Policy Committee Meeting Summary:

15-16 July 2007

Unidata Program Center, Boulder, Colorado

Participants

Members	UPC Staff
Steven Businger--Chair	Mohan Ramamurthy
Rich Clark	Tina Campbell, present for Strategic Plan discussion
Rudy Husar	Ben Domenico
Jim Koermer	Jo Hansen
Paul Ruscher	Linda Miller
Gene Takle	Terry Mitchell
David Tarboton	Russ Rew
Melanie Wetzel, present via phone hookup during day one afternoon	Tom Yoksas
Agency Representatives	Note: In addition to the Unidata Management Team, many Unidata staff members were present for the proposal development discussions on Monday afternoon and Tuesday morning.
Bernard Grant, (NSF/ATM) (present via access grid, day one)
Clifford Jacobs (NSF/ATM) (present via access grid, day one)	NCAR and UOP Representatives
Leroy Spayd (NOAA/NWS)	Jack Fellows, Director, UOP
	Tim Spangler, Director, COMET Program
Committee Representative	Roger Wakimoto, Director NCAR/EOL
Gary Lackmann (Users Committee Chair)	Steve Worley, NCAR/CISL/SCD
	Not in Attendance
	Jeff de La Beaujardiere, NASA

Monday, 16 July 2007

Administrative Matters (Businger)

• The Draft Agenda was approved
• The March 2007 meeting notes were approved.
• Next Meeting Dates: 15-16 October 2007 in Arlington, Virginia, subject to change in date and location.

See: Unidata Acronym/Glossary List
See: Meeting Index

Director's Report (Ramamurthy)

The Program Center's overall position has not changed much since the March meeting, i.e., the only areas of some concern are staffing level and long-term finances; overall, however there's a lot going on with many accomplishments and successful activities.

Mohan noted that the month is the 23rd anniversary of the submittal of the proposal to fund Unidata... nearly to the day. He also noted the departure of Sandra Petrie whose position has been ably filled by Tina Campbell.

Other highlights:

• IDD growth continues. Average data volumes are steady as well, with the exceptions of TIGGE and CONDUIT. TIGGE volumes flow in bursts up to 23 GB/hour. Unidata is in discussions with COSMIC to explore distribution of GPS data in real-time, on the LDM. [Subsequently a community survey has been conducted.]
• Discussions have been undertaken to see if NWS will implement an LDM-based terrestrial backup for NOAAPort. A draft letter from the Users Committee and Unidata was presented during the discussions. A CRAFT/CONDUIT meeting with users and other stakeholders to consider distributing other data in this manner will be held at AMS in New Orleans.
• Software:
  • IDV use continues to grow
  • GEMPAK continues to be the most used/downloaded application. NWS transition to AWIPS-II will impact GEMPAK. See: Spayd
  • McIDAS use is decreasing, and new users are increasingly from outside of the US
  • NetCDF-4 was released in beta in April. It supports full backward compatibility, while allowing users to benefit from some HDF5 features and performance.
• Training Workshops: are scheduled for 20 July to 10 August. There are 75 registrants with ten of them haling from outside the U.S. An Advanced IDV session is being offered, which is consistent with a wish expressed by this committee.
• A regional workshop to provide IDV training will be held at the University of Oklahoma in September, under its sponsorship. LEAD and THREDDS sessions will be included in it.
• Equipment Awards: A total of $100K was available for the awards against a total requested by 13 sites of $276K. A total of six awards were made. The Users Committee has recommended that Unidata seek additional funding in view of the expanded collaborations.
• Seminars: There've been two since the last committee meeting. Particularly well received was the netCDF best practices seminar.
• LEAD PIs will visit NSF August 30 to explore future funding options. They will meet with principals in ATM, OCI, and CISE (Computer and Information Science and Engineering). At the same time, collaborations within NCAR are also being explored.
• Support loads continue to grow, but not at an alarming rate.

FY06 year-end found Unidata close to projected marks with a surplus for FY07. If there's an increase in core funds for FY08 we should be in fairly good shape, although it is worthy of note that with LEAD funding ending, budget and staffing figures will be significantly impacted beginning in October 2008.

Discussion:

• The question, "does the budget report indicate Unidata is experiencing deficit spending?" elicited a qualified "yes" response; but the committee's feeling was that this does not require action on its part
• The UPC's spending of reserves may be considered an investment into human capital

Agency Reports

NSF (Jacobs)

Both the House and Senate appropriations to NSF recommend an increase: 10+% for the House and almost 11% for the Senate. All appropriations add up to $30B over the President's request. The White House has threatened a veto, but science funding bills support the President's goals. Should he veto the recommendations, however, there's the possibility of a veto override.

NSF-GEO will have a new Assistant Director effective October 1, 2007. Dr. Mark Abbot is currently a member of the NSB a position he will resign in August. Read more. [Update: Mark and NSF mutually agreed that he will not assume the position.]

NSF's Cyber-enabled Discovery and Innovation initiative (CDI) will begin in 2008. It's designed "to explore radically new concepts, approaches and tools at the intersection of computational and physical or biological worlds."

Discussion

• Committee members wondered how GEO would participate in the CDI initiative, and it appears that a "watch and wait" approach is indicated.
• Committee members also wondered if the CDI initiative is more concrete and domain specific than not, and again the answer is that is somewhat unclear.
• The UPC should respond strongly to the CDI initiative.
• Any LEAD proposal to CDI has the potential to address strongly all of the five themes

NWS (Spayd)

NCEP continues to move toward providing higher temporal and spatial resolution models while trying to balance the needs of operations and the rest of the community. Lessons learned during NCEP's participation in WRF model development include the observations that: Community directions and rate of evolution are not always in sync with operational needs while operational needs are not high priorities for community developers. Foreign partners are strong drivers in community model development At this time NCEP is moving to ESMF (Earth System Science Modeling Framework) for several important reasons one of which is the complexity of WRF code and the option to choose only the parts of ESMF that are wanted/needed. Moving into the future NCEP has to consider the most potential benefits across a wide spectrum of forecast services, ensemble composition, and product delivery. In the latter consideration time is critical because the product being delivered is perishable

CONDUIT's future, what it can and should provide needs to be the topic of a dialogue between the community and NOAA. [A meeting of CONDUIT users is planned for the 2008 AMS Annual meeting.]

MADIS: the Meteorological Assimilation Data Ingest System is a flexible, expandable, and interoperable architecture that is flexible, expandable, and interoperable. Transitioning MADIS to NWS operations will provide 24x7 maintenance support with offsite system backup.

AWIPS: AWIPS II development will impact Unidata. A transition will take place from mid-2009 to mid-2010. Not much will change during the transition, but as development of AWIPS II takes place other changes will be noticeable.

The presentation ended with a resounding endorsement of the Unidata program.

Resolution: The committee encourages the UPC to work closely with the National Weather Service to develop open-source standards for its new AWIPS release.

NASA (de La Beaujardiere)

Jeff was unable to attend the meeting, but provided the following input on the Strategic Plan:

Section 2, mashups: Most mashups work because data are combined visually (e.g., multiple overlaid layers). The OGC Web Map Service (WMS) specification provides an industry-standard way of doing that, and is supported by many GIS vendors, government projects, Google Earth, etc. Considering using that specification rather than inventing a separate Unidata API.

Section 10, Vis tools: Consider web service API for having IDV generate slices through multi-dimensional data, with IDV running in "batch" or "unattended" mode, rather than being driven by a human user. (Perhaps this is already possible).

Regarding NASA specifically:

- I have established a mailing list to enable NASA stakeholders to communicate successes and problems to me as input to Unidata, and for me to report Unidata developments to those stakeholders.

- Unidata and NASA should consider collaborating on a project involving 2D and 3D forecast data from the our GEOS model. I would be glad to discuss possible ideas should there be interest. This could involve IDV as mentioned above. I note, however, that I have not yet been able to identify a funding source for such a collaboration (but am working on it).

Users Committee Report (Lackmann)

The committee met in mid-May. In response to a request made during the 2006 summer Users Workshop, the committee has included a student representative on its membership roster. Sean Arms, University of Oklahoma, was appointed following a community-wide search process. The BAMS article resulting from the workshop has been published. BAMS has proposed a series of articles based on workshop presentations that include a point-counterpoint discussion of teach with models, specifically, how much depth of understanding is required of a model's structure before using it as a tool.

Other:

• The next Equipment Award focus should be THREDDS.
• A change of venue for the DeSouza award presentation will take place beginning next year. The award will take place in the context of a committee meeting except for Users Workshop years, in which case it will be presented during the workshop.
• Collaboration with the NCAR library was discussed in this context. A discussion of that topic can be found at the end of this document.

Strategic Plan (Ramamurthy)

Mohan noted at the outset that comments from Sarah Ruth, Bernard, and Cliff have been integrated into the plan together with staff and committee member input. NSF input resulted in the following change to the mission statement from: "To provide the data services, tools, and cyberinfrastructure leadership that advances Earth system science, education, and outreach" to the following: "To provide the data services, tools, and cyberinfrastructure leadership that advance Earth system science, enhance educational opportunities, and broaden participation."

Discussion was framed around the following points:

1. How does the plan look today, and
2. How will it inform the 5-year funding proposal

• Reviewers: Several suggestions focused on entraining reviewers, external reviewers, of the plan before it is finalized. Suggestions included:
  • Representatives of other disciplines and others not presently connected with Unidata
  • Another suggested set of reviewers was the panel members of the 2008 proposal.
• Layout: there was general agreement that the layout is considerably improved since the March meeting, but the possibility of revising fonts was mentioned.
  • One suggestion was to add captions to the graphics used in the document
  • There was a suggestion that text boxes that printed out with grid patterns should be revised, because this decreases text legibility.
• Target audience and collaboration:
  • Some wondered if it might be possible to reach grades lower than 6 to attain the "teachable moment" at that level and how to include that idea in the plan The response (which occurred later in the meeting) was that Unidata tries to "reach" this target group by providing University members with capabilities and tools that allows them to reach the pre-college group.
  • Others wondered if there are any specifics in the plan that target the Geodetic community.
  • Collaboration with the NCAR library came up during this discussion, and a resolution in that regard was made toward the end of the meeting, but because this discussion dominated e-mail correspondence following the meeting, it will be discussed in greater detail at the end of this document
  • Acknowledgement of collaboration with other divisions within UCAR/NCAR should be made.
• Communication and support: are there innovative methods that should be included here as "stretch" goals?
• Governance: how is Unidata management and governance going to evolve to meet the challenges of the next eight years?
  • The governance section should include some words to indicate "thoughtful support of the sponsoring agency's role." Suggested sentence to include in the governance section: "Unidata will be responsive to the evolving needs of NSF through the cooperative agreement."
  • Recognition of the agency's persistence in supporting the UPC was deemed to be important.
  • What are the goals to move forward? This is the hallmark of the program, therefore, there should be mention of future goals.
• Tools: some felt that this section should include a "best practices" statement.
• There were several comments on the lack of "how" statements in the plan, while others believed that the proposal document will illuminate the "how" plans
  • In this regard there was the thought that the plan is "organic" at this time, and such statements could be added later.
  • It was felt that the goals should be high (broad) enough at this point to define the proposal's focus
  • Some members felt that there should be more examples to back up statements of fact

Action Item: The committee requests that an announcement of the Strategic Plan's completion appear in a featured news item on the UPC home page or in the next CommunitE-letter. Policy Committee member final input should be received prior to that time.

Proposal Development (all)

Timeline: the proposal's due date has been pushed back from the current fiscal year-end September 30 to a date in early 2008, roughly defined as February-March.

A Strawman Outline was distributed in the meeting notebook. Committee members were asked to consider how to advance the goals stated in Strategic Plan. Some additional discussion emphasized:

• Unidata is a community facility with a 23-year history.
• What will the community be able to do in five years
• Will proposal scientific and technical merit persist through the next revision of NSF's Grant Proposal Guide
• What are Unidata's broader impacts... reference the metrics and appraisal document

The six Strategic Plan-defined areas defined the breakout groups: Community, Data Service, Tools, Communication and Support , Cyberinfrastructure, and Diversity.

Two committee members--one to facilitate and one to lead--one scribe (staff member), and at least one additional staff member made up breakout group membership. Three groups met the first day and the remaining three of them met the following day. There are links below to the notes for each group. Very brief summaries follow each one.

Community.

Three aspects of community were examined by the group: what it is today, how it will change, and how to broaden it. They identified four features of each of those aspects, democratize/ facilitate, empower and enable.

Data Services

Some ideas that emerged from this group's discussion: LDM-lite is an idea that would open LDM access to the Windows world. The challenge of addressing users's inertia when starting to work with the new systems needs to be addressed.

Tools

Data exchange through Unidata tools can facilitate community outreach. Integration is an important goal: integrating packages, data sources, and data types. Need to maintain compatibility to ensure future collaboration and interaction. In the interdisciplinary arena tools play a role in development.

Communication and Support

Support needs to be maintained at its existing high quality; but it must be recognized that support can detract from development. Since that is true the UPC needs to develop the means for maintaining the quality while giving developers more time to develop. Communicate with different users in different ways to reach them. Short training formats are a possibility. Need for different forms of documentation. Conducting walk-throughs of the web site during training workshops is a way to increase communication.

Cyberinfrastructure Leadership

An important Unidata CI-leadership role is making sure that Unidata tools are used in the CI arena. Partnering is important. Unidata has persistently provided leadership for standards and interoperability and must continue in this role. LEAD has demonstrated Unidata's strengths and capabilities. Similar opportunities in other fields that include hydrology, oceanography, Earth surface modeling, climate need to be explored. Much of the required expertise is unique to Unidata.

Diversity

The group developed some concrete suggestions for enhancing diversity in the community.

• Short-term visitor program-bring into program for short periods to work with staff.
• Dedicated equipment award pool for minority applicants.
• More regional workshops.
• Early in the proposal period workshop of minority groups to plot a plan of action in diversifying the community.
• Dedicated slots on governance committees

Wrap-up:

The following resolution was offered for consideration:

Resolution: Unidata will explore synergies with the NCAR library to develop dynamic virtual workspaces focusing on datasets, tools and services to enhance knowledge extraction and discovery by a network-enabled community (for example, expanding THREDDS to include shared community workspaces). [A modified version of the resolution has been circulated to the committee.]

No vote was taken on the resolution during the meeting; although it is not final at this time, a proposed revision: The Policy Committee resolves that Unidata explore synergies with the NCAR library where appropriate is under consideration by Policy Committee members at this time.

Two additional action items resulted from wrap-up discussions:

Action Item: The committee requests that there be a presentation on COSMIC data for the next meeting.

Action Item: Chair, Steven Businger, will work with Unidata staff Miller and Ramamurthy to (from the Users Committee Action Item 14): "develop a short survey to explore the Unidata community's interest in receiving COSMIC data in real-time using the LDM. The Users Committee will have an opportunity to review the survey prior to its distribution to the community." [NOTE: The survey was distributed to "dot edu" members of the Unidata community.]

Direct Comments and questions to jhansen@unidata.ucar.edu