- Timestamp:
- Aug 19, 2009 4:56:44 PM (15 years ago)
- Location:
- palm/trunk
- Files:
-
- 6 edited
Legend:
- Unmodified
- Added
- Removed
-
palm/trunk/DOC/app/appendix_a.html
r354 r359 1 1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> 2 <html> 3 <head> 2 <html><head> 4 3 5 4 … … 7 6 8 7 9 <title>PALM Appendix A</title> 10 </head> 11 12 13 <body> 8 <title>PALM Appendix A</title></head> 9 10 <body> 14 11 15 12 … … 3665 3662 3666 3663 <td> 3667 <p>**** contains numbers in the range from 1-4 28</p>3664 <p>**** contains numbers in the range from 1-444</p> 3668 3665 </td> 3669 3666 … … 3681 3678 3682 3679 3683 </body> 3684 </html> 3680 </body></html> -
palm/trunk/DOC/app/chapter_3.4.html
r270 r359 3349 3349 the subroutine <span style="font-family: Cumberland AMT;">init_grid</span> reads 3350 3350 the topography height information (in m above ground) for each grid point in 3351 a free floating point format. The data on file are laid out naturally,3351 a free floating point format. <font color="#000000">The ascii file format is <a href="http://en.wikipedia.org/wiki/ESRI_grid#ASCII" target="_blank">ESRI grid</a> - also known as <a href="http://en.wikipedia.org/wiki/ESRI_grid#ASCII" target="_blank">ARC/INFO ASCII GRID</a> <span style="font-style: italic;">- without the header</span>. </font>The data on file are laid out naturally, 3352 3352 i.e. in W-E orientation horizontally and in S-N orientation vertically, 3353 3353 they must thus be organized as follows:<br> … … 3372 3372 3373 3373 3374 </ul> 3375 3376 3377 Layout sketch:<br> 3374 </ul>Layout sketch:<br> 3378 3375 3379 3376 -
palm/trunk/SCRIPTS/mrun
r352 r359 185 185 # 08/07/09 - Siggi - default value for email_notification is none, 186 186 # option -e added to subjob 187 # 19/08/09 - Marcus - MPI debug option 187 188 188 189 … … 2866 2867 PATH=$PATH:$TEMPDIR 2867 2868 2869 # MPI debug option (argument checking, slows down execution due to increased latency) 2870 if [[ "$mpi_debug" = true ]] 2871 then 2872 export MPI_CHECK_ARGS=1 2873 printf "\n +++ MPI_CHECK_ARGS=$MPI_CHECK_ARGS" 2874 fi 2875 2868 2876 if [[ "$cond1" = debug || "$cond2" = debug ]] 2869 2877 then -
palm/trunk/SOURCE/CURRENT_MODIFICATIONS
r355 r359 51 51 Changed: 52 52 ------- 53 For extended NetCDF files, the updated title attribute includes an update of 54 time_average_text where appropriate. (netcdf) 55 53 56 In case of restart runs without extension, initial profiles are not written 54 57 to NetCDF-file anymore. (data_output_profiles, modules, read_var_list, write_var_list) … … 110 113 Errors: 111 114 ------ 115 Bugfix for extended NetCDF files: In order to avoid 'data mode' errors if 116 updated attributes are larger than their original size, NF90_PUT_ATT is called 117 in 'define mode' enclosed by NF90_REDEF and NF90_ENDDEF calls. This implies a 118 possible performance loss; an alternative strategy would be to ensure equal 119 attribute size in a job chain. (netcdf) 120 121 Bugfix: correction of initial volume flow for non-flat topography (init_3d_model) 122 Bugfix: zero initialization of arrays within buildings for 'cyclic_fill' (init_3d_model) 123 112 124 Bugfix: to_be_resorted => s_av for time-averaged scalars (data_output_2d, data_output_3d) 113 125 -
palm/trunk/SOURCE/init_3d_model.f90
r333 r359 7 7 ! Current revisions: 8 8 ! ----------------- 9 ! bugfix: correction of initial volume flow for non-flat topography 10 ! bugfix: zero initialization of arrays within buildings for 'cyclic_fill' 9 11 ! bugfix: avoid that ngp_2dh_s_inner becomes zero 10 12 ! initializing_actions='read_data_for_recycling' renamed to 'cyclic_fill', now … … 485 487 DO j = nys, nyn 486 488 DO k = nzb + 1, nzb_u_inner(j,nx) 487 u_nzb_p1_for_vfc(j) = u1d(k) * dzu(k) 489 u_nzb_p1_for_vfc(j) = u_nzb_p1_for_vfc(j) + & 490 u1d(k) * dzu(k) 488 491 ENDDO 489 492 ENDDO … … 492 495 DO i = nxl, nxr 493 496 DO k = nzb + 1, nzb_v_inner(ny,i) 494 v_nzb_p1_for_vfc(i) = v1d(k) * dzu(k) 497 v_nzb_p1_for_vfc(i) = v_nzb_p1_for_vfc(i) + & 498 v1d(k) * dzu(k) 495 499 ENDDO 496 500 ENDDO … … 558 562 DO j = nys, nyn 559 563 DO k = nzb + 1, nzb_u_inner(j,nx) + 1 560 u_nzb_p1_for_vfc(j) = u_init(k) * dzu(k) 564 u_nzb_p1_for_vfc(j) = u_nzb_p1_for_vfc(j) + & 565 u_init(k) * dzu(k) 561 566 ENDDO 562 567 ENDDO … … 565 570 DO i = nxl, nxr 566 571 DO k = nzb + 1, nzb_v_inner(ny,i) + 1 567 v_nzb_p1_for_vfc(i) = v_init(k) * dzu(k) 572 v_nzb_p1_for_vfc(i) = v_nzb_p1_for_vfc(i) + & 573 v_init(k) * dzu(k) 568 574 ENDDO 569 575 ENDDO … … 1028 1034 1029 1035 ! 1036 !-- Inside buildings set velocities and TKE back to zero 1037 IF ( TRIM( initializing_actions ) == 'cyclic_fill' .AND. & 1038 topography /= 'flat' ) THEN 1039 ! 1040 !-- Correction of initial volume flow 1041 IF ( conserve_volume_flow ) THEN 1042 IF ( nxr == nx ) THEN 1043 DO j = nys, nyn 1044 DO k = nzb + 1, nzb_u_inner(j,nx) 1045 u_nzb_p1_for_vfc(j) = u_nzb_p1_for_vfc(j) + & 1046 u(k,j,nx) * dzu(k) 1047 ENDDO 1048 ENDDO 1049 ENDIF 1050 IF ( nyn == ny ) THEN 1051 DO i = nxl, nxr 1052 DO k = nzb + 1, nzb_v_inner(ny,i) 1053 v_nzb_p1_for_vfc(i) = v_nzb_p1_for_vfc(i) + & 1054 v(k,ny,i) * dzu(k) 1055 ENDDO 1056 ENDDO 1057 ENDIF 1058 ENDIF 1059 1060 ! 1061 !-- Inside buildings set velocities and TKE back to zero. 1062 !-- Other scalars (pt, q, s, km, kh, p, sa, ...) are ignored at present, 1063 !-- maybe revise later. 1064 IF ( timestep_scheme(1:5) == 'runge' ) THEN 1065 DO i = nxl-1, nxr+1 1066 DO j = nys-1, nyn+1 1067 u (nzb:nzb_u_inner(j,i),j,i) = 0.0 1068 v (nzb:nzb_v_inner(j,i),j,i) = 0.0 1069 w (nzb:nzb_w_inner(j,i),j,i) = 0.0 1070 e (nzb:nzb_w_inner(j,i),j,i) = 0.0 1071 u_m(nzb:nzb_u_inner(j,i),j,i) = 0.0 1072 v_m(nzb:nzb_v_inner(j,i),j,i) = 0.0 1073 w_m(nzb:nzb_w_inner(j,i),j,i) = 0.0 1074 e_m(nzb:nzb_w_inner(j,i),j,i) = 0.0 1075 tu_m(nzb:nzb_u_inner(j,i),j,i) = 0.0 1076 tv_m(nzb:nzb_v_inner(j,i),j,i) = 0.0 1077 tw_m(nzb:nzb_w_inner(j,i),j,i) = 0.0 1078 te_m(nzb:nzb_w_inner(j,i),j,i) = 0.0 1079 tpt_m(nzb:nzb_w_inner(j,i),j,i) = 0.0 1080 ENDDO 1081 ENDDO 1082 ELSE 1083 DO i = nxl-1, nxr+1 1084 DO j = nys-1, nyn+1 1085 u (nzb:nzb_u_inner(j,i),j,i) = 0.0 1086 v (nzb:nzb_v_inner(j,i),j,i) = 0.0 1087 w (nzb:nzb_w_inner(j,i),j,i) = 0.0 1088 e (nzb:nzb_w_inner(j,i),j,i) = 0.0 1089 u_m(nzb:nzb_u_inner(j,i),j,i) = 0.0 1090 v_m(nzb:nzb_v_inner(j,i),j,i) = 0.0 1091 w_m(nzb:nzb_w_inner(j,i),j,i) = 0.0 1092 e_m(nzb:nzb_w_inner(j,i),j,i) = 0.0 1093 u_p(nzb:nzb_u_inner(j,i),j,i) = 0.0 1094 v_p(nzb:nzb_v_inner(j,i),j,i) = 0.0 1095 w_p(nzb:nzb_w_inner(j,i),j,i) = 0.0 1096 e_p(nzb:nzb_w_inner(j,i),j,i) = 0.0 1097 ENDDO 1098 ENDDO 1099 ENDIF 1100 1101 ENDIF 1102 1103 ! 1030 1104 !-- Calculate the initial volume flow at the right and north boundary 1031 1105 IF ( conserve_volume_flow .AND. & … … 1043 1117 ENDDO 1044 1118 ! 1045 !-- Correction if velocity at nzb+1 has been set zero further above1046 !-- Note: at present, u_nzb_p1_for_vfc is zero (maybe revise later)1119 !-- Correction if velocity inside buildings has been set to zero 1120 !-- further above 1047 1121 volume_flow_initial_l(1) = volume_flow_initial_l(1) + & 1048 1122 u_nzb_p1_for_vfc(j) … … 1058 1132 ENDDO 1059 1133 ! 1060 !-- Correction if velocity at nzb+1 has been set zero further above1061 !-- Note: at present, v_nzb_p1_for_vfc is zero (maybe revise later)1134 !-- Correction if velocity inside buildings has been set to zero 1135 !-- further above 1062 1136 volume_flow_initial_l(2) = volume_flow_initial_l(2) + & 1063 1137 v_nzb_p1_for_vfc(i) -
palm/trunk/SOURCE/netcdf.f90
r292 r359 7 7 ! Current revisions: 8 8 ! ------------------ 9 ! For extended NetCDF files, the updated title attribute includes an update of 10 ! time_average_text where appropriate. 11 ! Bugfix for extended NetCDF files: In order to avoid 'data mode' errors if 12 ! updated attributes are larger than their original size, NF90_PUT_ATT is called 13 ! in 'define mode' enclosed by NF90_REDEF and NF90_ENDDEF calls. This implies a 14 ! possible performance loss; an alternative strategy would be to ensure equal 15 ! attribute size in a job chain. 9 16 ! NetCDF unit attribute in timeseries output in case of statistic 10 17 ! regions added. … … 650 657 651 658 ! 652 !-- Change the title attribute on file 659 !-- Update the title attribute on file 660 !-- In order to avoid 'data mode' errors if updated attributes are larger 661 !-- than their original size, NF90_PUT_ATT is called in 'define mode' 662 !-- enclosed by NF90_REDEF and NF90_ENDDEF calls. This implies a possible 663 !-- performance loss due to data copying; an alternative strategy would be 664 !-- to ensure equal attribute size. Maybe revise later. 665 IF ( av == 0 ) THEN 666 time_average_text = ' ' 667 ELSE 668 WRITE (time_average_text, '('', '',F7.1,'' s average'')') & 669 averaging_interval 670 ENDIF 671 nc_stat = NF90_REDEF( id_set_3d(av) ) 672 CALL handle_netcdf_error( 'netcdf', 429 ) 653 673 nc_stat = NF90_PUT_ATT( id_set_3d(av), NF90_GLOBAL, 'title', & 654 TRIM( run_description_header ) ) 674 TRIM( run_description_header ) // & 675 TRIM( time_average_text ) ) 655 676 CALL handle_netcdf_error( 'netcdf', 96 ) 677 nc_stat = NF90_ENDDEF( id_set_3d(av) ) 678 CALL handle_netcdf_error( 'netcdf', 430 ) 656 679 message_string = 'NetCDF file for volume data ' // & 657 680 TRIM( var ) // ' from previous run found.' // & … … 1276 1299 1277 1300 ! 1278 !-- Change the title attribute on file 1301 !-- Update the title attribute on file 1302 !-- In order to avoid 'data mode' errors if updated attributes are larger 1303 !-- than their original size, NF90_PUT_ATT is called in 'define mode' 1304 !-- enclosed by NF90_REDEF and NF90_ENDDEF calls. This implies a possible 1305 !-- performance loss due to data copying; an alternative strategy would be 1306 !-- to ensure equal attribute size in a job chain. Maybe revise later. 1307 IF ( av == 0 ) THEN 1308 time_average_text = ' ' 1309 ELSE 1310 WRITE (time_average_text, '('', '',F7.1,'' s average'')') & 1311 averaging_interval 1312 ENDIF 1313 nc_stat = NF90_REDEF( id_set_xy(av) ) 1314 CALL handle_netcdf_error( 'netcdf', 431 ) 1279 1315 nc_stat = NF90_PUT_ATT( id_set_xy(av), NF90_GLOBAL, 'title', & 1280 TRIM( run_description_header ) ) 1316 TRIM( run_description_header ) // & 1317 TRIM( time_average_text ) ) 1281 1318 CALL handle_netcdf_error( 'netcdf', 139 ) 1319 nc_stat = NF90_ENDDEF( id_set_xy(av) ) 1320 CALL handle_netcdf_error( 'netcdf', 432 ) 1282 1321 message_string = 'NetCDF file for cross-sections ' // & 1283 1322 TRIM( var ) // ' from previous run found.' // & … … 1798 1837 1799 1838 ! 1800 !-- Change the title attribute on file 1839 !-- Update the title attribute on file 1840 !-- In order to avoid 'data mode' errors if updated attributes are larger 1841 !-- than their original size, NF90_PUT_ATT is called in 'define mode' 1842 !-- enclosed by NF90_REDEF and NF90_ENDDEF calls. This implies a possible 1843 !-- performance loss due to data copying; an alternative strategy would be 1844 !-- to ensure equal attribute size in a job chain. Maybe revise later. 1845 IF ( av == 0 ) THEN 1846 time_average_text = ' ' 1847 ELSE 1848 WRITE (time_average_text, '('', '',F7.1,'' s average'')') & 1849 averaging_interval 1850 ENDIF 1851 nc_stat = NF90_REDEF( id_set_xz(av) ) 1852 CALL handle_netcdf_error( 'netcdf', 433 ) 1801 1853 nc_stat = NF90_PUT_ATT( id_set_xz(av), NF90_GLOBAL, 'title', & 1802 TRIM( run_description_header ) ) 1854 TRIM( run_description_header ) // & 1855 TRIM( time_average_text ) ) 1803 1856 CALL handle_netcdf_error( 'netcdf', 178 ) 1857 nc_stat = NF90_ENDDEF( id_set_xz(av) ) 1858 CALL handle_netcdf_error( 'netcdf', 434 ) 1804 1859 message_string = 'NetCDF file for cross-sections ' // & 1805 1860 TRIM( var ) // ' from previous run found.' // & … … 2318 2373 2319 2374 ! 2320 !-- Change the title attribute on file 2375 !-- Update the title attribute on file 2376 !-- In order to avoid 'data mode' errors if updated attributes are larger 2377 !-- than their original size, NF90_PUT_ATT is called in 'define mode' 2378 !-- enclosed by NF90_REDEF and NF90_ENDDEF calls. This implies a possible 2379 !-- performance loss due to data copying; an alternative strategy would be 2380 !-- to ensure equal attribute size in a job chain. Maybe revise later. 2381 IF ( av == 0 ) THEN 2382 time_average_text = ' ' 2383 ELSE 2384 WRITE (time_average_text, '('', '',F7.1,'' s average'')') & 2385 averaging_interval 2386 ENDIF 2387 nc_stat = NF90_REDEF( id_set_yz(av) ) 2388 CALL handle_netcdf_error( 'netcdf', 435 ) 2321 2389 nc_stat = NF90_PUT_ATT( id_set_yz(av), NF90_GLOBAL, 'title', & 2322 TRIM( run_description_header ) ) 2390 TRIM( run_description_header ) // & 2391 TRIM( time_average_text ) ) 2323 2392 CALL handle_netcdf_error( 'netcdf', 217 ) 2393 nc_stat = NF90_ENDDEF( id_set_yz(av) ) 2394 CALL handle_netcdf_error( 'netcdf', 436 ) 2324 2395 message_string = 'NetCDF file for cross-sections ' // & 2325 2396 TRIM( var ) // ' from previous run found.' // & … … 2611 2682 2612 2683 ! 2613 !-- Change the title attribute on file 2684 !-- Update the title attribute on file 2685 !-- In order to avoid 'data mode' errors if updated attributes are larger 2686 !-- than their original size, NF90_PUT_ATT is called in 'define mode' 2687 !-- enclosed by NF90_REDEF and NF90_ENDDEF calls. This implies a possible 2688 !-- performance loss due to data copying; an alternative strategy would be 2689 !-- to ensure equal attribute size in a job chain. Maybe revise later. 2690 IF ( averaging_interval_pr == 0.0 ) THEN 2691 time_average_text = ' ' 2692 ELSE 2693 WRITE (time_average_text, '('', '',F7.1,'' s average'')') & 2694 averaging_interval_pr 2695 ENDIF 2696 nc_stat = NF90_REDEF( id_set_pr ) 2697 CALL handle_netcdf_error( 'netcdf', 437 ) 2614 2698 nc_stat = NF90_PUT_ATT( id_set_pr, NF90_GLOBAL, 'title', & 2615 TRIM( run_description_header ) ) 2699 TRIM( run_description_header ) // & 2700 TRIM( time_average_text ) ) 2616 2701 CALL handle_netcdf_error( 'netcdf', 248 ) 2702 nc_stat = NF90_ENDDEF( id_set_pr ) 2703 CALL handle_netcdf_error( 'netcdf', 438 ) 2617 2704 message_string = 'NetCDF file for vertical profiles ' // & 2618 2705 'from previous run found.' // & … … 2817 2904 2818 2905 ! 2819 !-- Change the title attribute on file 2906 !-- Update the title attribute on file 2907 !-- In order to avoid 'data mode' errors if updated attributes are larger 2908 !-- than their original size, NF90_PUT_ATT is called in 'define mode' 2909 !-- enclosed by NF90_REDEF and NF90_ENDDEF calls. This implies a possible 2910 !-- performance loss due to data copying; an alternative strategy would be 2911 !-- to ensure equal attribute size in a job chain. Maybe revise later. 2912 nc_stat = NF90_REDEF( id_set_ts ) 2913 CALL handle_netcdf_error( 'netcdf', 439 ) 2820 2914 nc_stat = NF90_PUT_ATT( id_set_ts, NF90_GLOBAL, 'title', & 2821 2915 TRIM( run_description_header ) ) 2822 2916 CALL handle_netcdf_error( 'netcdf', 267 ) 2917 nc_stat = NF90_ENDDEF( id_set_ts ) 2918 CALL handle_netcdf_error( 'netcdf', 440 ) 2823 2919 message_string = 'NetCDF file for time series ' // & 2824 2920 'from previous run found.' // & … … 3230 3326 3231 3327 ! 3232 !-- Change the title attribute on file 3328 !-- Update the title attribute on file 3329 !-- In order to avoid 'data mode' errors if updated attributes are larger 3330 !-- than their original size, NF90_PUT_ATT is called in 'define mode' 3331 !-- enclosed by NF90_REDEF and NF90_ENDDEF calls. This implies a possible 3332 !-- performance loss due to data copying; an alternative strategy would be 3333 !-- to ensure equal attribute size in a job chain. Maybe revise later. 3334 nc_stat = NF90_REDEF( id_set_sp ) 3335 CALL handle_netcdf_error( 'netcdf', 441 ) 3233 3336 IF ( averaging_interval_sp /= 0.0 ) THEN 3234 3337 WRITE (time_average_text,'('', '',F7.1,'' s average'')') & … … 3247 3350 ENDIF 3248 3351 CALL handle_netcdf_error( 'netcdf', 309 ) 3352 nc_stat = NF90_ENDDEF( id_set_sp ) 3353 CALL handle_netcdf_error( 'netcdf', 442 ) 3249 3354 message_string = 'NetCDF file for spectra ' // & 3250 3355 'from previous run found.' // & … … 3586 3691 3587 3692 ! 3588 !-- Change the title attribute on file 3693 !-- Update the title attribute on file 3694 !-- In order to avoid 'data mode' errors if updated attributes are larger 3695 !-- than their original size, NF90_PUT_ATT is called in 'define mode' 3696 !-- enclosed by NF90_REDEF and NF90_ENDDEF calls. This implies a possible 3697 !-- performance loss due to data copying; an alternative strategy would be 3698 !-- to ensure equal attribute size in a job chain. Maybe revise later. 3699 nc_stat = NF90_REDEF( id_set_pts ) 3700 CALL handle_netcdf_error( 'netcdf', 443 ) 3589 3701 nc_stat = NF90_PUT_ATT( id_set_pts, NF90_GLOBAL, 'title', & 3590 3702 TRIM( run_description_header ) ) 3591 3703 CALL handle_netcdf_error( 'netcdf', 411 ) 3704 nc_stat = NF90_ENDDEF( id_set_pts ) 3705 CALL handle_netcdf_error( 'netcdf', 444 ) 3592 3706 message_string = 'NetCDF file for particle time series ' // & 3593 3707 'from previous run found.' // &
Note: See TracChangeset
for help on using the changeset viewer.