Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 3117

Last change on this file since 3117 was 3052, checked in by raasch, 6 years ago
Do not open FORTRAN binary files in case of parallel netCDF I/O
Property svn:keywords set to `Id`
File size: 98.1 KB

Rev	Line
[1682]	1	!> @file data_output_2d.f90
[2000]	2	!------------------------------------------------------------------------------!
[2696]	3	! This file is part of the PALM model system.
[1036]	4	!
[2000]	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.
[1036]	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	!
[2718]	17	! Copyright 1997-2018 Leibniz Universitaet Hannover
[2000]	18	!------------------------------------------------------------------------------!
[1036]	19	!
[254]	20	! Current revisions:
[1]	21	! -----------------
[1961]	22	!
[3049]	23	!
[1552]	24	! Former revisions:
	25	! -----------------
	26	! $Id: data_output_2d.f90 3052 2018-05-31 06:11:20Z maronga $
[3052]	27	! Do not open FORTRAN binary files in case of parallel netCDF I/O
	28	!
	29	! 3049 2018-05-29 13:52:36Z Giersch
[3045]	30	! Error messages revised
	31	!
[3049]	32	! 3045 2018-05-28 07:55:41Z Giersch
	33	! Error messages revised
	34	!
[3045]	35	! 3014 2018-05-09 08:42:38Z maronga
[3014]	36	! Added nzb_do and nzt_do for some modules for 2d output
	37	!
	38	! 3004 2018-04-27 12:33:25Z Giersch
[3004]	39	! precipitation_rate removed, case prr*_xy removed, to_be_resorted have to point
	40	! to ql_vp_av and not to ql_vp, allocation checks implemented (averaged data
	41	! will be assigned to fill values if no allocation happened so far)
	42	!
	43	! 2963 2018-04-12 14:47:44Z suehring
[2963]	44	! Introduce index for vegetation/wall, pavement/green-wall and water/window
	45	! surfaces, for clearer access of surface fraction, albedo, emissivity, etc. .
	46	!
	47	! 2817 2018-02-19 16:32:21Z knoop
[2817]	48	! Preliminary gust module interface implemented
	49	!
	50	! 2805 2018-02-14 17:00:09Z suehring
[2798]	51	! Consider also default-type surfaces for surface temperature output.
	52	!
	53	! 2797 2018-02-08 13:24:35Z suehring
[2797]	54	! Enable output of ground-heat flux also at urban surfaces.
	55	!
	56	! 2743 2018-01-12 16:03:39Z suehring
[2743]	57	! In case of natural- and urban-type surfaces output surfaces fluxes in W/m2.
	58	!
	59	! 2742 2018-01-12 14:59:47Z suehring
[2742]	60	! Enable output of surface temperature
	61	!
	62	! 2735 2018-01-11 12:01:27Z suehring
[2735]	63	! output of r_a moved from land-surface to consider also urban-type surfaces
	64	!
	65	! 2718 2018-01-02 08:49:38Z maronga
[2716]	66	! Corrected "Former revisions" section
	67	!
	68	! 2696 2017-12-14 17:12:51Z kanani
	69	! Change in file header (GPL part)
[2696]	70	! Implementation of uv exposure model (FK)
	71	! Implementation of turbulence_closure_mod (TG)
	72	! Set fill values at topography grid points or e.g. non-natural-type surface
	73	! in case of LSM output (MS)
	74	!
	75	! 2512 2017-10-04 08:26:59Z raasch
[2512]	76	! upper bounds of cross section output changed from nx+1,ny+1 to nx,ny
	77	! no output of ghost layer data
	78	!
	79	! 2292 2017-06-20 09:51:42Z schwenkel
[2292]	80	! Implementation of new microphysic scheme: cloud_scheme = 'morrison'
	81	! includes two more prognostic equations for cloud drop concentration (nc)
	82	! and cloud water content (qc).
	83	!
	84	! 2277 2017-06-12 10:47:51Z kanani
[2277]	85	! Removed unused variables do2d_xy_n, do2d_xz_n, do2d_yz_n
	86	!
	87	! 2233 2017-05-30 18:08:54Z suehring
[1552]	88	!
[2233]	89	! 2232 2017-05-30 17:47:52Z suehring
	90	! Adjustments to new surface concept
	91	!
	92	!
[2191]	93	! 2190 2017-03-21 12:16:43Z raasch
	94	! bugfix for r2031: string rho replaced by rho_ocean
	95	!
[2032]	96	! 2031 2016-10-21 15:11:58Z knoop
	97	! renamed variable rho to rho_ocean and rho_av to rho_ocean_av
	98	!
[2001]	99	! 2000 2016-08-20 18:09:15Z knoop
	100	! Forced header and separation lines into 80 columns
	101	!
[1981]	102	! 1980 2016-07-29 15:51:57Z suehring
	103	! Bugfix, in order to steer user-defined output, setting flag found explicitly
	104	! to .F.
	105	!
[1977]	106	! 1976 2016-07-27 13:28:04Z maronga
	107	! Output of radiation quantities is now done directly in the respective module
	108	!
[1973]	109	! 1972 2016-07-26 07:52:02Z maronga
[1976]	110	! Output of land surface quantities is now done directly in the respective
	111	! module
[1973]	112	!
[1961]	113	! 1960 2016-07-12 16:34:24Z suehring
	114	! Scalar surface flux added
	115	! Rename INTEGER variable s into s_ind, as s is already assigned to scalar
	116	!
[1851]	117	! 1849 2016-04-08 11:33:18Z hoffmann
	118	! precipitation_amount, precipitation_rate, prr moved to arrays_3d
	119	!
[1823]	120	! 1822 2016-04-07 07:49:42Z hoffmann
	121	! Output of bulk cloud physics simplified.
	122	!
[1789]	123	! 1788 2016-03-10 11:01:04Z maronga
	124	! Added output of z0q
	125	!
[1784]	126	! 1783 2016-03-06 18:36:17Z raasch
	127	! name change of netcdf routines and module + related changes
	128	!
[1746]	129	! 1745 2016-02-05 13:06:51Z gronemeier
	130	! Bugfix: test if time axis limit exceeds moved to point after call of check_open
	131	!
[1704]	132	! 1703 2015-11-02 12:38:44Z raasch
	133	! bugfix for output of single (*) xy-sections in case of parallel netcdf I/O
	134	!
[1702]	135	! 1701 2015-11-02 07:43:04Z maronga
	136	! Bugfix in output of RRTGM data
	137	!
[1692]	138	! 1691 2015-10-26 16:17:44Z maronga
	139	! Added output of Obukhov length (ol) and radiative heating rates for RRTMG.
	140	! Formatting corrections.
	141	!
[1683]	142	! 1682 2015-10-07 23:56:08Z knoop
	143	! Code annotations made doxygen readable
	144	!
[1586]	145	! 1585 2015-04-30 07:05:52Z maronga
	146	! Added support for RRTMG
	147	!
[1556]	148	! 1555 2015-03-04 17:44:27Z maronga
	149	! Added output of r_a and r_s
	150	!
[1552]	151	! 1551 2015-03-03 14:18:16Z maronga
[1551]	152	! Added suppport for land surface model and radiation model output. In the course
	153	! of this action, the limits for vertical loops have been changed (from nzb and
	154	! nzt+1 to nzb_do and nzt_do, respectively in order to allow soil model output).
	155	! Moreover, a new vertical grid zs was introduced.
[1329]	156	!
[1360]	157	! 1359 2014-04-11 17:15:14Z hoffmann
	158	! New particle structure integrated.
	159	!
[1354]	160	! 1353 2014-04-08 15:21:23Z heinze
	161	! REAL constants provided with KIND-attribute
	162	!
[1329]	163	! 1327 2014-03-21 11:00:16Z raasch
	164	! parts concerning iso2d output removed,
	165	! -netcdf output queries
	166	!
[1321]	167	! 1320 2014-03-20 08:40:49Z raasch
[1320]	168	! ONLY-attribute added to USE-statements,
	169	! kind-parameters added to all INTEGER and REAL declaration statements,
	170	! kinds are defined in new module kinds,
	171	! revision history before 2012 removed,
	172	! comment fields (!:) to be used for variable explanations added to
	173	! all variable declaration statements
[1309]	174	!
[1319]	175	! 1318 2014-03-17 13:35:16Z raasch
	176	! barrier argument removed from cpu_log.
	177	! module interfaces removed
	178	!
[1312]	179	! 1311 2014-03-14 12:13:39Z heinze
	180	! bugfix: close #if defined( __netcdf )
	181	!
[1309]	182	! 1308 2014-03-13 14:58:42Z fricke
[1308]	183	! +local_2d_sections, local_2d_sections_l, ns
	184	! Check, if the limit of the time dimension is exceeded for parallel output
	185	! To increase the performance for parallel output, the following is done:
	186	! - Update of time axis is only done by PE0
	187	! - Cross sections are first stored on a local array and are written
	188	! collectively to the output file by all PEs.
[674]	189	!
[1116]	190	! 1115 2013-03-26 18:16:16Z hoffmann
	191	! ql is calculated by calc_liquid_water_content
	192	!
[1077]	193	! 1076 2012-12-05 08:30:18Z hoffmann
	194	! Bugfix in output of ql
	195	!
[1066]	196	! 1065 2012-11-22 17:42:36Z hoffmann
	197	! Bugfix: Output of cross sections of ql
	198	!
[1054]	199	! 1053 2012-11-13 17:11:03Z hoffmann
	200	! +qr, nr, qc and cross sections
	201	!
[1037]	202	! 1036 2012-10-22 13:43:42Z raasch
	203	! code put under GPL (PALM 3.9)
	204	!
[1035]	205	! 1031 2012-10-19 14:35:30Z raasch
	206	! netCDF4 without parallel file support implemented
	207	!
[1008]	208	! 1007 2012-09-19 14:30:36Z franke
	209	! Bugfix: missing calculation of ql_vp added
	210	!
[979]	211	! 978 2012-08-09 08:28:32Z fricke
	212	! +z0h
	213	!
[1]	214	! Revision 1.1 1997/08/11 06:24:09 raasch
	215	! Initial revision
	216	!
	217	!
	218	! Description:
	219	! ------------
[2512]	220	!> Data output of cross-sections in netCDF format or binary format
	221	!> to be later converted to NetCDF by helper routine combine_plot_fields.
[1682]	222	!> Attention: The position of the sectional planes is still not always computed
	223	!> --------- correctly. (zu is used always)!
[1]	224	!------------------------------------------------------------------------------!
