if' into 'ivar'
35	! - removed namelist LOCAL
36	! - removed variable rtext
37	!
38	! 3525 2018-11-14 16:06:14Z kanani
39	! Changes related to clean-up of biometeorology (dom_dwd_user)
40	!
41	! 3467 2018-10-30 19:05:21Z suehring
42	! Implementation of a new aerosol module salsa.
43	!
44	! 3448 2018-10-29 18:14:31Z kanani
45	! Add biometeorology
46	!
47	! 3421 2018-10-24 18:39:32Z gronemeier
48	! Renamed output variables
49	!
50	! 3419 2018-10-24 17:27:31Z gronemeier
51	! minor formatting (kanani)
52	! chem_data_output_2d subroutine added (basit)
53	!
54	! 3294 2018-10-01 02:37:10Z raasch
55	! changes concerning modularization of ocean option
56	!
57	! 3274 2018-09-24 15:42:55Z knoop
58	! Modularization of all bulk cloud physics code components
59	!
60	! 3241 2018-09-12 15:02:00Z raasch
61	! unused variables removed
62	!
63	! 3176 2018-07-26 17:12:48Z suehring
64	! Remove output of latent heat flux at urban-surfaces and set fill values
65	! instead
66	!
67	! 3052 2018-05-31 06:11:20Z raasch
68	! Do not open FORTRAN binary files in case of parallel netCDF I/O
69	!
70	! 3049 2018-05-29 13:52:36Z Giersch
71	! Error messages revised
72	!
73	! 3045 2018-05-28 07:55:41Z Giersch
74	! Error messages revised
75	!
76	! 3014 2018-05-09 08:42:38Z maronga
77	! Added nzb_do and nzt_do for some modules for 2d output
78	!
79	! 3004 2018-04-27 12:33:25Z Giersch
80	! precipitation_rate removed, case prr*_xy removed, to_be_resorted have to point
81	! to ql_vp_av and not to ql_vp, allocation checks implemented (averaged data
82	! will be assigned to fill values if no allocation happened so far)
83	!
84	! 2963 2018-04-12 14:47:44Z suehring
85	! Introduce index for vegetation/wall, pavement/green-wall and water/window
86	! surfaces, for clearer access of surface fraction, albedo, emissivity, etc. .
87	!
88	! 2817 2018-02-19 16:32:21Z knoop
89	! Preliminary gust module interface implemented
90	!
91	! 2805 2018-02-14 17:00:09Z suehring
92	! Consider also default-type surfaces for surface temperature output.
93	!
94	! 2797 2018-02-08 13:24:35Z suehring
95	! Enable output of ground-heat flux also at urban surfaces.
96	!
97	! 2743 2018-01-12 16:03:39Z suehring
98	! In case of natural- and urban-type surfaces output surfaces fluxes in W/m2.
99	!
100	! 2742 2018-01-12 14:59:47Z suehring
101	! Enable output of surface temperature
102	!
103	! 2735 2018-01-11 12:01:27Z suehring
104	! output of r_a moved from land-surface to consider also urban-type surfaces
105	!
106	! 2718 2018-01-02 08:49:38Z maronga
107	! Corrected "Former revisions" section
108	!
109	! 2696 2017-12-14 17:12:51Z kanani
110	! Change in file header (GPL part)
111	! Implementation of uv exposure model (FK)
112	! Implementation of turbulence_closure_mod (TG)
113	! Set fill values at topography grid points or e.g. non-natural-type surface
114	! in case of LSM output (MS)
115	!
116	! 2512 2017-10-04 08:26:59Z raasch
117	! upper bounds of cross section output changed from nx+1,ny+1 to nx,ny
118	! no output of ghost layer data
119	!
120	! 2292 2017-06-20 09:51:42Z schwenkel
121	! Implementation of new microphysic scheme: cloud_scheme = 'morrison'
122	! includes two more prognostic equations for cloud drop concentration (nc)
123	! and cloud water content (qc).
124	!
125	! 2277 2017-06-12 10:47:51Z kanani
126	! Removed unused variables do2d_xy_n, do2d_xz_n, do2d_yz_n
127	!
128	! 2233 2017-05-30 18:08:54Z suehring
129	!
130	! 2232 2017-05-30 17:47:52Z suehring
131	! Adjustments to new surface concept
132	!
133	!
134	! 2190 2017-03-21 12:16:43Z raasch
135	! bugfix for r2031: string rho replaced by rho_ocean
136	!
137	! 2031 2016-10-21 15:11:58Z knoop
138	! renamed variable rho to rho_ocean and rho_av to rho_ocean_av
139	!
140	! 2000 2016-08-20 18:09:15Z knoop
141	! Forced header and separation lines into 80 columns
142	!
143	! 1980 2016-07-29 15:51:57Z suehring
144	! Bugfix, in order to steer user-defined output, setting flag found explicitly
145	! to .F.
146	!
147	! 1976 2016-07-27 13:28:04Z maronga
148	! Output of radiation quantities is now done directly in the respective module
149	!
150	! 1972 2016-07-26 07:52:02Z maronga
151	! Output of land surface quantities is now done directly in the respective
152	! module
153	!
154	! 1960 2016-07-12 16:34:24Z suehring
155	! Scalar surface flux added
156	! Rename INTEGER variable s into s_ind, as s is already assigned to scalar
157	!
158	! 1849 2016-04-08 11:33:18Z hoffmann
159	! precipitation_amount, precipitation_rate, prr moved to arrays_3d
160	!
161	! 1822 2016-04-07 07:49:42Z hoffmann
162	! Output of bulk cloud physics simplified.
163	!
164	! 1788 2016-03-10 11:01:04Z maronga
165	! Added output of z0q
166	!
167	! 1783 2016-03-06 18:36:17Z raasch
168	! name change of netcdf routines and module + related changes
169	!
170	! 1745 2016-02-05 13:06:51Z gronemeier
171	! Bugfix: test if time axis limit exceeds moved to point after call of check_open
172	!
173	! 1703 2015-11-02 12:38:44Z raasch
174	! bugfix for output of single (*) xy-sections in case of parallel netcdf I/O
175	!
176	! 1701 2015-11-02 07:43:04Z maronga
177	! Bugfix in output of RRTGM data
178	!
179	! 1691 2015-10-26 16:17:44Z maronga
180	! Added output of Obukhov length (ol) and radiative heating rates for RRTMG.
181	! Formatting corrections.
182	!
183	! 1682 2015-10-07 23:56:08Z knoop
184	! Code annotations made doxygen readable
185	!
186	! 1585 2015-04-30 07:05:52Z maronga
187	! Added support for RRTMG
188	!
189	! 1555 2015-03-04 17:44:27Z maronga
190	! Added output of r_a and r_s
191	!
192	! 1551 2015-03-03 14:18:16Z maronga
193	! Added suppport for land surface model and radiation model output. In the course
194	! of this action, the limits for vertical loops have been changed (from nzb and
195	! nzt+1 to nzb_do and nzt_do, respectively in order to allow soil model output).
196	! Moreover, a new vertical grid zs was introduced.
197	!
198	! 1359 2014-04-11 17:15:14Z hoffmann
199	! New particle structure integrated.
200	!
201	! 1353 2014-04-08 15:21:23Z heinze
202	! REAL constants provided with KIND-attribute
203	!
204	! 1327 2014-03-21 11:00:16Z raasch
205	! parts concerning iso2d output removed,
206	! -netcdf output queries
207	!
208	! 1320 2014-03-20 08:40:49Z raasch
209	! ONLY-attribute added to USE-statements,
210	! kind-parameters added to all INTEGER and REAL declaration statements,
211	! kinds are defined in new module kinds,
212	! revision history before 2012 removed,
213	! comment fields (!:) to be used for variable explanations added to
214	! all variable declaration statements
215	!
216	! 1318 2014-03-17 13:35:16Z raasch
217	! barrier argument removed from cpu_log.
218	! module interfaces removed
219	!
220	! 1311 2014-03-14 12:13:39Z heinze
221	! bugfix: close #if defined( __netcdf )
222	!
223	! 1308 2014-03-13 14:58:42Z fricke
224	! +local_2d_sections, local_2d_sections_l, ns
225	! Check, if the limit of the time dimension is exceeded for parallel output
226	! To increase the performance for parallel output, the following is done:
227	! - Update of time axis is only done by PE0
228	! - Cross sections are first stored on a local array and are written
229	! collectively to the output file by all PEs.
230	!
231	! 1115 2013-03-26 18:16:16Z hoffmann
232	! ql is calculated by calc_liquid_water_content
233	!
234	! 1076 2012-12-05 08:30:18Z hoffmann
235	! Bugfix in output of ql
236	!
237	! 1065 2012-11-22 17:42:36Z hoffmann
238	! Bugfix: Output of cross sections of ql
239	!
240	! 1053 2012-11-13 17:11:03Z hoffmann
241	! +qr, nr, qc and cross sections
242	!
243	! 1036 2012-10-22 13:43:42Z raasch
244	! code put under GPL (PALM 3.9)
245	!
246	! 1031 2012-10-19 14:35:30Z raasch
247	! netCDF4 without parallel file support implemented
248	!
249	! 1007 2012-09-19 14:30:36Z franke
250	! Bugfix: missing calculation of ql_vp added
251	!
252	! 978 2012-08-09 08:28:32Z fricke
253	! +z0h
254	!
255	! Revision 1.1 1997/08/11 06:24:09 raasch
256	! Initial revision
257	!
258	!
259	! Description:
260	! ------------
261	!> Data output of cross-sections in netCDF format or binary format
262	!> to be later converted to NetCDF by helper routine combine_plot_fields.
263	!> Attention: The position of the sectional planes is still not always computed
264	!> --------- correctly. (zu is used always)!
265	!------------------------------------------------------------------------------!
266	SUBROUTINE data_output_2d( mode, av )
267
268
269	USE arrays_3d, &
270	ONLY: dzw, e, heatflux_output_conversion, nc, nr, p, pt, &
271	precipitation_amount, prr, q, qc, ql, ql_c, ql_v, qr, s, tend, &
272	u, v, vpt, w, zu, zw, waterflux_output_conversion, hyrho, d_exner
273
274	USE averaging
275
276	USE basic_constants_and_equations_mod, &
277	ONLY: c_p, lv_d_cp, l_v
278
279	USE biometeorology_mod, &
280	ONLY: bio_data_output_2d
281
282	USE bulk_cloud_model_mod, &
283	ONLY: bulk_cloud_model, bcm_data_output_2d
284
285	USE chemistry_model_mod, &
286	ONLY: chem_data_output_2d
287
288	USE control_parameters, &
289	ONLY: air_chemistry, biometeorology, data_output_2d_on_each_pe, &
290	data_output_xy, data_output_xz, data_output_yz, do2d, &
291	do2d_xy_last_time, do2d_xy_time_count, &
292	do2d_xz_last_time, do2d_xz_time_count, &
293	do2d_yz_last_time, do2d_yz_time_count, &
294	ibc_uv_b, io_blocks, io_group, land_surface, message_string, &
295	ntdim_2d_xy, ntdim_2d_xz, ntdim_2d_yz, &
296	ocean_mode, psolver, section, simulated_time, &
297	time_since_reference_point
298
299	USE cpulog, &
300	ONLY: cpu_log, log_point
301
302	USE gust_mod, &
303	ONLY: gust_data_output_2d, gust_module_enabled
304
305	USE indices, &
306	ONLY: nbgp, nx, nxl, nxlg, nxr, nxrg, ny, nyn, nyng, nys, nysg, &
307	nzb, nzt, wall_flags_0
308
309	USE kinds
310
311	USE land_surface_model_mod, &
312	ONLY: lsm_data_output_2d, zs
313
314	#if defined( __netcdf )
315	USE NETCDF
316	#endif
317
318	USE netcdf_interface, &
319	ONLY: fill_value, id_set_xy, id_set_xz, id_set_yz, id_var_do2d, &
320	id_var_time_xy, id_var_time_xz, id_var_time_yz, nc_stat, &
321	netcdf_data_format, netcdf_handle_error
322
323	USE ocean_mod, &
324	ONLY: ocean_data_output_2d
325
326	USE particle_attributes, &
327	ONLY: grid_particles, number_of_particles, particle_advection_start, &
328	particles, prt_count
329
330	USE pegrid
331
332	USE radiation_model_mod, &
333	ONLY: radiation, radiation_data_output_2d
334
335	USE salsa_mod, &
336	ONLY: salsa, salsa_data_output_2d
337
338	USE surface_mod, &
339	ONLY: ind_pav_green, ind_veg_wall, ind_wat_win, surf_def_h, &
340	surf_lsm_h, surf_usm_h
341
342	USE turbulence_closure_mod, &
343	ONLY: tcm_data_output_2d
344
345
346	IMPLICIT NONE
347
348	CHARACTER (LEN=2) :: do2d_mode !< output mode of variable ('xy', 'xz', 'yz')
349	CHARACTER (LEN=2) :: mode !< mode with which the routine is called ('xy', 'xz', 'yz')
350	CHARACTER (LEN=4) :: grid !< string defining the vertical grid
351
352	INTEGER(iwp) :: av !< flag for (non-)average output
353	INTEGER(iwp) :: ngp !< number of grid points of an output slice
354	INTEGER(iwp) :: file_id !< id of output files
355	INTEGER(iwp) :: flag_nr !< number of masking flag
356	INTEGER(iwp) :: i !< loop index
357	INTEGER(iwp) :: iis !< vertical index of a xy slice in array 'local_2d_sections'
358	INTEGER(iwp) :: is !< slice index
359	INTEGER(iwp) :: ivar !< variable index
360	INTEGER(iwp) :: j !< loop index
361	INTEGER(iwp) :: k !< loop index
362	INTEGER(iwp) :: l !< loop index
363	INTEGER(iwp) :: layer_xy !< vertical index of a xy slice in array 'local_pf'
364	INTEGER(iwp) :: m !< loop index
365	INTEGER(iwp) :: n !< loop index
366	INTEGER(iwp) :: nis !< number of vertical slices to be written via parallel NetCDF output
367	INTEGER(iwp) :: ns !< number of output slices
368	INTEGER(iwp) :: nzb_do !< lower limit of the data field (usually nzb)
369	INTEGER(iwp) :: nzt_do !< upper limit of the data field (usually nzt+1)
370	INTEGER(iwp) :: s_ind !< index of slice types (xy=1, xz=2, yz=3)
371	INTEGER(iwp) :: sender !< PE id of sending PE
372	INTEGER(iwp) :: ind(4) !< index limits (lower/upper bounds) of array 'local_2d'
373
374	LOGICAL :: found !< true if output variable was found
375	LOGICAL :: resorted !< true if variable is resorted
376	LOGICAL :: two_d !< true if variable is only two dimensional
377
378	REAL(wp) :: mean_r !< mean particle radius
379	REAL(wp) :: s_r2 !< sum( particle-radius**2 )
380	REAL(wp) :: s_r3 !< sum( particle-radius**3 )
381
382	REAL(wp), DIMENSION(:), ALLOCATABLE :: level_z !< z levels for output array
383	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d !< local 2-dimensional array containing output values
384	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d_l !< local 2-dimensional array containing output values
385
386	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !< output array
387	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections !< local array containing values at all slices
388	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections_l !< local array containing values at all slices
389
390	#if defined( __parallel )
391	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: total_2d !< same as local_2d
392	#endif
393	REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !< points to array which shall be output
394
395	!
396	!-- Immediate return, if no output is requested (no respective sections
397	!-- found in parameter data_output)
398	IF ( mode == 'xy' .AND. .NOT. data_output_xy(av) ) RETURN
399	IF ( mode == 'xz' .AND. .NOT. data_output_xz(av) ) RETURN
400	IF ( mode == 'yz' .AND. .NOT. data_output_yz(av) ) RETURN
401
402	CALL cpu_log (log_point(3),'data_output_2d','start')
403
404	two_d = .FALSE. ! local variable to distinguish between output of pure 2D
405	! arrays and cross-sections of 3D arrays.
406
407	!
408	!-- Depending on the orientation of the cross-section, the respective output
409	!-- files have to be opened.
410	SELECT CASE ( mode )
411
412	CASE ( 'xy' )
413	s_ind = 1
414	ALLOCATE( level_z(nzb:nzt+1), local_2d(nxl:nxr,nys:nyn) )
415
416	IF ( netcdf_data_format > 4 ) THEN
417	ns = 1
418	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
419	ns = ns + 1
420	ENDDO
421	ns = ns - 1
422	ALLOCATE( local_2d_sections(nxl:nxr,nys:nyn,1:ns) )
423	local_2d_sections = 0.0_wp
424	ENDIF
425
426	!
427	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
428	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
429	CALL check_open( 101+av*10 )
430	ENDIF
431	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
432	CALL check_open( 21 )
433	ELSE
434	IF ( myid == 0 ) THEN
435	#if defined( __parallel )
436	ALLOCATE( total_2d(0:nx,0:ny) )
437	#endif
438	ENDIF
439	ENDIF
440
441	CASE ( 'xz' )
442	s_ind = 2
443	ALLOCATE( local_2d(nxl:nxr,nzb:nzt+1) )
444
445	IF ( netcdf_data_format > 4 ) THEN
446	ns = 1
447	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
448	ns = ns + 1
449	ENDDO
450	ns = ns - 1
451	ALLOCATE( local_2d_sections(nxl:nxr,1:ns,nzb:nzt+1) )
452	ALLOCATE( local_2d_sections_l(nxl:nxr,1:ns,nzb:nzt+1) )
453	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
454	ENDIF
455
456	!
457	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
458	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
459	CALL check_open( 102+av*10 )
460	ENDIF
461
462	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
463	CALL check_open( 22 )
464	ELSE
465	IF ( myid == 0 ) THEN
466	#if defined( __parallel )
467	ALLOCATE( total_2d(0:nx,nzb:nzt+1) )
468	#endif
469	ENDIF
470	ENDIF
471
472	CASE ( 'yz' )
473	s_ind = 3
474	ALLOCATE( local_2d(nys:nyn,nzb:nzt+1) )
475
476	IF ( netcdf_data_format > 4 ) THEN
477	ns = 1
478	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
479	ns = ns + 1
480	ENDDO
481	ns = ns - 1
482	ALLOCATE( local_2d_sections(1:ns,nys:nyn,nzb:nzt+1) )
483	ALLOCATE( local_2d_sections_l(1:ns,nys:nyn,nzb:nzt+1) )
484	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
485	ENDIF
486
487	!
488	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
489	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
490	CALL check_open( 103+av*10 )
491	ENDIF
492
493	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
494	CALL check_open( 23 )
495	ELSE
496	IF ( myid == 0 ) THEN
497	#if defined( __parallel )
498	ALLOCATE( total_2d(0:ny,nzb:nzt+1) )
499	#endif
500	ENDIF
501	ENDIF
502
503	CASE DEFAULT
504	message_string = 'unknown cross-section: ' // TRIM( mode )
505	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
506
507	END SELECT
508
509	!
510	!-- For parallel netcdf output the time axis must be limited. Return, if this
511	!-- limit is exceeded. This could be the case, if the simulated time exceeds
512	!-- the given end time by the length of the given output interval.
513	IF ( netcdf_data_format > 4 ) THEN
514	IF ( mode == 'xy' .AND. do2d_xy_time_count(av) + 1 > &
515	ntdim_2d_xy(av) ) THEN
516	WRITE ( message_string, * ) 'Output of xy 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', 'PA0384', 0, 1, 0, 6, 0 )
520	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
521	RETURN
522	ENDIF
523	IF ( mode == 'xz' .AND. do2d_xz_time_count(av) + 1 > &
524	ntdim_2d_xz(av) ) THEN
525	WRITE ( message_string, * ) 'Output of xz 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', 'PA0385', 0, 1, 0, 6, 0 )
529	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
530	RETURN
531	ENDIF
532	IF ( mode == 'yz' .AND. do2d_yz_time_count(av) + 1 > &
533	ntdim_2d_yz(av) ) THEN
534	WRITE ( message_string, * ) 'Output of yz cross-sections is not ', &
535	'given at t=', simulated_time, '&because the', &
536	' maximum number of output time levels is exceeded.'
537	CALL message( 'data_output_2d', 'PA0386', 0, 1, 0, 6, 0 )
538	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
539	RETURN
540	ENDIF
541	ENDIF
542
543	!
544	!-- Allocate a temporary array for resorting (kji -> ijk).
545	ALLOCATE( local_pf(nxl:nxr,nys:nyn,nzb:nzt+1) )
546	local_pf = 0.0
547
548	!
549	!-- Loop of all variables to be written.
550	!-- Output dimensions chosen
551	ivar = 1
552	l = MAX( 2, LEN_TRIM( do2d(av,ivar) ) )
553	do2d_mode = do2d(av,ivar)(l-1:l)
554
555	DO WHILE ( do2d(av,ivar)(1:1) /= ' ' )
556
557	IF ( do2d_mode == mode ) THEN
558	!
559	!-- Set flag to steer output of radiation, land-surface, or user-defined
560	!-- quantities
561	found = .FALSE.
562
563	nzb_do = nzb
564	nzt_do = nzt+1
565	!
566	!-- Before each output, set array local_pf to fill value
567	local_pf = fill_value
568	!
569	!-- Set masking flag for topography for not resorted arrays
570	flag_nr = 0
571
572	!
573	!-- Store the array chosen on the temporary array.
574	resorted = .FALSE.
575	SELECT CASE ( TRIM( do2d(av,ivar) ) )
576	CASE ( 'e_xy', 'e_xz', 'e_yz' )
577	IF ( av == 0 ) THEN
578	to_be_resorted => e
579	ELSE
580	IF ( .NOT. ALLOCATED( e_av ) ) THEN
581	ALLOCATE( e_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
582	e_av = REAL( fill_value, KIND = wp )
583	ENDIF
584	to_be_resorted => e_av
585	ENDIF
586	IF ( mode == 'xy' ) level_z = zu
587
588	CASE ( 'thetal_xy', 'thetal_xz', 'thetal_yz' )
589	IF ( av == 0 ) THEN
590	to_be_resorted => pt
591	ELSE
592	IF ( .NOT. ALLOCATED( lpt_av ) ) THEN
593	ALLOCATE( lpt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
594	lpt_av = REAL( fill_value, KIND = wp )
595	ENDIF
596	to_be_resorted => lpt_av
597	ENDIF
598	IF ( mode == 'xy' ) level_z = zu
599
600	CASE ( 'lwp*_xy' ) ! 2d-array
601	IF ( av == 0 ) THEN
602	DO i = nxl, nxr
603	DO j = nys, nyn
604	local_pf(i,j,nzb+1) = SUM( ql(nzb:nzt,j,i) * &
605	dzw(1:nzt+1) )
606	ENDDO
607	ENDDO
608	ELSE
609	IF ( .NOT. ALLOCATED( lwp_av ) ) THEN
610	ALLOCATE( lwp_av(nysg:nyng,nxlg:nxrg) )
611	lwp_av = REAL( fill_value, KIND = wp )
612	ENDIF
613	DO i = nxl, nxr
614	DO j = nys, nyn
615	local_pf(i,j,nzb+1) = lwp_av(j,i)
616	ENDDO
617	ENDDO
618	ENDIF
619	resorted = .TRUE.
620	two_d = .TRUE.
621	level_z(nzb+1) = zu(nzb+1)
622
623	CASE ( 'ghf*_xy' ) ! 2d-array
624	IF ( av == 0 ) THEN
625	DO m = 1, surf_lsm_h%ns
626	i = surf_lsm_h%i(m)
627	j = surf_lsm_h%j(m)
628	local_pf(i,j,nzb+1) = surf_lsm_h%ghf(m)
629	ENDDO
630	DO m = 1, surf_usm_h%ns
631	i = surf_usm_h%i(m)
632	j = surf_usm_h%j(m)
633	local_pf(i,j,nzb+1) = surf_usm_h%frac(ind_veg_wall,m) * &
634	surf_usm_h%wghf_eb(m) + &
635	surf_usm_h%frac(ind_pav_green,m) * &
636	surf_usm_h%wghf_eb_green(m) + &
637	surf_usm_h%frac(ind_wat_win,m) * &
638	surf_usm_h%wghf_eb_window(m)
639	ENDDO
640	ELSE
641	IF ( .NOT. ALLOCATED( ghf_av ) ) THEN
642	ALLOCATE( ghf_av(nysg:nyng,nxlg:nxrg) )
643	ghf_av = REAL( fill_value, KIND = wp )
644	ENDIF
645	DO i = nxl, nxr
646	DO j = nys, nyn
647	local_pf(i,j,nzb+1) = ghf_av(j,i)
648	ENDDO
649	ENDDO
650	ENDIF
651
652	resorted = .TRUE.
653	two_d = .TRUE.
654	level_z(nzb+1) = zu(nzb+1)
655
656	CASE ( 'ol*_xy' ) ! 2d-array
657	IF ( av == 0 ) THEN
658	DO m = 1, surf_def_h(0)%ns
659	i = surf_def_h(0)%i(m)
660	j = surf_def_h(0)%j(m)
661	local_pf(i,j,nzb+1) = surf_def_h(0)%ol(m)
662	ENDDO
663	DO m = 1, surf_lsm_h%ns
664	i = surf_lsm_h%i(m)
665	j = surf_lsm_h%j(m)
666	local_pf(i,j,nzb+1) = surf_lsm_h%ol(m)
667	ENDDO
668	DO m = 1, surf_usm_h%ns
669	i = surf_usm_h%i(m)
670	j = surf_usm_h%j(m)
671	local_pf(i,j,nzb+1) = surf_usm_h%ol(m)
672	ENDDO
673	ELSE
674	IF ( .NOT. ALLOCATED( ol_av ) ) THEN
675	ALLOCATE( ol_av(nysg:nyng,nxlg:nxrg) )
676	ol_av = REAL( fill_value, KIND = wp )
677	ENDIF
678	DO i = nxl, nxr
679	DO j = nys, nyn
680	local_pf(i,j,nzb+1) = ol_av(j,i)
681	ENDDO
682	ENDDO
683	ENDIF
684	resorted = .TRUE.
685	two_d = .TRUE.
686	level_z(nzb+1) = zu(nzb+1)
687
688	CASE ( 'p_xy', 'p_xz', 'p_yz' )
689	IF ( av == 0 ) THEN
690	IF ( psolver /= 'sor' ) CALL exchange_horiz( p, nbgp )
691	to_be_resorted => p
692	ELSE
693	IF ( .NOT. ALLOCATED( p_av ) ) THEN
694	ALLOCATE( p_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
695	p_av = REAL( fill_value, KIND = wp )
696	ENDIF
697	IF ( psolver /= 'sor' ) CALL exchange_horiz( p_av, nbgp )
698	to_be_resorted => p_av
699	ENDIF
700	IF ( mode == 'xy' ) level_z = zu
701
702	CASE ( 'pc_xy', 'pc_xz', 'pc_yz' ) ! particle concentration
703	IF ( av == 0 ) THEN
704	IF ( simulated_time >= particle_advection_start ) THEN
705	tend = prt_count
706	! CALL exchange_horiz( tend, nbgp )
707	ELSE
708	tend = 0.0_wp
709	ENDIF
710	DO i = nxl, nxr
711	DO j = nys, nyn
712	DO k = nzb, nzt+1
713	local_pf(i,j,k) = tend(k,j,i)
714	ENDDO
715	ENDDO
716	ENDDO
717	resorted = .TRUE.
718	ELSE
719	IF ( .NOT. ALLOCATED( pc_av ) ) THEN
720	ALLOCATE( pc_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
721	pc_av = REAL( fill_value, KIND = wp )
722	ENDIF
723	! CALL exchange_horiz( pc_av, nbgp )
724	to_be_resorted => pc_av
725	ENDIF
726
727	CASE ( 'pr_xy', 'pr_xz', 'pr_yz' ) ! mean particle radius (effective radius)
728	IF ( av == 0 ) THEN
729	IF ( simulated_time >= particle_advection_start ) THEN
730	DO i = nxl, nxr
731	DO j = nys, nyn
732	DO k = nzb, nzt+1
733	number_of_particles = prt_count(k,j,i)
734	IF (number_of_particles <= 0) CYCLE
735	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
736	s_r2 = 0.0_wp
737	s_r3 = 0.0_wp
738	DO n = 1, number_of_particles
739	IF ( particles(n)%particle_mask ) THEN
740	s_r2 = s_r2 + particles(n)%radius*2 &
741	particles(n)%weight_factor
742	s_r3 = s_r3 + particles(n)%radius*3 &
743	particles(n)%weight_factor
744	ENDIF
745	ENDDO
746	IF ( s_r2 > 0.0_wp ) THEN
747	mean_r = s_r3 / s_r2
748	ELSE
749	mean_r = 0.0_wp
750	ENDIF
751	tend(k,j,i) = mean_r
752	ENDDO
753	ENDDO
754	ENDDO
755	! CALL exchange_horiz( tend, nbgp )
756	ELSE
757	tend = 0.0_wp
758	ENDIF
759	DO i = nxl, nxr
760	DO j = nys, nyn
761	DO k = nzb, nzt+1
762	local_pf(i,j,k) = tend(k,j,i)
763	ENDDO
764	ENDDO
765	ENDDO
766	resorted = .TRUE.
767	ELSE
768	IF ( .NOT. ALLOCATED( pr_av ) ) THEN
769	ALLOCATE( pr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
770	pr_av = REAL( fill_value, KIND = wp )
771	ENDIF
772	! CALL exchange_horiz( pr_av, nbgp )
773	to_be_resorted => pr_av
774	ENDIF
775
776	CASE ( 'theta_xy', 'theta_xz', 'theta_yz' )
777	IF ( av == 0 ) THEN
778	IF ( .NOT. bulk_cloud_model ) THEN
779	to_be_resorted => pt
780	ELSE
781	DO i = nxl, nxr
782	DO j = nys, nyn
783	DO k = nzb, nzt+1
784	local_pf(i,j,k) = pt(k,j,i) + lv_d_cp * &
785	d_exner(k) * &
786	ql(k,j,i)
787	ENDDO
788	ENDDO
789	ENDDO
790	resorted = .TRUE.
791	ENDIF
792	ELSE
793	IF ( .NOT. ALLOCATED( pt_av ) ) THEN
794	ALLOCATE( pt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
795	pt_av = REAL( fill_value, KIND = wp )
796	ENDIF
797	to_be_resorted => pt_av
798	ENDIF
799	IF ( mode == 'xy' ) level_z = zu
800
801	CASE ( 'q_xy', 'q_xz', 'q_yz' )
802	IF ( av == 0 ) THEN
803	to_be_resorted => q
804	ELSE
805	IF ( .NOT. ALLOCATED( q_av ) ) THEN
806	ALLOCATE( q_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
807	q_av = REAL( fill_value, KIND = wp )
808	ENDIF
809	to_be_resorted => q_av
810	ENDIF
811	IF ( mode == 'xy' ) level_z = zu
812
813	CASE ( 'ql_xy', 'ql_xz', 'ql_yz' )
814	IF ( av == 0 ) THEN
815	to_be_resorted => ql
816	ELSE
817	IF ( .NOT. ALLOCATED( ql_av ) ) THEN
818	ALLOCATE( ql_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
819	ql_av = REAL( fill_value, KIND = wp )
820	ENDIF
821	to_be_resorted => ql_av
822	ENDIF
823	IF ( mode == 'xy' ) level_z = zu
824
825	CASE ( 'ql_c_xy', 'ql_c_xz', 'ql_c_yz' )
826	IF ( av == 0 ) THEN
827	to_be_resorted => ql_c
828	ELSE
829	IF ( .NOT. ALLOCATED( ql_c_av ) ) THEN
830	ALLOCATE( ql_c_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
831	ql_c_av = REAL( fill_value, KIND = wp )
832	ENDIF
833	to_be_resorted => ql_c_av
834	ENDIF
835	IF ( mode == 'xy' ) level_z = zu
836
837	CASE ( 'ql_v_xy', 'ql_v_xz', 'ql_v_yz' )
838	IF ( av == 0 ) THEN
839	to_be_resorted => ql_v
840	ELSE
841	IF ( .NOT. ALLOCATED( ql_v_av ) ) THEN
842	ALLOCATE( ql_v_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
843	ql_v_av = REAL( fill_value, KIND = wp )
844	ENDIF
845	to_be_resorted => ql_v_av
846	ENDIF
847	IF ( mode == 'xy' ) level_z = zu
848
849	CASE ( 'ql_vp_xy', 'ql_vp_xz', 'ql_vp_yz' )
850	IF ( av == 0 ) THEN
851	IF ( simulated_time >= particle_advection_start ) THEN
852	DO i = nxl, nxr
853	DO j = nys, nyn
854	DO k = nzb, nzt+1
855	number_of_particles = prt_count(k,j,i)
856	IF (number_of_particles <= 0) CYCLE
857	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
858	DO n = 1, number_of_particles
859	IF ( particles(n)%particle_mask ) THEN
860	tend(k,j,i) = tend(k,j,i) + &
861	particles(n)%weight_factor / &
862	prt_count(k,j,i)
863	ENDIF
864	ENDDO
865	ENDDO
866	ENDDO
867	ENDDO
868	! CALL exchange_horiz( tend, nbgp )
869	ELSE
870	tend = 0.0_wp
871	ENDIF
872	DO i = nxl, nxr
873	DO j = nys, nyn
874	DO k = nzb, nzt+1
875	local_pf(i,j,k) = tend(k,j,i)
876	ENDDO
877	ENDDO
878	ENDDO
879	resorted = .TRUE.
880	ELSE
881	IF ( .NOT. ALLOCATED( ql_vp_av ) ) THEN
882	ALLOCATE( ql_vp_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
883	ql_vp_av = REAL( fill_value, KIND = wp )
884	ENDIF
885	! CALL exchange_horiz( ql_vp_av, nbgp )
886	to_be_resorted => ql_vp_av
887	ENDIF
888	IF ( mode == 'xy' ) level_z = zu
889
890	CASE ( 'qsws*_xy' ) ! 2d-array
891	IF ( av == 0 ) THEN
892	local_pf(:,:,nzb+1) = REAL( fill_value, KIND = wp )
893	!
894	!-- In case of default surfaces, clean-up flux by density.
895	!-- In case of land-surfaces, convert fluxes into
896	!-- dynamic units
897	DO m = 1, surf_def_h(0)%ns
898	i = surf_def_h(0)%i(m)
899	j = surf_def_h(0)%j(m)
900	k = surf_def_h(0)%k(m)
901	local_pf(i,j,nzb+1) = surf_def_h(0)%qsws(m) * &
902	waterflux_output_conversion(k)
903	ENDDO
904	DO m = 1, surf_lsm_h%ns
905	i = surf_lsm_h%i(m)
906	j = surf_lsm_h%j(m)
907	k = surf_lsm_h%k(m)
908	local_pf(i,j,nzb+1) = surf_lsm_h%qsws(m) * l_v
909	ENDDO
910	ELSE
911	IF ( .NOT. ALLOCATED( qsws_av ) ) THEN
912	ALLOCATE( qsws_av(nysg:nyng,nxlg:nxrg) )
913	qsws_av = REAL( fill_value, KIND = wp )
914	ENDIF
915	DO i = nxl, nxr
916	DO j = nys, nyn
917	local_pf(i,j,nzb+1) = qsws_av(j,i)
918	ENDDO
919	ENDDO
920	ENDIF
921	resorted = .TRUE.
922	two_d = .TRUE.
923	level_z(nzb+1) = zu(nzb+1)
924
925	CASE ( 'qv_xy', 'qv_xz', 'qv_yz' )
926	IF ( av == 0 ) THEN
927	DO i = nxl, nxr
928	DO j = nys, nyn
929	DO k = nzb, nzt+1
930	local_pf(i,j,k) = q(k,j,i) - ql(k,j,i)
931	ENDDO
932	ENDDO
933	ENDDO
934	resorted = .TRUE.
935	ELSE
936	IF ( .NOT. ALLOCATED( qv_av ) ) THEN
937	ALLOCATE( qv_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
938	qv_av = REAL( fill_value, KIND = wp )
939	ENDIF
940	to_be_resorted => qv_av
941	ENDIF
942	IF ( mode == 'xy' ) level_z = zu
943
944	CASE ( 'r_a*_xy' ) ! 2d-array
945	IF ( av == 0 ) THEN
946	DO m = 1, surf_lsm_h%ns
947	i = surf_lsm_h%i(m)
948	j = surf_lsm_h%j(m)
949	local_pf(i,j,nzb+1) = surf_lsm_h%r_a(m)
950	ENDDO
951
952	DO m = 1, surf_usm_h%ns
953	i = surf_usm_h%i(m)
954	j = surf_usm_h%j(m)
955	local_pf(i,j,nzb+1) = &
956	( surf_usm_h%frac(ind_veg_wall,m) * &
957	surf_usm_h%r_a(m) + &
958	surf_usm_h%frac(ind_pav_green,m) * &
959	surf_usm_h%r_a_green(m) + &
960	surf_usm_h%frac(ind_wat_win,m) * &
961	surf_usm_h%r_a_window(m) )
962	ENDDO
963	ELSE
964	IF ( .NOT. ALLOCATED( r_a_av ) ) THEN
965	ALLOCATE( r_a_av(nysg:nyng,nxlg:nxrg) )
966	r_a_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) = r_a_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 ( 's_xy', 's_xz', 's_yz' )
979	IF ( av == 0 ) THEN
980	to_be_resorted => s
981	ELSE
982	IF ( .NOT. ALLOCATED( s_av ) ) THEN
983	ALLOCATE( s_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
984	s_av = REAL( fill_value, KIND = wp )
985	ENDIF
986	to_be_resorted => s_av
987	ENDIF
988
989	CASE ( 'shf*_xy' ) ! 2d-array
990	IF ( av == 0 ) THEN
991	!
992	!-- In case of default surfaces, clean-up flux by density.
993	!-- In case of land- and urban-surfaces, convert fluxes into
994	!-- dynamic units.
995	DO m = 1, surf_def_h(0)%ns
996	i = surf_def_h(0)%i(m)
997	j = surf_def_h(0)%j(m)
998	k = surf_def_h(0)%k(m)
999	local_pf(i,j,nzb+1) = surf_def_h(0)%shf(m) * &
1000	heatflux_output_conversion(k)
1001	ENDDO
1002	DO m = 1, surf_lsm_h%ns
1003	i = surf_lsm_h%i(m)
1004	j = surf_lsm_h%j(m)
1005	k = surf_lsm_h%k(m)
1006	local_pf(i,j,nzb+1) = surf_lsm_h%shf(m) * c_p
1007	ENDDO
1008	DO m = 1, surf_usm_h%ns
1009	i = surf_usm_h%i(m)
1010	j = surf_usm_h%j(m)
1011	k = surf_usm_h%k(m)
1012	local_pf(i,j,nzb+1) = surf_usm_h%shf(m) * c_p
1013	ENDDO
1014	ELSE
1015	IF ( .NOT. ALLOCATED( shf_av ) ) THEN
1016	ALLOCATE( shf_av(nysg:nyng,nxlg:nxrg) )
1017	shf_av = REAL( fill_value, KIND = wp )
1018	ENDIF
1019	DO i = nxl, nxr
1020	DO j = nys, nyn
1021	local_pf(i,j,nzb+1) = shf_av(j,i)
1022	ENDDO
1023	ENDDO
1024	ENDIF
1025	resorted = .TRUE.
1026	two_d = .TRUE.
1027	level_z(nzb+1) = zu(nzb+1)
1028
1029	CASE ( 'ssws*_xy' ) ! 2d-array
1030	IF ( av == 0 ) THEN
1031	DO m = 1, surf_def_h(0)%ns
1032	i = surf_def_h(0)%i(m)
1033	j = surf_def_h(0)%j(m)
1034	local_pf(i,j,nzb+1) = surf_def_h(0)%ssws(m)
1035	ENDDO
1036	DO m = 1, surf_lsm_h%ns
1037	i = surf_lsm_h%i(m)
1038	j = surf_lsm_h%j(m)
1039	local_pf(i,j,nzb+1) = surf_lsm_h%ssws(m)
1040	ENDDO
1041	DO m = 1, surf_usm_h%ns
1042	i = surf_usm_h%i(m)
1043	j = surf_usm_h%j(m)
1044	local_pf(i,j,nzb+1) = surf_usm_h%ssws(m)
1045	ENDDO
1046	ELSE
1047	IF ( .NOT. ALLOCATED( ssws_av ) ) THEN
1048	ALLOCATE( ssws_av(nysg:nyng,nxlg:nxrg) )
1049	ssws_av = REAL( fill_value, KIND = wp )
1050	ENDIF
1051	DO i = nxl, nxr
1052	DO j = nys, nyn
1053	local_pf(i,j,nzb+1) = ssws_av(j,i)
1054	ENDDO
1055	ENDDO
1056	ENDIF
1057	resorted = .TRUE.
1058	two_d = .TRUE.
1059	level_z(nzb+1) = zu(nzb+1)
1060
1061	CASE ( 't*_xy' ) ! 2d-array
1062	IF ( av == 0 ) THEN
1063	DO m = 1, surf_def_h(0)%ns
1064	i = surf_def_h(0)%i(m)
1065	j = surf_def_h(0)%j(m)
1066	local_pf(i,j,nzb+1) = surf_def_h(0)%ts(m)
1067	ENDDO
1068	DO m = 1, surf_lsm_h%ns
1069	i = surf_lsm_h%i(m)
1070	j = surf_lsm_h%j(m)
1071	local_pf(i,j,nzb+1) = surf_lsm_h%ts(m)
1072	ENDDO
1073	DO m = 1, surf_usm_h%ns
1074	i = surf_usm_h%i(m)
1075	j = surf_usm_h%j(m)
1076	local_pf(i,j,nzb+1) = surf_usm_h%ts(m)
1077	ENDDO
1078	ELSE
1079	IF ( .NOT. ALLOCATED( ts_av ) ) THEN
1080	ALLOCATE( ts_av(nysg:nyng,nxlg:nxrg) )
1081	ts_av = REAL( fill_value, KIND = wp )
1082	ENDIF
1083	DO i = nxl, nxr
1084	DO j = nys, nyn
1085	local_pf(i,j,nzb+1) = ts_av(j,i)
1086	ENDDO
1087	ENDDO
1088	ENDIF
1089	resorted = .TRUE.
1090	two_d = .TRUE.
1091	level_z(nzb+1) = zu(nzb+1)
1092
1093	CASE ( 'tsurf*_xy' ) ! 2d-array
1094	IF ( av == 0 ) THEN
1095	DO m = 1, surf_def_h(0)%ns
1096	i = surf_def_h(0)%i(m)
1097	j = surf_def_h(0)%j(m)
1098	local_pf(i,j,nzb+1) = surf_def_h(0)%pt_surface(m)
1099	ENDDO
1100
1101	DO m = 1, surf_lsm_h%ns
1102	i = surf_lsm_h%i(m)
1103	j = surf_lsm_h%j(m)
1104	local_pf(i,j,nzb+1) = surf_lsm_h%pt_surface(m)
1105	ENDDO
1106
1107	DO m = 1, surf_usm_h%ns
1108	i = surf_usm_h%i(m)
1109	j = surf_usm_h%j(m)
1110	local_pf(i,j,nzb+1) = surf_usm_h%pt_surface(m)
1111	ENDDO
1112
1113	ELSE
1114	IF ( .NOT. ALLOCATED( tsurf_av ) ) THEN
1115	ALLOCATE( tsurf_av(nysg:nyng,nxlg:nxrg) )
1116	tsurf_av = REAL( fill_value, KIND = wp )
1117	ENDIF
1118	DO i = nxl, nxr
1119	DO j = nys, nyn
1120	local_pf(i,j,nzb+1) = tsurf_av(j,i)
1121	ENDDO
1122	ENDDO
1123	ENDIF
1124	resorted = .TRUE.
1125	two_d = .TRUE.
1126	level_z(nzb+1) = zu(nzb+1)
1127
1128	CASE ( 'u_xy', 'u_xz', 'u_yz' )
1129	flag_nr = 1
1130	IF ( av == 0 ) THEN
1131	to_be_resorted => u
1132	ELSE
1133	IF ( .NOT. ALLOCATED( u_av ) ) THEN
1134	ALLOCATE( u_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1135	u_av = REAL( fill_value, KIND = wp )
1136	ENDIF
1137	to_be_resorted => u_av
1138	ENDIF
1139	IF ( mode == 'xy' ) level_z = zu
1140	!
1141	!-- Substitute the values generated by "mirror" boundary condition
1142	!-- at the bottom boundary by the real surface values.
1143	IF ( do2d(av,ivar) == 'u_xz' .OR. do2d(av,ivar) == 'u_yz' ) THEN
1144	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
1145	ENDIF
1146
1147	CASE ( 'us*_xy' ) ! 2d-array
1148	IF ( av == 0 ) THEN
1149	DO m = 1, surf_def_h(0)%ns
1150	i = surf_def_h(0)%i(m)
1151	j = surf_def_h(0)%j(m)
1152	local_pf(i,j,nzb+1) = surf_def_h(0)%us(m)
1153	ENDDO
1154	DO m = 1, surf_lsm_h%ns
1155	i = surf_lsm_h%i(m)
1156	j = surf_lsm_h%j(m)
1157	local_pf(i,j,nzb+1) = surf_lsm_h%us(m)
1158	ENDDO
1159	DO m = 1, surf_usm_h%ns
1160	i = surf_usm_h%i(m)
1161	j = surf_usm_h%j(m)
1162	local_pf(i,j,nzb+1) = surf_usm_h%us(m)
1163	ENDDO
1164	ELSE
1165	IF ( .NOT. ALLOCATED( us_av ) ) THEN
1166	ALLOCATE( us_av(nysg:nyng,nxlg:nxrg) )
1167	us_av = REAL( fill_value, KIND = wp )
1168	ENDIF
1169	DO i = nxl, nxr
1170	DO j = nys, nyn
1171	local_pf(i,j,nzb+1) = us_av(j,i)
1172	ENDDO
1173	ENDDO
1174	ENDIF
1175	resorted = .TRUE.
1176	two_d = .TRUE.
1177	level_z(nzb+1) = zu(nzb+1)
1178
1179	CASE ( 'v_xy', 'v_xz', 'v_yz' )
1180	flag_nr = 2
1181	IF ( av == 0 ) THEN
1182	to_be_resorted => v
1183	ELSE
1184	IF ( .NOT. ALLOCATED( v_av ) ) THEN
1185	ALLOCATE( v_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1186	v_av = REAL( fill_value, KIND = wp )
1187	ENDIF
1188	to_be_resorted => v_av
1189	ENDIF
1190	IF ( mode == 'xy' ) level_z = zu
1191	!
1192	!-- Substitute the values generated by "mirror" boundary condition
1193	!-- at the bottom boundary by the real surface values.
1194	IF ( do2d(av,ivar) == 'v_xz' .OR. do2d(av,ivar) == 'v_yz' ) THEN
1195	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
1196	ENDIF
1197
1198	CASE ( 'thetav_xy', 'thetav_xz', 'thetav_yz' )
1199	IF ( av == 0 ) THEN
1200	to_be_resorted => vpt
1201	ELSE
1202	IF ( .NOT. ALLOCATED( vpt_av ) ) THEN
1203	ALLOCATE( vpt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1204	vpt_av = REAL( fill_value, KIND = wp )
1205	ENDIF
1206	to_be_resorted => vpt_av
1207	ENDIF
1208	IF ( mode == 'xy' ) level_z = zu
1209
1210	CASE ( 'w_xy', 'w_xz', 'w_yz' )
1211	flag_nr = 3
1212	IF ( av == 0 ) THEN
1213	to_be_resorted => w
1214	ELSE
1215	IF ( .NOT. ALLOCATED( w_av ) ) THEN
1216	ALLOCATE( w_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
1217	w_av = REAL( fill_value, KIND = wp )
1218	ENDIF
1219	to_be_resorted => w_av
1220	ENDIF
1221	IF ( mode == 'xy' ) level_z = zw
1222
1223	CASE ( 'z0*_xy' ) ! 2d-array
1224	IF ( av == 0 ) THEN
1225	DO m = 1, surf_def_h(0)%ns
1226	i = surf_def_h(0)%i(m)
1227	j = surf_def_h(0)%j(m)
1228	local_pf(i,j,nzb+1) = surf_def_h(0)%z0(m)
1229	ENDDO
1230	DO m = 1, surf_lsm_h%ns
1231	i = surf_lsm_h%i(m)
1232	j = surf_lsm_h%j(m)
1233	local_pf(i,j,nzb+1) = surf_lsm_h%z0(m)
1234	ENDDO
1235	DO m = 1, surf_usm_h%ns
1236	i = surf_usm_h%i(m)
1237	j = surf_usm_h%j(m)
1238	local_pf(i,j,nzb+1) = surf_usm_h%z0(m)
1239	ENDDO
1240	ELSE
1241	IF ( .NOT. ALLOCATED( z0_av ) ) THEN
1242	ALLOCATE( z0_av(nysg:nyng,nxlg:nxrg) )
1243	z0_av = REAL( fill_value, KIND = wp )
1244	ENDIF
1245	DO i = nxl, nxr
1246	DO j = nys, nyn
1247	local_pf(i,j,nzb+1) = z0_av(j,i)
1248	ENDDO
1249	ENDDO
1250	ENDIF
1251	resorted = .TRUE.
1252	two_d = .TRUE.
1253	level_z(nzb+1) = zu(nzb+1)
1254
1255	CASE ( 'z0h*_xy' ) ! 2d-array
1256	IF ( av == 0 ) THEN
1257	DO m = 1, surf_def_h(0)%ns
1258	i = surf_def_h(0)%i(m)
1259	j = surf_def_h(0)%j(m)
1260	local_pf(i,j,nzb+1) = surf_def_h(0)%z0h(m)
1261	ENDDO
1262	DO m = 1, surf_lsm_h%ns
1263	i = surf_lsm_h%i(m)
1264	j = surf_lsm_h%j(m)
1265	local_pf(i,j,nzb+1) = surf_lsm_h%z0h(m)
1266	ENDDO
1267	DO m = 1, surf_usm_h%ns
1268	i = surf_usm_h%i(m)
1269	j = surf_usm_h%j(m)
1270	local_pf(i,j,nzb+1) = surf_usm_h%z0h(m)
1271	ENDDO
1272	ELSE
1273	IF ( .NOT. ALLOCATED( z0h_av ) ) THEN
1274	ALLOCATE( z0h_av(nysg:nyng,nxlg:nxrg) )
1275	z0h_av = REAL( fill_value, KIND = wp )
1276	ENDIF
1277	DO i = nxl, nxr
1278	DO j = nys, nyn
1279	local_pf(i,j,nzb+1) = z0h_av(j,i)
1280	ENDDO
1281	ENDDO
1282	ENDIF
1283	resorted = .TRUE.
1284	two_d = .TRUE.
1285	level_z(nzb+1) = zu(nzb+1)
1286
1287	CASE ( 'z0q*_xy' ) ! 2d-array
1288	IF ( av == 0 ) THEN
1289	DO m = 1, surf_def_h(0)%ns
1290	i = surf_def_h(0)%i(m)
1291	j = surf_def_h(0)%j(m)
1292	local_pf(i,j,nzb+1) = surf_def_h(0)%z0q(m)
1293	ENDDO
1294	DO m = 1, surf_lsm_h%ns
1295	i = surf_lsm_h%i(m)
1296	j = surf_lsm_h%j(m)
1297	local_pf(i,j,nzb+1) = surf_lsm_h%z0q(m)
1298	ENDDO
1299	DO m = 1, surf_usm_h%ns
1300	i = surf_usm_h%i(m)
1301	j = surf_usm_h%j(m)
1302	local_pf(i,j,nzb+1) = surf_usm_h%z0q(m)
1303	ENDDO
1304	ELSE
1305	IF ( .NOT. ALLOCATED( z0q_av ) ) THEN
1306	ALLOCATE( z0q_av(nysg:nyng,nxlg:nxrg) )
1307	z0q_av = REAL( fill_value, KIND = wp )
1308	ENDIF
1309	DO i = nxl, nxr
1310	DO j = nys, nyn
1311	local_pf(i,j,nzb+1) = z0q_av(j,i)
1312	ENDDO
1313	ENDDO
1314	ENDIF
1315	resorted = .TRUE.
1316	two_d = .TRUE.
1317	level_z(nzb+1) = zu(nzb+1)
1318
1319	CASE DEFAULT
1320
1321	!
1322	!-- Quantities of other modules
1323	IF ( .NOT. found .AND. bulk_cloud_model ) THEN
1324	CALL bcm_data_output_2d( av, do2d(av,ivar), found, grid, mode,&
1325	local_pf, two_d, nzb_do, nzt_do )
1326	ENDIF
1327
1328	IF ( .NOT. found .AND. gust_module_enabled ) THEN
1329	CALL gust_data_output_2d( av, do2d(av,ivar), found, grid, &
1330	local_pf, two_d, nzb_do, nzt_do )
1331	ENDIF
1332
1333	IF ( .NOT. found .AND. biometeorology ) THEN
1334	CALL bio_data_output_2d( av, do2d(av,ivar), found, grid, &
1335	local_pf, two_d, nzb_do, nzt_do )
1336	ENDIF
1337
1338	IF ( .NOT. found .AND. land_surface ) THEN
1339	CALL lsm_data_output_2d( av, do2d(av,ivar), found, grid, mode,&
1340	local_pf, two_d, nzb_do, nzt_do )
1341	ENDIF
1342
1343	IF ( .NOT. found .AND. ocean_mode ) THEN
1344	CALL ocean_data_output_2d( av, do2d(av,ivar), found, grid, &
1345	mode, local_pf, nzb_do, nzt_do )
1346	ENDIF
1347
1348	IF ( .NOT. found .AND. radiation ) THEN
1349	CALL radiation_data_output_2d( av, do2d(av,ivar), found, grid,&
1350	mode, local_pf, two_d, &
1351	nzb_do, nzt_do )
1352	ENDIF
1353
1354	IF ( .NOT. found .AND. salsa ) THEN
1355	CALL salsa_data_output_2d( av, do2d(av,ivar), found, grid, &
1356	mode, local_pf, two_d )
1357	ENDIF
1358
1359	IF ( .NOT. found .AND. air_chemistry ) THEN
1360	CALL chem_data_output_2d( av, do2d(av,ivar), found, grid, mode, &
1361	local_pf, two_d, nzb_do, nzt_do, fill_value )
1362	ENDIF
1363	!
1364	!-- User defined quantities
1365	IF ( .NOT. found ) THEN
1366	CALL user_data_output_2d( av, do2d(av,ivar), found, grid, &
1367	local_pf, two_d, nzb_do, nzt_do )
1368	ENDIF
1369
1370	resorted = .TRUE.
1371
1372	IF ( grid == 'zu' ) THEN
1373	IF ( mode == 'xy' ) level_z = zu
1374	ELSEIF ( grid == 'zw' ) THEN
1375	IF ( mode == 'xy' ) level_z = zw
1376	ELSEIF ( grid == 'zu1' ) THEN
1377	IF ( mode == 'xy' ) level_z(nzb+1) = zu(nzb+1)
1378	ELSEIF ( grid == 'zs' ) THEN
1379	IF ( mode == 'xy' ) level_z = zs
1380	ENDIF
1381
1382	IF ( .NOT. found ) THEN
1383	message_string = 'no output provided for: ' // &
1384	TRIM( do2d(av,ivar) )
1385	CALL message( 'data_output_2d', 'PA0181', 0, 0, 0, 6, 0 )
1386	ENDIF
1387
1388	END SELECT
1389
1390	!
1391	!-- Resort the array to be output, if not done above. Flag topography
1392	!-- grid points with fill values, using the corresponding maksing flag.
