email from Xu 25 oct 2010:
I found that some problems exist in the mete file on 2047/07/26 03:00 
(which has also been revealed by Moeko on 
http://wiki.seas.harvard.edu/geos-chem/index.php/File_I/O_with_GAMAP). 
This is maybe caused by the regriding. As a result, I made a new 20470726.a6.4x5. 
The 03:00 values of Q and MOISTQ (which have problems) are recalculated as the 
mean values at 2047/07/25 21:00 and 2047/07/26 09:00. 

I also evaluated the new data file and found it reasonable. 
You may find the new "20470726.a6.4x5" in " ~xyue/fire/geos-chem/run/gcap/a1b" and move 
it to "/as/pub/ftp/pub/gcap/GCAP_4x5/AGRID/2047/07".