[1682]	225	SUBROUTINE data_output_2d( mode, av )
	226
[1]	227
[1320]	228	USE arrays_3d, &
[2743]	229	ONLY: dzw, e, heatflux_output_conversion, nc, nr, p, pt, &
[3004]	230	precipitation_amount, prr, q, qc, ql, ql_c, ql_v, ql_vp, qr, &
	231	rho_ocean, s, sa, tend, u, v, vpt, w, zu, zw, &
	232	waterflux_output_conversion
[1320]	233
[1]	234	USE averaging
[1320]	235
	236	USE cloud_parameters, &
[2743]	237	ONLY: cp, hyrho, l_d_cp, l_v, pt_d_t
[1320]	238
	239	USE control_parameters, &
	240	ONLY: cloud_physics, data_output_2d_on_each_pe, data_output_xy, &
	241	data_output_xz, data_output_yz, do2d, &
[2277]	242	do2d_xy_last_time, do2d_xy_time_count, &
	243	do2d_xz_last_time, do2d_xz_time_count, &
	244	do2d_yz_last_time, do2d_yz_time_count, &
[2232]	245	ibc_uv_b, io_blocks, io_group, land_surface, message_string, &
[1822]	246	ntdim_2d_xy, ntdim_2d_xz, ntdim_2d_yz, &
	247	psolver, section, simulated_time, simulated_time_chr, &
[2696]	248	time_since_reference_point, uv_exposure
[1320]	249
	250	USE cpulog, &
	251	ONLY: cpu_log, log_point
	252
	253	USE grid_variables, &
	254	ONLY: dx, dy
[2817]	255
	256	USE gust_mod, &
	257	ONLY: gust_data_output_2d, gust_module_enabled
[1320]	258
	259	USE indices, &
[3004]	260	ONLY: nbgp, nx, nxl, nxlg, nxr, nxrg, ny, nyn, nyng, nys, nysg, nz, &
	261	nzb, nzt, wall_flags_0
[1320]	262
	263	USE kinds
[1551]	264
	265	USE land_surface_model_mod, &
[2232]	266	ONLY: lsm_data_output_2d, zs
[1551]	267
[1783]	268	#if defined( __netcdf )
	269	USE NETCDF
	270	#endif
[1320]	271
[1783]	272	USE netcdf_interface, &
[2696]	273	ONLY: fill_value, id_set_xy, id_set_xz, id_set_yz, id_var_do2d, &
	274	id_var_time_xy, id_var_time_xz, id_var_time_yz, nc_stat, &
	275	netcdf_data_format, netcdf_handle_error
[1783]	276
[1320]	277	USE particle_attributes, &
[1359]	278	ONLY: grid_particles, number_of_particles, particle_advection_start, &
	279	particles, prt_count
[1320]	280
[1]	281	USE pegrid
	282
[1551]	283	USE radiation_model_mod, &
[1976]	284	ONLY: radiation, radiation_data_output_2d
[1551]	285
[2232]	286	USE surface_mod, &
[2963]	287	ONLY: ind_pav_green, ind_veg_wall, ind_wat_win, surf_def_h, &
	288	surf_lsm_h, surf_usm_h
[2232]	289
[2696]	290	USE turbulence_closure_mod, &
	291	ONLY: tcm_data_output_2d
	292
	293	USE uv_exposure_model_mod, &
	294	ONLY: uvem_data_output_2d
	295
	296
[1]	297	IMPLICIT NONE
	298
[1682]	299	CHARACTER (LEN=2) :: do2d_mode !<
	300	CHARACTER (LEN=2) :: mode !<
	301	CHARACTER (LEN=4) :: grid !<
	302	CHARACTER (LEN=25) :: section_chr !<
	303	CHARACTER (LEN=50) :: rtext !<
[1320]	304
[1682]	305	INTEGER(iwp) :: av !<
	306	INTEGER(iwp) :: ngp !<
	307	INTEGER(iwp) :: file_id !<
[2696]	308	INTEGER(iwp) :: flag_nr !< number of masking flag
[1682]	309	INTEGER(iwp) :: i !<
	310	INTEGER(iwp) :: if !<
	311	INTEGER(iwp) :: is !<
	312	INTEGER(iwp) :: iis !<
	313	INTEGER(iwp) :: j !<
	314	INTEGER(iwp) :: k !<
	315	INTEGER(iwp) :: l !<
	316	INTEGER(iwp) :: layer_xy !<
[2232]	317	INTEGER(iwp) :: m !<
[1682]	318	INTEGER(iwp) :: n !<
[1703]	319	INTEGER(iwp) :: nis !<
[1682]	320	INTEGER(iwp) :: ns !<
	321	INTEGER(iwp) :: nzb_do !< lower limit of the data field (usually nzb)
	322	INTEGER(iwp) :: nzt_do !< upper limit of the data field (usually nzt+1)
	323	INTEGER(iwp) :: psi !<
[1960]	324	INTEGER(iwp) :: s_ind !<
[1682]	325	INTEGER(iwp) :: sender !<
	326	INTEGER(iwp) :: ind(4) !<
[1320]	327
[1682]	328	LOGICAL :: found !<
	329	LOGICAL :: resorted !<
	330	LOGICAL :: two_d !<
[1320]	331
[1682]	332	REAL(wp) :: mean_r !<
	333	REAL(wp) :: s_r2 !<
	334	REAL(wp) :: s_r3 !<
[1320]	335
[1682]	336	REAL(wp), DIMENSION(:), ALLOCATABLE :: level_z !<
	337	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d !<
	338	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d_l !<
[2232]	339
[1682]	340	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !<
	341	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections !<
	342	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections_l !<
[1359]	343
[1]	344	#if defined( __parallel )
[1682]	345	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: total_2d !<
[1]	346	#endif
[1682]	347	REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !<
[1]	348
	349	NAMELIST /LOCAL/ rtext
	350
	351	!
	352	!-- Immediate return, if no output is requested (no respective sections
	353	!-- found in parameter data_output)
	354	IF ( mode == 'xy' .AND. .NOT. data_output_xy(av) ) RETURN
	355	IF ( mode == 'xz' .AND. .NOT. data_output_xz(av) ) RETURN
	356	IF ( mode == 'yz' .AND. .NOT. data_output_yz(av) ) RETURN
	357
[1308]	358	CALL cpu_log (log_point(3),'data_output_2d','start')
	359
[1]	360	two_d = .FALSE. ! local variable to distinguish between output of pure 2D
	361	! arrays and cross-sections of 3D arrays.
	362
	363	!
	364	!-- Depending on the orientation of the cross-section, the respective output
	365	!-- files have to be opened.
	366	SELECT CASE ( mode )
	367
	368	CASE ( 'xy' )
[1960]	369	s_ind = 1
[2512]	370	ALLOCATE( level_z(nzb:nzt+1), local_2d(nxl:nxr,nys:nyn) )
[1]	371
[1308]	372	IF ( netcdf_data_format > 4 ) THEN
	373	ns = 1
[1960]	374	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	375	ns = ns + 1
	376	ENDDO
	377	ns = ns - 1
[2512]	378	ALLOCATE( local_2d_sections(nxl:nxr,nys:nyn,1:ns) )
[1353]	379	local_2d_sections = 0.0_wp
[1308]	380	ENDIF
	381
[493]	382	!
[1031]	383	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	384	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	385	CALL check_open( 101+av*10 )
	386	ENDIF
[3052]	387	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
[1]	388	CALL check_open( 21 )
	389	ELSE
	390	IF ( myid == 0 ) THEN
	391	#if defined( __parallel )
[2512]	392	ALLOCATE( total_2d(0:nx,0:ny) )
[1]	393	#endif
	394	ENDIF
	395	ENDIF
	396
	397	CASE ( 'xz' )
[1960]	398	s_ind = 2
[2512]	399	ALLOCATE( local_2d(nxl:nxr,nzb:nzt+1) )
[1]	400
[1308]	401	IF ( netcdf_data_format > 4 ) THEN
	402	ns = 1
[1960]	403	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	404	ns = ns + 1
	405	ENDDO
	406	ns = ns - 1
[2512]	407	ALLOCATE( local_2d_sections(nxl:nxr,1:ns,nzb:nzt+1) )
	408	ALLOCATE( local_2d_sections_l(nxl:nxr,1:ns,nzb:nzt+1) )
[1353]	409	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
[1308]	410	ENDIF
	411
[493]	412	!
[1031]	413	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	414	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	415	CALL check_open( 102+av*10 )
	416	ENDIF
[1]	417
[3052]	418	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
[1]	419	CALL check_open( 22 )
	420	ELSE
	421	IF ( myid == 0 ) THEN
	422	#if defined( __parallel )
[2512]	423	ALLOCATE( total_2d(0:nx,nzb:nzt+1) )
[1]	424	#endif
	425	ENDIF
	426	ENDIF
	427
	428	CASE ( 'yz' )
[1960]	429	s_ind = 3
[2512]	430	ALLOCATE( local_2d(nys:nyn,nzb:nzt+1) )
[1]	431
[1308]	432	IF ( netcdf_data_format > 4 ) THEN
	433	ns = 1
[1960]	434	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	435	ns = ns + 1
	436	ENDDO
	437	ns = ns - 1
[2512]	438	ALLOCATE( local_2d_sections(1:ns,nys:nyn,nzb:nzt+1) )
	439	ALLOCATE( local_2d_sections_l(1:ns,nys:nyn,nzb:nzt+1) )
[1353]	440	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
[1308]	441	ENDIF
	442
[493]	443	!
[1031]	444	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	445	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	446	CALL check_open( 103+av*10 )
	447	ENDIF
[1]	448
[3052]	449	IF ( data_output_2d_on_each_pe .AND. netcdf_data_format < 5 ) THEN
[1]	450	CALL check_open( 23 )
	451	ELSE
	452	IF ( myid == 0 ) THEN
	453	#if defined( __parallel )
[2512]	454	ALLOCATE( total_2d(0:ny,nzb:nzt+1) )
[1]	455	#endif
	456	ENDIF
	457	ENDIF
	458
	459	CASE DEFAULT
[254]	460	message_string = 'unknown cross-section: ' // TRIM( mode )
	461	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
[1]	462
	463	END SELECT
	464
	465	!
[1745]	466	!-- For parallel netcdf output the time axis must be limited. Return, if this
	467	!-- limit is exceeded. This could be the case, if the simulated time exceeds
	468	!-- the given end time by the length of the given output interval.
	469	IF ( netcdf_data_format > 4 ) THEN
	470	IF ( mode == 'xy' .AND. do2d_xy_time_count(av) + 1 > &
	471	ntdim_2d_xy(av) ) THEN
	472	WRITE ( message_string, * ) 'Output of xy cross-sections is not ', &
[3046]	473	'given at t=', simulated_time, '&because the', &
[1745]	474	' maximum number of output time levels is exceeded.'
	475	CALL message( 'data_output_2d', 'PA0384', 0, 1, 0, 6, 0 )
	476	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	477	RETURN
	478	ENDIF
	479	IF ( mode == 'xz' .AND. do2d_xz_time_count(av) + 1 > &
	480	ntdim_2d_xz(av) ) THEN
	481	WRITE ( message_string, * ) 'Output of xz cross-sections is not ', &
[3046]	482	'given at t=', simulated_time, '&because the', &
[1745]	483	' maximum number of output time levels is exceeded.'
	484	CALL message( 'data_output_2d', 'PA0385', 0, 1, 0, 6, 0 )
	485	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	486	RETURN
	487	ENDIF
	488	IF ( mode == 'yz' .AND. do2d_yz_time_count(av) + 1 > &
	489	ntdim_2d_yz(av) ) THEN
	490	WRITE ( message_string, * ) 'Output of yz cross-sections is not ', &
[3046]	491	'given at t=', simulated_time, '&because the', &
[1745]	492	' maximum number of output time levels is exceeded.'
	493	CALL message( 'data_output_2d', 'PA0386', 0, 1, 0, 6, 0 )
	494	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	495	RETURN
	496	ENDIF
	497	ENDIF
	498
	499	!
[1]	500	!-- Allocate a temporary array for resorting (kji -> ijk).
[2512]	501	ALLOCATE( local_pf(nxl:nxr,nys:nyn,nzb:nzt+1) )
[2232]	502	local_pf = 0.0
[1]	503
	504	!
	505	!-- Loop of all variables to be written.
	506	!-- Output dimensions chosen
	507	if = 1
	508	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
	509	do2d_mode = do2d(av,if)(l-1:l)
	510
	511	DO WHILE ( do2d(av,if)(1:1) /= ' ' )
	512
	513	IF ( do2d_mode == mode ) THEN
[1980]	514	!
	515	!-- Set flag to steer output of radiation, land-surface, or user-defined
	516	!-- quantities
	517	found = .FALSE.
[1551]	518
	519	nzb_do = nzb
	520	nzt_do = nzt+1
[1]	521	!
[2696]	522	!-- Before each output, set array local_pf to fill value
	523	local_pf = fill_value
	524	!
	525	!-- Set masking flag for topography for not resorted arrays
	526	flag_nr = 0
	527
	528	!
[1]	529	!-- Store the array chosen on the temporary array.
	530	resorted = .FALSE.
	531	SELECT CASE ( TRIM( do2d(av,if) ) )
	532	CASE ( 'e_xy', 'e_xz', 'e_yz' )
	533	IF ( av == 0 ) THEN
	534	to_be_resorted => e
	535	ELSE
[3004]	536	IF ( .NOT. ALLOCATED( e_av ) ) THEN
	537	ALLOCATE( e_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	538	e_av = REAL( fill_value, KIND = wp )
	539	ENDIF
[1]	540	to_be_resorted => e_av
	541	ENDIF
	542	IF ( mode == 'xy' ) level_z = zu
	543
[771]	544	CASE ( 'lpt_xy', 'lpt_xz', 'lpt_yz' )
	545	IF ( av == 0 ) THEN
	546	to_be_resorted => pt
	547	ELSE
[3004]	548	IF ( .NOT. ALLOCATED( lpt_av ) ) THEN
	549	ALLOCATE( lpt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	550	lpt_av = REAL( fill_value, KIND = wp )
	551	ENDIF
[771]	552	to_be_resorted => lpt_av
	553	ENDIF
	554	IF ( mode == 'xy' ) level_z = zu
	555
[1]	556	CASE ( 'lwp*_xy' ) ! 2d-array
	557	IF ( av == 0 ) THEN
[2512]	558	DO i = nxl, nxr
	559	DO j = nys, nyn
[1320]	560	local_pf(i,j,nzb+1) = SUM( ql(nzb:nzt,j,i) * &
[1]	561	dzw(1:nzt+1) )
	562	ENDDO
	563	ENDDO
	564	ELSE
[3004]	565	IF ( .NOT. ALLOCATED( lwp_av ) ) THEN
	566	ALLOCATE( lwp_av(nysg:nyng,nxlg:nxrg) )
	567	lwp_av = REAL( fill_value, KIND = wp )
	568	ENDIF
[2512]	569	DO i = nxl, nxr
	570	DO j = nys, nyn
[1]	571	local_pf(i,j,nzb+1) = lwp_av(j,i)
	572	ENDDO
	573	ENDDO
	574	ENDIF
	575	resorted = .TRUE.
	576	two_d = .TRUE.
	577	level_z(nzb+1) = zu(nzb+1)
	578
[2292]	579	CASE ( 'nc_xy', 'nc_xz', 'nc_yz' )
	580	IF ( av == 0 ) THEN
	581	to_be_resorted => nc
	582	ELSE
[3004]	583	IF ( .NOT. ALLOCATED( nc_av ) ) THEN
	584	ALLOCATE( nc_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	585	nc_av = REAL( fill_value, KIND = wp )
	586	ENDIF
[2292]	587	to_be_resorted => nc_av
	588	ENDIF
	589	IF ( mode == 'xy' ) level_z = zu
	590
[1053]	591	CASE ( 'nr_xy', 'nr_xz', 'nr_yz' )
	592	IF ( av == 0 ) THEN
	593	to_be_resorted => nr
	594	ELSE
[3004]	595	IF ( .NOT. ALLOCATED( nr_av ) ) THEN
	596	ALLOCATE( nr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	597	nr_av = REAL( fill_value, KIND = wp )
	598	ENDIF
[1053]	599	to_be_resorted => nr_av
	600	ENDIF
	601	IF ( mode == 'xy' ) level_z = zu
	602
[2797]	603	CASE ( 'ghf*_xy' ) ! 2d-array
	604	IF ( av == 0 ) THEN
	605	DO m = 1, surf_lsm_h%ns
	606	i = surf_lsm_h%i(m)
	607	j = surf_lsm_h%j(m)
	608	local_pf(i,j,nzb+1) = surf_lsm_h%ghf(m)
	609	ENDDO
	610	DO m = 1, surf_usm_h%ns
	611	i = surf_usm_h%i(m)
	612	j = surf_usm_h%j(m)
[2963]	613	local_pf(i,j,nzb+1) = surf_usm_h%frac(ind_veg_wall,m) * &
[2797]	614	surf_usm_h%wghf_eb(m) + &
[2963]	615	surf_usm_h%frac(ind_pav_green,m) * &
[2797]	616	surf_usm_h%wghf_eb_green(m) + &
[2963]	617	surf_usm_h%frac(ind_wat_win,m) * &
[2797]	618	surf_usm_h%wghf_eb_window(m)
	619	ENDDO
	620	ELSE
[3004]	621	IF ( .NOT. ALLOCATED( ghf_av ) ) THEN
	622	ALLOCATE( ghf_av(nysg:nyng,nxlg:nxrg) )
	623	ghf_av = REAL( fill_value, KIND = wp )
	624	ENDIF
[2797]	625	DO i = nxl, nxr
	626	DO j = nys, nyn
	627	local_pf(i,j,nzb+1) = ghf_av(j,i)
	628	ENDDO
	629	ENDDO
	630	ENDIF
	631
	632	resorted = .TRUE.
	633	two_d = .TRUE.
	634	level_z(nzb+1) = zu(nzb+1)
	635
[1691]	636	CASE ( 'ol*_xy' ) ! 2d-array
	637	IF ( av == 0 ) THEN
[2232]	638	DO m = 1, surf_def_h(0)%ns
	639	i = surf_def_h(0)%i(m)
	640	j = surf_def_h(0)%j(m)
[2512]	641	local_pf(i,j,nzb+1) = surf_def_h(0)%ol(m)
[2232]	642	ENDDO
	643	DO m = 1, surf_lsm_h%ns
	644	i = surf_lsm_h%i(m)
	645	j = surf_lsm_h%j(m)
[2512]	646	local_pf(i,j,nzb+1) = surf_lsm_h%ol(m)
[2232]	647	ENDDO
	648	DO m = 1, surf_usm_h%ns
	649	i = surf_usm_h%i(m)
	650	j = surf_usm_h%j(m)
[2512]	651	local_pf(i,j,nzb+1) = surf_usm_h%ol(m)
[2232]	652	ENDDO
[1691]	653	ELSE
[3004]	654	IF ( .NOT. ALLOCATED( ol_av ) ) THEN
	655	ALLOCATE( ol_av(nysg:nyng,nxlg:nxrg) )
	656	ol_av = REAL( fill_value, KIND = wp )
	657	ENDIF
[2512]	658	DO i = nxl, nxr
	659	DO j = nys, nyn
[1691]	660	local_pf(i,j,nzb+1) = ol_av(j,i)
	661	ENDDO
	662	ENDDO
	663	ENDIF
	664	resorted = .TRUE.
	665	two_d = .TRUE.
	666	level_z(nzb+1) = zu(nzb+1)
	667
[1]	668	CASE ( 'p_xy', 'p_xz', 'p_yz' )
	669	IF ( av == 0 ) THEN
[729]	670	IF ( psolver /= 'sor' ) CALL exchange_horiz( p, nbgp )
[1]	671	to_be_resorted => p
	672	ELSE
[3004]	673	IF ( .NOT. ALLOCATED( p_av ) ) THEN
	674	ALLOCATE( p_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	675	p_av = REAL( fill_value, KIND = wp )
	676	ENDIF
[729]	677	IF ( psolver /= 'sor' ) CALL exchange_horiz( p_av, nbgp )
[1]	678	to_be_resorted => p_av
	679	ENDIF
	680	IF ( mode == 'xy' ) level_z = zu
	681
	682	CASE ( 'pc_xy', 'pc_xz', 'pc_yz' ) ! particle concentration
	683	IF ( av == 0 ) THEN
[215]	684	IF ( simulated_time >= particle_advection_start ) THEN
	685	tend = prt_count
[2512]	686	! CALL exchange_horiz( tend, nbgp )
[215]	687	ELSE
[1353]	688	tend = 0.0_wp
[215]	689	ENDIF
[2512]	690	DO i = nxl, nxr
	691	DO j = nys, nyn
[1]	692	DO k = nzb, nzt+1
	693	local_pf(i,j,k) = tend(k,j,i)
	694	ENDDO
	695	ENDDO
	696	ENDDO
	697	resorted = .TRUE.
	698	ELSE
[3004]	699	IF ( .NOT. ALLOCATED( pc_av ) ) THEN
	700	ALLOCATE( pc_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	701	pc_av = REAL( fill_value, KIND = wp )
	702	ENDIF
[2512]	703	! CALL exchange_horiz( pc_av, nbgp )
[1]	704	to_be_resorted => pc_av
	705	ENDIF
	706
[1359]	707	CASE ( 'pr_xy', 'pr_xz', 'pr_yz' ) ! mean particle radius (effective radius)
[1]	708	IF ( av == 0 ) THEN
[215]	709	IF ( simulated_time >= particle_advection_start ) THEN
	710	DO i = nxl, nxr
	711	DO j = nys, nyn
	712	DO k = nzb, nzt+1
[1359]	713	number_of_particles = prt_count(k,j,i)
	714	IF (number_of_particles <= 0) CYCLE
	715	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
	716	s_r2 = 0.0_wp
[1353]	717	s_r3 = 0.0_wp
[1359]	718	DO n = 1, number_of_particles
	719	IF ( particles(n)%particle_mask ) THEN
	720	s_r2 = s_r2 + particles(n)%radius*2 &
	721	particles(n)%weight_factor
	722	s_r3 = s_r3 + particles(n)%radius*3 &
	723	particles(n)%weight_factor
	724	ENDIF
[215]	725	ENDDO
[1359]	726	IF ( s_r2 > 0.0_wp ) THEN
	727	mean_r = s_r3 / s_r2
[215]	728	ELSE
[1353]	729	mean_r = 0.0_wp
[215]	730	ENDIF
	731	tend(k,j,i) = mean_r
[1]	732	ENDDO
	733	ENDDO
	734	ENDDO
[2512]	735	! CALL exchange_horiz( tend, nbgp )
[215]	736	ELSE
[1353]	737	tend = 0.0_wp
[1359]	738	ENDIF
[2512]	739	DO i = nxl, nxr
	740	DO j = nys, nyn
[1]	741	DO k = nzb, nzt+1
	742	local_pf(i,j,k) = tend(k,j,i)
	743	ENDDO
	744	ENDDO
	745	ENDDO
	746	resorted = .TRUE.
	747	ELSE
[3004]	748	IF ( .NOT. ALLOCATED( pr_av ) ) THEN
	749	ALLOCATE( pr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	750	pr_av = REAL( fill_value, KIND = wp )
	751	ENDIF
[2512]	752	! CALL exchange_horiz( pr_av, nbgp )
[1]	753	to_be_resorted => pr_av
	754	ENDIF
	755
[72]	756	CASE ( 'pra*_xy' ) ! 2d-array / integral quantity => no av
[2512]	757	! CALL exchange_horiz_2d( precipitation_amount )
	758	DO i = nxl, nxr
	759	DO j = nys, nyn
[72]	760	local_pf(i,j,nzb+1) = precipitation_amount(j,i)
	761	ENDDO
	762	ENDDO
[1353]	763	precipitation_amount = 0.0_wp ! reset for next integ. interval
[72]	764	resorted = .TRUE.
	765	two_d = .TRUE.
	766	level_z(nzb+1) = zu(nzb+1)
	767
[1053]	768	CASE ( 'prr_xy', 'prr_xz', 'prr_yz' )
[72]	769	IF ( av == 0 ) THEN
[2512]	770	! CALL exchange_horiz( prr, nbgp )
	771	DO i = nxl, nxr
	772	DO j = nys, nyn
[1053]	773	DO k = nzb, nzt+1
[1822]	774	local_pf(i,j,k) = prr(k,j,i) * hyrho(nzb+1)
[1053]	775	ENDDO
[72]	776	ENDDO
	777	ENDDO
	778	ELSE
[3004]	779	IF ( .NOT. ALLOCATED( prr_av ) ) THEN
	780	ALLOCATE( prr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	781	prr_av = REAL( fill_value, KIND = wp )
	782	ENDIF
[2512]	783	! CALL exchange_horiz( prr_av, nbgp )
	784	DO i = nxl, nxr
	785	DO j = nys, nyn
[1053]	786	DO k = nzb, nzt+1
[1822]	787	local_pf(i,j,k) = prr_av(k,j,i) * hyrho(nzb+1)
[1053]	788	ENDDO
[72]	789	ENDDO
	790	ENDDO
	791	ENDIF
	792	resorted = .TRUE.
[1053]	793	IF ( mode == 'xy' ) level_z = zu
[72]	794
[1]	795	CASE ( 'pt_xy', 'pt_xz', 'pt_yz' )
	796	IF ( av == 0 ) THEN
	797	IF ( .NOT. cloud_physics ) THEN
	798	to_be_resorted => pt
	799	ELSE
[2512]	800	DO i = nxl, nxr
	801	DO j = nys, nyn
[1]	802	DO k = nzb, nzt+1
[1320]	803	local_pf(i,j,k) = pt(k,j,i) + l_d_cp * &
	804	pt_d_t(k) * &
[1]	805	ql(k,j,i)
	806	ENDDO
	807	ENDDO
	808	ENDDO
	809	resorted = .TRUE.
	810	ENDIF
	811	ELSE
[3004]	812	IF ( .NOT. ALLOCATED( pt_av ) ) THEN
	813	ALLOCATE( pt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	814	pt_av = REAL( fill_value, KIND = wp )
	815	ENDIF
[1]	816	to_be_resorted => pt_av
	817	ENDIF
	818	IF ( mode == 'xy' ) level_z = zu
	819
	820	CASE ( 'q_xy', 'q_xz', 'q_yz' )
	821	IF ( av == 0 ) THEN
	822	to_be_resorted => q
	823	ELSE
[3004]	824	IF ( .NOT. ALLOCATED( q_av ) ) THEN
	825	ALLOCATE( q_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	826	q_av = REAL( fill_value, KIND = wp )
	827	ENDIF
[1]	828	to_be_resorted => q_av
	829	ENDIF
	830	IF ( mode == 'xy' ) level_z = zu
	831
[1053]	832	CASE ( 'qc_xy', 'qc_xz', 'qc_yz' )
[1]	833	IF ( av == 0 ) THEN
[1115]	834	to_be_resorted => qc
[1]	835	ELSE
[3004]	836	IF ( .NOT. ALLOCATED( qc_av ) ) THEN
	837	ALLOCATE( qc_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	838	qc_av = REAL( fill_value, KIND = wp )
	839	ENDIF
[1115]	840	to_be_resorted => qc_av
[1]	841	ENDIF
	842	IF ( mode == 'xy' ) level_z = zu
	843
[1053]	844	CASE ( 'ql_xy', 'ql_xz', 'ql_yz' )
	845	IF ( av == 0 ) THEN
[1115]	846	to_be_resorted => ql
[1053]	847	ELSE
[3004]	848	IF ( .NOT. ALLOCATED( ql_av ) ) THEN
	849	ALLOCATE( ql_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	850	ql_av = REAL( fill_value, KIND = wp )
	851	ENDIF
[1115]	852	to_be_resorted => ql_av
[1053]	853	ENDIF
	854	IF ( mode == 'xy' ) level_z = zu
	855
[1]	856	CASE ( 'ql_c_xy', 'ql_c_xz', 'ql_c_yz' )
	857	IF ( av == 0 ) THEN
	858	to_be_resorted => ql_c
	859	ELSE
[3004]	860	IF ( .NOT. ALLOCATED( ql_c_av ) ) THEN
	861	ALLOCATE( ql_c_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	862	ql_c_av = REAL( fill_value, KIND = wp )
	863	ENDIF
[1]	864	to_be_resorted => ql_c_av
	865	ENDIF
	866	IF ( mode == 'xy' ) level_z = zu
	867
	868	CASE ( 'ql_v_xy', 'ql_v_xz', 'ql_v_yz' )
	869	IF ( av == 0 ) THEN
	870	to_be_resorted => ql_v
	871	ELSE
[3004]	872	IF ( .NOT. ALLOCATED( ql_v_av ) ) THEN
	873	ALLOCATE( ql_v_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	874	ql_v_av = REAL( fill_value, KIND = wp )
	875	ENDIF
[1]	876	to_be_resorted => ql_v_av
	877	ENDIF
	878	IF ( mode == 'xy' ) level_z = zu
	879
	880	CASE ( 'ql_vp_xy', 'ql_vp_xz', 'ql_vp_yz' )
	881	IF ( av == 0 ) THEN
[1007]	882	IF ( simulated_time >= particle_advection_start ) THEN
	883	DO i = nxl, nxr
	884	DO j = nys, nyn
	885	DO k = nzb, nzt+1
[1359]	886	number_of_particles = prt_count(k,j,i)
	887	IF (number_of_particles <= 0) CYCLE
	888	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
	889	DO n = 1, number_of_particles
	890	IF ( particles(n)%particle_mask ) THEN
	891	tend(k,j,i) = tend(k,j,i) + &
	892	particles(n)%weight_factor / &
	893	prt_count(k,j,i)
	894	ENDIF
[1007]	895	ENDDO
	896	ENDDO
	897	ENDDO
	898	ENDDO
[2512]	899	! CALL exchange_horiz( tend, nbgp )
[1007]	900	ELSE
[1353]	901	tend = 0.0_wp
[1359]	902	ENDIF
[2512]	903	DO i = nxl, nxr
	904	DO j = nys, nyn
[1007]	905	DO k = nzb, nzt+1
	906	local_pf(i,j,k) = tend(k,j,i)
	907	ENDDO
	908	ENDDO
	909	ENDDO
	910	resorted = .TRUE.
	911	ELSE
[3004]	912	IF ( .NOT. ALLOCATED( ql_vp_av ) ) THEN
	913	ALLOCATE( ql_vp_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	914	ql_vp_av = REAL( fill_value, KIND = wp )
	915	ENDIF
[2512]	916	! CALL exchange_horiz( ql_vp_av, nbgp )
[3004]	917	to_be_resorted => ql_vp_av
[1]	918	ENDIF
	919	IF ( mode == 'xy' ) level_z = zu
	920
[1053]	921	CASE ( 'qr_xy', 'qr_xz', 'qr_yz' )
	922	IF ( av == 0 ) THEN
	923	to_be_resorted => qr
	924	ELSE
[3004]	925	IF ( .NOT. ALLOCATED( qr_av ) ) THEN
	926	ALLOCATE( qr_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	927	qr_av = REAL( fill_value, KIND = wp )
	928	ENDIF
[1053]	929	to_be_resorted => qr_av
	930	ENDIF
	931	IF ( mode == 'xy' ) level_z = zu
	932
[354]	933	CASE ( 'qsws*_xy' ) ! 2d-array
	934	IF ( av == 0 ) THEN
[2743]	935	!
	936	!-- In case of default surfaces, clean-up flux by density.
	937	!-- In case of land- and urban-surfaces, convert fluxes into
	938	!-- dynamic units
[2232]	939	DO m = 1, surf_def_h(0)%ns
	940	i = surf_def_h(0)%i(m)
	941	j = surf_def_h(0)%j(m)
[2743]	942	k = surf_def_h(0)%k(m)
	943	local_pf(i,j,nzb+1) = surf_def_h(0)%qsws(m) * &
	944	waterflux_output_conversion(k)
[2232]	945	ENDDO
	946	DO m = 1, surf_lsm_h%ns
	947	i = surf_lsm_h%i(m)
	948	j = surf_lsm_h%j(m)
[2743]	949	k = surf_lsm_h%k(m)
	950	local_pf(i,j,nzb+1) = surf_lsm_h%qsws(m) * l_v
[2232]	951	ENDDO
	952	DO m = 1, surf_usm_h%ns
	953	i = surf_usm_h%i(m)
	954	j = surf_usm_h%j(m)
[2743]	955	k = surf_usm_h%k(m)
	956	local_pf(i,j,nzb+1) = surf_usm_h%qsws(m) * l_v
[2232]	957	ENDDO
[354]	958	ELSE
[3004]	959	IF ( .NOT. ALLOCATED( qsws_av ) ) THEN
	960	ALLOCATE( qsws_av(nysg:nyng,nxlg:nxrg) )
	961	qsws_av = REAL( fill_value, KIND = wp )
	962	ENDIF
[2512]	963	DO i = nxl, nxr
	964	DO j = nys, nyn
[354]	965	local_pf(i,j,nzb+1) = qsws_av(j,i)
	966	ENDDO
	967	ENDDO
	968	ENDIF
	969	resorted = .TRUE.
	970	two_d = .TRUE.
	971	level_z(nzb+1) = zu(nzb+1)
	972
[1]	973	CASE ( 'qv_xy', 'qv_xz', 'qv_yz' )
	974	IF ( av == 0 ) THEN
[2512]	975	DO i = nxl, nxr
	976	DO j = nys, nyn
[1]	977	DO k = nzb, nzt+1
	978	local_pf(i,j,k) = q(k,j,i) - ql(k,j,i)
	979	ENDDO
	980	ENDDO
	981	ENDDO
	982	resorted = .TRUE.
	983	ELSE
[3004]	984	IF ( .NOT. ALLOCATED( qv_av ) ) THEN
	985	ALLOCATE( qv_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	986	qv_av = REAL( fill_value, KIND = wp )
	987	ENDIF
[1]	988	to_be_resorted => qv_av
	989	ENDIF
	990	IF ( mode == 'xy' ) level_z = zu
	991
[2735]	992	CASE ( 'r_a*_xy' ) ! 2d-array
	993	IF ( av == 0 ) THEN
	994	DO m = 1, surf_lsm_h%ns
	995	i = surf_lsm_h%i(m)
	996	j = surf_lsm_h%j(m)
	997	local_pf(i,j,nzb+1) = surf_lsm_h%r_a(m)
	998	ENDDO
[1551]	999
[2735]	1000	DO m = 1, surf_usm_h%ns
	1001	i = surf_usm_h%i(m)
	1002	j = surf_usm_h%j(m)
	1003	local_pf(i,j,nzb+1) = &
[2963]	1004	( surf_usm_h%frac(ind_veg_wall,m) * &
	1005	surf_usm_h%r_a(m) + &
	1006	surf_usm_h%frac(ind_pav_green,m) * &
	1007	surf_usm_h%r_a_green(m) + &
	1008	surf_usm_h%frac(ind_wat_win,m) * &
	1009	surf_usm_h%r_a_window(m) )
[2735]	1010	ENDDO
	1011	ELSE
[3004]	1012	IF ( .NOT. ALLOCATED( r_a_av ) ) THEN
	1013	ALLOCATE( r_a_av(nysg:nyng,nxlg:nxrg) )
	1014	r_a_av = REAL( fill_value, KIND = wp )
	1015	ENDIF
[2735]	1016	DO i = nxl, nxr
	1017	DO j = nys, nyn
	1018	local_pf(i,j,nzb+1) = r_a_av(j,i)
	1019	ENDDO
	1020	ENDDO
	1021	ENDIF
	1022	resorted = .TRUE.
	1023	two_d = .TRUE.
	1024	level_z(nzb+1) = zu(nzb+1)
	1025
[2190]	1026	CASE ( 'rho_ocean_xy', 'rho_ocean_xz', 'rho_ocean_yz' )
[96]	1027	IF ( av == 0 ) THEN
[2031]	1028	to_be_resorted => rho_ocean
[96]	1029	ELSE
[3004]	1030	IF ( .NOT. ALLOCATED( rho_ocean_av ) ) THEN
	1031	ALLOCATE( rho_ocean_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	1032	rho_ocean_av = REAL( fill_value, KIND = wp )
	1033	ENDIF
[2031]	1034	to_be_resorted => rho_ocean_av
[96]	1035	ENDIF
	1036
[1]	1037	CASE ( 's_xy', 's_xz', 's_yz' )
	1038	IF ( av == 0 ) THEN
[1960]	1039	to_be_resorted => s
[1]	1040	ELSE
[3004]	1041	IF ( .NOT. ALLOCATED( s_av ) ) THEN
	1042	ALLOCATE( s_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	1043	s_av = REAL( fill_value, KIND = wp )
	1044	ENDIF
[355]	1045	to_be_resorted => s_av
[1]	1046	ENDIF
	1047
[96]	1048	CASE ( 'sa_xy', 'sa_xz', 'sa_yz' )
	1049	IF ( av == 0 ) THEN
	1050	to_be_resorted => sa
	1051	ELSE
[3004]	1052	IF ( .NOT. ALLOCATED( sa_av ) ) THEN
	1053	ALLOCATE( sa_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	1054	sa_av = REAL( fill_value, KIND = wp )
	1055	ENDIF
[96]	1056	to_be_resorted => sa_av
	1057	ENDIF
	1058
[354]	1059	CASE ( 'shf*_xy' ) ! 2d-array
	1060	IF ( av == 0 ) THEN
[2743]	1061	!
	1062	!-- In case of default surfaces, clean-up flux by density.
	1063	!-- In case of land- and urban-surfaces, convert fluxes into
	1064	!-- dynamic units.
[2232]	1065	DO m = 1, surf_def_h(0)%ns
	1066	i = surf_def_h(0)%i(m)
	1067	j = surf_def_h(0)%j(m)
[2743]	1068	k = surf_def_h(0)%k(m)
	1069	local_pf(i,j,nzb+1) = surf_def_h(0)%shf(m) * &
	1070	heatflux_output_conversion(k)
[2232]	1071	ENDDO
	1072	DO m = 1, surf_lsm_h%ns
	1073	i = surf_lsm_h%i(m)
	1074	j = surf_lsm_h%j(m)
[2743]	1075	k = surf_lsm_h%k(m)
	1076	local_pf(i,j,nzb+1) = surf_lsm_h%shf(m) * cp
[2232]	1077	ENDDO
	1078	DO m = 1, surf_usm_h%ns
	1079	i = surf_usm_h%i(m)
	1080	j = surf_usm_h%j(m)
[2743]	1081	k = surf_usm_h%k(m)
	1082	local_pf(i,j,nzb+1) = surf_usm_h%shf(m) * cp
[2232]	1083	ENDDO
[354]	1084	ELSE
[3004]	1085	IF ( .NOT. ALLOCATED( shf_av ) ) THEN
	1086	ALLOCATE( shf_av(nysg:nyng,nxlg:nxrg) )
	1087	shf_av = REAL( fill_value, KIND = wp )
	1088	ENDIF
[2512]	1089	DO i = nxl, nxr
	1090	DO j = nys, nyn
[354]	1091	local_pf(i,j,nzb+1) = shf_av(j,i)
	1092	ENDDO
	1093	ENDDO
	1094	ENDIF
	1095	resorted = .TRUE.
	1096	two_d = .TRUE.
	1097	level_z(nzb+1) = zu(nzb+1)
[1960]	1098
	1099	CASE ( 'ssws*_xy' ) ! 2d-array
	1100	IF ( av == 0 ) THEN
[2232]	1101	DO m = 1, surf_def_h(0)%ns
	1102	i = surf_def_h(0)%i(m)
	1103	j = surf_def_h(0)%j(m)
[2512]	1104	local_pf(i,j,nzb+1) = surf_def_h(0)%ssws(m)
[2232]	1105	ENDDO
	1106	DO m = 1, surf_lsm_h%ns
	1107	i = surf_lsm_h%i(m)
	1108	j = surf_lsm_h%j(m)
[2512]	1109	local_pf(i,j,nzb+1) = surf_lsm_h%ssws(m)
[2232]	1110	ENDDO
	1111	DO m = 1, surf_usm_h%ns
	1112	i = surf_usm_h%i(m)
	1113	j = surf_usm_h%j(m)
[2512]	1114	local_pf(i,j,nzb+1) = surf_usm_h%ssws(m)
[2232]	1115	ENDDO
[1960]	1116	ELSE
[3004]	1117	IF ( .NOT. ALLOCATED( ssws_av ) ) THEN
	1118	ALLOCATE( ssws_av(nysg:nyng,nxlg:nxrg) )
	1119	ssws_av = REAL( fill_value, KIND = wp )
	1120	ENDIF
[2512]	1121	DO i = nxl, nxr
	1122	DO j = nys, nyn
[1960]	1123	local_pf(i,j,nzb+1) = ssws_av(j,i)
	1124	ENDDO
	1125	ENDDO
	1126	ENDIF
	1127	resorted = .TRUE.
	1128	two_d = .TRUE.
	1129	level_z(nzb+1) = zu(nzb+1)
[1551]	1130
[1]	1131	CASE ( 't*_xy' ) ! 2d-array
	1132	IF ( av == 0 ) THEN
[2232]	1133	DO m = 1, surf_def_h(0)%ns
	1134	i = surf_def_h(0)%i(m)
	1135	j = surf_def_h(0)%j(m)
[2512]	1136	local_pf(i,j,nzb+1) = surf_def_h(0)%ts(m)
[2232]	1137	ENDDO
	1138	DO m = 1, surf_lsm_h%ns
	1139	i = surf_lsm_h%i(m)
	1140	j = surf_lsm_h%j(m)
[2512]	1141	local_pf(i,j,nzb+1) = surf_lsm_h%ts(m)
[2232]	1142	ENDDO
	1143	DO m = 1, surf_usm_h%ns
	1144	i = surf_usm_h%i(m)
	1145	j = surf_usm_h%j(m)
[2512]	1146	local_pf(i,j,nzb+1) = surf_usm_h%ts(m)
[2232]	1147	ENDDO
[1]	1148	ELSE
[3004]	1149	IF ( .NOT. ALLOCATED( ts_av ) ) THEN
	1150	ALLOCATE( ts_av(nysg:nyng,nxlg:nxrg) )
	1151	ts_av = REAL( fill_value, KIND = wp )
	1152	ENDIF
[2512]	1153	DO i = nxl, nxr
	1154	DO j = nys, nyn
[1]	1155	local_pf(i,j,nzb+1) = ts_av(j,i)
	1156	ENDDO
	1157	ENDDO
	1158	ENDIF
	1159	resorted = .TRUE.
	1160	two_d = .TRUE.
	1161	level_z(nzb+1) = zu(nzb+1)
	1162
[2742]	1163	CASE ( 'tsurf*_xy' ) ! 2d-array
	1164	IF ( av == 0 ) THEN
[2798]	1165	DO m = 1, surf_def_h(0)%ns
	1166	i = surf_def_h(0)%i(m)
	1167	j = surf_def_h(0)%j(m)
	1168	local_pf(i,j,nzb+1) = surf_def_h(0)%pt_surface(m)
	1169	ENDDO
	1170
[2742]	1171	DO m = 1, surf_lsm_h%ns
	1172	i = surf_lsm_h%i(m)
	1173	j = surf_lsm_h%j(m)
	1174	local_pf(i,j,nzb+1) = surf_lsm_h%pt_surface(m)
	1175	ENDDO
	1176
	1177	DO m = 1, surf_usm_h%ns
	1178	i = surf_usm_h%i(m)
	1179	j = surf_usm_h%j(m)
	1180	local_pf(i,j,nzb+1) = surf_usm_h%pt_surface(m)
	1181	ENDDO
	1182
	1183	ELSE
[3004]	1184	IF ( .NOT. ALLOCATED( tsurf_av ) ) THEN
	1185	ALLOCATE( tsurf_av(nysg:nyng,nxlg:nxrg) )
	1186	tsurf_av = REAL( fill_value, KIND = wp )
	1187	ENDIF
[2742]	1188	DO i = nxl, nxr
	1189	DO j = nys, nyn
	1190	local_pf(i,j,nzb+1) = tsurf_av(j,i)
	1191	ENDDO
	1192	ENDDO
	1193	ENDIF
	1194	resorted = .TRUE.
	1195	two_d = .TRUE.
	1196	level_z(nzb+1) = zu(nzb+1)
	1197
[1]	1198	CASE ( 'u_xy', 'u_xz', 'u_yz' )
[2696]	1199	flag_nr = 1
[1]	1200	IF ( av == 0 ) THEN
	1201	to_be_resorted => u
	1202	ELSE
[3004]	1203	IF ( .NOT. ALLOCATED( u_av ) ) THEN
	1204	ALLOCATE( u_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	1205	u_av = REAL( fill_value, KIND = wp )
	1206	ENDIF
[1]	1207	to_be_resorted => u_av
	1208	ENDIF
	1209	IF ( mode == 'xy' ) level_z = zu
	1210	!
	1211	!-- Substitute the values generated by "mirror" boundary condition
	1212	!-- at the bottom boundary by the real surface values.
	1213	IF ( do2d(av,if) == 'u_xz' .OR. do2d(av,if) == 'u_yz' ) THEN
[1353]	1214	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
[1]	1215	ENDIF
	1216
	1217	CASE ( 'u*_xy' ) ! 2d-array
	1218	IF ( av == 0 ) THEN
[2232]	1219	DO m = 1, surf_def_h(0)%ns
	1220	i = surf_def_h(0)%i(m)
	1221	j = surf_def_h(0)%j(m)
[2512]	1222	local_pf(i,j,nzb+1) = surf_def_h(0)%us(m)
[2232]	1223	ENDDO
	1224	DO m = 1, surf_lsm_h%ns
	1225	i = surf_lsm_h%i(m)
	1226	j = surf_lsm_h%j(m)
[2512]	1227	local_pf(i,j,nzb+1) = surf_lsm_h%us(m)
[2232]	1228	ENDDO
	1229	DO m = 1, surf_usm_h%ns
	1230	i = surf_usm_h%i(m)
	1231	j = surf_usm_h%j(m)
[2512]	1232	local_pf(i,j,nzb+1) = surf_usm_h%us(m)
[2232]	1233	ENDDO
[1]	1234	ELSE
[3004]	1235	IF ( .NOT. ALLOCATED( us_av ) ) THEN
	1236	ALLOCATE( us_av(nysg:nyng,nxlg:nxrg) )
	1237	us_av = REAL( fill_value, KIND = wp )
	1238	ENDIF
[2512]	1239	DO i = nxl, nxr
	1240	DO j = nys, nyn
[1]	1241	local_pf(i,j,nzb+1) = us_av(j,i)
	1242	ENDDO
	1243	ENDDO
	1244	ENDIF
	1245	resorted = .TRUE.
	1246	two_d = .TRUE.
	1247	level_z(nzb+1) = zu(nzb+1)
	1248
	1249	CASE ( 'v_xy', 'v_xz', 'v_yz' )
[2696]	1250	flag_nr = 2
[1]	1251	IF ( av == 0 ) THEN
	1252	to_be_resorted => v
	1253	ELSE
[3004]	1254	IF ( .NOT. ALLOCATED( v_av ) ) THEN
	1255	ALLOCATE( v_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	1256	v_av = REAL( fill_value, KIND = wp )
	1257	ENDIF
[1]	1258	to_be_resorted => v_av
	1259	ENDIF
	1260	IF ( mode == 'xy' ) level_z = zu
	1261	!
	1262	!-- Substitute the values generated by "mirror" boundary condition
	1263	!-- at the bottom boundary by the real surface values.
	1264	IF ( do2d(av,if) == 'v_xz' .OR. do2d(av,if) == 'v_yz' ) THEN
[1353]	1265	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
[1]	1266	ENDIF
	1267
	1268	CASE ( 'vpt_xy', 'vpt_xz', 'vpt_yz' )
	1269	IF ( av == 0 ) THEN
	1270	to_be_resorted => vpt
	1271	ELSE
[3004]	1272	IF ( .NOT. ALLOCATED( vpt_av ) ) THEN
	1273	ALLOCATE( vpt_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	1274	vpt_av = REAL( fill_value, KIND = wp )
	1275	ENDIF
[1]	1276	to_be_resorted => vpt_av
	1277	ENDIF
	1278	IF ( mode == 'xy' ) level_z = zu
	1279
	1280	CASE ( 'w_xy', 'w_xz', 'w_yz' )
[2696]	1281	flag_nr = 3
[1]	1282	IF ( av == 0 ) THEN
	1283	to_be_resorted => w
	1284	ELSE
[3004]	1285	IF ( .NOT. ALLOCATED( w_av ) ) THEN
	1286	ALLOCATE( w_av(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	1287	w_av = REAL( fill_value, KIND = wp )
	1288	ENDIF
[1]	1289	to_be_resorted => w_av
	1290	ENDIF
	1291	IF ( mode == 'xy' ) level_z = zw
	1292
[72]	1293	CASE ( 'z0*_xy' ) ! 2d-array
	1294	IF ( av == 0 ) THEN
[2232]	1295	DO m = 1, surf_def_h(0)%ns
	1296	i = surf_def_h(0)%i(m)
	1297	j = surf_def_h(0)%j(m)
[2512]	1298	local_pf(i,j,nzb+1) = surf_def_h(0)%z0(m)
[2232]	1299	ENDDO
	1300	DO m = 1, surf_lsm_h%ns
	1301	i = surf_lsm_h%i(m)
	1302	j = surf_lsm_h%j(m)
[2512]	1303	local_pf(i,j,nzb+1) = surf_lsm_h%z0(m)
[2232]	1304	ENDDO
	1305	DO m = 1, surf_usm_h%ns
	1306	i = surf_usm_h%i(m)
	1307	j = surf_usm_h%j(m)
[2512]	1308	local_pf(i,j,nzb+1) = surf_usm_h%z0(m)
[2232]	1309	ENDDO
[72]	1310	ELSE
[3004]	1311	IF ( .NOT. ALLOCATED( z0_av ) ) THEN
	1312	ALLOCATE( z0_av(nysg:nyng,nxlg:nxrg) )
	1313	z0_av = REAL( fill_value, KIND = wp )
	1314	ENDIF
[2512]	1315	DO i = nxl, nxr
	1316	DO j = nys, nyn
[72]	1317	local_pf(i,j,nzb+1) = z0_av(j,i)
	1318	ENDDO
	1319	ENDDO
	1320	ENDIF
	1321	resorted = .TRUE.
	1322	two_d = .TRUE.
	1323	level_z(nzb+1) = zu(nzb+1)
	1324
[978]	1325	CASE ( 'z0h*_xy' ) ! 2d-array
	1326	IF ( av == 0 ) THEN
[2232]	1327	DO m = 1, surf_def_h(0)%ns
	1328	i = surf_def_h(0)%i(m)
	1329	j = surf_def_h(0)%j(m)
[2512]	1330	local_pf(i,j,nzb+1) = surf_def_h(0)%z0h(m)
[2232]	1331	ENDDO
	1332	DO m = 1, surf_lsm_h%ns
	1333	i = surf_lsm_h%i(m)
	1334	j = surf_lsm_h%j(m)
[2512]	1335	local_pf(i,j,nzb+1) = surf_lsm_h%z0h(m)
[2232]	1336	ENDDO
	1337	DO m = 1, surf_usm_h%ns
	1338	i = surf_usm_h%i(m)
	1339	j = surf_usm_h%j(m)
[2512]	1340	local_pf(i,j,nzb+1) = surf_usm_h%z0h(m)
[2232]	1341	ENDDO
[978]	1342	ELSE
[3004]	1343	IF ( .NOT. ALLOCATED( z0h_av ) ) THEN
	1344	ALLOCATE( z0h_av(nysg:nyng,nxlg:nxrg) )
	1345	z0h_av = REAL( fill_value, KIND = wp )
	1346	ENDIF
[2512]	1347	DO i = nxl, nxr
	1348	DO j = nys, nyn
[978]	1349	local_pf(i,j,nzb+1) = z0h_av(j,i)
	1350	ENDDO
	1351	ENDDO
	1352	ENDIF
	1353	resorted = .TRUE.
	1354	two_d = .TRUE.
	1355	level_z(nzb+1) = zu(nzb+1)
	1356
[1788]	1357	CASE ( 'z0q*_xy' ) ! 2d-array
	1358	IF ( av == 0 ) THEN
[2232]	1359	DO m = 1, surf_def_h(0)%ns
	1360	i = surf_def_h(0)%i(m)
	1361	j = surf_def_h(0)%j(m)
[2512]	1362	local_pf(i,j,nzb+1) = surf_def_h(0)%z0q(m)
[2232]	1363	ENDDO
	1364	DO m = 1, surf_lsm_h%ns
	1365	i = surf_lsm_h%i(m)
	1366	j = surf_lsm_h%j(m)
[2512]	1367	local_pf(i,j,nzb+1) = surf_lsm_h%z0q(m)
[2232]	1368	ENDDO
	1369	DO m = 1, surf_usm_h%ns
	1370	i = surf_usm_h%i(m)
	1371	j = surf_usm_h%j(m)
[2512]	1372	local_pf(i,j,nzb+1) = surf_usm_h%z0q(m)
[2232]	1373	ENDDO
[1788]	1374	ELSE
[3004]	1375	IF ( .NOT. ALLOCATED( z0q_av ) ) THEN
	1376	ALLOCATE( z0q_av(nysg:nyng,nxlg:nxrg) )
	1377	z0q_av = REAL( fill_value, KIND = wp )
	1378	ENDIF
[2512]	1379	DO i = nxl, nxr
	1380	DO j = nys, nyn
[1788]	1381	local_pf(i,j,nzb+1) = z0q_av(j,i)
	1382	ENDDO
	1383	ENDDO
	1384	ENDIF
	1385	resorted = .TRUE.
	1386	two_d = .TRUE.
	1387	level_z(nzb+1) = zu(nzb+1)
	1388
[1]	1389	CASE DEFAULT
[1972]	1390
[1]	1391	!
[1972]	1392	!-- Land surface model quantity
	1393	IF ( land_surface ) THEN
	1394	CALL lsm_data_output_2d( av, do2d(av,if), found, grid, mode,&
	1395	local_pf, two_d, nzb_do, nzt_do )
	1396	ENDIF
	1397
	1398	!
[2696]	1399	!-- Turbulence closure variables
	1400	IF ( .NOT. found ) THEN
	1401	CALL tcm_data_output_2d( av, do2d(av,if), found, grid, mode,&
	1402	local_pf, two_d, nzb_do, nzt_do )
	1403	ENDIF
	1404
	1405	!
[1976]	1406	!-- Radiation quantity
	1407	IF ( .NOT. found .AND. radiation ) THEN
	1408	CALL radiation_data_output_2d( av, do2d(av,if), found, grid,&
[3014]	1409	mode, local_pf, two_d, &
	1410	nzb_do, nzt_do )
[1976]	1411	ENDIF
	1412
	1413	!
[2817]	1414	!-- Gust module quantities
	1415	IF ( .NOT. found .AND. gust_module_enabled ) THEN
	1416	CALL gust_data_output_2d( av, do2d(av,if), found, grid, &
	1417	local_pf, two_d, nzb_do, nzt_do )
	1418	ENDIF
	1419
	1420	!
[2696]	1421	!-- UV exposure model quantity
	1422	IF ( uv_exposure ) THEN
	1423	CALL uvem_data_output_2d( av, do2d(av,if), found, grid, mode,&
	1424	local_pf, two_d, nzb_do, nzt_do )
	1425	ENDIF
	1426
	1427	!
[1]	1428	!-- User defined quantity
[1972]	1429	IF ( .NOT. found ) THEN
	1430	CALL user_data_output_2d( av, do2d(av,if), found, grid, &
	1431	local_pf, two_d, nzb_do, nzt_do )
	1432	ENDIF
	1433
[1]	1434	resorted = .TRUE.
	1435
	1436	IF ( grid == 'zu' ) THEN
	1437	IF ( mode == 'xy' ) level_z = zu
	1438	ELSEIF ( grid == 'zw' ) THEN
	1439	IF ( mode == 'xy' ) level_z = zw
[343]	1440	ELSEIF ( grid == 'zu1' ) THEN
	1441	IF ( mode == 'xy' ) level_z(nzb+1) = zu(nzb+1)
[1551]	1442	ELSEIF ( grid == 'zs' ) THEN
	1443	IF ( mode == 'xy' ) level_z = zs
[1]	1444	ENDIF
	1445
	1446	IF ( .NOT. found ) THEN
[1320]	1447	message_string = 'no output provided for: ' // &
[274]	1448	TRIM( do2d(av,if) )
[254]	1449	CALL message( 'data_output_2d', 'PA0181', 0, 0, 0, 6, 0 )
[1]	1450	ENDIF
	1451
	1452	END SELECT
	1453
	1454	!
[2696]	1455	!-- Resort the array to be output, if not done above. Flag topography
	1456	!-- grid points with fill values, using the corresponding maksing flag.
[1]	1457	IF ( .NOT. resorted ) THEN
[2512]	1458	DO i = nxl, nxr
	1459	DO j = nys, nyn
[1551]	1460	DO k = nzb_do, nzt_do
[2696]	1461	local_pf(i,j,k) = MERGE( to_be_resorted(k,j,i), &
	1462	REAL( fill_value, KIND = wp ), &
	1463	BTEST( wall_flags_0(k,j,i), &
	1464	flag_nr ) )
[1]	1465	ENDDO
	1466	ENDDO
	1467	ENDDO
	1468	ENDIF
	1469
	1470	!
	1471	!-- Output of the individual cross-sections, depending on the cross-
	1472	!-- section mode chosen.
	1473	is = 1
[1960]	1474	loop1: DO WHILE ( section(is,s_ind) /= -9999 .OR. two_d )
[1]	1475
	1476	SELECT CASE ( mode )
	1477
	1478	CASE ( 'xy' )
	1479	!
	1480	!-- Determine the cross section index
	1481	IF ( two_d ) THEN
	1482	layer_xy = nzb+1
	1483	ELSE
[1960]	1484	layer_xy = section(is,s_ind)
[1]	1485	ENDIF
	1486
	1487	!
[1551]	1488	!-- Exit the loop for layers beyond the data output domain
	1489	!-- (used for soil model)
[1691]	1490	IF ( layer_xy > nzt_do ) THEN
[1551]	1491	EXIT loop1
	1492	ENDIF
	1493
	1494	!
[1308]	1495	!-- Update the netCDF xy cross section time axis.
	1496	!-- In case of parallel output, this is only done by PE0
	1497	!-- to increase the performance.
	1498	IF ( simulated_time /= do2d_xy_last_time(av) ) THEN
	1499	do2d_xy_time_count(av) = do2d_xy_time_count(av) + 1
	1500	do2d_xy_last_time(av) = simulated_time
	1501	IF ( myid == 0 ) THEN
[1327]	1502	IF ( .NOT. data_output_2d_on_each_pe &
	1503	.OR. netcdf_data_format > 4 ) &
[493]	1504	THEN
[1]	1505	#if defined( __netcdf )
	1506	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1507	id_var_time_xy(av), &
[291]	1508	(/ time_since_reference_point /), &
[1]	1509	start = (/ do2d_xy_time_count(av) /), &
	1510	count = (/ 1 /) )
[1783]	1511	CALL netcdf_handle_error( 'data_output_2d', 53 )
[1]	1512	#endif
	1513	ENDIF
	1514	ENDIF
	1515	ENDIF
	1516	!
	1517	!-- If required, carry out averaging along z
[1960]	1518	IF ( section(is,s_ind) == -1 .AND. .NOT. two_d ) THEN
[1]	1519
[1353]	1520	local_2d = 0.0_wp
[1]	1521	!
	1522	!-- Carry out the averaging (all data are on the PE)
[1551]	1523	DO k = nzb_do, nzt_do
[2512]	1524	DO j = nys, nyn
	1525	DO i = nxl, nxr
[1]	1526	local_2d(i,j) = local_2d(i,j) + local_pf(i,j,k)
	1527	ENDDO
	1528	ENDDO
	1529	ENDDO
	1530
[1551]	1531	local_2d = local_2d / ( nzt_do - nzb_do + 1.0_wp)
[1]	1532
	1533	ELSE
	1534	!
	1535	!-- Just store the respective section on the local array
	1536	local_2d = local_pf(:,:,layer_xy)
	1537
	1538	ENDIF
	1539
	1540	#if defined( __parallel )
[1327]	1541	IF ( netcdf_data_format > 4 ) THEN
[1]	1542	!
[1031]	1543	!-- Parallel output in netCDF4/HDF5 format.
[493]	1544	IF ( two_d ) THEN
	1545	iis = 1
	1546	ELSE
	1547	iis = is
	1548	ENDIF
	1549
[1]	1550	#if defined( __netcdf )
[1308]	1551	!
	1552	!-- For parallel output, all cross sections are first stored
	1553	!-- here on a local array and will be written to the output
	1554	!-- file afterwards to increase the performance.
[2512]	1555	DO i = nxl, nxr
	1556	DO j = nys, nyn
[1308]	1557	local_2d_sections(i,j,iis) = local_2d(i,j)
	1558	ENDDO
	1559	ENDDO
[1]	1560	#endif
[493]	1561	ELSE
[1]	1562
[493]	1563	IF ( data_output_2d_on_each_pe ) THEN
[1]	1564	!
[493]	1565	!-- Output of partial arrays on each PE
	1566	#if defined( __netcdf )
[1327]	1567	IF ( myid == 0 ) THEN
[1320]	1568	WRITE ( 21 ) time_since_reference_point, &
[493]	1569	do2d_xy_time_count(av), av
	1570	ENDIF
	1571	#endif
[759]	1572	DO i = 0, io_blocks-1
	1573	IF ( i == io_group ) THEN
[2512]	1574	WRITE ( 21 ) nxl, nxr, nys, nyn, nys, nyn
[759]	1575	WRITE ( 21 ) local_2d
	1576	ENDIF
	1577	#if defined( __parallel )
	1578	CALL MPI_BARRIER( comm2d, ierr )
	1579	#endif
	1580	ENDDO
[559]	1581
[493]	1582	ELSE
[1]	1583	!
[493]	1584	!-- PE0 receives partial arrays from all processors and
	1585	!-- then outputs them. Here a barrier has to be set,
	1586	!-- because otherwise "-MPI- FATAL: Remote protocol queue
	1587	!-- full" may occur.
	1588	CALL MPI_BARRIER( comm2d, ierr )
	1589
[2512]	1590	ngp = ( nxr-nxl+1 ) * ( nyn-nys+1 )
[493]	1591	IF ( myid == 0 ) THEN
[1]	1592	!
[493]	1593	!-- Local array can be relocated directly.
[2512]	1594	total_2d(nxl:nxr,nys:nyn) = local_2d
[1]	1595	!
[493]	1596	!-- Receive data from all other PEs.
	1597	DO n = 1, numprocs-1
[1]	1598	!
[493]	1599	!-- Receive index limits first, then array.
	1600	!-- Index limits are received in arbitrary order from
	1601	!-- the PEs.
[1320]	1602	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1603	MPI_ANY_SOURCE, 0, comm2d, &
[493]	1604	status, ierr )
	1605	sender = status(MPI_SOURCE)
	1606	DEALLOCATE( local_2d )
	1607	ALLOCATE( local_2d(ind(1):ind(2),ind(3):ind(4)) )
[1320]	1608	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1609	MPI_REAL, sender, 1, comm2d, &
[493]	1610	status, ierr )
	1611	total_2d(ind(1):ind(2),ind(3):ind(4)) = local_2d
	1612	ENDDO
[1]	1613	!
[493]	1614	!-- Relocate the local array for the next loop increment
	1615	DEALLOCATE( local_2d )
[2512]	1616	ALLOCATE( local_2d(nxl:nxr,nys:nyn) )
[1]	1617
	1618	#if defined( __netcdf )
[1327]	1619	IF ( two_d ) THEN
	1620	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1621	id_var_do2d(av,if), &
[2512]	1622	total_2d(0:nx,0:ny), &
[1327]	1623	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
[2512]	1624	count = (/ nx+1, ny+1, 1, 1 /) )
[1327]	1625	ELSE
	1626	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1627	id_var_do2d(av,if), &
[2512]	1628	total_2d(0:nx,0:ny), &
[1327]	1629	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
[2512]	1630	count = (/ nx+1, ny+1, 1, 1 /) )
[1]	1631	ENDIF
[1783]	1632	CALL netcdf_handle_error( 'data_output_2d', 54 )
[1]	1633	#endif
	1634
[493]	1635	ELSE
[1]	1636	!
[493]	1637	!-- First send the local index limits to PE0
[2512]	1638	ind(1) = nxl; ind(2) = nxr
	1639	ind(3) = nys; ind(4) = nyn
[1320]	1640	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1641	comm2d, ierr )
[1]	1642	!
[493]	1643	!-- Send data to PE0
[2512]	1644	CALL MPI_SEND( local_2d(nxl,nys), ngp, &
[493]	1645	MPI_REAL, 0, 1, comm2d, ierr )
	1646	ENDIF
	1647	!
	1648	!-- A barrier has to be set, because otherwise some PEs may
	1649	!-- proceed too fast so that PE0 may receive wrong data on
	1650	!-- tag 0
	1651	CALL MPI_BARRIER( comm2d, ierr )
[1]	1652	ENDIF
[493]	1653
[1]	1654	ENDIF
	1655	#else
	1656	#if defined( __netcdf )
[1327]	1657	IF ( two_d ) THEN
	1658	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1659	id_var_do2d(av,if), &
[2512]	1660	local_2d(nxl:nxr,nys:nyn), &
[1327]	1661	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
[2512]	1662	count = (/ nx+1, ny+1, 1, 1 /) )
[1327]	1663	ELSE
	1664	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1665	id_var_do2d(av,if), &
[2512]	1666	local_2d(nxl:nxr,nys:nyn), &
[1327]	1667	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
[2512]	1668	count = (/ nx+1, ny+1, 1, 1 /) )
[1]	1669	ENDIF
[1783]	1670	CALL netcdf_handle_error( 'data_output_2d', 447 )
[1]	1671	#endif
	1672	#endif
[2277]	1673
[1]	1674	!
	1675	!-- For 2D-arrays (e.g. u*) only one cross-section is available.
	1676	!-- Hence exit loop of output levels.
	1677	IF ( two_d ) THEN
[1703]	1678	IF ( netcdf_data_format < 5 ) two_d = .FALSE.
[1]	1679	EXIT loop1
	1680	ENDIF
	1681
	1682	CASE ( 'xz' )
	1683	!
[1308]	1684	!-- Update the netCDF xz cross section time axis.
	1685	!-- In case of parallel output, this is only done by PE0
	1686	!-- to increase the performance.
	1687	IF ( simulated_time /= do2d_xz_last_time(av) ) THEN
	1688	do2d_xz_time_count(av) = do2d_xz_time_count(av) + 1
	1689	do2d_xz_last_time(av) = simulated_time
	1690	IF ( myid == 0 ) THEN
[1327]	1691	IF ( .NOT. data_output_2d_on_each_pe &
	1692	.OR. netcdf_data_format > 4 ) &
[493]	1693	THEN
[1]	1694	#if defined( __netcdf )
	1695	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1696	id_var_time_xz(av), &
[291]	1697	(/ time_since_reference_point /), &
[1]	1698	start = (/ do2d_xz_time_count(av) /), &
	1699	count = (/ 1 /) )
[1783]	1700	CALL netcdf_handle_error( 'data_output_2d', 56 )
[1]	1701	#endif
	1702	ENDIF
	1703	ENDIF
	1704	ENDIF
[667]	1705
[1]	1706	!
	1707	!-- If required, carry out averaging along y
[1960]	1708	IF ( section(is,s_ind) == -1 ) THEN
[1]	1709
[2512]	1710	ALLOCATE( local_2d_l(nxl:nxr,nzb_do:nzt_do) )
[1353]	1711	local_2d_l = 0.0_wp
[2512]	1712	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
[1]	1713	!
	1714	!-- First local averaging on the PE
[1551]	1715	DO k = nzb_do, nzt_do
[1]	1716	DO j = nys, nyn
[2512]	1717	DO i = nxl, nxr
[1320]	1718	local_2d_l(i,k) = local_2d_l(i,k) + &
[1]	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 y
[622]	1726	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[2512]	1727	CALL MPI_ALLREDUCE( local_2d_l(nxl,nzb_do), &
	1728	local_2d(nxl,nzb_do), ngp, MPI_REAL, &
[1]	1729	MPI_SUM, comm1dy, ierr )
	1730	#else
	1731	local_2d = local_2d_l
	1732	#endif
[1353]	1733	local_2d = local_2d / ( ny + 1.0_wp )
[1]	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!)
[1960]	1741	IF ( section(is,s_ind) >= nys .AND. section(is,s_ind) <= nyn ) &
[1]	1742	THEN
[1960]	1743	local_2d = local_pf(:,section(is,s_ind),nzb_do:nzt_do)
[1]	1744	ENDIF
	1745
	1746	ENDIF
	1747
	1748	#if defined( __parallel )
[1327]	1749	IF ( netcdf_data_format > 4 ) THEN
[1]	1750	!
[1031]	1751	!-- Output in netCDF4/HDF5 format.
[493]	1752	!-- Output only on those PEs where the respective cross
	1753	!-- sections reside. Cross sections averaged along y are
	1754	!-- output on the respective first PE along y (myidy=0).
[1960]	1755	IF ( ( section(is,s_ind) >= nys .AND. &
	1756	section(is,s_ind) <= nyn ) .OR. &
	1757	( section(is,s_ind) == -1 .AND. myidy == 0 ) ) THEN
[1]	1758	#if defined( __netcdf )
[493]	1759	!
[1308]	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.
[2512]	1763	DO i = nxl, nxr
[1551]	1764	DO k = nzb_do, nzt_do
[1308]	1765	local_2d_sections_l(i,is,k) = local_2d(i,k)
	1766	ENDDO
	1767	ENDDO
[1]	1768	#endif
	1769	ENDIF
	1770
	1771	ELSE
	1772
[493]	1773	IF ( data_output_2d_on_each_pe ) THEN
[1]	1774	!
[493]	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 )
[1327]	1779	IF ( myid == 0 ) THEN
[1320]	1780	WRITE ( 22 ) time_since_reference_point, &
[493]	1781	do2d_xz_time_count(av), av
	1782	ENDIF
	1783	#endif
[759]	1784	DO i = 0, io_blocks-1
	1785	IF ( i == io_group ) THEN
[1960]	1786	IF ( ( section(is,s_ind) >= nys .AND. &
	1787	section(is,s_ind) <= nyn ) .OR. &
	1788	( section(is,s_ind) == -1 .AND. &
[1320]	1789	nys-1 == -1 ) ) &
[759]	1790	THEN
[2512]	1791	WRITE (22) nxl, nxr, nzb_do, nzt_do, nzb, nzt+1
[759]	1792	WRITE (22) local_2d
	1793	ELSE
[1551]	1794	WRITE (22) -1, -1, -1, -1, -1, -1
[759]	1795	ENDIF
	1796	ENDIF
	1797	#if defined( __parallel )
	1798	CALL MPI_BARRIER( comm2d, ierr )
	1799	#endif
	1800	ENDDO
[493]	1801
	1802	ELSE
[1]	1803	!
[493]	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
[2512]	1810	ngp = ( nxr-nxl + 1 ) * ( nzt_do-nzb_do + 1 )
[493]	1811	IF ( myid == 0 ) THEN
[1]	1812	!
[493]	1813	!-- Local array can be relocated directly.
[1960]	1814	IF ( ( section(is,s_ind) >= nys .AND. &
	1815	section(is,s_ind) <= nyn ) .OR. &
	1816	( section(is,s_ind) == -1 .AND. &
	1817	nys-1 == -1 ) ) THEN
[2512]	1818	total_2d(nxl:nxr,nzb_do:nzt_do) = local_2d
[493]	1819	ENDIF
[1]	1820	!
[493]	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.
[1320]	1827	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1828	MPI_ANY_SOURCE, 0, comm2d, &
[1]	1829	status, ierr )
[493]	1830	!
	1831	!-- Not all PEs have data for XZ-cross-section.
	1832	IF ( ind(1) /= -9999 ) THEN
	1833	sender = status(MPI_SOURCE)
	1834	DEALLOCATE( local_2d )
[1320]	1835	ALLOCATE( local_2d(ind(1):ind(2), &
[493]	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 )
[1320]	1840	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
[493]	1841	local_2d
	1842	ENDIF
	1843	ENDDO
	1844	!
	1845	!-- Relocate the local array for the next loop increment
	1846	DEALLOCATE( local_2d )
[2512]	1847	ALLOCATE( local_2d(nxl:nxr,nzb_do:nzt_do) )
[1]	1848
	1849	#if defined( __netcdf )
[2512]	1850	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1851	id_var_do2d(av,if), &
	1852	total_2d(0:nx,nzb_do:nzt_do), &
	1853	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
	1854	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
[1783]	1855	CALL netcdf_handle_error( 'data_output_2d', 58 )
[1]	1856	#endif
	1857
[493]	1858	ELSE
[1]	1859	!
[493]	1860	!-- If the cross section resides on the PE, send the
	1861	!-- local index limits, otherwise send -9999 to PE0.
[1960]	1862	IF ( ( section(is,s_ind) >= nys .AND. &
	1863	section(is,s_ind) <= nyn ) .OR. &
	1864	( section(is,s_ind) == -1 .AND. nys-1 == -1 ) ) &
[493]	1865	THEN
[2512]	1866	ind(1) = nxl; ind(2) = nxr
[1551]	1867	ind(3) = nzb_do; ind(4) = nzt_do
[493]	1868	ELSE
	1869	ind(1) = -9999; ind(2) = -9999
	1870	ind(3) = -9999; ind(4) = -9999
	1871	ENDIF
[1320]	1872	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1873	comm2d, ierr )
	1874	!
	1875	!-- If applicable, send data to PE0.
	1876	IF ( ind(1) /= -9999 ) THEN
[2512]	1877	CALL MPI_SEND( local_2d(nxl,nzb_do), ngp, &
[493]	1878	MPI_REAL, 0, 1, comm2d, ierr )
	1879	ENDIF
[1]	1880	ENDIF
	1881	!
[493]	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 )
[1]	1886	ENDIF
[493]	1887
[1]	1888	ENDIF
	1889	#else
	1890	#if defined( __netcdf )
[1327]	1891	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1892	id_var_do2d(av,if), &
[2512]	1893	local_2d(nxl:nxr,nzb_do:nzt_do), &
[1327]	1894	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
[2512]	1895	count = (/ nx+1, 1, nzt_do-nzb_do+1, 1 /) )
[1783]	1896	CALL netcdf_handle_error( 'data_output_2d', 451 )
[1]	1897	#endif
	1898	#endif
	1899
	1900	CASE ( 'yz' )
	1901	!
[1308]	1902	!-- Update the netCDF yz cross section time axis.
	1903	!-- In case of parallel output, this is only done by PE0
	1904	!-- to increase the performance.
	1905	IF ( simulated_time /= do2d_yz_last_time(av) ) THEN
	1906	do2d_yz_time_count(av) = do2d_yz_time_count(av) + 1
	1907	do2d_yz_last_time(av) = simulated_time
	1908	IF ( myid == 0 ) THEN
[1327]	1909	IF ( .NOT. data_output_2d_on_each_pe &
	1910	.OR. netcdf_data_format > 4 ) &
[493]	1911	THEN
[1]	1912	#if defined( __netcdf )
	1913	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1914	id_var_time_yz(av), &
[291]	1915	(/ time_since_reference_point /), &
[1]	1916	start = (/ do2d_yz_time_count(av) /), &
	1917	count = (/ 1 /) )
[1783]	1918	CALL netcdf_handle_error( 'data_output_2d', 59 )
[1]	1919	#endif
	1920	ENDIF
	1921	ENDIF
[1308]	1922	ENDIF
[493]	1923
[1]	1924	!
	1925	!-- If required, carry out averaging along x
[1960]	1926	IF ( section(is,s_ind) == -1 ) THEN
[1]	1927
[2512]	1928	ALLOCATE( local_2d_l(nys:nyn,nzb_do:nzt_do) )
[1353]	1929	local_2d_l = 0.0_wp
[2512]	1930	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
[1]	1931	!
	1932	!-- First local averaging on the PE
[1551]	1933	DO k = nzb_do, nzt_do
[2512]	1934	DO j = nys, nyn
[1]	1935	DO i = nxl, nxr
[1320]	1936	local_2d_l(j,k) = local_2d_l(j,k) + &
[1]	1937	local_pf(i,j,k)
	1938	ENDDO
	1939	ENDDO
	1940	ENDDO
	1941	#if defined( __parallel )
	1942	!
	1943	!-- Now do the averaging over all PEs along x
[622]	1944	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[2512]	1945	CALL MPI_ALLREDUCE( local_2d_l(nys,nzb_do), &
	1946	local_2d(nys,nzb_do), ngp, MPI_REAL, &
[1]	1947	MPI_SUM, comm1dx, ierr )
	1948	#else
	1949	local_2d = local_2d_l
	1950	#endif
[1353]	1951	local_2d = local_2d / ( nx + 1.0_wp )
[1]	1952
	1953	DEALLOCATE( local_2d_l )
	1954
	1955	ELSE
	1956	!
	1957	!-- Just store the respective section on the local array
	1958	!-- (but only if it is available on this PE!)
[1960]	1959	IF ( section(is,s_ind) >= nxl .AND. section(is,s_ind) <= nxr ) &
[1]	1960	THEN
[1960]	1961	local_2d = local_pf(section(is,s_ind),:,nzb_do:nzt_do)
[1]	1962	ENDIF
	1963
	1964	ENDIF
	1965
	1966	#if defined( __parallel )
[1327]	1967	IF ( netcdf_data_format > 4 ) THEN
[1]	1968	!
[1031]	1969	!-- Output in netCDF4/HDF5 format.
[493]	1970	!-- Output only on those PEs where the respective cross
	1971	!-- sections reside. Cross sections averaged along x are
	1972	!-- output on the respective first PE along x (myidx=0).
[1960]	1973	IF ( ( section(is,s_ind) >= nxl .AND. &
	1974	section(is,s_ind) <= nxr ) .OR. &
	1975	( section(is,s_ind) == -1 .AND. myidx == 0 ) ) THEN
[1]	1976	#if defined( __netcdf )
[493]	1977	!
[1308]	1978	!-- For parallel output, all cross sections are first
	1979	!-- stored here on a local array and will be written to the
	1980	!-- output file afterwards to increase the performance.
[2512]	1981	DO j = nys, nyn
[1551]	1982	DO k = nzb_do, nzt_do
[1308]	1983	local_2d_sections_l(is,j,k) = local_2d(j,k)
	1984	ENDDO
	1985	ENDDO
[1]	1986	#endif
	1987	ENDIF
	1988
	1989	ELSE
	1990
[493]	1991	IF ( data_output_2d_on_each_pe ) THEN
[1]	1992	!
[493]	1993	!-- Output of partial arrays on each PE. If the cross
	1994	!-- section does not reside on the PE, output special
	1995	!-- index values.
	1996	#if defined( __netcdf )
[1327]	1997	IF ( myid == 0 ) THEN
[1320]	1998	WRITE ( 23 ) time_since_reference_point, &
[493]	1999	do2d_yz_time_count(av), av
	2000	ENDIF
	2001	#endif
[759]	2002	DO i = 0, io_blocks-1
	2003	IF ( i == io_group ) THEN
[1960]	2004	IF ( ( section(is,s_ind) >= nxl .AND. &
	2005	section(is,s_ind) <= nxr ) .OR. &
	2006	( section(is,s_ind) == -1 .AND. &
[1320]	2007	nxl-1 == -1 ) ) &
[759]	2008	THEN
[2512]	2009	WRITE (23) nys, nyn, nzb_do, nzt_do, nzb, nzt+1
[759]	2010	WRITE (23) local_2d
	2011	ELSE
[1551]	2012	WRITE (23) -1, -1, -1, -1, -1, -1
[759]	2013	ENDIF
	2014	ENDIF
	2015	#if defined( __parallel )
	2016	CALL MPI_BARRIER( comm2d, ierr )
	2017	#endif
	2018	ENDDO
[493]	2019
	2020	ELSE
[1]	2021	!
[493]	2022	!-- PE0 receives partial arrays from all processors of the
	2023	!-- respective cross section and outputs them. Here a
	2024	!-- barrier has to be set, because otherwise
	2025	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
	2026	CALL MPI_BARRIER( comm2d, ierr )
	2027
[2512]	2028	ngp = ( nyn-nys+1 ) * ( nzt_do-nzb_do+1 )
[493]	2029	IF ( myid == 0 ) THEN
[1]	2030	!
[493]	2031	!-- Local array can be relocated directly.
[1960]	2032	IF ( ( section(is,s_ind) >= nxl .AND. &
	2033	section(is,s_ind) <= nxr ) .OR. &
	2034	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
[493]	2035	THEN
[2512]	2036	total_2d(nys:nyn,nzb_do:nzt_do) = local_2d
[493]	2037	ENDIF
[1]	2038	!
[493]	2039	!-- Receive data from all other PEs.
	2040	DO n = 1, numprocs-1
	2041	!
	2042	!-- Receive index limits first, then array.
	2043	!-- Index limits are received in arbitrary order from
	2044	!-- the PEs.
[1320]	2045	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	2046	MPI_ANY_SOURCE, 0, comm2d, &
[1]	2047	status, ierr )
[493]	2048	!
	2049	!-- Not all PEs have data for YZ-cross-section.
	2050	IF ( ind(1) /= -9999 ) THEN
	2051	sender = status(MPI_SOURCE)
	2052	DEALLOCATE( local_2d )
[1320]	2053	ALLOCATE( local_2d(ind(1):ind(2), &
[493]	2054	ind(3):ind(4)) )
	2055	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	2056	MPI_REAL, sender, 1, comm2d, &
	2057	status, ierr )
[1320]	2058	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
[493]	2059	local_2d
	2060	ENDIF
	2061	ENDDO
	2062	!
	2063	!-- Relocate the local array for the next loop increment
	2064	DEALLOCATE( local_2d )
[2512]	2065	ALLOCATE( local_2d(nys:nyn,nzb_do:nzt_do) )
[1]	2066
	2067	#if defined( __netcdf )
[2512]	2068	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	2069	id_var_do2d(av,if), &
	2070	total_2d(0:ny,nzb_do:nzt_do), &
	2071	start = (/ is, 1, 1, do2d_yz_time_count(av) /), &
	2072	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
[1783]	2073	CALL netcdf_handle_error( 'data_output_2d', 61 )
[1]	2074	#endif
	2075
[493]	2076	ELSE
[1]	2077	!
[493]	2078	!-- If the cross section resides on the PE, send the
	2079	!-- local index limits, otherwise send -9999 to PE0.
[1960]	2080	IF ( ( section(is,s_ind) >= nxl .AND. &
	2081	section(is,s_ind) <= nxr ) .OR. &
	2082	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
[493]	2083	THEN
[2512]	2084	ind(1) = nys; ind(2) = nyn
[1551]	2085	ind(3) = nzb_do; ind(4) = nzt_do
[493]	2086	ELSE
	2087	ind(1) = -9999; ind(2) = -9999
	2088	ind(3) = -9999; ind(4) = -9999
	2089	ENDIF
[1320]	2090	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	2091	comm2d, ierr )
	2092	!
	2093	!-- If applicable, send data to PE0.
	2094	IF ( ind(1) /= -9999 ) THEN
[2512]	2095	CALL MPI_SEND( local_2d(nys,nzb_do), ngp, &
[493]	2096	MPI_REAL, 0, 1, comm2d, ierr )
	2097	ENDIF
[1]	2098	ENDIF
	2099	!
[493]	2100	!-- A barrier has to be set, because otherwise some PEs may
	2101	!-- proceed too fast so that PE0 may receive wrong data on
	2102	!-- tag 0
	2103	CALL MPI_BARRIER( comm2d, ierr )
[1]	2104	ENDIF
[493]	2105
[1]	2106	ENDIF
	2107	#else
	2108	#if defined( __netcdf )
[1327]	2109	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	2110	id_var_do2d(av,if), &
[2512]	2111	local_2d(nys:nyn,nzb_do:nzt_do), &
[1327]	2112	start = (/ is, 1, 1, do2d_xz_time_count(av) /), &
[2512]	2113	count = (/ 1, ny+1, nzt_do-nzb_do+1, 1 /) )
[1783]	2114	CALL netcdf_handle_error( 'data_output_2d', 452 )
[1]	2115	#endif
	2116	#endif
	2117
	2118	END SELECT
	2119
	2120	is = is + 1
	2121	ENDDO loop1
	2122
[1308]	2123	!
	2124	!-- For parallel output, all data were collected before on a local array
	2125	!-- and are written now to the netcdf file. This must be done to increase
	2126	!-- the performance of the parallel output.
	2127	#if defined( __netcdf )
[1327]	2128	IF ( netcdf_data_format > 4 ) THEN
[1308]	2129
	2130	SELECT CASE ( mode )
	2131
	2132	CASE ( 'xy' )
	2133	IF ( two_d ) THEN
[1703]	2134	nis = 1
	2135	two_d = .FALSE.
[1308]	2136	ELSE
[1703]	2137	nis = ns
[1308]	2138	ENDIF
	2139	!
	2140	!-- Do not output redundant ghost point data except for the
	2141	!-- boundaries of the total domain.
[2512]	2142	! IF ( nxr == nx .AND. nyn /= ny ) THEN
	2143	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	2144	! id_var_do2d(av,if), &
	2145	! local_2d_sections(nxl:nxr+1, &
	2146	! nys:nyn,1:nis), &
	2147	! start = (/ nxl+1, nys+1, 1, &
	2148	! do2d_xy_time_count(av) /), &
	2149	! count = (/ nxr-nxl+2, &
	2150	! nyn-nys+1, nis, 1 &
	2151	! /) )
	2152	! ELSEIF ( nxr /= nx .AND. nyn == ny ) THEN
	2153	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	2154	! id_var_do2d(av,if), &
	2155	! local_2d_sections(nxl:nxr, &
	2156	! nys:nyn+1,1:nis), &
	2157	! start = (/ nxl+1, nys+1, 1, &
	2158	! do2d_xy_time_count(av) /), &
	2159	! count = (/ nxr-nxl+1, &
	2160	! nyn-nys+2, nis, 1 &
	2161	! /) )
	2162	! ELSEIF ( nxr == nx .AND. nyn == ny ) THEN
	2163	! nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	2164	! id_var_do2d(av,if), &
	2165	! local_2d_sections(nxl:nxr+1, &
	2166	! nys:nyn+1,1:nis), &
	2167	! start = (/ nxl+1, nys+1, 1, &
	2168	! do2d_xy_time_count(av) /), &
	2169	! count = (/ nxr-nxl+2, &
	2170	! nyn-nys+2, nis, 1 &
	2171	! /) )
	2172	! ELSE
[1308]	2173	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	2174	id_var_do2d(av,if), &
	2175	local_2d_sections(nxl:nxr, &
[1703]	2176	nys:nyn,1:nis), &
[1308]	2177	start = (/ nxl+1, nys+1, 1, &
	2178	do2d_xy_time_count(av) /), &
	2179	count = (/ nxr-nxl+1, &
[1703]	2180	nyn-nys+1, nis, 1 &
[1308]	2181	/) )
[2512]	2182	! ENDIF
[1308]	2183
[1783]	2184	CALL netcdf_handle_error( 'data_output_2d', 55 )
[1308]	2185
	2186	CASE ( 'xz' )
	2187	!
	2188	!-- First, all PEs get the information of all cross-sections.
	2189	!-- Then the data are written to the output file by all PEs
	2190	!-- while NF90_COLLECTIVE is set in subroutine
	2191	!-- define_netcdf_header. Although redundant information are
	2192	!-- written to the output file in that case, the performance
	2193	!-- is significantly better compared to the case where only
	2194	!-- the first row of PEs in x-direction (myidx = 0) is given
	2195	!-- the output while NF90_INDEPENDENT is set.
	2196	IF ( npey /= 1 ) THEN
	2197
	2198	#if defined( __parallel )
	2199	!
	2200	!-- Distribute data over all PEs along y
[2512]	2201	ngp = ( nxr-nxl+1 ) * ( nzt_do-nzb_do+1 ) * ns
[1308]	2202	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[2512]	2203	CALL MPI_ALLREDUCE( local_2d_sections_l(nxl,1,nzb_do), &
	2204	local_2d_sections(nxl,1,nzb_do), &
[1308]	2205	ngp, MPI_REAL, MPI_SUM, comm1dy, &
	2206	ierr )
	2207	#else
	2208	local_2d_sections = local_2d_sections_l
	2209	#endif
	2210	ENDIF
	2211	!
	2212	!-- Do not output redundant ghost point data except for the
	2213	!-- boundaries of the total domain.
[2512]	2214	! IF ( nxr == nx ) THEN
	2215	! nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	2216	! id_var_do2d(av,if), &
	2217	! local_2d_sections(nxl:nxr+1,1:ns, &
	2218	! nzb_do:nzt_do), &
	2219	! start = (/ nxl+1, 1, 1, &
	2220	! do2d_xz_time_count(av) /), &
	2221	! count = (/ nxr-nxl+2, ns, nzt_do-nzb_do+1, &
	2222	! 1 /) )
	2223	! ELSE
[1308]	2224	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	2225	id_var_do2d(av,if), &
	2226	local_2d_sections(nxl:nxr,1:ns, &
[1551]	2227	nzb_do:nzt_do), &
[1308]	2228	start = (/ nxl+1, 1, 1, &
	2229	do2d_xz_time_count(av) /), &
[1551]	2230	count = (/ nxr-nxl+1, ns, nzt_do-nzb_do+1, &
[1308]	2231	1 /) )
[2512]	2232	! ENDIF
[1308]	2233
[1783]	2234	CALL netcdf_handle_error( 'data_output_2d', 57 )
[1308]	2235
	2236	CASE ( 'yz' )
	2237	!
	2238	!-- First, all PEs get the information of all cross-sections.
	2239	!-- Then the data are written to the output file by all PEs
	2240	!-- while NF90_COLLECTIVE is set in subroutine
	2241	!-- define_netcdf_header. Although redundant information are
	2242	!-- written to the output file in that case, the performance
	2243	!-- is significantly better compared to the case where only
	2244	!-- the first row of PEs in y-direction (myidy = 0) is given
	2245	!-- the output while NF90_INDEPENDENT is set.
	2246	IF ( npex /= 1 ) THEN
	2247
	2248	#if defined( __parallel )
	2249	!
	2250	!-- Distribute data over all PEs along x
[2512]	2251	ngp = ( nyn-nys+1 ) * ( nzt-nzb + 2 ) * ns
[1308]	2252	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[2512]	2253	CALL MPI_ALLREDUCE( local_2d_sections_l(1,nys,nzb_do), &
	2254	local_2d_sections(1,nys,nzb_do), &
[1308]	2255	ngp, MPI_REAL, MPI_SUM, comm1dx, &
	2256	ierr )
	2257	#else
	2258	local_2d_sections = local_2d_sections_l
	2259	#endif
	2260	ENDIF
	2261	!
	2262	!-- Do not output redundant ghost point data except for the
	2263	!-- boundaries of the total domain.
[2512]	2264	! IF ( nyn == ny ) THEN
	2265	! nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	2266	! id_var_do2d(av,if), &
	2267	! local_2d_sections(1:ns, &
	2268	! nys:nyn+1,nzb_do:nzt_do), &
	2269	! start = (/ 1, nys+1, 1, &
	2270	! do2d_yz_time_count(av) /), &
	2271	! count = (/ ns, nyn-nys+2, &
	2272	! nzt_do-nzb_do+1, 1 /) )
	2273	! ELSE
[1308]	2274	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	2275	id_var_do2d(av,if), &
	2276	local_2d_sections(1:ns,nys:nyn, &
[1551]	2277	nzb_do:nzt_do), &
[1308]	2278	start = (/ 1, nys+1, 1, &
	2279	do2d_yz_time_count(av) /), &
	2280	count = (/ ns, nyn-nys+1, &
[1551]	2281	nzt_do-nzb_do+1, 1 /) )
[2512]	2282	! ENDIF
[1308]	2283
[1783]	2284	CALL netcdf_handle_error( 'data_output_2d', 60 )
[1308]	2285
	2286	CASE DEFAULT
	2287	message_string = 'unknown cross-section: ' // TRIM( mode )
	2288	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
	2289
	2290	END SELECT
	2291
	2292	ENDIF
[1311]	2293	#endif
[1]	2294	ENDIF
	2295
	2296	if = if + 1
	2297	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
	2298	do2d_mode = do2d(av,if)(l-1:l)
	2299
	2300	ENDDO
	2301
	2302	!
	2303	!-- Deallocate temporary arrays.
	2304	IF ( ALLOCATED( level_z ) ) DEALLOCATE( level_z )
[1308]	2305	IF ( netcdf_data_format > 4 ) THEN
	2306	DEALLOCATE( local_pf, local_2d, local_2d_sections )
	2307	IF( mode == 'xz' .OR. mode == 'yz' ) DEALLOCATE( local_2d_sections_l )
	2308	ENDIF
[1]	2309	#if defined( __parallel )
	2310	IF ( .NOT. data_output_2d_on_each_pe .AND. myid == 0 ) THEN
	2311	DEALLOCATE( total_2d )
	2312	ENDIF
	2313	#endif
	2314
	2315	!
	2316	!-- Close plot output file.
[1960]	2317	file_id = 20 + s_ind
[1]	2318
	2319	IF ( data_output_2d_on_each_pe ) THEN
[759]	2320	DO i = 0, io_blocks-1
	2321	IF ( i == io_group ) THEN
	2322	CALL close_file( file_id )
	2323	ENDIF
	2324	#if defined( __parallel )
	2325	CALL MPI_BARRIER( comm2d, ierr )
	2326	#endif
	2327	ENDDO
[1]	2328	ELSE
	2329	IF ( myid == 0 ) CALL close_file( file_id )
	2330	ENDIF
	2331
[1318]	2332	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
[1]	2333
	2334	END SUBROUTINE data_output_2d

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |