Hi Isabelle,
if you look into the monthly files for that catalog
(http://www.ncdc.noaa.gov/thredds/catalog/gpcp/monthly/catalog.html)
there are to types of files gpcp_v2.2_esg.YYYY.gz and
gpcp_v2.2_psg.YYYY.gz.
Both types of files range from 1979 to 2010 and contain a variable with
the same name (precipitation) so I am afraid that the aggregation
aggregates both in one single series that goes from 1979 to 2010 and
then again from 1979 to 2010 and when you request data from any date up
to the last available date (2010-12-01) the Netcdf Subset Service finds
the first occurrence of 2010-12-01 in the series but the next date
available is 1979-01-01 that is before than the last date requested so
it keeps writing data into the response until the second occurrence and
you end up getting that file that contains your requested times + 384.
One trick, but you'll miss the 2010-12-01 date, is request data from
1991-01-01 to 2010-11-01.
Since we don't have control on that THREDDS server I'm ccing Steve
Ansari who is the contact that appears in the WMS capabilities document
and perhaps he is more helpful.
Cheers,
Marcos.
On 04/17/2012 03:31 AM, isabelle.charon wrote:
Dear all,
Thanks for the very useful tool NetcdfSubset.
http://www.ncdc.noaa.gov/thredds/ncss/grid/gpcp_agg/dataset.html
I try to use it to extract monthly mean GPCP precipitation, for years
from 1991 to 2010, in netcdf format.
I obtain a file which contains 624 time steps. I've expected 240: 12
months * 20 years. I think I obtain first the 240 time steps I've
required in the file, then the entire archive from 1979 to 2010 (384
time steps more).
Can you explain me this think? Is it possible to extract only the 240
time steps?
Thank you very much.
Isabelle Charon (Meteo-France, Frebch meteorology service)
_______________________________________________
thredds mailing list
thredds@xxxxxxxxxxxxxxxx
For list information or to unsubscribe, visit:
http://www.unidata.ucar.edu/mailing_lists/
