20030611: top level IDD relays needed for NEXRAD Level II data (cont.)

>From: "Arthur A. Person" <address@hidden>
>Organization: Penn State
>Keywords: 200306111647.h5BGloLd001053 IDD NEXRAD Level II CRAFT relay

Hi Art,

re: request for IDD top level relays for NWS Level II data

>If my math is right, that would be a peak of 1120 MB/hr to 4480 MB/hr,
>projecting it into the future, which is quite a range.

That is correct.  The last increase factor of 4 to 8, however, should
be a ways down the road.  The sequence of data increase will basically
be:

- the current CRAFT stream has up to ~140 MB/hour

- the full set of NEXRADs could bump the content up to ~420 MB/hour
  at peak

- the halving of the range gate couple with the doubling of the number
  of radials will quadruple the ~420 MB/hour peak rate.  When the NWS
  actually implements this is open to speculation

- the relaying of polarization data will increase the volume an
  unknown amount, but it should lie within a factor of 4 to 8 of
  the last figure

>If one assumes the
>worst case, that's more than twice the current CONDUIT peak.

That is correct.  That should be a ways off in the future, however.

>Also, the
>CONDUIT feed has busy and quiet times during the day, whereas the NEXRAD
>II stream (I assume) would be more continuous all day, leading to a much
>larger data flow per day.

The volume of NEXRAD Level II data will, of course, depend on the weather
being experienced around the country.

>Is there any way you can be a bit more specific
>about the rates and, especially, the per-day amount of data that would
>have to be handled?

Good question! No, I can't really be a lot more specific.  I can't, for
instance, predict what fraction of radars will be operating in
storm/precip modes versus clear air mode.  This will be the largest
factor in overall volume of data.

As to when the NWS actually makes the changes to range gates, radials,
and polarization, your guess is as good as mine.  A number I heard
tossed around at the last AMS meeting was that the volume of data
_could_ increase by a factor of 27 within a 10 year time frame.  So, if
the average volume of data increased to 250 MB/hour by the additions of
the non-CRAFT radars, the full volume could rise to 6.7 GB/hour.  Many
of us don't see this happening for quite awhile.  After all, the NWS
has to move all of that data to its collection point at the University
of Maryland AND to NCDC reliably.

One last comment: it may be the case that no one Unidata university
will be able to handle the full volume of NEXRAD Level II data.  They
could, however, handle some fraction of it.  Our request for
participation in this "adventure" was an attempt to identify
universities that want and are able to participate at some level.

Cheers,

Tom