1393	IF ( .NOT. resorted ) THEN
1394	DO i = nxl, nxr
1395	DO j = nys, nyn
1396	DO k = nzb_do, nzt_do
1397	local_pf(i,j,k) = MERGE( to_be_resorted(k,j,i), &
1398	REAL( fill_value, KIND = wp ), &
1399	BTEST( wall_flags_0(k,j,i), &
1400	flag_nr ) )
1401	ENDDO
1402	ENDDO
1403	ENDDO
1404	ENDIF
1405
1406	!
1407	!-- Output of the individual cross-sections, depending on the cross-
1408	!-- section mode chosen.
1409	is = 1
1410	loop1: DO WHILE ( section(is,s_ind) /= -9999 .OR. two_d )
1411
1412	SELECT CASE ( mode )
1413
1414	CASE ( 'xy' )
1415	!
1416	!-- Determine the cross section index
1417	IF ( two_d ) THEN
1418	layer_xy = nzb+1
1419	ELSE
1420	layer_xy = section(is,s_ind)
1421	ENDIF
1422
1423	!
1424	!-- Exit the loop for layers beyond the data output domain
1425	!-- (used for soil model)
1426	IF ( layer_xy > nzt_do ) THEN
1427	EXIT loop1
1428	ENDIF
1429
1430	!
1431	!-- Update the netCDF xy cross section time axis.
1432	!-- In case of parallel output, this is only done by PE0
1433	!-- to increase the performance.
1434	IF ( simulated_time /= do2d_xy_last_time(av) ) THEN
1435	do2d_xy_time_count(av) = do2d_xy_time_count(av) + 1
1436	do2d_xy_last_time(av) = simulated_time
1437	IF ( myid == 0 ) THEN
1438	IF ( .NOT. data_output_2d_on_each_pe &
1439	.OR. netcdf_data_format > 4 ) &
1440	THEN
1441	#if defined( __netcdf )
1442	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1443	id_var_time_xy(av), &
1444	(/ time_since_reference_point /), &
1445	start = (/ do2d_xy_time_count(av) /), &
1446	count = (/ 1 /) )
1447	CALL netcdf_handle_error( 'data_output_2d', 53 )
1448	#endif
1449	ENDIF
1450	ENDIF
1451	ENDIF
1452	!
1453	!-- If required, carry out averaging along z
1454	IF ( section(is,s_ind) == -1 .AND. .NOT. two_d ) THEN
1455
1456	local_2d = 0.0_wp
1457	!
1458	!-- Carry out the averaging (all data are on the PE)
1459	DO k = nzb_do, nzt_do
1460	DO j = nys, nyn
1461	DO i = nxl, nxr
1462	local_2d(i,j) = local_2d(i,j) + local_pf(i,j,k)
1463	ENDDO
1464	ENDDO
1465	ENDDO
1466
1467	local_2d = local_2d / ( nzt_do - nzb_do + 1.0_wp)
1468
1469	ELSE
1470	!
1471	!-- Just store the respective section on the local array
1472	local_2d = local_pf(:,:,layer_xy)
1473
1474	ENDIF
1475
1476	#if defined( __parallel )
1477	IF ( netcdf_data_format > 4 ) THEN
1478	!
1479	!-- Parallel output in netCDF4/HDF5 format.
1480	IF ( two_d ) THEN
1481	iis = 1
1482	ELSE
1483	iis = is
1484	ENDIF
1485
1486	#if defined( __netcdf )
1487	!
1488	!-- For parallel output, all cross sections are first stored
1489	!-- here on a local array and will be written to the output
1490	!-- file afterwards to increase the performance.
1491	DO i = nxl, nxr
1492	DO j = nys, nyn
1493	local_2d_sections(i,j,iis) = local_2d(i,j)
1494	ENDDO
1495	ENDDO
1496	#endif
1497	ELSE
1498
1499	IF ( data_output_2d_on_each_pe ) THEN
1500	!
1501	!-- Output of partial arrays on each PE
1502	#if defined( __netcdf )
1503	IF ( myid == 0 ) THEN
1504	WRITE ( 21 ) time_since_reference_point, &
1505	do2d_xy_time_count(av), av
1506	ENDIF
1507	#endif
1508	DO i = 0, io_blocks-1
1509	IF ( i == io_group ) THEN
1510	WRITE ( 21 ) nxl, nxr, nys, nyn, nys, nyn
1511	WRITE ( 21 ) local_2d
1512	ENDIF
1513	#if defined( __parallel )
1514	CALL MPI_BARRIER( comm2d, ierr )
1515	#endif
1516	ENDDO
1517
1518	ELSE
1519	!
1520	!-- PE0 receives partial arrays from all processors and
1521	!-- then outputs them. Here a barrier has to be set,
1522	!-- because otherwise "-MPI- FATAL: Remote protocol queue
1523	!-- full" may occur.
1524	CALL MPI_BARRIER( comm2d, ierr )
1525
1526	ngp = ( nxr-nxl+1 ) * ( nyn-nys+1 )
1527	IF ( myid == 0 ) THEN
1528	!
1529	!-- Local array can be relocated directly.
1530	total_2d(nxl:nxr,nys:nyn) = local_2d
1531	!
1532	!-- Receive data from all other PEs.
1533	DO n = 1, numprocs-1
1534	!
1535	!-- Receive index limits first, then array.
1536	!-- Index limits are received in arbitrary order from
1537	!-- the PEs.
1538	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1539	MPI_ANY_SOURCE, 0, comm2d, &
1540	status, ierr )
1541	sender = status(MPI_SOURCE)
1542	DEALLOCATE( local_2d )
1543	ALLOCATE( local_2d(ind(1):ind(2),ind(3):ind(4)) )
1544	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1545	MPI_REAL, sender, 1, comm2d, &
1546	status, ierr )
1547	total_2d(ind(1):ind(2),ind(3):ind(4)) = local_2d
1548	ENDDO
1549	!
1550	!-- Relocate the local array for the next loop increment
1551	DEALLOCATE( local_2d )
1552	ALLOCATE( local_2d(nxl:nxr,nys:nyn) )
1553
1554	#if defined( __netcdf )
1555	IF ( two_d ) THEN
1556	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1557	id_var_do2d(av,ivar), &
1558	total_2d(0:nx,0:ny), &
1559	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
1560	count = (/ nx+1, ny+1, 1, 1 /) )
1561	ELSE
1562	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1563	id_var_do2d(av,ivar), &
1564	total_2d(0:nx,0:ny), &
1565	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
1566	count = (/ nx+1, ny+1, 1, 1 /) )
1567	ENDIF
1568	CALL netcdf_handle_error( 'data_output_2d', 54 )
1569	#endif
1570
1571	ELSE
1572	!
1573	!-- First send the local index limits to PE0
1574	ind(1) = nxl; ind(2) = nxr
1575	ind(3) = nys; ind(4) = nyn
1576	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1577	comm2d, ierr )
1578	!
1579	!-- Send data to PE0
1580	CALL MPI_SEND( local_2d(nxl,nys), ngp, &
1581	MPI_REAL, 0, 1, comm2d, ierr )
1582	ENDIF
1583	!
1584	!-- A barrier has to be set, because otherwise some PEs may
1585	!-- proceed too fast so that PE0 may receive wrong data on
1586	!-- tag 0
1587	CALL MPI_BARRIER( comm2d, ierr )
1588	ENDIF
1589
1590	ENDIF
1591	#else
1592	#if defined( __netcdf )
1593	IF ( two_d ) THEN
1594	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1595	id_var_do2d(av,ivar), &
1596	local_2d(nxl:nxr,nys:nyn), &
1597	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
1598	count = (/ nx+1, ny+1, 1, 1 /) )
1599	ELSE
1600	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
1601	id_var_do2d(av,ivar), &
1602	local_2d(nxl:nxr,nys:nyn), &
1603	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
1604	count = (/ nx+1, ny+1, 1, 1 /) )
1605	ENDIF
1606	CALL netcdf_handle_error( 'data_output_2d', 447 )
1607	#endif
1608	#endif
1609
1610	!
1611	!-- For 2D-arrays (e.g. u*) only one cross-section is available.
1612	!-- Hence exit loop of output levels.
1613	IF ( two_d ) THEN
1614	IF ( netcdf_data_format < 5 ) two_d = .FALSE.
1615	EXIT loop1
1616	ENDIF
1617
1618	CASE ( 'xz' )
1619	!
1620	!-- Update the netCDF xz cross section time axis.
1621	!-- In case of parallel output, this is only done by PE0
1622	!-- to increase the performance.
1623	IF ( simulated_time /= do2d_xz_last_time(av) ) THEN
1624	do2d_xz_time_count(av) = do2d_xz_time_count(av) + 1
1625	do2d_xz_last_time(av) = simulated_time
1626	IF ( myid == 0 ) THEN
1627	IF ( .NOT. data_output_2d_on_each_pe &
1628	.OR. netcdf_data_format > 4 ) &
1629	THEN
1630	#if defined( __netcdf )
1631	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1632	id_var_time_xz(av), &
1633	(/ time_since_reference_point /), &
1634	start = (/ do2d_xz_time_count(av) /), &
1635	count = (/ 1 /) )
1636	CALL netcdf_handle_error( 'data_output_2d', 56 )
1637	#endif
1638	ENDIF
1639	ENDIF
1640	ENDIF
1641
1642	!
1643	!-- If required, carry out averaging along y
1644	IF ( section(is,s_ind) == -1 ) THEN
1645
1646	ALLOCATE( local_2d_l(nxl:nxr,nzb_do:nzt_do) )
1647	local_2d_l = 0.0_wp
1648	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
1649	!
1650	!-- First local averaging on the PE
1651	DO k = nzb_do, nzt_do
1652	DO j = nys, nyn
1653	DO i = nxl, nxr
1654	local_2d_l(i,k) = local_2d_l(i,k) + &
1655	local_pf(i,j,k)
1656	ENDDO
1657	ENDDO
1658	ENDDO
1659	#if defined( __parallel )
1660	!
1661	!-- Now do the averaging over all PEs along y
1662	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1663	CALL MPI_ALLREDUCE( local_2d_l(nxl,nzb_do), &
1664	local_2d(nxl,nzb_do), ngp, MPI_REAL, &
1665	MPI_SUM, comm1dy, ierr )
1666	#else
1667	local_2d = local_2d_l
1668	#endif
1669	local_2d = local_2d / ( ny + 1.0_wp )
1670
1671	DEALLOCATE( local_2d_l )
1672
1673	ELSE
1674	!
1675	!-- Just store the respective section on the local array
1676	!-- (but only if it is available on this PE!)
1677	IF ( section(is,s_ind) >= nys .AND. section(is,s_ind) <= nyn ) &
1678	THEN
1679	local_2d = local_pf(:,section(is,s_ind),nzb_do:nzt_do)
1680	ENDIF
1681
1682	ENDIF
1683
1684	#if defined( __parallel )
1685	IF ( netcdf_data_format > 4 ) THEN
1686	!
1687	!-- Output in netCDF4/HDF5 format.
1688	!-- Output only on those PEs where the respective cross
1689	!-- sections reside. Cross sections averaged along y are
1690	!-- output on the respective first PE along y (myidy=0).
1691	IF ( ( section(is,s_ind) >= nys .AND. &
1692	section(is,s_ind) <= nyn ) .OR. &
1693	( section(is,s_ind) == -1 .AND. myidy == 0 ) ) THEN
1694	#if defined( __netcdf )
1695	!
1696	!-- For parallel output, all cross sections are first
1697	!-- stored here on a local array and will be written to the
1698	!-- output file afterwards to increase the performance.
1699	DO i = nxl, nxr
1700	DO k = nzb_do, nzt_do
1701	local_2d_sections_l(i,is,k) = local_2d(i,k)
1702	ENDDO
1703	ENDDO
1704	#endif
1705	ENDIF
1706
1707	ELSE
1708
1709	IF ( data_output_2d_on_each_pe ) THEN
1710	!
1711	!-- Output of partial arrays on each PE. If the cross
1712	!-- section does not reside on the PE, output special
1713	!-- index values.
1714	#if defined( __netcdf )
1715	IF ( myid == 0 ) THEN
1716	WRITE ( 22 ) time_since_reference_point, &
1717	do2d_xz_time_count(av), av
1718	ENDIF
1719	#endif
1720	DO i = 0, io_blocks-1
1721	IF ( i == io_group ) THEN
1722	IF ( ( section(is,s_ind) >= nys .AND. &
1723	section(is,s_ind) <= nyn ) .OR. &
1724	( section(is,s_ind) == -1 .AND. &
1725	nys-1 == -1 ) ) &
1726	THEN
1727	WRITE (22) nxl, nxr, nzb_do, nzt_do, nzb, nzt+1
1728	WRITE (22) local_2d
1729	ELSE
1730	WRITE (22) -1, -1, -1, -1, -1, -1
1731	ENDIF
1732	ENDIF
1733	#if defined( __parallel )
1734	CALL MPI_BARRIER( comm2d, ierr )
1735	#endif
1736	ENDDO
1737
1738	ELSE
1739	!
1740	!-- PE0 receives partial arrays from all processors of the
1741	!-- respective cross section and outputs them. Here a
1742	!-- barrier has to be set, because otherwise
1743	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
1744	CALL MPI_BARRIER( comm2d, ierr )
1745
1746	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
1747	IF ( myid == 0 ) THEN
1748	!
1749	!-- Local array can be relocated directly.
1750	IF ( ( section(is,s_ind) >= nys .AND. &
1751	section(is,s_ind) <= nyn ) .OR. &
1752	( section(is,s_ind) == -1 .AND. &
1753	nys-1 == -1 ) ) THEN
1754	total_2d(nxl:nxr,nzb_do:nzt_do) = local_2d
1755	ENDIF
1756	!
1757	!-- Receive data from all other PEs.
1758	DO n = 1, numprocs-1
1759	!
1760	!-- Receive index limits first, then array.
1761	!-- Index limits are received in arbitrary order from
1762	!-- the PEs.
1763	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1764	MPI_ANY_SOURCE, 0, comm2d, &
1765	status, ierr )
1766	!
1767	!-- Not all PEs have data for XZ-cross-section.
1768	IF ( ind(1) /= -9999 ) THEN
1769	sender = status(MPI_SOURCE)
1770	DEALLOCATE( local_2d )
1771	ALLOCATE( local_2d(ind(1):ind(2), &
1772	ind(3):ind(4)) )
1773	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1774	MPI_REAL, sender, 1, comm2d, &
1775	status, ierr )
1776	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
1777	local_2d
1778	ENDIF
1779	ENDDO
1780	!
1781	!-- Relocate the local array for the next loop increment
1782	DEALLOCATE( local_2d )
1783	ALLOCATE( local_2d(nxl:nxr,nzb_do:nzt_do) )
1784
1785	#if defined( __netcdf )
1786	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1787	id_var_do2d(av,ivar), &
1788	total_2d(0:nx,nzb_do:nzt_do), &
1789	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
1790	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
1791	CALL netcdf_handle_error( 'data_output_2d', 58 )
1792	#endif
1793
1794	ELSE
1795	!
1796	!-- If the cross section resides on the PE, send the
1797	!-- local index limits, otherwise send -9999 to PE0.
1798	IF ( ( section(is,s_ind) >= nys .AND. &
1799	section(is,s_ind) <= nyn ) .OR. &
1800	( section(is,s_ind) == -1 .AND. nys-1 == -1 ) ) &
1801	THEN
1802	ind(1) = nxl; ind(2) = nxr
1803	ind(3) = nzb_do; ind(4) = nzt_do
1804	ELSE
1805	ind(1) = -9999; ind(2) = -9999
1806	ind(3) = -9999; ind(4) = -9999
1807	ENDIF
1808	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
1809	comm2d, ierr )
1810	!
1811	!-- If applicable, send data to PE0.
1812	IF ( ind(1) /= -9999 ) THEN
1813	CALL MPI_SEND( local_2d(nxl,nzb_do), ngp, &
1814	MPI_REAL, 0, 1, comm2d, ierr )
1815	ENDIF
1816	ENDIF
1817	!
1818	!-- A barrier has to be set, because otherwise some PEs may
1819	!-- proceed too fast so that PE0 may receive wrong data on
1820	!-- tag 0
1821	CALL MPI_BARRIER( comm2d, ierr )
1822	ENDIF
1823
1824	ENDIF
1825	#else
1826	#if defined( __netcdf )
1827	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
1828	id_var_do2d(av,ivar), &
1829	local_2d(nxl:nxr,nzb_do:nzt_do), &
1830	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
1831	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
1832	CALL netcdf_handle_error( 'data_output_2d', 451 )
1833	#endif
1834	#endif
1835
1836	CASE ( 'yz' )
1837	!
1838	!-- Update the netCDF yz cross section time axis.
1839	!-- In case of parallel output, this is only done by PE0
1840	!-- to increase the performance.
1841	IF ( simulated_time /= do2d_yz_last_time(av) ) THEN
1842	do2d_yz_time_count(av) = do2d_yz_time_count(av) + 1
1843	do2d_yz_last_time(av) = simulated_time
1844	IF ( myid == 0 ) THEN
1845	IF ( .NOT. data_output_2d_on_each_pe &
1846	.OR. netcdf_data_format > 4 ) &
1847	THEN
1848	#if defined( __netcdf )
1849	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
1850	id_var_time_yz(av), &
1851	(/ time_since_reference_point /), &
1852	start = (/ do2d_yz_time_count(av) /), &
1853	count = (/ 1 /) )
1854	CALL netcdf_handle_error( 'data_output_2d', 59 )
1855	#endif
1856	ENDIF
1857	ENDIF
1858	ENDIF
1859
1860	!
1861	!-- If required, carry out averaging along x
1862	IF ( section(is,s_ind) == -1 ) THEN
1863
1864	ALLOCATE( local_2d_l(nys:nyn,nzb_do:nzt_do) )
1865	local_2d_l = 0.0_wp
1866	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
1867	!
1868	!-- First local averaging on the PE
1869	DO k = nzb_do, nzt_do
1870	DO j = nys, nyn
1871	DO i = nxl, nxr
1872	local_2d_l(j,k) = local_2d_l(j,k) + &
1873	local_pf(i,j,k)
1874	ENDDO
1875	ENDDO
1876	ENDDO
1877	#if defined( __parallel )
1878	!
1879	!-- Now do the averaging over all PEs along x
1880	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
1881	CALL MPI_ALLREDUCE( local_2d_l(nys,nzb_do), &
1882	local_2d(nys,nzb_do), ngp, MPI_REAL, &
1883	MPI_SUM, comm1dx, ierr )
1884	#else
1885	local_2d = local_2d_l
1886	#endif
1887	local_2d = local_2d / ( nx + 1.0_wp )
1888
1889	DEALLOCATE( local_2d_l )
1890
1891	ELSE
1892	!
1893	!-- Just store the respective section on the local array
1894	!-- (but only if it is available on this PE!)
1895	IF ( section(is,s_ind) >= nxl .AND. section(is,s_ind) <= nxr ) &
1896	THEN
1897	local_2d = local_pf(section(is,s_ind),:,nzb_do:nzt_do)
1898	ENDIF
1899
1900	ENDIF
1901
1902	#if defined( __parallel )
1903	IF ( netcdf_data_format > 4 ) THEN
1904	!
1905	!-- Output in netCDF4/HDF5 format.
1906	!-- Output only on those PEs where the respective cross
1907	!-- sections reside. Cross sections averaged along x are
1908	!-- output on the respective first PE along x (myidx=0).
1909	IF ( ( section(is,s_ind) >= nxl .AND. &
1910	section(is,s_ind) <= nxr ) .OR. &
1911	( section(is,s_ind) == -1 .AND. myidx == 0 ) ) THEN
1912	#if defined( __netcdf )
1913	!
1914	!-- For parallel output, all cross sections are first
1915	!-- stored here on a local array and will be written to the
1916	!-- output file afterwards to increase the performance.
1917	DO j = nys, nyn
1918	DO k = nzb_do, nzt_do
1919	local_2d_sections_l(is,j,k) = local_2d(j,k)
1920	ENDDO
1921	ENDDO
1922	#endif
1923	ENDIF
1924
1925	ELSE
1926
1927	IF ( data_output_2d_on_each_pe ) THEN
1928	!
1929	!-- Output of partial arrays on each PE. If the cross
1930	!-- section does not reside on the PE, output special
1931	!-- index values.
1932	#if defined( __netcdf )
1933	IF ( myid == 0 ) THEN
1934	WRITE ( 23 ) time_since_reference_point, &
1935	do2d_yz_time_count(av), av
1936	ENDIF
1937	#endif
1938	DO i = 0, io_blocks-1
1939	IF ( i == io_group ) THEN
1940	IF ( ( section(is,s_ind) >= nxl .AND. &
1941	section(is,s_ind) <= nxr ) .OR. &
1942	( section(is,s_ind) == -1 .AND. &
1943	nxl-1 == -1 ) ) &
1944	THEN
1945	WRITE (23) nys, nyn, nzb_do, nzt_do, nzb, nzt+1
1946	WRITE (23) local_2d
1947	ELSE
1948	WRITE (23) -1, -1, -1, -1, -1, -1
1949	ENDIF
1950	ENDIF
1951	#if defined( __parallel )
1952	CALL MPI_BARRIER( comm2d, ierr )
1953	#endif
1954	ENDDO
1955
1956	ELSE
1957	!
1958	!-- PE0 receives partial arrays from all processors of the
1959	!-- respective cross section and outputs them. Here a
1960	!-- barrier has to be set, because otherwise
1961	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
1962	CALL MPI_BARRIER( comm2d, ierr )
1963
1964	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
1965	IF ( myid == 0 ) THEN
1966	!
1967	!-- Local array can be relocated directly.
1968	IF ( ( section(is,s_ind) >= nxl .AND. &
1969	section(is,s_ind) <= nxr ) .OR. &
1970	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
1971	THEN
1972	total_2d(nys:nyn,nzb_do:nzt_do) = local_2d
1973	ENDIF
1974	!
1975	!-- Receive data from all other PEs.
1976	DO n = 1, numprocs-1
1977	!
1978	!-- Receive index limits first, then array.
1979	!-- Index limits are received in arbitrary order from
1980	!-- the PEs.
1981	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
1982	MPI_ANY_SOURCE, 0, comm2d, &
1983	status, ierr )
1984	!
1985	!-- Not all PEs have data for YZ-cross-section.
1986	IF ( ind(1) /= -9999 ) THEN
1987	sender = status(MPI_SOURCE)
1988	DEALLOCATE( local_2d )
1989	ALLOCATE( local_2d(ind(1):ind(2), &
1990	ind(3):ind(4)) )
1991	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
1992	MPI_REAL, sender, 1, comm2d, &
1993	status, ierr )
1994	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
1995	local_2d
1996	ENDIF
1997	ENDDO
1998	!
1999	!-- Relocate the local array for the next loop increment
2000	DEALLOCATE( local_2d )
2001	ALLOCATE( local_2d(nys:nyn,nzb_do:nzt_do) )
2002
2003	#if defined( __netcdf )
2004	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2005	id_var_do2d(av,ivar), &
2006	total_2d(0:ny,nzb_do:nzt_do), &
2007	start = (/ is, 1, 1, do2d_yz_time_count(av) /), &
2008	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
2009	CALL netcdf_handle_error( 'data_output_2d', 61 )
2010	#endif
2011
2012	ELSE
2013	!
2014	!-- If the cross section resides on the PE, send the
2015	!-- local index limits, otherwise send -9999 to PE0.
2016	IF ( ( section(is,s_ind) >= nxl .AND. &
2017	section(is,s_ind) <= nxr ) .OR. &
2018	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
2019	THEN
2020	ind(1) = nys; ind(2) = nyn
2021	ind(3) = nzb_do; ind(4) = nzt_do
2022	ELSE
2023	ind(1) = -9999; ind(2) = -9999
2024	ind(3) = -9999; ind(4) = -9999
2025	ENDIF
2026	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
2027	comm2d, ierr )
2028	!
2029	!-- If applicable, send data to PE0.
2030	IF ( ind(1) /= -9999 ) THEN
2031	CALL MPI_SEND( local_2d(nys,nzb_do), ngp, &
2032	MPI_REAL, 0, 1, comm2d, ierr )
2033	ENDIF
2034	ENDIF
2035	!
2036	!-- A barrier has to be set, because otherwise some PEs may
2037	!-- proceed too fast so that PE0 may receive wrong data on
2038	!-- tag 0
2039	CALL MPI_BARRIER( comm2d, ierr )
2040	ENDIF
2041
2042	ENDIF
2043	#else
2044	#if defined( __netcdf )
2045	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2046	id_var_do2d(av,ivar), &
2047	local_2d(nys:nyn,nzb_do:nzt_do), &
2048	start = (/ is, 1, 1, do2d_xz_time_count(av) /), &
2049	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
2050	CALL netcdf_handle_error( 'data_output_2d', 452 )
2051	#endif
2052	#endif
2053
2054	END SELECT
2055
2056	is = is + 1
2057	ENDDO loop1
2058
2059	!
2060	!-- For parallel output, all data were collected before on a local array
2061	!-- and are written now to the netcdf file. This must be done to increase
2062	!-- the performance of the parallel output.
2063	#if defined( __netcdf )
2064	IF ( netcdf_data_format > 4 ) THEN
2065
2066	SELECT CASE ( mode )
2067
2068	CASE ( 'xy' )
2069	IF ( two_d ) THEN
2070	nis = 1
2071	two_d = .FALSE.
2072	ELSE
2073	nis = ns
2074	ENDIF
2075	!
2076	!-- Do not output redundant ghost point data except for the
2077	!-- boundaries of the total domain.
2078	! IF ( nxr == nx .AND. nyn /= ny ) THEN
2079	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2080	! id_var_do2d(av,ivar), &
2081	! local_2d_sections(nxl:nxr+1, &
2082	! nys:nyn,1:nis), &
2083	! start = (/ nxl+1, nys+1, 1, &
2084	! do2d_xy_time_count(av) /), &
2085	! count = (/ nxr-nxl+2, &
2086	! nyn-nys+1, nis, 1 &
2087	! /) )
2088	! ELSEIF ( nxr /= nx .AND. nyn == ny ) THEN
2089	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2090	! id_var_do2d(av,ivar), &
2091	! local_2d_sections(nxl:nxr, &
2092	! nys:nyn+1,1:nis), &
2093	! start = (/ nxl+1, nys+1, 1, &
2094	! do2d_xy_time_count(av) /), &
2095	! count = (/ nxr-nxl+1, &
2096	! nyn-nys+2, nis, 1 &
2097	! /) )
2098	! ELSEIF ( nxr == nx .AND. nyn == ny ) THEN
2099	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2100	! id_var_do2d(av,ivar), &
2101	! local_2d_sections(nxl:nxr+1, &
2102	! nys:nyn+1,1:nis), &
2103	! start = (/ nxl+1, nys+1, 1, &
2104	! do2d_xy_time_count(av) /), &
2105	! count = (/ nxr-nxl+2, &
2106	! nyn-nys+2, nis, 1 &
2107	! /) )
2108	! ELSE
2109	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
2110	id_var_do2d(av,ivar), &
2111	local_2d_sections(nxl:nxr, &
2112	nys:nyn,1:nis), &
2113	start = (/ nxl+1, nys+1, 1, &
2114	do2d_xy_time_count(av) /), &
2115	count = (/ nxr-nxl+1, &
2116	nyn-nys+1, nis, 1 &
2117	/) )
2118	! ENDIF
2119
2120	CALL netcdf_handle_error( 'data_output_2d', 55 )
2121
2122	CASE ( 'xz' )
2123	!
2124	!-- First, all PEs get the information of all cross-sections.
2125	!-- Then the data are written to the output file by all PEs
2126	!-- while NF90_COLLECTIVE is set in subroutine
2127	!-- define_netcdf_header. Although redundant information are
2128	!-- written to the output file in that case, the performance
2129	!-- is significantly better compared to the case where only
2130	!-- the first row of PEs in x-direction (myidx = 0) is given
2131	!-- the output while NF90_INDEPENDENT is set.
2132	IF ( npey /= 1 ) THEN
2133
2134	#if defined( __parallel )
2135	!
2136	!-- Distribute data over all PEs along y
2137	ngp = ( nxr-nxl+1 ) * ( nzt_do-nzb_do+1 ) * ns
2138	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
2139	CALL MPI_ALLREDUCE( local_2d_sections_l(nxl,1,nzb_do), &
2140	local_2d_sections(nxl,1,nzb_do), &
2141	ngp, MPI_REAL, MPI_SUM, comm1dy, &
2142	ierr )
2143	#else
2144	local_2d_sections = local_2d_sections_l
2145	#endif
2146	ENDIF
2147	!
2148	!-- Do not output redundant ghost point data except for the
2149	!-- boundaries of the total domain.
2150	! IF ( nxr == nx ) THEN
2151	! nc_stat = NF90_PUT_VAR( id_set_xz(av), &
2152	! id_var_do2d(av,ivar), &
2153	! local_2d_sections(nxl:nxr+1,1:ns, &
2154	! nzb_do:nzt_do), &
2155	! start = (/ nxl+1, 1, 1, &
2156	! do2d_xz_time_count(av) /), &
2157	! count = (/ nxr-nxl+2, ns, nzt_do-nzb_do+1, &
2158	! 1 /) )
2159	! ELSE
2160	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
2161	id_var_do2d(av,ivar), &
2162	local_2d_sections(nxl:nxr,1:ns, &
2163	nzb_do:nzt_do), &
2164	start = (/ nxl+1, 1, 1, &
2165	do2d_xz_time_count(av) /), &
2166	count = (/ nxr-nxl+1, ns, nzt_do-nzb_do+1, &
2167	1 /) )
2168	! ENDIF
2169
2170	CALL netcdf_handle_error( 'data_output_2d', 57 )
2171
2172	CASE ( 'yz' )
2173	!
2174	!-- First, all PEs get the information of all cross-sections.
2175	!-- Then the data are written to the output file by all PEs
2176	!-- while NF90_COLLECTIVE is set in subroutine
2177	!-- define_netcdf_header. Although redundant information are
2178	!-- written to the output file in that case, the performance
2179	!-- is significantly better compared to the case where only
2180	!-- the first row of PEs in y-direction (myidy = 0) is given
2181	!-- the output while NF90_INDEPENDENT is set.
2182	IF ( npex /= 1 ) THEN
2183
2184	#if defined( __parallel )
2185	!
2186	!-- Distribute data over all PEs along x
2187	ngp = ( nyn-nys+1 ) * ( nzt-nzb + 2 ) * ns
2188	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
2189	CALL MPI_ALLREDUCE( local_2d_sections_l(1,nys,nzb_do), &
2190	local_2d_sections(1,nys,nzb_do), &
2191	ngp, MPI_REAL, MPI_SUM, comm1dx, &
2192	ierr )
2193	#else
2194	local_2d_sections = local_2d_sections_l
2195	#endif
2196	ENDIF
2197	!
2198	!-- Do not output redundant ghost point data except for the
2199	!-- boundaries of the total domain.
2200	! IF ( nyn == ny ) THEN
2201	! nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2202	! id_var_do2d(av,ivar), &
2203	! local_2d_sections(1:ns, &
2204	! nys:nyn+1,nzb_do:nzt_do), &
2205	! start = (/ 1, nys+1, 1, &
2206	! do2d_yz_time_count(av) /), &
2207	! count = (/ ns, nyn-nys+2, &
2208	! nzt_do-nzb_do+1, 1 /) )
2209	! ELSE
2210	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
2211	id_var_do2d(av,ivar), &
2212	local_2d_sections(1:ns,nys:nyn, &
2213	nzb_do:nzt_do), &
2214	start = (/ 1, nys+1, 1, &
2215	do2d_yz_time_count(av) /), &
2216	count = (/ ns, nyn-nys+1, &
2217	nzt_do-nzb_do+1, 1 /) )
2218	! ENDIF
2219
2220	CALL netcdf_handle_error( 'data_output_2d', 60 )
2221
2222	CASE DEFAULT
2223	message_string = 'unknown cross-section: ' // TRIM( mode )
2224	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
2225
2226	END SELECT
2227
2228	ENDIF
2229	#endif
2230	ENDIF
2231
2232	ivar = ivar + 1
2233	l = MAX( 2, LEN_TRIM( do2d(av,ivar) ) )
2234	do2d_mode = do2d(av,ivar)(l-1:l)
2235
2236	ENDDO
2237
2238	!
2239	!-- Deallocate temporary arrays.
2240	IF ( ALLOCATED( level_z ) ) DEALLOCATE( level_z )
2241	IF ( netcdf_data_format > 4 ) THEN
2242	DEALLOCATE( local_pf, local_2d, local_2d_sections )
2243	IF( mode == 'xz' .OR. mode == 'yz' ) DEALLOCATE( local_2d_sections_l )
2244	ENDIF
2245	#if defined( __parallel )
2246	IF ( .NOT. data_output_2d_on_each_pe .AND. myid == 0 ) THEN
2247	DEALLOCATE( total_2d )
2248	ENDIF
2249	#endif
2250
2251	!
2252	!-- Close plot output file.
2253	file_id = 20 + s_ind
2254
2255	IF ( data_output_2d_on_each_pe ) THEN
2256	DO i = 0, io_blocks-1
2257	IF ( i == io_group ) THEN
2258	CALL close_file( file_id )
2259	ENDIF
2260	#if defined( __parallel )
2261	CALL MPI_BARRIER( comm2d, ierr )
2262	#endif
2263	ENDDO
2264	ELSE
2265	IF ( myid == 0 ) CALL close_file( file_id )
2266	ENDIF
2267
2268	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
2269
2270	END SUBROUTINE data_output_2d

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |