Home

Context Navigation

source: palm/trunk/SOURCE/data_output_2d.f90 @ 2286

Last change on this file since 2286 was 2277, checked in by kanani, 7 years ago
code documentation and cleanup
Property svn:keywords set to `Id`
File size: 84.3 KB

Rev	Line
[1682]	1	!> @file data_output_2d.f90
[2000]	2	!------------------------------------------------------------------------------!
[1036]	3	! This file is part of PALM.
	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	!
[2101]	17	! Copyright 1997-2017 Leibniz Universitaet Hannover
[2000]	18	!------------------------------------------------------------------------------!
[1036]	19	!
[254]	20	! Current revisions:
[1]	21	! -----------------
[1961]	22	!
[2233]	23	!
[1552]	24	! Former revisions:
	25	! -----------------
	26	! $Id: data_output_2d.f90 2277 2017-06-12 10:47:51Z raasch $
[2277]	27	! Removed unused variables do2d_xy_n, do2d_xz_n, do2d_yz_n
	28	!
	29	! 2233 2017-05-30 18:08:54Z suehring
[1552]	30	!
[2233]	31	! 2232 2017-05-30 17:47:52Z suehring
	32	! Adjustments to new surface concept
	33	!
	34	!
[2191]	35	! 2190 2017-03-21 12:16:43Z raasch
	36	! bugfix for r2031: string rho replaced by rho_ocean
	37	!
[2032]	38	! 2031 2016-10-21 15:11:58Z knoop
	39	! renamed variable rho to rho_ocean and rho_av to rho_ocean_av
	40	!
[2001]	41	! 2000 2016-08-20 18:09:15Z knoop
	42	! Forced header and separation lines into 80 columns
	43	!
[1981]	44	! 1980 2016-07-29 15:51:57Z suehring
	45	! Bugfix, in order to steer user-defined output, setting flag found explicitly
	46	! to .F.
	47	!
[1977]	48	! 1976 2016-07-27 13:28:04Z maronga
	49	! Output of radiation quantities is now done directly in the respective module
	50	!
[1973]	51	! 1972 2016-07-26 07:52:02Z maronga
[1976]	52	! Output of land surface quantities is now done directly in the respective
	53	! module
[1973]	54	!
[1961]	55	! 1960 2016-07-12 16:34:24Z suehring
	56	! Scalar surface flux added
	57	! Rename INTEGER variable s into s_ind, as s is already assigned to scalar
	58	!
[1851]	59	! 1849 2016-04-08 11:33:18Z hoffmann
	60	! precipitation_amount, precipitation_rate, prr moved to arrays_3d
	61	!
[1823]	62	! 1822 2016-04-07 07:49:42Z hoffmann
	63	! Output of bulk cloud physics simplified.
	64	!
[1789]	65	! 1788 2016-03-10 11:01:04Z maronga
	66	! Added output of z0q
	67	!
[1784]	68	! 1783 2016-03-06 18:36:17Z raasch
	69	! name change of netcdf routines and module + related changes
	70	!
[1746]	71	! 1745 2016-02-05 13:06:51Z gronemeier
	72	! Bugfix: test if time axis limit exceeds moved to point after call of check_open
	73	!
[1704]	74	! 1703 2015-11-02 12:38:44Z raasch
	75	! bugfix for output of single (*) xy-sections in case of parallel netcdf I/O
	76	!
[1702]	77	! 1701 2015-11-02 07:43:04Z maronga
	78	! Bugfix in output of RRTGM data
	79	!
[1692]	80	! 1691 2015-10-26 16:17:44Z maronga
	81	! Added output of Obukhov length (ol) and radiative heating rates for RRTMG.
	82	! Formatting corrections.
	83	!
[1683]	84	! 1682 2015-10-07 23:56:08Z knoop
	85	! Code annotations made doxygen readable
	86	!
[1586]	87	! 1585 2015-04-30 07:05:52Z maronga
	88	! Added support for RRTMG
	89	!
[1556]	90	! 1555 2015-03-04 17:44:27Z maronga
	91	! Added output of r_a and r_s
	92	!
[1552]	93	! 1551 2015-03-03 14:18:16Z maronga
[1551]	94	! Added suppport for land surface model and radiation model output. In the course
	95	! of this action, the limits for vertical loops have been changed (from nzb and
	96	! nzt+1 to nzb_do and nzt_do, respectively in order to allow soil model output).
	97	! Moreover, a new vertical grid zs was introduced.
[1329]	98	!
[1360]	99	! 1359 2014-04-11 17:15:14Z hoffmann
	100	! New particle structure integrated.
	101	!
[1354]	102	! 1353 2014-04-08 15:21:23Z heinze
	103	! REAL constants provided with KIND-attribute
	104	!
[1329]	105	! 1327 2014-03-21 11:00:16Z raasch
	106	! parts concerning iso2d output removed,
	107	! -netcdf output queries
	108	!
[1321]	109	! 1320 2014-03-20 08:40:49Z raasch
[1320]	110	! ONLY-attribute added to USE-statements,
	111	! kind-parameters added to all INTEGER and REAL declaration statements,
	112	! kinds are defined in new module kinds,
	113	! revision history before 2012 removed,
	114	! comment fields (!:) to be used for variable explanations added to
	115	! all variable declaration statements
[1309]	116	!
[1319]	117	! 1318 2014-03-17 13:35:16Z raasch
	118	! barrier argument removed from cpu_log.
	119	! module interfaces removed
	120	!
[1312]	121	! 1311 2014-03-14 12:13:39Z heinze
	122	! bugfix: close #if defined( __netcdf )
	123	!
[1309]	124	! 1308 2014-03-13 14:58:42Z fricke
[1308]	125	! +local_2d_sections, local_2d_sections_l, ns
	126	! Check, if the limit of the time dimension is exceeded for parallel output
	127	! To increase the performance for parallel output, the following is done:
	128	! - Update of time axis is only done by PE0
	129	! - Cross sections are first stored on a local array and are written
	130	! collectively to the output file by all PEs.
[674]	131	!
[1116]	132	! 1115 2013-03-26 18:16:16Z hoffmann
	133	! ql is calculated by calc_liquid_water_content
	134	!
[1077]	135	! 1076 2012-12-05 08:30:18Z hoffmann
	136	! Bugfix in output of ql
	137	!
[1066]	138	! 1065 2012-11-22 17:42:36Z hoffmann
	139	! Bugfix: Output of cross sections of ql
	140	!
[1054]	141	! 1053 2012-11-13 17:11:03Z hoffmann
	142	! +qr, nr, qc and cross sections
	143	!
[1037]	144	! 1036 2012-10-22 13:43:42Z raasch
	145	! code put under GPL (PALM 3.9)
	146	!
[1035]	147	! 1031 2012-10-19 14:35:30Z raasch
	148	! netCDF4 without parallel file support implemented
	149	!
[1008]	150	! 1007 2012-09-19 14:30:36Z franke
	151	! Bugfix: missing calculation of ql_vp added
	152	!
[979]	153	! 978 2012-08-09 08:28:32Z fricke
	154	! +z0h
	155	!
[1]	156	! Revision 1.1 1997/08/11 06:24:09 raasch
	157	! Initial revision
	158	!
	159	!
	160	! Description:
	161	! ------------
[1682]	162	!> Data output of horizontal cross-sections in netCDF format or binary format
	163	!> compatible to old graphic software iso2d.
	164	!> Attention: The position of the sectional planes is still not always computed
	165	!> --------- correctly. (zu is used always)!
[1]	166	!------------------------------------------------------------------------------!
[1682]	167	SUBROUTINE data_output_2d( mode, av )
	168
[1]	169
[1320]	170	USE arrays_3d, &
[2232]	171	ONLY: dzw, e, nr, p, pt, precipitation_amount, precipitation_rate, &
	172	prr, q, qc, ql, ql_c, ql_v, ql_vp, qr, rho_ocean, s, sa, &
	173	tend, u, v, vpt, w, zu, zw
[1320]	174
[1]	175	USE averaging
[1320]	176
	177	USE cloud_parameters, &
[1849]	178	ONLY: hyrho, l_d_cp, pt_d_t
[1320]	179
	180	USE control_parameters, &
	181	ONLY: cloud_physics, data_output_2d_on_each_pe, data_output_xy, &
	182	data_output_xz, data_output_yz, do2d, &
[2277]	183	do2d_xy_last_time, do2d_xy_time_count, &
	184	do2d_xz_last_time, do2d_xz_time_count, &
	185	do2d_yz_last_time, do2d_yz_time_count, &
[2232]	186	ibc_uv_b, io_blocks, io_group, land_surface, message_string, &
[1822]	187	ntdim_2d_xy, ntdim_2d_xz, ntdim_2d_yz, &
	188	psolver, section, simulated_time, simulated_time_chr, &
	189	time_since_reference_point
[1320]	190
	191	USE cpulog, &
	192	ONLY: cpu_log, log_point
	193
	194	USE grid_variables, &
	195	ONLY: dx, dy
	196
	197	USE indices, &
	198	ONLY: nbgp, nx, nxl, nxlg, nxr, nxrg, ny, nyn, nyng, nys, nysg, &
	199	nz, nzb, nzt
	200
	201	USE kinds
[1551]	202
	203	USE land_surface_model_mod, &
[2232]	204	ONLY: lsm_data_output_2d, zs
[1551]	205
[1783]	206	#if defined( __netcdf )
	207	USE NETCDF
	208	#endif
[1320]	209
[1783]	210	USE netcdf_interface, &
	211	ONLY: id_set_xy, id_set_xz, id_set_yz, id_var_do2d, id_var_time_xy, &
	212	id_var_time_xz, id_var_time_yz, nc_stat, netcdf_data_format, &
	213	netcdf_handle_error
	214
[1320]	215	USE particle_attributes, &
[1359]	216	ONLY: grid_particles, number_of_particles, particle_advection_start, &
	217	particles, prt_count
[1320]	218
[1]	219	USE pegrid
	220
[1551]	221	USE radiation_model_mod, &
[1976]	222	ONLY: radiation, radiation_data_output_2d
[1551]	223
[2232]	224	USE surface_mod, &
	225	ONLY: surf_def_h, surf_lsm_h, surf_usm_h
	226
[1]	227	IMPLICIT NONE
	228
[1682]	229	CHARACTER (LEN=2) :: do2d_mode !<
	230	CHARACTER (LEN=2) :: mode !<
	231	CHARACTER (LEN=4) :: grid !<
	232	CHARACTER (LEN=25) :: section_chr !<
	233	CHARACTER (LEN=50) :: rtext !<
[1320]	234
[1682]	235	INTEGER(iwp) :: av !<
	236	INTEGER(iwp) :: ngp !<
	237	INTEGER(iwp) :: file_id !<
	238	INTEGER(iwp) :: i !<
	239	INTEGER(iwp) :: if !<
	240	INTEGER(iwp) :: is !<
	241	INTEGER(iwp) :: iis !<
	242	INTEGER(iwp) :: j !<
	243	INTEGER(iwp) :: k !<
	244	INTEGER(iwp) :: l !<
	245	INTEGER(iwp) :: layer_xy !<
[2232]	246	INTEGER(iwp) :: m !<
[1682]	247	INTEGER(iwp) :: n !<
[1703]	248	INTEGER(iwp) :: nis !<
[1682]	249	INTEGER(iwp) :: ns !<
	250	INTEGER(iwp) :: nzb_do !< lower limit of the data field (usually nzb)
	251	INTEGER(iwp) :: nzt_do !< upper limit of the data field (usually nzt+1)
	252	INTEGER(iwp) :: psi !<
[1960]	253	INTEGER(iwp) :: s_ind !<
[1682]	254	INTEGER(iwp) :: sender !<
	255	INTEGER(iwp) :: ind(4) !<
[1320]	256
[1682]	257	LOGICAL :: found !<
	258	LOGICAL :: resorted !<
	259	LOGICAL :: two_d !<
[1320]	260
[1682]	261	REAL(wp) :: mean_r !<
	262	REAL(wp) :: s_r2 !<
	263	REAL(wp) :: s_r3 !<
[1320]	264
[1682]	265	REAL(wp), DIMENSION(:), ALLOCATABLE :: level_z !<
	266	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d !<
	267	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: local_2d_l !<
[2232]	268	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: tmp_2d !< temporary field used to exchange surface-related quantities
	269
[1682]	270	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_pf !<
	271	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections !<
	272	REAL(wp), DIMENSION(:,:,:), ALLOCATABLE :: local_2d_sections_l !<
[1359]	273
[1]	274	#if defined( __parallel )
[1682]	275	REAL(wp), DIMENSION(:,:), ALLOCATABLE :: total_2d !<
[1]	276	#endif
[1682]	277	REAL(wp), DIMENSION(:,:,:), POINTER :: to_be_resorted !<
[1]	278
	279	NAMELIST /LOCAL/ rtext
	280
	281	!
	282	!-- Immediate return, if no output is requested (no respective sections
	283	!-- found in parameter data_output)
	284	IF ( mode == 'xy' .AND. .NOT. data_output_xy(av) ) RETURN
	285	IF ( mode == 'xz' .AND. .NOT. data_output_xz(av) ) RETURN
	286	IF ( mode == 'yz' .AND. .NOT. data_output_yz(av) ) RETURN
	287
[1308]	288	CALL cpu_log (log_point(3),'data_output_2d','start')
	289
[1]	290	two_d = .FALSE. ! local variable to distinguish between output of pure 2D
	291	! arrays and cross-sections of 3D arrays.
	292
	293	!
	294	!-- Depending on the orientation of the cross-section, the respective output
	295	!-- files have to be opened.
	296	SELECT CASE ( mode )
	297
	298	CASE ( 'xy' )
[1960]	299	s_ind = 1
[667]	300	ALLOCATE( level_z(nzb:nzt+1), local_2d(nxlg:nxrg,nysg:nyng) )
[1]	301
[1308]	302	IF ( netcdf_data_format > 4 ) THEN
	303	ns = 1
[1960]	304	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	305	ns = ns + 1
	306	ENDDO
	307	ns = ns - 1
	308	ALLOCATE( local_2d_sections(nxlg:nxrg,nysg:nyng,1:ns) )
[1353]	309	local_2d_sections = 0.0_wp
[1308]	310	ENDIF
	311
[493]	312	!
[1031]	313	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	314	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	315	CALL check_open( 101+av*10 )
	316	ENDIF
[1]	317	IF ( data_output_2d_on_each_pe ) THEN
	318	CALL check_open( 21 )
	319	ELSE
	320	IF ( myid == 0 ) THEN
	321	#if defined( __parallel )
[667]	322	ALLOCATE( total_2d(-nbgp:nx+nbgp,-nbgp:ny+nbgp) )
[1]	323	#endif
	324	ENDIF
	325	ENDIF
	326
	327	CASE ( 'xz' )
[1960]	328	s_ind = 2
[667]	329	ALLOCATE( local_2d(nxlg:nxrg,nzb:nzt+1) )
[1]	330
[1308]	331	IF ( netcdf_data_format > 4 ) THEN
	332	ns = 1
[1960]	333	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	334	ns = ns + 1
	335	ENDDO
	336	ns = ns - 1
	337	ALLOCATE( local_2d_sections(nxlg:nxrg,1:ns,nzb:nzt+1) )
	338	ALLOCATE( local_2d_sections_l(nxlg:nxrg,1:ns,nzb:nzt+1) )
[1353]	339	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
[1308]	340	ENDIF
	341
[493]	342	!
[1031]	343	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	344	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	345	CALL check_open( 102+av*10 )
	346	ENDIF
[1]	347
	348	IF ( data_output_2d_on_each_pe ) THEN
	349	CALL check_open( 22 )
	350	ELSE
	351	IF ( myid == 0 ) THEN
	352	#if defined( __parallel )
[667]	353	ALLOCATE( total_2d(-nbgp:nx+nbgp,nzb:nzt+1) )
[1]	354	#endif
	355	ENDIF
	356	ENDIF
	357
	358	CASE ( 'yz' )
[1960]	359	s_ind = 3
[667]	360	ALLOCATE( local_2d(nysg:nyng,nzb:nzt+1) )
[1]	361
[1308]	362	IF ( netcdf_data_format > 4 ) THEN
	363	ns = 1
[1960]	364	DO WHILE ( section(ns,s_ind) /= -9999 .AND. ns <= 100 )
[1308]	365	ns = ns + 1
	366	ENDDO
	367	ns = ns - 1
	368	ALLOCATE( local_2d_sections(1:ns,nysg:nyng,nzb:nzt+1) )
	369	ALLOCATE( local_2d_sections_l(1:ns,nysg:nyng,nzb:nzt+1) )
[1353]	370	local_2d_sections = 0.0_wp; local_2d_sections_l = 0.0_wp
[1308]	371	ENDIF
	372
[493]	373	!
[1031]	374	!-- Parallel netCDF4/HDF5 output is done on all PEs, all other on PE0 only
[1327]	375	IF ( myid == 0 .OR. netcdf_data_format > 4 ) THEN
[493]	376	CALL check_open( 103+av*10 )
	377	ENDIF
[1]	378
	379	IF ( data_output_2d_on_each_pe ) THEN
	380	CALL check_open( 23 )
	381	ELSE
	382	IF ( myid == 0 ) THEN
	383	#if defined( __parallel )
[667]	384	ALLOCATE( total_2d(-nbgp:ny+nbgp,nzb:nzt+1) )
[1]	385	#endif
	386	ENDIF
	387	ENDIF
	388
	389	CASE DEFAULT
[254]	390	message_string = 'unknown cross-section: ' // TRIM( mode )
	391	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
[1]	392
	393	END SELECT
	394
	395	!
[1745]	396	!-- For parallel netcdf output the time axis must be limited. Return, if this
	397	!-- limit is exceeded. This could be the case, if the simulated time exceeds
	398	!-- the given end time by the length of the given output interval.
	399	IF ( netcdf_data_format > 4 ) THEN
	400	IF ( mode == 'xy' .AND. do2d_xy_time_count(av) + 1 > &
	401	ntdim_2d_xy(av) ) THEN
	402	WRITE ( message_string, * ) 'Output of xy cross-sections is not ', &
	403	'given at t=', simulated_time, '&because the', &
	404	' maximum number of output time levels is exceeded.'
	405	CALL message( 'data_output_2d', 'PA0384', 0, 1, 0, 6, 0 )
	406	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	407	RETURN
	408	ENDIF
	409	IF ( mode == 'xz' .AND. do2d_xz_time_count(av) + 1 > &
	410	ntdim_2d_xz(av) ) THEN
	411	WRITE ( message_string, * ) 'Output of xz cross-sections is not ', &
	412	'given at t=', simulated_time, '&because the', &
	413	' maximum number of output time levels is exceeded.'
	414	CALL message( 'data_output_2d', 'PA0385', 0, 1, 0, 6, 0 )
	415	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	416	RETURN
	417	ENDIF
	418	IF ( mode == 'yz' .AND. do2d_yz_time_count(av) + 1 > &
	419	ntdim_2d_yz(av) ) THEN
	420	WRITE ( message_string, * ) 'Output of yz cross-sections is not ', &
	421	'given at t=', simulated_time, '&because the', &
	422	' maximum number of output time levels is exceeded.'
	423	CALL message( 'data_output_2d', 'PA0386', 0, 1, 0, 6, 0 )
	424	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
	425	RETURN
	426	ENDIF
	427	ENDIF
	428
	429	!
[1]	430	!-- Allocate a temporary array for resorting (kji -> ijk).
[667]	431	ALLOCATE( local_pf(nxlg:nxrg,nysg:nyng,nzb:nzt+1) )
[2232]	432	local_pf = 0.0
	433	!
	434	!-- Allocate temporary array used for exchanging ghoist points of surface-data
	435	ALLOCATE( tmp_2d(nysg:nyng,nxlg:nxrg) )
	436	tmp_2d = 0.0
[1]	437
	438	!
	439	!-- Loop of all variables to be written.
	440	!-- Output dimensions chosen
	441	if = 1
	442	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
	443	do2d_mode = do2d(av,if)(l-1:l)
	444
	445	DO WHILE ( do2d(av,if)(1:1) /= ' ' )
	446
	447	IF ( do2d_mode == mode ) THEN
[1980]	448	!
	449	!-- Set flag to steer output of radiation, land-surface, or user-defined
	450	!-- quantities
	451	found = .FALSE.
[1551]	452
	453	nzb_do = nzb
	454	nzt_do = nzt+1
[1]	455	!
	456	!-- Store the array chosen on the temporary array.
	457	resorted = .FALSE.
	458	SELECT CASE ( TRIM( do2d(av,if) ) )
	459
	460	CASE ( 'e_xy', 'e_xz', 'e_yz' )
	461	IF ( av == 0 ) THEN
	462	to_be_resorted => e
	463	ELSE
	464	to_be_resorted => e_av
	465	ENDIF
	466	IF ( mode == 'xy' ) level_z = zu
	467
[771]	468	CASE ( 'lpt_xy', 'lpt_xz', 'lpt_yz' )
	469	IF ( av == 0 ) THEN
	470	to_be_resorted => pt
	471	ELSE
	472	to_be_resorted => lpt_av
	473	ENDIF
	474	IF ( mode == 'xy' ) level_z = zu
	475
[1]	476	CASE ( 'lwp*_xy' ) ! 2d-array
	477	IF ( av == 0 ) THEN
[667]	478	DO i = nxlg, nxrg
	479	DO j = nysg, nyng
[1320]	480	local_pf(i,j,nzb+1) = SUM( ql(nzb:nzt,j,i) * &
[1]	481	dzw(1:nzt+1) )
	482	ENDDO
	483	ENDDO
	484	ELSE
[667]	485	DO i = nxlg, nxrg
	486	DO j = nysg, nyng
[1]	487	local_pf(i,j,nzb+1) = lwp_av(j,i)
	488	ENDDO
	489	ENDDO
	490	ENDIF
	491	resorted = .TRUE.
	492	two_d = .TRUE.
	493	level_z(nzb+1) = zu(nzb+1)
	494
[1053]	495	CASE ( 'nr_xy', 'nr_xz', 'nr_yz' )
	496	IF ( av == 0 ) THEN
	497	to_be_resorted => nr
	498	ELSE
	499	to_be_resorted => nr_av
	500	ENDIF
	501	IF ( mode == 'xy' ) level_z = zu
	502
[1691]	503	CASE ( 'ol*_xy' ) ! 2d-array
	504	IF ( av == 0 ) THEN
[2232]	505	DO m = 1, surf_def_h(0)%ns
	506	i = surf_def_h(0)%i(m)
	507	j = surf_def_h(0)%j(m)
	508	tmp_2d(j,i) = surf_def_h(0)%ol(m)
	509	ENDDO
	510	DO m = 1, surf_lsm_h%ns
	511	i = surf_lsm_h%i(m)
	512	j = surf_lsm_h%j(m)
	513	tmp_2d(j,i) = surf_lsm_h%ol(m)
	514	ENDDO
	515	DO m = 1, surf_usm_h%ns
	516	i = surf_usm_h%i(m)
	517	j = surf_usm_h%j(m)
	518	tmp_2d(j,i) = surf_usm_h%ol(m)
	519	ENDDO
	520
	521	CALL exchange_horiz_2d( tmp_2d, nbgp )
	522
[1691]	523	DO i = nxlg, nxrg
	524	DO j = nysg, nyng
[2232]	525	local_pf(i,j,nzb+1) = tmp_2d(j,i)
[1691]	526	ENDDO
	527	ENDDO
[2232]	528
[1691]	529	ELSE
	530	DO i = nxlg, nxrg
	531	DO j = nysg, nyng
	532	local_pf(i,j,nzb+1) = ol_av(j,i)
	533	ENDDO
	534	ENDDO
	535	ENDIF
	536	resorted = .TRUE.
	537	two_d = .TRUE.
	538	level_z(nzb+1) = zu(nzb+1)
	539
[1]	540	CASE ( 'p_xy', 'p_xz', 'p_yz' )
	541	IF ( av == 0 ) THEN
[729]	542	IF ( psolver /= 'sor' ) CALL exchange_horiz( p, nbgp )
[1]	543	to_be_resorted => p
	544	ELSE
[729]	545	IF ( psolver /= 'sor' ) CALL exchange_horiz( p_av, nbgp )
[1]	546	to_be_resorted => p_av
	547	ENDIF
	548	IF ( mode == 'xy' ) level_z = zu
	549
	550	CASE ( 'pc_xy', 'pc_xz', 'pc_yz' ) ! particle concentration
	551	IF ( av == 0 ) THEN
[215]	552	IF ( simulated_time >= particle_advection_start ) THEN
	553	tend = prt_count
[667]	554	CALL exchange_horiz( tend, nbgp )
[215]	555	ELSE
[1353]	556	tend = 0.0_wp
[215]	557	ENDIF
[667]	558	DO i = nxlg, nxrg
	559	DO j = nysg, nyng
[1]	560	DO k = nzb, nzt+1
	561	local_pf(i,j,k) = tend(k,j,i)
	562	ENDDO
	563	ENDDO
	564	ENDDO
	565	resorted = .TRUE.
	566	ELSE
[667]	567	CALL exchange_horiz( pc_av, nbgp )
[1]	568	to_be_resorted => pc_av
	569	ENDIF
	570
[1359]	571	CASE ( 'pr_xy', 'pr_xz', 'pr_yz' ) ! mean particle radius (effective radius)
[1]	572	IF ( av == 0 ) THEN
[215]	573	IF ( simulated_time >= particle_advection_start ) THEN
	574	DO i = nxl, nxr
	575	DO j = nys, nyn
	576	DO k = nzb, nzt+1
[1359]	577	number_of_particles = prt_count(k,j,i)
	578	IF (number_of_particles <= 0) CYCLE
	579	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
	580	s_r2 = 0.0_wp
[1353]	581	s_r3 = 0.0_wp
[1359]	582	DO n = 1, number_of_particles
	583	IF ( particles(n)%particle_mask ) THEN
	584	s_r2 = s_r2 + particles(n)%radius*2 &
	585	particles(n)%weight_factor
	586	s_r3 = s_r3 + particles(n)%radius*3 &
	587	particles(n)%weight_factor
	588	ENDIF
[215]	589	ENDDO
[1359]	590	IF ( s_r2 > 0.0_wp ) THEN
	591	mean_r = s_r3 / s_r2
[215]	592	ELSE
[1353]	593	mean_r = 0.0_wp
[215]	594	ENDIF
	595	tend(k,j,i) = mean_r
[1]	596	ENDDO
	597	ENDDO
	598	ENDDO
[667]	599	CALL exchange_horiz( tend, nbgp )
[215]	600	ELSE
[1353]	601	tend = 0.0_wp
[1359]	602	ENDIF
[667]	603	DO i = nxlg, nxrg
	604	DO j = nysg, nyng
[1]	605	DO k = nzb, nzt+1
	606	local_pf(i,j,k) = tend(k,j,i)
	607	ENDDO
	608	ENDDO
	609	ENDDO
	610	resorted = .TRUE.
	611	ELSE
[667]	612	CALL exchange_horiz( pr_av, nbgp )
[1]	613	to_be_resorted => pr_av
	614	ENDIF
	615
[72]	616	CASE ( 'pra*_xy' ) ! 2d-array / integral quantity => no av
	617	CALL exchange_horiz_2d( precipitation_amount )
[667]	618	DO i = nxlg, nxrg
	619	DO j = nysg, nyng
[72]	620	local_pf(i,j,nzb+1) = precipitation_amount(j,i)
	621	ENDDO
	622	ENDDO
[1353]	623	precipitation_amount = 0.0_wp ! reset for next integ. interval
[72]	624	resorted = .TRUE.
	625	two_d = .TRUE.
	626	level_z(nzb+1) = zu(nzb+1)
	627
	628	CASE ( 'prr*_xy' ) ! 2d-array
[1822]	629	IF ( av == 0 ) THEN
	630	CALL exchange_horiz_2d( prr(nzb+1,:,:) )
	631	DO i = nxlg, nxrg
	632	DO j = nysg, nyng
	633	local_pf(i,j,nzb+1) = prr(nzb+1,j,i) * hyrho(nzb+1)
[1053]	634	ENDDO
[1822]	635	ENDDO
[1053]	636	ELSE
[1822]	637	CALL exchange_horiz_2d( prr_av(nzb+1,:,:) )
	638	DO i = nxlg, nxrg
	639	DO j = nysg, nyng
	640	local_pf(i,j,nzb+1) = prr_av(nzb+1,j,i) * hyrho(nzb+1)
[1053]	641	ENDDO
[1822]	642	ENDDO
[1053]	643	ENDIF
	644	resorted = .TRUE.
	645	two_d = .TRUE.
	646	level_z(nzb+1) = zu(nzb+1)
	647
	648	CASE ( 'prr_xy', 'prr_xz', 'prr_yz' )
[72]	649	IF ( av == 0 ) THEN
[1053]	650	CALL exchange_horiz( prr, nbgp )
[667]	651	DO i = nxlg, nxrg
	652	DO j = nysg, nyng
[1053]	653	DO k = nzb, nzt+1
[1822]	654	local_pf(i,j,k) = prr(k,j,i) * hyrho(nzb+1)
[1053]	655	ENDDO
[72]	656	ENDDO
	657	ENDDO
	658	ELSE
[1053]	659	CALL exchange_horiz( prr_av, nbgp )
[667]	660	DO i = nxlg, nxrg
	661	DO j = nysg, nyng
[1053]	662	DO k = nzb, nzt+1
[1822]	663	local_pf(i,j,k) = prr_av(k,j,i) * hyrho(nzb+1)
[1053]	664	ENDDO
[72]	665	ENDDO
	666	ENDDO
	667	ENDIF
	668	resorted = .TRUE.
[1053]	669	IF ( mode == 'xy' ) level_z = zu
[72]	670
[1]	671	CASE ( 'pt_xy', 'pt_xz', 'pt_yz' )
	672	IF ( av == 0 ) THEN
	673	IF ( .NOT. cloud_physics ) THEN
	674	to_be_resorted => pt
	675	ELSE
[667]	676	DO i = nxlg, nxrg
	677	DO j = nysg, nyng
[1]	678	DO k = nzb, nzt+1
[1320]	679	local_pf(i,j,k) = pt(k,j,i) + l_d_cp * &
	680	pt_d_t(k) * &
[1]	681	ql(k,j,i)
	682	ENDDO
	683	ENDDO
	684	ENDDO
	685	resorted = .TRUE.
	686	ENDIF
	687	ELSE
	688	to_be_resorted => pt_av
	689	ENDIF
	690	IF ( mode == 'xy' ) level_z = zu
	691
	692	CASE ( 'q_xy', 'q_xz', 'q_yz' )
	693	IF ( av == 0 ) THEN
	694	to_be_resorted => q
	695	ELSE
	696	to_be_resorted => q_av
	697	ENDIF
	698	IF ( mode == 'xy' ) level_z = zu
	699
[1053]	700	CASE ( 'qc_xy', 'qc_xz', 'qc_yz' )
[1]	701	IF ( av == 0 ) THEN
[1115]	702	to_be_resorted => qc
[1]	703	ELSE
[1115]	704	to_be_resorted => qc_av
[1]	705	ENDIF
	706	IF ( mode == 'xy' ) level_z = zu
	707
[1053]	708	CASE ( 'ql_xy', 'ql_xz', 'ql_yz' )
	709	IF ( av == 0 ) THEN
[1115]	710	to_be_resorted => ql
[1053]	711	ELSE
[1115]	712	to_be_resorted => ql_av
[1053]	713	ENDIF
	714	IF ( mode == 'xy' ) level_z = zu
	715
[1]	716	CASE ( 'ql_c_xy', 'ql_c_xz', 'ql_c_yz' )
	717	IF ( av == 0 ) THEN
	718	to_be_resorted => ql_c
	719	ELSE
	720	to_be_resorted => ql_c_av
	721	ENDIF
	722	IF ( mode == 'xy' ) level_z = zu
	723
	724	CASE ( 'ql_v_xy', 'ql_v_xz', 'ql_v_yz' )
	725	IF ( av == 0 ) THEN
	726	to_be_resorted => ql_v
	727	ELSE
	728	to_be_resorted => ql_v_av
	729	ENDIF
	730	IF ( mode == 'xy' ) level_z = zu
	731
	732	CASE ( 'ql_vp_xy', 'ql_vp_xz', 'ql_vp_yz' )
	733	IF ( av == 0 ) THEN
[1007]	734	IF ( simulated_time >= particle_advection_start ) THEN
	735	DO i = nxl, nxr
	736	DO j = nys, nyn
	737	DO k = nzb, nzt+1
[1359]	738	number_of_particles = prt_count(k,j,i)
	739	IF (number_of_particles <= 0) CYCLE
	740	particles => grid_particles(k,j,i)%particles(1:number_of_particles)
	741	DO n = 1, number_of_particles
	742	IF ( particles(n)%particle_mask ) THEN
	743	tend(k,j,i) = tend(k,j,i) + &
	744	particles(n)%weight_factor / &
	745	prt_count(k,j,i)
	746	ENDIF
[1007]	747	ENDDO
	748	ENDDO
	749	ENDDO
	750	ENDDO
	751	CALL exchange_horiz( tend, nbgp )
	752	ELSE
[1353]	753	tend = 0.0_wp
[1359]	754	ENDIF
[1007]	755	DO i = nxlg, nxrg
	756	DO j = nysg, nyng
	757	DO k = nzb, nzt+1
	758	local_pf(i,j,k) = tend(k,j,i)
	759	ENDDO
	760	ENDDO
	761	ENDDO
	762	resorted = .TRUE.
	763	ELSE
	764	CALL exchange_horiz( ql_vp_av, nbgp )
[1]	765	to_be_resorted => ql_vp
	766	ENDIF
	767	IF ( mode == 'xy' ) level_z = zu
	768
[1053]	769	CASE ( 'qr_xy', 'qr_xz', 'qr_yz' )
	770	IF ( av == 0 ) THEN
	771	to_be_resorted => qr
	772	ELSE
	773	to_be_resorted => qr_av
	774	ENDIF
	775	IF ( mode == 'xy' ) level_z = zu
	776
[354]	777	CASE ( 'qsws*_xy' ) ! 2d-array
	778	IF ( av == 0 ) THEN
[2232]	779	DO m = 1, surf_def_h(0)%ns
	780	i = surf_def_h(0)%i(m)
	781	j = surf_def_h(0)%j(m)
	782	tmp_2d(j,i) = surf_def_h(0)%qsws(m)
	783	ENDDO
	784	DO m = 1, surf_lsm_h%ns
	785	i = surf_lsm_h%i(m)
	786	j = surf_lsm_h%j(m)
	787	tmp_2d(j,i) = surf_lsm_h%qsws(m)
	788	ENDDO
	789	DO m = 1, surf_usm_h%ns
	790	i = surf_usm_h%i(m)
	791	j = surf_usm_h%j(m)
	792	tmp_2d(j,i) = surf_usm_h%qsws(m)
	793	ENDDO
	794
	795	CALL exchange_horiz_2d( tmp_2d, nbgp )
	796
[667]	797	DO i = nxlg, nxrg
	798	DO j = nysg, nyng
[2232]	799	local_pf(i,j,nzb+1) = tmp_2d(j,i)
[354]	800	ENDDO
	801	ENDDO
	802	ELSE
[667]	803	DO i = nxlg, nxrg
	804	DO j = nysg, nyng
[354]	805	local_pf(i,j,nzb+1) = qsws_av(j,i)
	806	ENDDO
	807	ENDDO
	808	ENDIF
	809	resorted = .TRUE.
	810	two_d = .TRUE.
	811	level_z(nzb+1) = zu(nzb+1)
	812
[1]	813	CASE ( 'qv_xy', 'qv_xz', 'qv_yz' )
	814	IF ( av == 0 ) THEN
[667]	815	DO i = nxlg, nxrg
	816	DO j = nysg, nyng
[1]	817	DO k = nzb, nzt+1
	818	local_pf(i,j,k) = q(k,j,i) - ql(k,j,i)
	819	ENDDO
	820	ENDDO
	821	ENDDO
	822	resorted = .TRUE.
	823	ELSE
	824	to_be_resorted => qv_av
	825	ENDIF
	826	IF ( mode == 'xy' ) level_z = zu
	827
[1551]	828
[1585]	829
[2190]	830	CASE ( 'rho_ocean_xy', 'rho_ocean_xz', 'rho_ocean_yz' )
[96]	831	IF ( av == 0 ) THEN
[2031]	832	to_be_resorted => rho_ocean
[96]	833	ELSE
[2031]	834	to_be_resorted => rho_ocean_av
[96]	835	ENDIF
	836
[1]	837	CASE ( 's_xy', 's_xz', 's_yz' )
	838	IF ( av == 0 ) THEN
[1960]	839	to_be_resorted => s
[1]	840	ELSE
[355]	841	to_be_resorted => s_av
[1]	842	ENDIF
	843
[96]	844	CASE ( 'sa_xy', 'sa_xz', 'sa_yz' )
	845	IF ( av == 0 ) THEN
	846	to_be_resorted => sa
	847	ELSE
	848	to_be_resorted => sa_av
	849	ENDIF
	850
[354]	851	CASE ( 'shf*_xy' ) ! 2d-array
	852	IF ( av == 0 ) THEN
[2232]	853	DO m = 1, surf_def_h(0)%ns
	854	i = surf_def_h(0)%i(m)
	855	j = surf_def_h(0)%j(m)
	856	tmp_2d(j,i) = surf_def_h(0)%shf(m)
	857	ENDDO
	858	DO m = 1, surf_lsm_h%ns
	859	i = surf_lsm_h%i(m)
	860	j = surf_lsm_h%j(m)
	861	tmp_2d(j,i) = surf_lsm_h%shf(m)
	862	ENDDO
	863	DO m = 1, surf_usm_h%ns
	864	i = surf_usm_h%i(m)
	865	j = surf_usm_h%j(m)
	866	tmp_2d(j,i) = surf_usm_h%shf(m)
	867	ENDDO
	868
	869	CALL exchange_horiz_2d( tmp_2d, nbgp )
	870
[667]	871	DO i = nxlg, nxrg
	872	DO j = nysg, nyng
[2232]	873	local_pf(i,j,nzb+1) = tmp_2d(j,i)
[354]	874	ENDDO
	875	ENDDO
	876	ELSE
[667]	877	DO i = nxlg, nxrg
	878	DO j = nysg, nyng
[354]	879	local_pf(i,j,nzb+1) = shf_av(j,i)
	880	ENDDO
	881	ENDDO
	882	ENDIF
	883	resorted = .TRUE.
	884	two_d = .TRUE.
	885	level_z(nzb+1) = zu(nzb+1)
[1960]	886
	887	CASE ( 'ssws*_xy' ) ! 2d-array
	888	IF ( av == 0 ) THEN
[2232]	889	DO m = 1, surf_def_h(0)%ns
	890	i = surf_def_h(0)%i(m)
	891	j = surf_def_h(0)%j(m)
	892	tmp_2d(j,i) = surf_def_h(0)%ssws(m)
	893	ENDDO
	894	DO m = 1, surf_lsm_h%ns
	895	i = surf_lsm_h%i(m)
	896	j = surf_lsm_h%j(m)
	897	tmp_2d(j,i) = surf_lsm_h%ssws(m)
	898	ENDDO
	899	DO m = 1, surf_usm_h%ns
	900	i = surf_usm_h%i(m)
	901	j = surf_usm_h%j(m)
	902	tmp_2d(j,i) = surf_usm_h%ssws(m)
	903	ENDDO
	904
	905	CALL exchange_horiz_2d( tmp_2d, nbgp )
	906
[1960]	907	DO i = nxlg, nxrg
	908	DO j = nysg, nyng
[2232]	909	local_pf(i,j,nzb+1) = tmp_2d(j,i)
[1960]	910	ENDDO
	911	ENDDO
	912	ELSE
	913	DO i = nxlg, nxrg
	914	DO j = nysg, nyng
	915	local_pf(i,j,nzb+1) = ssws_av(j,i)
	916	ENDDO
	917	ENDDO
	918	ENDIF
	919	resorted = .TRUE.
	920	two_d = .TRUE.
	921	level_z(nzb+1) = zu(nzb+1)
[1551]	922
[1]	923	CASE ( 't*_xy' ) ! 2d-array
	924	IF ( av == 0 ) THEN
[2232]	925	DO m = 1, surf_def_h(0)%ns
	926	i = surf_def_h(0)%i(m)
	927	j = surf_def_h(0)%j(m)
	928	tmp_2d(j,i) = surf_def_h(0)%ts(m)
	929	ENDDO
	930	DO m = 1, surf_lsm_h%ns
	931	i = surf_lsm_h%i(m)
	932	j = surf_lsm_h%j(m)
	933	tmp_2d(j,i) = surf_lsm_h%ts(m)
	934	ENDDO
	935	DO m = 1, surf_usm_h%ns
	936	i = surf_usm_h%i(m)
	937	j = surf_usm_h%j(m)
	938	tmp_2d(j,i) = surf_usm_h%ts(m)
	939	ENDDO
	940
	941	CALL exchange_horiz_2d( tmp_2d, nbgp )
	942
[667]	943	DO i = nxlg, nxrg
	944	DO j = nysg, nyng
[2232]	945	local_pf(i,j,nzb+1) = tmp_2d(j,i)
[1]	946	ENDDO
	947	ENDDO
[2232]	948
[1]	949	ELSE
[667]	950	DO i = nxlg, nxrg
	951	DO j = nysg, nyng
[1]	952	local_pf(i,j,nzb+1) = ts_av(j,i)
	953	ENDDO
	954	ENDDO
	955	ENDIF
	956	resorted = .TRUE.
	957	two_d = .TRUE.
	958	level_z(nzb+1) = zu(nzb+1)
	959
	960	CASE ( 'u_xy', 'u_xz', 'u_yz' )
	961	IF ( av == 0 ) THEN
	962	to_be_resorted => u
	963	ELSE
	964	to_be_resorted => u_av
	965	ENDIF
	966	IF ( mode == 'xy' ) level_z = zu
	967	!
	968	!-- Substitute the values generated by "mirror" boundary condition
	969	!-- at the bottom boundary by the real surface values.
	970	IF ( do2d(av,if) == 'u_xz' .OR. do2d(av,if) == 'u_yz' ) THEN
[1353]	971	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
[1]	972	ENDIF
	973
	974	CASE ( 'u*_xy' ) ! 2d-array
	975	IF ( av == 0 ) THEN
[2232]	976	DO m = 1, surf_def_h(0)%ns
	977	i = surf_def_h(0)%i(m)
	978	j = surf_def_h(0)%j(m)
	979	tmp_2d(j,i) = surf_def_h(0)%us(m)
	980	ENDDO
	981	DO m = 1, surf_lsm_h%ns
	982	i = surf_lsm_h%i(m)
	983	j = surf_lsm_h%j(m)
	984	tmp_2d(j,i) = surf_lsm_h%us(m)
	985	ENDDO
	986	DO m = 1, surf_usm_h%ns
	987	i = surf_usm_h%i(m)
	988	j = surf_usm_h%j(m)
	989	tmp_2d(j,i) = surf_usm_h%us(m)
	990	ENDDO
	991
	992	CALL exchange_horiz_2d( tmp_2d, nbgp )
	993
[667]	994	DO i = nxlg, nxrg
	995	DO j = nysg, nyng
[2232]	996	local_pf(i,j,nzb+1) = tmp_2d(j,i)
[1]	997	ENDDO
	998	ENDDO
	999	ELSE
[667]	1000	DO i = nxlg, nxrg
	1001	DO j = nysg, nyng
[1]	1002	local_pf(i,j,nzb+1) = us_av(j,i)
	1003	ENDDO
	1004	ENDDO
	1005	ENDIF
	1006	resorted = .TRUE.
	1007	two_d = .TRUE.
	1008	level_z(nzb+1) = zu(nzb+1)
	1009
	1010	CASE ( 'v_xy', 'v_xz', 'v_yz' )
	1011	IF ( av == 0 ) THEN
	1012	to_be_resorted => v
	1013	ELSE
	1014	to_be_resorted => v_av
	1015	ENDIF
	1016	IF ( mode == 'xy' ) level_z = zu
	1017	!
	1018	!-- Substitute the values generated by "mirror" boundary condition
	1019	!-- at the bottom boundary by the real surface values.
	1020	IF ( do2d(av,if) == 'v_xz' .OR. do2d(av,if) == 'v_yz' ) THEN
[1353]	1021	IF ( ibc_uv_b == 0 ) local_pf(:,:,nzb) = 0.0_wp
[1]	1022	ENDIF
	1023
	1024	CASE ( 'vpt_xy', 'vpt_xz', 'vpt_yz' )
	1025	IF ( av == 0 ) THEN
	1026	to_be_resorted => vpt
	1027	ELSE
	1028	to_be_resorted => vpt_av
	1029	ENDIF
	1030	IF ( mode == 'xy' ) level_z = zu
	1031
	1032	CASE ( 'w_xy', 'w_xz', 'w_yz' )
	1033	IF ( av == 0 ) THEN
	1034	to_be_resorted => w
	1035	ELSE
	1036	to_be_resorted => w_av
	1037	ENDIF
	1038	IF ( mode == 'xy' ) level_z = zw
	1039
[72]	1040	CASE ( 'z0*_xy' ) ! 2d-array
	1041	IF ( av == 0 ) THEN
[2232]	1042	DO m = 1, surf_def_h(0)%ns
	1043	i = surf_def_h(0)%i(m)
	1044	j = surf_def_h(0)%j(m)
	1045	tmp_2d(j,i) = surf_def_h(0)%z0(m)
	1046	ENDDO
	1047	DO m = 1, surf_lsm_h%ns
	1048	i = surf_lsm_h%i(m)
	1049	j = surf_lsm_h%j(m)
	1050	tmp_2d(j,i) = surf_lsm_h%z0(m)
	1051	ENDDO
	1052	DO m = 1, surf_usm_h%ns
	1053	i = surf_usm_h%i(m)
	1054	j = surf_usm_h%j(m)
	1055	tmp_2d(j,i) = surf_usm_h%z0(m)
	1056	ENDDO
	1057
	1058	CALL exchange_horiz_2d( tmp_2d, nbgp )
	1059
[667]	1060	DO i = nxlg, nxrg
	1061	DO j = nysg, nyng
[2232]	1062	local_pf(i,j,nzb+1) = tmp_2d(j,i)
[72]	1063	ENDDO
	1064	ENDDO
[2232]	1065
[72]	1066	ELSE
[667]	1067	DO i = nxlg, nxrg
	1068	DO j = nysg, nyng
[72]	1069	local_pf(i,j,nzb+1) = z0_av(j,i)
	1070	ENDDO
	1071	ENDDO
	1072	ENDIF
	1073	resorted = .TRUE.
	1074	two_d = .TRUE.
	1075	level_z(nzb+1) = zu(nzb+1)
	1076
[978]	1077	CASE ( 'z0h*_xy' ) ! 2d-array
	1078	IF ( av == 0 ) THEN
[2232]	1079	DO m = 1, surf_def_h(0)%ns
	1080	i = surf_def_h(0)%i(m)
	1081	j = surf_def_h(0)%j(m)
	1082	tmp_2d(j,i) = surf_def_h(0)%z0h(m)
	1083	ENDDO
	1084	DO m = 1, surf_lsm_h%ns
	1085	i = surf_lsm_h%i(m)
	1086	j = surf_lsm_h%j(m)
	1087	tmp_2d(j,i) = surf_lsm_h%z0h(m)
	1088	ENDDO
	1089	DO m = 1, surf_usm_h%ns
	1090	i = surf_usm_h%i(m)
	1091	j = surf_usm_h%j(m)
	1092	tmp_2d(j,i) = surf_usm_h%z0h(m)
	1093	ENDDO
	1094
	1095	CALL exchange_horiz_2d( tmp_2d, nbgp )
	1096
[978]	1097	DO i = nxlg, nxrg
	1098	DO j = nysg, nyng
[2232]	1099	local_pf(i,j,nzb+1) = tmp_2d(j,i)
[978]	1100	ENDDO
	1101	ENDDO
	1102	ELSE
	1103	DO i = nxlg, nxrg
	1104	DO j = nysg, nyng
	1105	local_pf(i,j,nzb+1) = z0h_av(j,i)
	1106	ENDDO
	1107	ENDDO
	1108	ENDIF
	1109	resorted = .TRUE.
	1110	two_d = .TRUE.
	1111	level_z(nzb+1) = zu(nzb+1)
	1112
[1788]	1113	CASE ( 'z0q*_xy' ) ! 2d-array
	1114	IF ( av == 0 ) THEN
[2232]	1115	DO m = 1, surf_def_h(0)%ns
	1116	i = surf_def_h(0)%i(m)
	1117	j = surf_def_h(0)%j(m)
	1118	tmp_2d(j,i) = surf_def_h(0)%z0q(m)
	1119	ENDDO
	1120	DO m = 1, surf_lsm_h%ns
	1121	i = surf_lsm_h%i(m)
	1122	j = surf_lsm_h%j(m)
	1123	tmp_2d(j,i) = surf_lsm_h%z0q(m)
	1124	ENDDO
	1125	DO m = 1, surf_usm_h%ns
	1126	i = surf_usm_h%i(m)
	1127	j = surf_usm_h%j(m)
	1128	tmp_2d(j,i) = surf_usm_h%z0q(m)
	1129	ENDDO
	1130
	1131	CALL exchange_horiz_2d( tmp_2d, nbgp )
	1132
[1788]	1133	DO i = nxlg, nxrg
	1134	DO j = nysg, nyng
[2232]	1135	local_pf(i,j,nzb+1) = tmp_2d(j,i)
[1788]	1136	ENDDO
	1137	ENDDO
	1138	ELSE
	1139	DO i = nxlg, nxrg
	1140	DO j = nysg, nyng
	1141	local_pf(i,j,nzb+1) = z0q_av(j,i)
	1142	ENDDO
	1143	ENDDO
	1144	ENDIF
	1145	resorted = .TRUE.
	1146	two_d = .TRUE.
	1147	level_z(nzb+1) = zu(nzb+1)
	1148
[1]	1149	CASE DEFAULT
[1972]	1150
[1]	1151	!
[1972]	1152	!-- Land surface model quantity
	1153	IF ( land_surface ) THEN
	1154	CALL lsm_data_output_2d( av, do2d(av,if), found, grid, mode,&
	1155	local_pf, two_d, nzb_do, nzt_do )
	1156	ENDIF
	1157
	1158	!
[1976]	1159	!-- Radiation quantity
	1160	IF ( .NOT. found .AND. radiation ) THEN
	1161	CALL radiation_data_output_2d( av, do2d(av,if), found, grid,&
	1162	mode, local_pf, two_d )
	1163	ENDIF
	1164
	1165	!
[1]	1166	!-- User defined quantity
[1972]	1167	IF ( .NOT. found ) THEN
	1168	CALL user_data_output_2d( av, do2d(av,if), found, grid, &
	1169	local_pf, two_d, nzb_do, nzt_do )
	1170	ENDIF
	1171
[1]	1172	resorted = .TRUE.
	1173
	1174	IF ( grid == 'zu' ) THEN
	1175	IF ( mode == 'xy' ) level_z = zu
	1176	ELSEIF ( grid == 'zw' ) THEN
	1177	IF ( mode == 'xy' ) level_z = zw
[343]	1178	ELSEIF ( grid == 'zu1' ) THEN
	1179	IF ( mode == 'xy' ) level_z(nzb+1) = zu(nzb+1)
[1551]	1180	ELSEIF ( grid == 'zs' ) THEN
	1181	IF ( mode == 'xy' ) level_z = zs
[1]	1182	ENDIF
	1183
	1184	IF ( .NOT. found ) THEN
[1320]	1185	message_string = 'no output provided for: ' // &
[274]	1186	TRIM( do2d(av,if) )
[254]	1187	CALL message( 'data_output_2d', 'PA0181', 0, 0, 0, 6, 0 )
[1]	1188	ENDIF
	1189
	1190	END SELECT
	1191
	1192	!
	1193	!-- Resort the array to be output, if not done above
	1194	IF ( .NOT. resorted ) THEN
[667]	1195	DO i = nxlg, nxrg
	1196	DO j = nysg, nyng
[1551]	1197	DO k = nzb_do, nzt_do
[1]	1198	local_pf(i,j,k) = to_be_resorted(k,j,i)
	1199	ENDDO
	1200	ENDDO
	1201	ENDDO
	1202	ENDIF
	1203
	1204	!
	1205	!-- Output of the individual cross-sections, depending on the cross-
	1206	!-- section mode chosen.
	1207	is = 1
[1960]	1208	loop1: DO WHILE ( section(is,s_ind) /= -9999 .OR. two_d )
[1]	1209
	1210	SELECT CASE ( mode )
	1211
	1212	CASE ( 'xy' )
	1213	!
	1214	!-- Determine the cross section index
	1215	IF ( two_d ) THEN
	1216	layer_xy = nzb+1
	1217	ELSE
[1960]	1218	layer_xy = section(is,s_ind)
[1]	1219	ENDIF
	1220
	1221	!
[1551]	1222	!-- Exit the loop for layers beyond the data output domain
	1223	!-- (used for soil model)
[1691]	1224	IF ( layer_xy > nzt_do ) THEN
[1551]	1225	EXIT loop1
	1226	ENDIF
	1227
	1228	!
[1308]	1229	!-- Update the netCDF xy cross section time axis.
	1230	!-- In case of parallel output, this is only done by PE0
	1231	!-- to increase the performance.
	1232	IF ( simulated_time /= do2d_xy_last_time(av) ) THEN
	1233	do2d_xy_time_count(av) = do2d_xy_time_count(av) + 1
	1234	do2d_xy_last_time(av) = simulated_time
	1235	IF ( myid == 0 ) THEN
[1327]	1236	IF ( .NOT. data_output_2d_on_each_pe &
	1237	.OR. netcdf_data_format > 4 ) &
[493]	1238	THEN
[1]	1239	#if defined( __netcdf )
	1240	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1241	id_var_time_xy(av), &
[291]	1242	(/ time_since_reference_point /), &
[1]	1243	start = (/ do2d_xy_time_count(av) /), &
	1244	count = (/ 1 /) )
[1783]	1245	CALL netcdf_handle_error( 'data_output_2d', 53 )
[1]	1246	#endif
	1247	ENDIF
	1248	ENDIF
	1249	ENDIF
	1250	!
	1251	!-- If required, carry out averaging along z
[1960]	1252	IF ( section(is,s_ind) == -1 .AND. .NOT. two_d ) THEN
[1]	1253
[1353]	1254	local_2d = 0.0_wp
[1]	1255	!
	1256	!-- Carry out the averaging (all data are on the PE)
[1551]	1257	DO k = nzb_do, nzt_do
[667]	1258	DO j = nysg, nyng
	1259	DO i = nxlg, nxrg
[1]	1260	local_2d(i,j) = local_2d(i,j) + local_pf(i,j,k)
	1261	ENDDO
	1262	ENDDO
	1263	ENDDO
	1264
[1551]	1265	local_2d = local_2d / ( nzt_do - nzb_do + 1.0_wp)
[1]	1266
	1267	ELSE
	1268	!
	1269	!-- Just store the respective section on the local array
	1270	local_2d = local_pf(:,:,layer_xy)
	1271
	1272	ENDIF
	1273
	1274	#if defined( __parallel )
[1327]	1275	IF ( netcdf_data_format > 4 ) THEN
[1]	1276	!
[1031]	1277	!-- Parallel output in netCDF4/HDF5 format.
[493]	1278	IF ( two_d ) THEN
	1279	iis = 1
	1280	ELSE
	1281	iis = is
	1282	ENDIF
	1283
[1]	1284	#if defined( __netcdf )
[1308]	1285	!
	1286	!-- For parallel output, all cross sections are first stored
	1287	!-- here on a local array and will be written to the output
	1288	!-- file afterwards to increase the performance.
	1289	DO i = nxlg, nxrg
	1290	DO j = nysg, nyng
	1291	local_2d_sections(i,j,iis) = local_2d(i,j)
	1292	ENDDO
	1293	ENDDO
[1]	1294	#endif
[493]	1295	ELSE
[1]	1296
[493]	1297	IF ( data_output_2d_on_each_pe ) THEN
[1]	1298	!
[493]	1299	!-- Output of partial arrays on each PE
	1300	#if defined( __netcdf )
[1327]	1301	IF ( myid == 0 ) THEN
[1320]	1302	WRITE ( 21 ) time_since_reference_point, &
[493]	1303	do2d_xy_time_count(av), av
	1304	ENDIF
	1305	#endif
[759]	1306	DO i = 0, io_blocks-1
	1307	IF ( i == io_group ) THEN
[1551]	1308	WRITE ( 21 ) nxlg, nxrg, nysg, nyng, nysg, nyng
[759]	1309	WRITE ( 21 ) local_2d
	1310	ENDIF
	1311	#if defined( __parallel )
	1312	CALL MPI_BARRIER( comm2d, ierr )
	1313	#endif
	1314	ENDDO
[559]	1315
[493]	1316	ELSE
[1]	1317	!
[493]	1318	!-- PE0 receives partial arrays from all processors and
	1319	!-- then outputs them. Here a barrier has to be set,
	1320	!-- because otherwise "-MPI- FATAL: Remote protocol queue
	1321	!-- full" may occur.
	1322	CALL MPI_BARRIER( comm2d, ierr )
	1323
[667]	1324	ngp = ( nxrg-nxlg+1 ) * ( nyng-nysg+1 )
[493]	1325	IF ( myid == 0 ) THEN
[1]	1326	!
[493]	1327	!-- Local array can be relocated directly.
[667]	1328	total_2d(nxlg:nxrg,nysg:nyng) = local_2d
[1]	1329	!
[493]	1330	!-- Receive data from all other PEs.
	1331	DO n = 1, numprocs-1
[1]	1332	!
[493]	1333	!-- Receive index limits first, then array.
	1334	!-- Index limits are received in arbitrary order from
	1335	!-- the PEs.
[1320]	1336	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1337	MPI_ANY_SOURCE, 0, comm2d, &
[493]	1338	status, ierr )
	1339	sender = status(MPI_SOURCE)
	1340	DEALLOCATE( local_2d )
	1341	ALLOCATE( local_2d(ind(1):ind(2),ind(3):ind(4)) )
[1320]	1342	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1343	MPI_REAL, sender, 1, comm2d, &
[493]	1344	status, ierr )
	1345	total_2d(ind(1):ind(2),ind(3):ind(4)) = local_2d
	1346	ENDDO
[1]	1347	!
[493]	1348	!-- Relocate the local array for the next loop increment
	1349	DEALLOCATE( local_2d )
[667]	1350	ALLOCATE( local_2d(nxlg:nxrg,nysg:nyng) )
[1]	1351
	1352	#if defined( __netcdf )
[1327]	1353	IF ( two_d ) THEN
	1354	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1355	id_var_do2d(av,if), &
	1356	total_2d(0:nx+1,0:ny+1), &
	1357	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
	1358	count = (/ nx+2, ny+2, 1, 1 /) )
	1359	ELSE
	1360	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1361	id_var_do2d(av,if), &
	1362	total_2d(0:nx+1,0:ny+1), &
	1363	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
	1364	count = (/ nx+2, ny+2, 1, 1 /) )
[1]	1365	ENDIF
[1783]	1366	CALL netcdf_handle_error( 'data_output_2d', 54 )
[1]	1367	#endif
	1368
[493]	1369	ELSE
[1]	1370	!
[493]	1371	!-- First send the local index limits to PE0
[667]	1372	ind(1) = nxlg; ind(2) = nxrg
	1373	ind(3) = nysg; ind(4) = nyng
[1320]	1374	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1375	comm2d, ierr )
[1]	1376	!
[493]	1377	!-- Send data to PE0
[1320]	1378	CALL MPI_SEND( local_2d(nxlg,nysg), ngp, &
[493]	1379	MPI_REAL, 0, 1, comm2d, ierr )
	1380	ENDIF
	1381	!
	1382	!-- A barrier has to be set, because otherwise some PEs may
	1383	!-- proceed too fast so that PE0 may receive wrong data on
	1384	!-- tag 0
	1385	CALL MPI_BARRIER( comm2d, ierr )
[1]	1386	ENDIF
[493]	1387
[1]	1388	ENDIF
	1389	#else
	1390	#if defined( __netcdf )
[1327]	1391	IF ( two_d ) THEN
	1392	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1393	id_var_do2d(av,if), &
	1394	local_2d(nxl:nxr+1,nys:nyn+1), &
	1395	start = (/ 1, 1, 1, do2d_xy_time_count(av) /), &
	1396	count = (/ nx+2, ny+2, 1, 1 /) )
	1397	ELSE
	1398	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1399	id_var_do2d(av,if), &
	1400	local_2d(nxl:nxr+1,nys:nyn+1), &
	1401	start = (/ 1, 1, is, do2d_xy_time_count(av) /), &
	1402	count = (/ nx+2, ny+2, 1, 1 /) )
[1]	1403	ENDIF
[1783]	1404	CALL netcdf_handle_error( 'data_output_2d', 447 )
[1]	1405	#endif
	1406	#endif
[2277]	1407
[1]	1408	!
	1409	!-- For 2D-arrays (e.g. u*) only one cross-section is available.
	1410	!-- Hence exit loop of output levels.
	1411	IF ( two_d ) THEN
[1703]	1412	IF ( netcdf_data_format < 5 ) two_d = .FALSE.
[1]	1413	EXIT loop1
	1414	ENDIF
	1415
	1416	CASE ( 'xz' )
	1417	!
[1308]	1418	!-- Update the netCDF xz cross section time axis.
	1419	!-- In case of parallel output, this is only done by PE0
	1420	!-- to increase the performance.
	1421	IF ( simulated_time /= do2d_xz_last_time(av) ) THEN
	1422	do2d_xz_time_count(av) = do2d_xz_time_count(av) + 1
	1423	do2d_xz_last_time(av) = simulated_time
	1424	IF ( myid == 0 ) THEN
[1327]	1425	IF ( .NOT. data_output_2d_on_each_pe &
	1426	.OR. netcdf_data_format > 4 ) &
[493]	1427	THEN
[1]	1428	#if defined( __netcdf )
	1429	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1430	id_var_time_xz(av), &
[291]	1431	(/ time_since_reference_point /), &
[1]	1432	start = (/ do2d_xz_time_count(av) /), &
	1433	count = (/ 1 /) )
[1783]	1434	CALL netcdf_handle_error( 'data_output_2d', 56 )
[1]	1435	#endif
	1436	ENDIF
	1437	ENDIF
	1438	ENDIF
[667]	1439
[1]	1440	!
	1441	!-- If required, carry out averaging along y
[1960]	1442	IF ( section(is,s_ind) == -1 ) THEN
[1]	1443
[1551]	1444	ALLOCATE( local_2d_l(nxlg:nxrg,nzb_do:nzt_do) )
[1353]	1445	local_2d_l = 0.0_wp
[1551]	1446	ngp = ( nxrg-nxlg + 1 ) * ( nzt_do-nzb_do + 1 )
[1]	1447	!
	1448	!-- First local averaging on the PE
[1551]	1449	DO k = nzb_do, nzt_do
[1]	1450	DO j = nys, nyn
[667]	1451	DO i = nxlg, nxrg
[1320]	1452	local_2d_l(i,k) = local_2d_l(i,k) + &
[1]	1453	local_pf(i,j,k)
	1454	ENDDO
	1455	ENDDO
	1456	ENDDO
	1457	#if defined( __parallel )
	1458	!
	1459	!-- Now do the averaging over all PEs along y
[622]	1460	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1461	CALL MPI_ALLREDUCE( local_2d_l(nxlg,nzb_do), &
	1462	local_2d(nxlg,nzb_do), ngp, MPI_REAL, &
[1]	1463	MPI_SUM, comm1dy, ierr )
	1464	#else
	1465	local_2d = local_2d_l
	1466	#endif
[1353]	1467	local_2d = local_2d / ( ny + 1.0_wp )
[1]	1468
	1469	DEALLOCATE( local_2d_l )
	1470
	1471	ELSE
	1472	!
	1473	!-- Just store the respective section on the local array
	1474	!-- (but only if it is available on this PE!)
[1960]	1475	IF ( section(is,s_ind) >= nys .AND. section(is,s_ind) <= nyn ) &
[1]	1476	THEN
[1960]	1477	local_2d = local_pf(:,section(is,s_ind),nzb_do:nzt_do)
[1]	1478	ENDIF
	1479
	1480	ENDIF
	1481
	1482	#if defined( __parallel )
[1327]	1483	IF ( netcdf_data_format > 4 ) THEN
[1]	1484	!
[1031]	1485	!-- Output in netCDF4/HDF5 format.
[493]	1486	!-- Output only on those PEs where the respective cross
	1487	!-- sections reside. Cross sections averaged along y are
	1488	!-- output on the respective first PE along y (myidy=0).
[1960]	1489	IF ( ( section(is,s_ind) >= nys .AND. &
	1490	section(is,s_ind) <= nyn ) .OR. &
	1491	( section(is,s_ind) == -1 .AND. myidy == 0 ) ) THEN
[1]	1492	#if defined( __netcdf )
[493]	1493	!
[1308]	1494	!-- For parallel output, all cross sections are first
	1495	!-- stored here on a local array and will be written to the
	1496	!-- output file afterwards to increase the performance.
	1497	DO i = nxlg, nxrg
[1551]	1498	DO k = nzb_do, nzt_do
[1308]	1499	local_2d_sections_l(i,is,k) = local_2d(i,k)
	1500	ENDDO
	1501	ENDDO
[1]	1502	#endif
	1503	ENDIF
	1504
	1505	ELSE
	1506
[493]	1507	IF ( data_output_2d_on_each_pe ) THEN
[1]	1508	!
[493]	1509	!-- Output of partial arrays on each PE. If the cross
	1510	!-- section does not reside on the PE, output special
	1511	!-- index values.
	1512	#if defined( __netcdf )
[1327]	1513	IF ( myid == 0 ) THEN
[1320]	1514	WRITE ( 22 ) time_since_reference_point, &
[493]	1515	do2d_xz_time_count(av), av
	1516	ENDIF
	1517	#endif
[759]	1518	DO i = 0, io_blocks-1
	1519	IF ( i == io_group ) THEN
[1960]	1520	IF ( ( section(is,s_ind) >= nys .AND. &
	1521	section(is,s_ind) <= nyn ) .OR. &
	1522	( section(is,s_ind) == -1 .AND. &
[1320]	1523	nys-1 == -1 ) ) &
[759]	1524	THEN
[1551]	1525	WRITE (22) nxlg, nxrg, nzb_do, nzt_do, nzb, nzt+1
[759]	1526	WRITE (22) local_2d
	1527	ELSE
[1551]	1528	WRITE (22) -1, -1, -1, -1, -1, -1
[759]	1529	ENDIF
	1530	ENDIF
	1531	#if defined( __parallel )
	1532	CALL MPI_BARRIER( comm2d, ierr )
	1533	#endif
	1534	ENDDO
[493]	1535
	1536	ELSE
[1]	1537	!
[493]	1538	!-- PE0 receives partial arrays from all processors of the
	1539	!-- respective cross section and outputs them. Here a
	1540	!-- barrier has to be set, because otherwise
	1541	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
	1542	CALL MPI_BARRIER( comm2d, ierr )
	1543
[1551]	1544	ngp = ( nxrg-nxlg + 1 ) * ( nzt_do-nzb_do + 1 )
[493]	1545	IF ( myid == 0 ) THEN
[1]	1546	!
[493]	1547	!-- Local array can be relocated directly.
[1960]	1548	IF ( ( section(is,s_ind) >= nys .AND. &
	1549	section(is,s_ind) <= nyn ) .OR. &
	1550	( section(is,s_ind) == -1 .AND. &
	1551	nys-1 == -1 ) ) THEN
[1551]	1552	total_2d(nxlg:nxrg,nzb_do:nzt_do) = local_2d
[493]	1553	ENDIF
[1]	1554	!
[493]	1555	!-- Receive data from all other PEs.
	1556	DO n = 1, numprocs-1
	1557	!
	1558	!-- Receive index limits first, then array.
	1559	!-- Index limits are received in arbitrary order from
	1560	!-- the PEs.
[1320]	1561	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1562	MPI_ANY_SOURCE, 0, comm2d, &
[1]	1563	status, ierr )
[493]	1564	!
	1565	!-- Not all PEs have data for XZ-cross-section.
	1566	IF ( ind(1) /= -9999 ) THEN
	1567	sender = status(MPI_SOURCE)
	1568	DEALLOCATE( local_2d )
[1320]	1569	ALLOCATE( local_2d(ind(1):ind(2), &
[493]	1570	ind(3):ind(4)) )
	1571	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1572	MPI_REAL, sender, 1, comm2d, &
	1573	status, ierr )
[1320]	1574	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
[493]	1575	local_2d
	1576	ENDIF
	1577	ENDDO
	1578	!
	1579	!-- Relocate the local array for the next loop increment
	1580	DEALLOCATE( local_2d )
[1551]	1581	ALLOCATE( local_2d(nxlg:nxrg,nzb_do:nzt_do) )
[1]	1582
	1583	#if defined( __netcdf )
[1327]	1584	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1585	id_var_do2d(av,if), &
[1551]	1586	total_2d(0:nx+1,nzb_do:nzt_do),&
[1327]	1587	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
[1551]	1588	count = (/ nx+2, 1, nzt_do-nzb_do+1, 1 /) )
[1783]	1589	CALL netcdf_handle_error( 'data_output_2d', 58 )
[1]	1590	#endif
	1591
[493]	1592	ELSE
[1]	1593	!
[493]	1594	!-- If the cross section resides on the PE, send the
	1595	!-- local index limits, otherwise send -9999 to PE0.
[1960]	1596	IF ( ( section(is,s_ind) >= nys .AND. &
	1597	section(is,s_ind) <= nyn ) .OR. &
	1598	( section(is,s_ind) == -1 .AND. nys-1 == -1 ) ) &
[493]	1599	THEN
[667]	1600	ind(1) = nxlg; ind(2) = nxrg
[1551]	1601	ind(3) = nzb_do; ind(4) = nzt_do
[493]	1602	ELSE
	1603	ind(1) = -9999; ind(2) = -9999
	1604	ind(3) = -9999; ind(4) = -9999
	1605	ENDIF
[1320]	1606	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1607	comm2d, ierr )
	1608	!
	1609	!-- If applicable, send data to PE0.
	1610	IF ( ind(1) /= -9999 ) THEN
[1551]	1611	CALL MPI_SEND( local_2d(nxlg,nzb_do), ngp, &
[493]	1612	MPI_REAL, 0, 1, comm2d, ierr )
	1613	ENDIF
[1]	1614	ENDIF
	1615	!
[493]	1616	!-- A barrier has to be set, because otherwise some PEs may
	1617	!-- proceed too fast so that PE0 may receive wrong data on
	1618	!-- tag 0
	1619	CALL MPI_BARRIER( comm2d, ierr )
[1]	1620	ENDIF
[493]	1621
[1]	1622	ENDIF
	1623	#else
	1624	#if defined( __netcdf )
[1327]	1625	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1626	id_var_do2d(av,if), &
[1551]	1627	local_2d(nxl:nxr+1,nzb_do:nzt_do), &
[1327]	1628	start = (/ 1, is, 1, do2d_xz_time_count(av) /), &
[1551]	1629	count = (/ nx+2, 1, nzt_do-nzb_do+1, 1 /) )
[1783]	1630	CALL netcdf_handle_error( 'data_output_2d', 451 )
[1]	1631	#endif
	1632	#endif
	1633
	1634	CASE ( 'yz' )
	1635	!
[1308]	1636	!-- Update the netCDF yz cross section time axis.
	1637	!-- In case of parallel output, this is only done by PE0
	1638	!-- to increase the performance.
	1639	IF ( simulated_time /= do2d_yz_last_time(av) ) THEN
	1640	do2d_yz_time_count(av) = do2d_yz_time_count(av) + 1
	1641	do2d_yz_last_time(av) = simulated_time
	1642	IF ( myid == 0 ) THEN
[1327]	1643	IF ( .NOT. data_output_2d_on_each_pe &
	1644	.OR. netcdf_data_format > 4 ) &
[493]	1645	THEN
[1]	1646	#if defined( __netcdf )
	1647	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1648	id_var_time_yz(av), &
[291]	1649	(/ time_since_reference_point /), &
[1]	1650	start = (/ do2d_yz_time_count(av) /), &
	1651	count = (/ 1 /) )
[1783]	1652	CALL netcdf_handle_error( 'data_output_2d', 59 )
[1]	1653	#endif
	1654	ENDIF
	1655	ENDIF
[1308]	1656	ENDIF
[493]	1657
[1]	1658	!
	1659	!-- If required, carry out averaging along x
[1960]	1660	IF ( section(is,s_ind) == -1 ) THEN
[1]	1661
[1551]	1662	ALLOCATE( local_2d_l(nysg:nyng,nzb_do:nzt_do) )
[1353]	1663	local_2d_l = 0.0_wp
[1551]	1664	ngp = ( nyng-nysg+1 ) * ( nzt_do-nzb_do+1 )
[1]	1665	!
	1666	!-- First local averaging on the PE
[1551]	1667	DO k = nzb_do, nzt_do
[667]	1668	DO j = nysg, nyng
[1]	1669	DO i = nxl, nxr
[1320]	1670	local_2d_l(j,k) = local_2d_l(j,k) + &
[1]	1671	local_pf(i,j,k)
	1672	ENDDO
	1673	ENDDO
	1674	ENDDO
	1675	#if defined( __parallel )
	1676	!
	1677	!-- Now do the averaging over all PEs along x
[622]	1678	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1679	CALL MPI_ALLREDUCE( local_2d_l(nysg,nzb_do), &
	1680	local_2d(nysg,nzb_do), ngp, MPI_REAL, &
[1]	1681	MPI_SUM, comm1dx, ierr )
	1682	#else
	1683	local_2d = local_2d_l
	1684	#endif
[1353]	1685	local_2d = local_2d / ( nx + 1.0_wp )
[1]	1686
	1687	DEALLOCATE( local_2d_l )
	1688
	1689	ELSE
	1690	!
	1691	!-- Just store the respective section on the local array
	1692	!-- (but only if it is available on this PE!)
[1960]	1693	IF ( section(is,s_ind) >= nxl .AND. section(is,s_ind) <= nxr ) &
[1]	1694	THEN
[1960]	1695	local_2d = local_pf(section(is,s_ind),:,nzb_do:nzt_do)
[1]	1696	ENDIF
	1697
	1698	ENDIF
	1699
	1700	#if defined( __parallel )
[1327]	1701	IF ( netcdf_data_format > 4 ) THEN
[1]	1702	!
[1031]	1703	!-- Output in netCDF4/HDF5 format.
[493]	1704	!-- Output only on those PEs where the respective cross
	1705	!-- sections reside. Cross sections averaged along x are
	1706	!-- output on the respective first PE along x (myidx=0).
[1960]	1707	IF ( ( section(is,s_ind) >= nxl .AND. &
	1708	section(is,s_ind) <= nxr ) .OR. &
	1709	( section(is,s_ind) == -1 .AND. myidx == 0 ) ) THEN
[1]	1710	#if defined( __netcdf )
[493]	1711	!
[1308]	1712	!-- For parallel output, all cross sections are first
	1713	!-- stored here on a local array and will be written to the
	1714	!-- output file afterwards to increase the performance.
	1715	DO j = nysg, nyng
[1551]	1716	DO k = nzb_do, nzt_do
[1308]	1717	local_2d_sections_l(is,j,k) = local_2d(j,k)
	1718	ENDDO
	1719	ENDDO
[1]	1720	#endif
	1721	ENDIF
	1722
	1723	ELSE
	1724
[493]	1725	IF ( data_output_2d_on_each_pe ) THEN
[1]	1726	!
[493]	1727	!-- Output of partial arrays on each PE. If the cross
	1728	!-- section does not reside on the PE, output special
	1729	!-- index values.
	1730	#if defined( __netcdf )
[1327]	1731	IF ( myid == 0 ) THEN
[1320]	1732	WRITE ( 23 ) time_since_reference_point, &
[493]	1733	do2d_yz_time_count(av), av
	1734	ENDIF
	1735	#endif
[759]	1736	DO i = 0, io_blocks-1
	1737	IF ( i == io_group ) THEN
[1960]	1738	IF ( ( section(is,s_ind) >= nxl .AND. &
	1739	section(is,s_ind) <= nxr ) .OR. &
	1740	( section(is,s_ind) == -1 .AND. &
[1320]	1741	nxl-1 == -1 ) ) &
[759]	1742	THEN
[1551]	1743	WRITE (23) nysg, nyng, nzb_do, nzt_do, nzb, nzt+1
[759]	1744	WRITE (23) local_2d
	1745	ELSE
[1551]	1746	WRITE (23) -1, -1, -1, -1, -1, -1
[759]	1747	ENDIF
	1748	ENDIF
	1749	#if defined( __parallel )
	1750	CALL MPI_BARRIER( comm2d, ierr )
	1751	#endif
	1752	ENDDO
[493]	1753
	1754	ELSE
[1]	1755	!
[493]	1756	!-- PE0 receives partial arrays from all processors of the
	1757	!-- respective cross section and outputs them. Here a
	1758	!-- barrier has to be set, because otherwise
	1759	!-- "-MPI- FATAL: Remote protocol queue full" may occur.
	1760	CALL MPI_BARRIER( comm2d, ierr )
	1761
[1551]	1762	ngp = ( nyng-nysg+1 ) * ( nzt_do-nzb_do+1 )
[493]	1763	IF ( myid == 0 ) THEN
[1]	1764	!
[493]	1765	!-- Local array can be relocated directly.
[1960]	1766	IF ( ( section(is,s_ind) >= nxl .AND. &
	1767	section(is,s_ind) <= nxr ) .OR. &
	1768	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
[493]	1769	THEN
[1551]	1770	total_2d(nysg:nyng,nzb_do:nzt_do) = local_2d
[493]	1771	ENDIF
[1]	1772	!
[493]	1773	!-- Receive data from all other PEs.
	1774	DO n = 1, numprocs-1
	1775	!
	1776	!-- Receive index limits first, then array.
	1777	!-- Index limits are received in arbitrary order from
	1778	!-- the PEs.
[1320]	1779	CALL MPI_RECV( ind(1), 4, MPI_INTEGER, &
	1780	MPI_ANY_SOURCE, 0, comm2d, &
[1]	1781	status, ierr )
[493]	1782	!
	1783	!-- Not all PEs have data for YZ-cross-section.
	1784	IF ( ind(1) /= -9999 ) THEN
	1785	sender = status(MPI_SOURCE)
	1786	DEALLOCATE( local_2d )
[1320]	1787	ALLOCATE( local_2d(ind(1):ind(2), &
[493]	1788	ind(3):ind(4)) )
	1789	CALL MPI_RECV( local_2d(ind(1),ind(3)), ngp, &
	1790	MPI_REAL, sender, 1, comm2d, &
	1791	status, ierr )
[1320]	1792	total_2d(ind(1):ind(2),ind(3):ind(4)) = &
[493]	1793	local_2d
	1794	ENDIF
	1795	ENDDO
	1796	!
	1797	!-- Relocate the local array for the next loop increment
	1798	DEALLOCATE( local_2d )
[1551]	1799	ALLOCATE( local_2d(nysg:nyng,nzb_do:nzt_do) )
[1]	1800
	1801	#if defined( __netcdf )
[1327]	1802	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1803	id_var_do2d(av,if), &
[1551]	1804	total_2d(0:ny+1,nzb_do:nzt_do),&
[1327]	1805	start = (/ is, 1, 1, do2d_yz_time_count(av) /), &
[1551]	1806	count = (/ 1, ny+2, nzt_do-nzb_do+1, 1 /) )
[1783]	1807	CALL netcdf_handle_error( 'data_output_2d', 61 )
[1]	1808	#endif
	1809
[493]	1810	ELSE
[1]	1811	!
[493]	1812	!-- If the cross section resides on the PE, send the
	1813	!-- local index limits, otherwise send -9999 to PE0.
[1960]	1814	IF ( ( section(is,s_ind) >= nxl .AND. &
	1815	section(is,s_ind) <= nxr ) .OR. &
	1816	( section(is,s_ind) == -1 .AND. nxl-1 == -1 ) ) &
[493]	1817	THEN
[667]	1818	ind(1) = nysg; ind(2) = nyng
[1551]	1819	ind(3) = nzb_do; ind(4) = nzt_do
[493]	1820	ELSE
	1821	ind(1) = -9999; ind(2) = -9999
	1822	ind(3) = -9999; ind(4) = -9999
	1823	ENDIF
[1320]	1824	CALL MPI_SEND( ind(1), 4, MPI_INTEGER, 0, 0, &
[493]	1825	comm2d, ierr )
	1826	!
	1827	!-- If applicable, send data to PE0.
	1828	IF ( ind(1) /= -9999 ) THEN
[1551]	1829	CALL MPI_SEND( local_2d(nysg,nzb_do), ngp, &
[493]	1830	MPI_REAL, 0, 1, comm2d, ierr )
	1831	ENDIF
[1]	1832	ENDIF
	1833	!
[493]	1834	!-- A barrier has to be set, because otherwise some PEs may
	1835	!-- proceed too fast so that PE0 may receive wrong data on
	1836	!-- tag 0
	1837	CALL MPI_BARRIER( comm2d, ierr )
[1]	1838	ENDIF
[493]	1839
[1]	1840	ENDIF
	1841	#else
	1842	#if defined( __netcdf )
[1327]	1843	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	1844	id_var_do2d(av,if), &
[1551]	1845	local_2d(nys:nyn+1,nzb_do:nzt_do), &
[1327]	1846	start = (/ is, 1, 1, do2d_xz_time_count(av) /), &
[1551]	1847	count = (/ 1, ny+2, nzt_do-nzb_do+1, 1 /) )
[1783]	1848	CALL netcdf_handle_error( 'data_output_2d', 452 )
[1]	1849	#endif
	1850	#endif
	1851
	1852	END SELECT
	1853
	1854	is = is + 1
	1855	ENDDO loop1
	1856
[1308]	1857	!
	1858	!-- For parallel output, all data were collected before on a local array
	1859	!-- and are written now to the netcdf file. This must be done to increase
	1860	!-- the performance of the parallel output.
	1861	#if defined( __netcdf )
[1327]	1862	IF ( netcdf_data_format > 4 ) THEN
[1308]	1863
	1864	SELECT CASE ( mode )
	1865
	1866	CASE ( 'xy' )
	1867	IF ( two_d ) THEN
[1703]	1868	nis = 1
	1869	two_d = .FALSE.
[1308]	1870	ELSE
[1703]	1871	nis = ns
[1308]	1872	ENDIF
	1873	!
	1874	!-- Do not output redundant ghost point data except for the
	1875	!-- boundaries of the total domain.
	1876	IF ( nxr == nx .AND. nyn /= ny ) THEN
	1877	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1878	id_var_do2d(av,if), &
	1879	local_2d_sections(nxl:nxr+1, &
[1703]	1880	nys:nyn,1:nis), &
[1308]	1881	start = (/ nxl+1, nys+1, 1, &
	1882	do2d_xy_time_count(av) /), &
	1883	count = (/ nxr-nxl+2, &
[1703]	1884	nyn-nys+1, nis, 1 &
[1308]	1885	/) )
	1886	ELSEIF ( nxr /= nx .AND. nyn == ny ) THEN
	1887	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1888	id_var_do2d(av,if), &
	1889	local_2d_sections(nxl:nxr, &
[1703]	1890	nys:nyn+1,1:nis), &
[1308]	1891	start = (/ nxl+1, nys+1, 1, &
	1892	do2d_xy_time_count(av) /), &
	1893	count = (/ nxr-nxl+1, &
[1703]	1894	nyn-nys+2, nis, 1 &
[1308]	1895	/) )
	1896	ELSEIF ( nxr == nx .AND. nyn == ny ) THEN
	1897	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1898	id_var_do2d(av,if), &
	1899	local_2d_sections(nxl:nxr+1, &
[1703]	1900	nys:nyn+1,1:nis), &
[1308]	1901	start = (/ nxl+1, nys+1, 1, &
	1902	do2d_xy_time_count(av) /), &
	1903	count = (/ nxr-nxl+2, &
[1703]	1904	nyn-nys+2, nis, 1 &
[1308]	1905	/) )
	1906	ELSE
	1907	nc_stat = NF90_PUT_VAR( id_set_xy(av), &
	1908	id_var_do2d(av,if), &
	1909	local_2d_sections(nxl:nxr, &
[1703]	1910	nys:nyn,1:nis), &
[1308]	1911	start = (/ nxl+1, nys+1, 1, &
	1912	do2d_xy_time_count(av) /), &
	1913	count = (/ nxr-nxl+1, &
[1703]	1914	nyn-nys+1, nis, 1 &
[1308]	1915	/) )
	1916	ENDIF
	1917
[1783]	1918	CALL netcdf_handle_error( 'data_output_2d', 55 )
[1308]	1919
	1920	CASE ( 'xz' )
	1921	!
	1922	!-- First, all PEs get the information of all cross-sections.
	1923	!-- Then the data are written to the output file by all PEs
	1924	!-- while NF90_COLLECTIVE is set in subroutine
	1925	!-- define_netcdf_header. Although redundant information are
	1926	!-- written to the output file in that case, the performance
	1927	!-- is significantly better compared to the case where only
	1928	!-- the first row of PEs in x-direction (myidx = 0) is given
	1929	!-- the output while NF90_INDEPENDENT is set.
	1930	IF ( npey /= 1 ) THEN
	1931
	1932	#if defined( __parallel )
	1933	!
	1934	!-- Distribute data over all PEs along y
[1551]	1935	ngp = ( nxrg-nxlg+1 ) * ( nzt_do-nzb_do+1 ) * ns
[1308]	1936	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1937	CALL MPI_ALLREDUCE( local_2d_sections_l(nxlg,1,nzb_do), &
	1938	local_2d_sections(nxlg,1,nzb_do), &
[1308]	1939	ngp, MPI_REAL, MPI_SUM, comm1dy, &
	1940	ierr )
	1941	#else
	1942	local_2d_sections = local_2d_sections_l
	1943	#endif
	1944	ENDIF
	1945	!
	1946	!-- Do not output redundant ghost point data except for the
	1947	!-- boundaries of the total domain.
	1948	IF ( nxr == nx ) THEN
	1949	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1950	id_var_do2d(av,if), &
	1951	local_2d_sections(nxl:nxr+1,1:ns, &
[1551]	1952	nzb_do:nzt_do), &
[1308]	1953	start = (/ nxl+1, 1, 1, &
	1954	do2d_xz_time_count(av) /), &
[1551]	1955	count = (/ nxr-nxl+2, ns, nzt_do-nzb_do+1, &
[1308]	1956	1 /) )
	1957	ELSE
	1958	nc_stat = NF90_PUT_VAR( id_set_xz(av), &
	1959	id_var_do2d(av,if), &
	1960	local_2d_sections(nxl:nxr,1:ns, &
[1551]	1961	nzb_do:nzt_do), &
[1308]	1962	start = (/ nxl+1, 1, 1, &
	1963	do2d_xz_time_count(av) /), &
[1551]	1964	count = (/ nxr-nxl+1, ns, nzt_do-nzb_do+1, &
[1308]	1965	1 /) )
	1966	ENDIF
	1967
[1783]	1968	CALL netcdf_handle_error( 'data_output_2d', 57 )
[1308]	1969
	1970	CASE ( 'yz' )
	1971	!
	1972	!-- First, all PEs get the information of all cross-sections.
	1973	!-- Then the data are written to the output file by all PEs
	1974	!-- while NF90_COLLECTIVE is set in subroutine
	1975	!-- define_netcdf_header. Although redundant information are
	1976	!-- written to the output file in that case, the performance
	1977	!-- is significantly better compared to the case where only
	1978	!-- the first row of PEs in y-direction (myidy = 0) is given
	1979	!-- the output while NF90_INDEPENDENT is set.
	1980	IF ( npex /= 1 ) THEN
	1981
	1982	#if defined( __parallel )
	1983	!
	1984	!-- Distribute data over all PEs along x
	1985	ngp = ( nyng-nysg+1 ) * ( nzt-nzb + 2 ) * ns
	1986	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1551]	1987	CALL MPI_ALLREDUCE( local_2d_sections_l(1,nysg,nzb_do), &
	1988	local_2d_sections(1,nysg,nzb_do), &
[1308]	1989	ngp, MPI_REAL, MPI_SUM, comm1dx, &
	1990	ierr )
	1991	#else
	1992	local_2d_sections = local_2d_sections_l
	1993	#endif
	1994	ENDIF
	1995	!
	1996	!-- Do not output redundant ghost point data except for the
	1997	!-- boundaries of the total domain.
	1998	IF ( nyn == ny ) THEN
	1999	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	2000	id_var_do2d(av,if), &
	2001	local_2d_sections(1:ns, &
[1551]	2002	nys:nyn+1,nzb_do:nzt_do), &
[1308]	2003	start = (/ 1, nys+1, 1, &
	2004	do2d_yz_time_count(av) /), &
	2005	count = (/ ns, nyn-nys+2, &
[1551]	2006	nzt_do-nzb_do+1, 1 /) )
[1308]	2007	ELSE
	2008	nc_stat = NF90_PUT_VAR( id_set_yz(av), &
	2009	id_var_do2d(av,if), &
	2010	local_2d_sections(1:ns,nys:nyn, &
[1551]	2011	nzb_do:nzt_do), &
[1308]	2012	start = (/ 1, nys+1, 1, &
	2013	do2d_yz_time_count(av) /), &
	2014	count = (/ ns, nyn-nys+1, &
[1551]	2015	nzt_do-nzb_do+1, 1 /) )
[1308]	2016	ENDIF
	2017
[1783]	2018	CALL netcdf_handle_error( 'data_output_2d', 60 )
[1308]	2019
	2020	CASE DEFAULT
	2021	message_string = 'unknown cross-section: ' // TRIM( mode )
	2022	CALL message( 'data_output_2d', 'PA0180', 1, 2, 0, 6, 0 )
	2023
	2024	END SELECT
	2025
	2026	ENDIF
[1311]	2027	#endif
[1]	2028	ENDIF
	2029
	2030	if = if + 1
	2031	l = MAX( 2, LEN_TRIM( do2d(av,if) ) )
	2032	do2d_mode = do2d(av,if)(l-1:l)
	2033
	2034	ENDDO
	2035
	2036	!
	2037	!-- Deallocate temporary arrays.
	2038	IF ( ALLOCATED( level_z ) ) DEALLOCATE( level_z )
[1308]	2039	IF ( netcdf_data_format > 4 ) THEN
	2040	DEALLOCATE( local_pf, local_2d, local_2d_sections )
	2041	IF( mode == 'xz' .OR. mode == 'yz' ) DEALLOCATE( local_2d_sections_l )
	2042	ENDIF
[1]	2043	#if defined( __parallel )
	2044	IF ( .NOT. data_output_2d_on_each_pe .AND. myid == 0 ) THEN
	2045	DEALLOCATE( total_2d )
	2046	ENDIF
	2047	#endif
	2048
	2049	!
	2050	!-- Close plot output file.
[1960]	2051	file_id = 20 + s_ind
[1]	2052
	2053	IF ( data_output_2d_on_each_pe ) THEN
[759]	2054	DO i = 0, io_blocks-1
	2055	IF ( i == io_group ) THEN
	2056	CALL close_file( file_id )
	2057	ENDIF
	2058	#if defined( __parallel )
	2059	CALL MPI_BARRIER( comm2d, ierr )
	2060	#endif
	2061	ENDDO
[1]	2062	ELSE
	2063	IF ( myid == 0 ) CALL close_file( file_id )
	2064	ENDIF
	2065
[1318]	2066	CALL cpu_log( log_point(3), 'data_output_2d', 'stop' )
[1]	2067
	2068	END SUBROUTINE data_output_2d

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

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |