Changes between Version 27 and Version 28 of DevelopmentActivities/MergeHydro


Ignore:
Timestamp:
2012-06-06T08:00:05+02:00 (12 years ago)
Author:
maignan
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • DevelopmentActivities/MergeHydro

    v27 v28  
    7979Nicolas Vuichard (March 2012) - off-line tests at FLUXNET sites using STOMATE : ongoing 
    8080 
    81 Fabienne Maignan (Feb-May 2012) - global off-line simulation 
     81Fabienne Maignan (Feb-June2012) - global off-line simulation 
    8282 * forcing: ERA-Interim regular grid 0.72° 3-hourly 
    8383 * computer: LSCE/obelix* 
    8484 * 16 procs 
    85  * CWRR + STOMATE 
     85 * CWRR + STOMATE + ROUTING 
    8686 * segmentation fault SIGSEGV here: 
    8787{{{ 
     
    138138... 
    139139}}} 
    140  * With ROUTING, stop in ''routing_findbasins'': 
    141 {{{ 
    142 Why are we here and can not find a trip larger than 96 
    143  Does this mean that the basin does not have any outflow            1 -2147483648 
    144 }}} 
    145  This means that there is an invalid basin number equal to -2147483648 (integer variable ''bname'' copied from the call variable ''basin''/''basin_bx'').[[BR]] 
    146  Before the call to ''routing_findbasins'' there is a call to ''routing_getgrid'', with a real to integer conversion: 
    147 {{{ 
    148 basin_bx(iloc, jloc) = NINT(basins(sub_index(ib, ip, 1), sub_index(ib,ip,2))) 
    149 }}} 
    150  ''basins'' is a real variable read from the file ''routing.nc''. The missing value is 1.e20 (equal to ''sechiba_undef'') and NINT (1.e20) = -2147483648[[BR]] 
    151  This hampers the filtering against an invalid basin number in ''routing_findbasins'': 
    152 {{{ 
    153 IF ( basin(ip,jp) .LT. undef_int) THEN 
    154 }}} 
    155  (undef_int = 999999999) 
    156   - First test: the filter is temporarily replaced with: 
    157 {{{ 
    158 IF  (basin (ip, jp). GT. 0) THEN 
    159 }}} 
    160   - Second test proposed by Jan: change the unvalid value just after the reading of the ''basins'' variable in ''routing_basins'': 
    161 {{{ 
    162     CALL flinget(fid, 'basins', iml, jml, lml, tml, 1, 1, basins) 
    163     WHERE ( basins(:,:) .GT. undef_sechiba-1)   !@FM_120531 
    164         basins(:,:) = undef_int+1               !@FM_120531 
    165     ENDWHERE                                    !@FM_120531 
    166 }}} 
    167  -> In both cases everything seems to be fine...except that:  
    168  * With ROUTING, ''NaN'' values are detected in ''sechiba_restart.nc'', ''stomate_restart.nc'' and ''stomate_history.nc'' files.[[BR]] 
    169  Under investigation... 
    170140 
    171141Frédérique Cheruy - global simulation coupled to LMDz : to be done