by Tom Baltzer and Brian Kelly

As part of the Linked Environments for Atmospheric Discovery (LEAD) project, Unidata has been building out a powerful storage and compute “testbed.” Figure 1 shows the plan for the final testbed configuration. The large storage cluster will ultimately consist of a 40 TB dual RAID assuring the integrity of the data. This system that will house a many-month archive of data sets from the Internet Data Distribution (IDD) streams (to include: Surface METAR Data, Upper Air Rawinsondes, Level II and III RADAR, GOES Satellite Imagery, NAM Forecast Model and Upper Air ACARS data) as well as products generated by the processing cluster (Data Mining products from ADaM, Assimilations from ADAS and NWP from WRF), other LEAD testbed systems and ultimately the Teragrid. The products housed on the Unidata LEAD testbed will be made available to the Unidata community via OPeNDAP technologies and cataloged using THREDDS tools, giving community members access to a longer term archive than they may be able to house at their own institution or is currently available on the Unidata motherlode system.

The processing cluster being built will ultimately represent a roughly 30 Gigaflop capability that will be part of the LEAD Grid, a set of testbed systems being constructed at 5 of the LEAD institutions. Currently, it is used to generate mesoscale Numerical Weather Prediction (NWP) products using the Weather & Research Forecast (WRF) Model. The first instantiation of capability for the processing cluster is shown in Figure 2, and involves the LEAD concept of Dynamic Adaptability. The domain of the NWP is determined using a Precipitation Locator algorithm that centers both the Unidata run Workstation Eta and the WRF run on the LEAD testbed. The results of both of these models are made available from the LEAD testbed system.

Sitting in front of the storage and processing clusters will be a group of machines to provide service to the LEAD and Unidata communities. There will be an LDM ingest system for acquiring the IDD streams, a machine to provide access to the data storage cluster through THREDDS and OPeNDAP as well as to provide GRID enabled access to other LEAD testbed systems and users (e.g. gridFTP and GRAM) and a test system to allow for experimentation with new ideas and technologies that will be rolled out to the communities once they have been integrated and hardened.

We envision that these capabilities will be a resource that can be leveraged by the Unidata and LEAD communities for their use in research and education both in the short term and ultimately as part of the overall LEAD vision.

by Gabe Langbauer

Using data distributed through Unidata's Internet Data Distribution Ohio State's Atmospheric Science and Climatic Studies Program provides a wealth of information to Ohio's general public.

At any given time, on any given day you would find an amazing variety of browsers on line with you if you logged into The Ohio State University Weather Server. Site administrators estimate that approximately two new users access the page nearly every minute and remain connected to it for around ten minutes each. Although it's probable that most site users are ordinary every-day type folk, there's an eclectic mix apart from that group that includes: fire and weather prediction professionals, power companies, water conservationists, recreational sailers, and professionals investigating mine fires and sealings. Forecasters, pilots, and navy personnel seem to have the highest hit ratios.

Ohio State uses Unidata's IDD CONDUIT feed to initialize its very own MM5 runs.At this point they are operational models, run in real-time and used by whomever wants the information. NAM model output, since it covers North America, is initialized to run the weather forecast for the State of Ohio, Lake Erie, and their vicinities The Byrd Polar Research Center (BPRC) runs the Polar MM5 over the Arctic, Antarctic, and Greenland domains.

Polar MM5 is a modified version of PSU/NCAR MM5 for use in polar regions based on the experience of mesoscale modeling in polar regions by the Polar Meteorology Group of the BPRC.

The BPRC will play a key role during the The International Polar Year (see sidebar).

There is the hope of increasing the Greenland model run to do some mass balance research in real time. Antarctic Internet Data Distribution (AIDD) is still in its infant stage. We plan to add many more sites to this group and to add a lot more data as bandwidth allows.

SOARS Student Honored Shanna-Shaye Forbes received the only computer science award at the SACNAS (Society for Advancment of Chicanos and Native Americans in Science) conference in Denver earlier this month. The award was based on the work that Shanna completed during her summer here in the program center working under the guidance of her science mentor, Unidata staff member Anne Wilson. Of 500 posters presented, only 39 received awards. We're happy for Shanna, and we know that you join us in extending congratulations to her.

Hurricane Events The UPC is interested in work you've done using Unidata tools, software, and/or data about hurricane events of this 2005 season. We're particularly interested in public outreach materials that demonstrate the value and use of realtime meteorological data. An example can be found in this news posting to the Unidata site. Contact Jo Hansen for answers to questions or to submit materials. Thanks. We look forward to hearing from you.

Equipment Awards Award recipients from the 2004 Award round have submitted write-ups as specified in the RFP. You'll find the reports interesting reading. The University of Oklahoma purchased equipment to set up an additional OPeNDAP/ADDE server. Valparaiso University's purchase targeted equipment to address student needs and aid faculty and student research, data archiving, and teaching. San Jose State University replaced 12 outdated personal computers in our synoptic lab with modern workstations which allowed them to deploy and test Unidata's Integrated Data Viewer (IDV). Central Michigan University purchased equipment to upgrade a computer lab. Watch this spot for the 2006 RFP.

THREDDS Data Server Announcement of the new new THREDDS/IDD Data Server (TDS) occurred October 24. You can access the data catalog as HTML or XML. We'll post a detailed article to the Unidata portal about the TDS in the coming days.