20060921: CONDUIT data stream changes

Pete,

I sent a note to the CONDUIT email list regarding the transition of
ensemble products on Sep 27.

The updated data subset patterns will be:
nccf.*prod/(gfs.*pgrb2|ruc2a.*grib2|gefs).*[09]$
nccf.*prod/(gfs.*pgrb2|ruc2a.*grib2|gefs).*[18]$
nccf.*prod/(gfs.*pgrb2|ruc2a.*grib2|gefs).*[27]$
nccf.*prod/(gfs.*pgrb2|ruc2a.*grib2|gefs).*[36]$
nccf.*prod/(gfs.*pgrb2|ruc2a.*grib2|gefs).*[45]$

If you have any schedule conflicts for the 27th, we can work around that
by setting up
a request to the ATM/NSF computer for the nccf.*gefs products ahead of
time and you will get those products when the change is made to that
host until you are
able to make the change to the ncepldm host on your end.

The volume of the data will be somewhat smaller than the current volume
initially due to the use of GRIB2. Over the next few weeks, NCEP will
add several additional
members and parameters. The rough calculation that NCEP provided was
just a little under what the current volume is, so I'm assuming a net
neutral volume/bandwidth change.

Let me know what works for you regarding this change.

Thanks,

Steve




-- 
Steve Chiswell <address@hidden>
Unidata