Meeting to discuss access & distribution of WSR-88D
(NEXRAD) Level II data
February 4, 1998
DRAFT Summary, Linda Miller
Participants:
Harry Edmon, Univ of WA
Rit Carbone, USWRP
Woody Roberts, FSL
Richard Chinman, IITA
Dave Fulker, Unidata
Linda Miller, Unidata
Dave Fulker called Ken Crawford, a Unidata Policy Committee member, to
invite him or a OU or CAPS representative to attend the meeting, but there
was an important CAPS meeting that conflicted with the date.
Background:
Dave provided an overview of activities leading to the meeting. He
emphasized the Users Committee recommendation that Unidata act on
behalf of the Unidata sites to access Level II data from the radars and
make it available to all institutions that are interested in using the
data.
Fulker and Miller met with several NWS officials to discuss the access
and distribution of the data and were encouraged by the NWS to pursue the
use of the Level II data among university institutions. During that
meeting, it was agreed that an amendment would be made to the current
RIDDS arrangement to allow redistribution of the Level II data for use in
education and research activities.
Direct Connection to Local WSR-88D:
The process for obtaining the direct connection to a local radar is a
letter to the Regional NWS Headquarters requesting the connection with
a short proposal on a collaborative project using the data.
RIDDS is the system that is used to provide access to the local radar
and distribution of the data to a single point. It requires two
Sun workstations placed at the WSFO site. One is referred to as the
"RIDDS" station and is directly connected to the RPG and a Ethernet hub.
The second station is the one that can be modified by the institution and
is connected to the Ethernet hub.
The following university sites have a RIDDS agreement for direct
connection to their local WSR-88D radars:
Univ of WA
Univ of CA, Santa Cruz
Univ of NV, DRI
Naval Post Graduate School
NCSU
MIT
TAMU
Penn State
SUNY-Albany
Radar Interface & Data Distribution System (RIDDS)
Harry defined the characteristics of the RIDDS system in Seattle.
The two Suns are described above. The station running the RIDDS has
about 32 MB of memory. The site is archiving about 3 weeks of
data.
The experience
in Seattle has been about 9.8 MB of data with a volume scan every 6
minutes in interesting weather and every 10 minutes in normal
conditions.
There are 4 products with 17 tilts. Bob House is using Zebra
(NCAR
product) to view the data. Various compression schemes were tested
with worst being 4:1 and best being 50:1. The average is 8:1 and
using GZIP, it is 10:1. The RIDDS is a common interface to the Radar
Products Generator (RPG), created by the Operational Support Facility
(OSF) in Norman, OK.
Woody reported that in 18 months the RIDDS will become obsolete.
Bob Saffle, NWS, will be moving the system to an open RPG system with 4
ports on the Ethernet with Level II data. They will be using the
TCP/IP protocol.
Forecast System Lab (FSL)
Woody Roberts provided an overview of FSL related activities. FSL
is accessing three radars; Cheyenne, Goodland and Denver. The RIDDS
system has cost them about 13K for each connection. FSL is using a
SPARC2 and runs a 5km Cartesian grid. FSL ignores the spectrum
width.
They are creating a mosaic from the 3 radar systems. He indicated
that modelers want the radial winds. The Cartesian grid is 500m using 2km
nested grid. He mentioned that AWIPS would carry a subset of products
from the NEXRAD sites.
Next Steps
It was agreed that a co-sponsored workshop should
be held in the September-October
time frame to discuss product definition and how to
proceed.
The co-sponsors would be the USWRP and Unidata. Rit Carbone pointed
out that there is a plan for 120M for a 5 year
research initiative between 2000-2005
There are several groups and communities who are interested in the
products from the RPG, including hydrology, NCAR's RAP, ATD, MMM etc.
We will begin a proof of concept which will include the
University
of Washington, FSL,
and perhaps others. It was suggested that we contact Richard
Oye, ATD, to gather information about related activities.
Fulker emphasized the use of the vBNS or Next Generation Internet (NGI)
for experimenting. Harry said that Washington has 2 or 3 T3s
and their
connection has been very effective in moving the data using the
LDM/IDD technology thus far.
Institutions or contacts mentioned for a prototype experiment were:
Princeton, Jim Smith
Florida State
UCLA
San Jose
Post Scripts to this report:
Exchange between Woody and Dave proposing:
-
Access for FSL & Washington to FSL high speed network
-
Porting/installation of the Washington code at FSL
-
Washington & FSL ingest routines developed to handle each
other's data
-
Create rectilinear products at the point of ingest that
are free of information loss, but that permit recipients to subscribe to
_subsets_ of each volume scan (Rit's description)
-
Investigate the applicability of Dick Oye's software
Mike Carelli, NWS, called to say that testing can begin, but there is a
need to be very cautious with the use of the data for education and
research activities. The NWS and Tri-Agency group will provide
a formal amendment to the RIDDS agreement soon. It will be passed to
Unidata, in draft form, for review and comment.
Linda Miller
External
Liaison
Telephone: (303) 497-8646
UCAR/Unidata Program
Center
Fax: (303) 497-8690
P.O. Box
3000
E-mail: lmiller@unidata.ucar.edu
Boulder, CO 80307-3000
http://my.unidata.ucar.edu/