Policy Committee Meeting Summary:

5-6 February 2004

Boulder, Colorado

Participants

Administrative Matters

The draft agenda was approved.

Melanie Wetzel was named to the position of Chair Designate effective the February meeting. Melanie will chair the May meeting. The creation of a voting technical position on the committee was announced. This position has a three-year term as do other committee member appointments. Phil Bogden, who has accepted the committee appointment to the position, was unable to attend the meeting.

New committee member, David Maidment (UT-Austin) was introduced. Dr. Maidment is a hydrologist with established working relationships with the GIS community. His membership on the committee representative of Unidata's expansion into new communities. Jim Steenburgh (University of Utah), Users Committee Chair, who was unable to attend the October meeting was introduced. Jim is a long-time Unidata user, a longstanding member of the Users Committee, and a several-time participant as a presenter at Unidata's Users workshops.We welcome both David and Jim.

October Summary Notes approved after correction of point 2, Additional Items, in the Users Committee report on page 3. Specifically, "SSEC at the University of Wisconsin is actively exploring," instead of "will be using." Also correction of a typographical error on page one was requested.

Next Meeting Dates:

Spring Meeting, 24-25 May 2004, location TBD (probably Boulder)
Fall Meeting, 23-24 September 2004, location TBD (probably Arlington)

Discussion and review of October meeting action items yielded the following:

Items two and three were completed. Item 1 (an item resulting from the 9-10 October 2003 Users Committee meeting stated: The Policy Committee will compose a letter to the Deans of the IDD Relay sites with copies to IT and site representatives. The letter should explain the relay's importance to the Earth sciences community, and, the reason for the size and the importance of the data being moved. The letter is in draft form; completing it and determining to whom it should be addressed are the next steps.

Thanks were again expressed to SCD for retrospective data access, specifically to Steve Worley representing SCD at this meeting.

Status Reports

Director's Report (Ramamurthy)

Unidata's public face has been enhanced through the construction of a greatly-needed new booth which was unveiled at the AMS annual meeting. In addition to the booth Unidata's presence at the meeting was greater than usual in the number of presentations (15) and in the Users Committee award of the fourth Russell L. DeSouza award to Dan Vietor for developing the WXP software and for his active participation in the Unidata community's e-mail activity.

In addition to the above Unidata participated in the MeteoForum meeting which was organized by COMET, included 18 participants from WMO, WMO RMTCs, COMET, and Unidata. There was also a Unidata presence during the CRAFT-CONDUIT meeting (later agenda item).

Staffing issues include two software engineers for the LEAD project, a replacement software engineer/web developer, and a replacement for the front office administrative assistant position.

Finally, the space issue which has long been on the table, has changed in focus, because the FL4 remodel has been put on hold; but Unidata will achieve long hoped for collocation of all staff by expanding into the south wing and occupying the entire east wing, first floor.

As a result of LDM-6's greatly improved performance, more sites are receiving CONDUIT and Level II data and the weather service will use LDM-6 to distribute data operationally.

The e-mail support load remains high, but appears to vary seasonally.

The health of the program at this time is reasonably good with some areas of concern with software development (impacted by the increased support load) and budget concerns.

Budget Report (Graham)

Total new funding received for FY03 was $4.32M against expenditures of $4.36M for the same period. NSF core funds made up the largest percentage of the funds received. A portion of the total was earmarked for the equipment awards process. Funding for the netCDF/HDF merger, THREDDS, Case Studies (carryover), and DODS/OpenDAP made up the difference.

The first year (FY04) budget request for our NSF-core proposal (Unidata 2008) was 4.096M. The NSF target for FY04 funding of that proposal is 3.4M. We were able to offset that 696K deficit by bringing in new awards that in part will leverage funding for some of the proposed activities. We will still need to make cuts of about 293K in our core in order to balance those activities. In FY04 we will not fill a new position earmarked to work on on-line training and web-functionality with regard to community support activities, as well as a visitor to help us expand into new disciplines. We will, in addition, cut funding to some of our subcontracts as well as daily expenses such as travel and materials.

Projected funding for FY04 from all sources combined is $4.53M.

The outyears of FY05 and 06 will bring new challenges because of a projected flat budget scenario. Funding for both the THREDDS and NASA-AIST awards ends with FY04's end.

In summary FY04 projections are not rosy but with various cost-cutting measures (and some staff attrition) we will be in reasonable condition at the end of the fiscal year. Multiple funding sources alleviate the problems to some extent.

Discussion:

• Cliff suggested that once a year we present progress made toward accomplishing our proposed activities rather than just our financial status. That report could also focus on what we are not doing because of finances. We discussed the need for Unidata management to take a hard look at our finances and what we can/can’t do given the negative budget picture and present that for discussion at an upcoming Policy Committee meeting.
• There was a suggestion that Unidata might consider developing a proposal to refine data streams for a larger community, specifically CUAHSI. In addition, searching the NSF RFPs to identify opportunities for additional funding was suggested.

Committee Report

Users Committee (Steenburgh)

The Users Committee has not met since the last Policy Committee meeting. Its activities are ongoing, however. Noteworthy are the following:

• The committee and the Unidata community remain concerned about the future of the COMET case study library. The Usercomm has asked the UPC to explore options for the continuation of the case-study library. It should also be noted that community members are interested in using AWIPS/FX-NET, with this issue a regular discussion at UserComm meetings.
  No decisions have been made at this time.
• A number of presenters at the summer Expanding Horizons have been invited to submit a collection of articles for BAMS. Committee members Anton Kruger, Mark Laufersweiler, and Michael Morgan and UPC staff member Jo Hansen are coordinating the collection.
• The 2004 Russell DeSouza Award for Outstanding Community Service was awarded to Dan Vietor, Unisys, for his development and support of WXP and service to the Unidata community
• The Usercomm is beginning the process of regular phone conversations with Unidata institutions (6/year) to gauge their satisfaction with Unidata services and solicit other input from their “constituents.”
• A new member orientation was held for new Users committee members prior to the fall meeting and was extremely successful.
• At the request of Unidata, NCAR/SCD is developing an SCD Portal-Delay Model to IDD Data Products. The Usercom believes this will enhance education and research at member institutions, and have asked SCD to expand product availability. The committee passed a resolution commending NCAR/SCD for this effort.

Agency Reports

NSF (Jacobs)

NSF's 2004 budget appropriation is 5% over that of 2003. Research and related, Major Research Equipment and Facilities Construction (MREFC), and EHR are marked for increases in FY04. In 05, however, EHR's proposed budget will see a decrease because the Math and Science Partnerships program is targeted for elimination at NSF, and it will be transferred to the U.S. Department of Education.

Although the agency has seen a 60% growth increase since 1998, 2005 does not look promising.

FY05's budget's request focuses on:

1. Organizational excellence: strengthen NSF management to maximize effectiveness and performance
2. Improve productivity of researchers and expanding opportunities for students (people)
3. Strengthening the nation's performance with world class instruments and facilities (tools)

The first refers to efficient managing of NSF and is 6% of the proposed budget. The second, people, is 19% of the proposed budget, and the third, tools, is 26%.

In the people category, more people are getting more money. The average fellowship has increased from 5K to 5.5K Stipends maintained at $30,000 annually. In the stipends are three categories; Graduate Research Fellowships (GRF), Graduate Teaching Fellowships in K-12 Education (GK-12), Integrative Graduate Education and Research Trainee ships (IGERT).

Ideas, the category occupying the largest piece of the budget pie, 49%, NSF's international program will receive a 6M increase. The growth in average award size (also part of the Ideas category) has been significant. Since 1998 it is up 71%, but the increase from 04 to 05 is modest. Funding for science and technology centers, science of learning centers, and others is also marked for a modest increase.

Investment in tools increase is 104M. Included in this category are cyberinfrastructure and Major Research Equipment and Facilities Construction (MREFC). A new MREFC project is the Scientific Ocean Drilling Vessel (SODV) whose funding request is at 41M. FY04 marks the end of the ITR initiatives. This frees up some funds for other large projects.

NOAA/NWS (Spayd)

The National Digital Forecast Database effort, its accomplishments since September 2003, and some concerns about it were the focus of the report.

Public comment for the following elements ended December 2003.

Maximum Temperature
Minimum Temperature
Probability of Precipitation (12 hr forecast)
Weather
Wind speed and direction
Significant wave height

A decision on implementing these elements will be made at the end of March. At the same time public comment is being invited on the temperature, dewpoint, and quantitative precipitation (QFP), and snow amounts.

Satisfaction with NDFD graphics is at the 95% level.

NWS is concerned that it has received little feedback from its customers on the use of gridded data. This input is needed to make requirements to develop criteria to make the grids operational.

As of January 2004, most feedback indicated:

Data formats in addition to GRIB 2: XML and netCDF (85% of respondents made this request)
Addition of NOMADS to dissemination system
Additional protocols including non-ftp access and OPeNDAP

Respondents were unanimous in their request for access to verification statistics to be available on the public web site.

Some verification issues persist.

Level II Data Distribution (Miller, Ramamurthy)

Unidata's involvement in the Level II data access and distribution effort spans over six years and was endorsed by the Policy Committee in May 1998. The topic has been a frequent Policy Committee agenda item since then. The CRAFT project furthered the effort by creating a successful collaboration with diverse stakeholders. Unidata's LDM software was chosen to move data from radars to NCDC and NCEP and for data distribution to interested stakeholders. At this time 102 radars broadcast data.

CRAFT became an appropriate project for technology transfer to NWS operations, and plans for NWS distribution of the data are nearing completion. Four sites will serve as top-tier nodes in the data distribution system. While NWS preselected two sites the Mid-Atlantic Crossroads (Max) GigaPop[collocated with the University of Maryland] and the ERC GigaPop [collocated with NCDC Asheville]) the other two sites are being selected through a competitive process.

Proposed NCAR Reorganization, UCAR, and UOP (Fellows)  Rick Anthes was present for this item.

NCAR's proposed reorganization was presented to the broader UCAR/NCAR community a month ago. Town meetings to discuss the proposal were held, but questions regarding UCAR generally and UOP's place specifically in the new structure have arisen.  The NCAR directors and the President's Council created the final plan, after receiving input from NCAR leadership and staff over the year.

As a non-profit consortium UCAR was formed to focus on scientific problems beyond the scope of individual member institutions. UCAR's membership (68 universities with doctoral programs in the field, and 20 non-doctorate programs) provides governance of the consortium. The organization receives guidance from and is reviewed by its community. Its mission of extending the capabilities of the university community, knowledge and technology transfers, has been successfully met.

A fresh look at the present NCAR organizational structure was suggested by the NSF Review panel of NCAR and UCAR in 2002. NCAR management took this suggestion seriously. The objectives of the proposed reorganization are to reorganize to address significant current national needs and new ones that may arise. A need to maintain disciplinary alignments is another objective.

Criteria for the decision to realign included, among others,

• Consistency with the Strategic Plan
• Community service and strategic partnerships
• Efficient management structure with fewer direct reports
• Enabling/identifying new scientific frontiers

The proposed structure would create five science laboratories:

1. Earth Observing
2. Advanced Studies Education and Training
3. Sun, Weather, and Climate
4. Computational and Information Systems
5. Science Applications

And, it would create three institutes: 1) ESS: the Earth System Studies Institute; 2) IMAGe: the Institute for Mathematics Applied to the Geosciences, and 3) ISSE: the Institute for the Study of Society and Environment.

The reorganization proposal has been developed by the NCAR Directors and Presidents Councils and will be presented to the UCAR Board of Trustees this month.

The overarching goal of the reorganization is collaborative integrated research.

A refinement of the proposed realignment would have positioned the UCAR Office of Programs entities under one of the laboratories. [NOTE: at this time the merger of UOP and NCAR has been deferred until much later. And, in fact, if may never happen.]

Discussion and comments:

Although broad, the discussion centered around two questions: 1) What is the realignment all about, and 2) How does it impact Unidata.

The reorganization will result in an emphasis on large programs: unique, large programs. Community interactions will be enhanced by an increase in visitors, both to and from NCAR, and through frequent evaluation.

The UCAR community's reaction to the reorganization has been positive.

About the Laboratories

The new management layer created by the laboratories actually consists of FTEs that already exist within the NCAR organization. The laboratories reflect a "thematic approach" which is indicated by their names. The heads of the five laboratories would be at the senior management level and be compensated accordingly.

New "themed" laboratories may be added, but

• Alignment of Unidata with IMAGe: the Institute for Mathematics Applied to the Geosciences might enhance and foster collaborative activity in math and statistics within the community.
• More remote sensing studies would occur within the Earth Observing Lab (EOL)
• In a merger scenario the UPC's home would likely be the Computational and Information Systems Laboratory.
• The reorganization will require the development of metrics to effectively gauge its success.
• COSMIC's alignment with the EOL, scheduled for implementation in 2006 makes sense, but COSMIC may value its autonomy. 
• There is some thought that COMET would fit in the Advanced Studies, Education and Training Laboratory or the Science Applications Lab.

The Institutes

The three proposed Institutes will draw staff members from all of the science groups. This will increase opportunities for collaboration, both internal to NCAR and external as well. There might be ways that staff from UOP could also participate in the institutes.

There was a suggestion that the UPC could itself be an institute.

The Unidata Community

Unidata is community driven. The economics and logistics of a merger potentially could affect this. The assimilation of UOP into the NCAR culture bringing with it increased management layers could destroy access to the Program Center by the participating entities that comprise the Unidata community; thus community input to any merger is viewed as critical. Community governance is pivotal to Unidata. Through it our community participates directly in the governing process. 

Easier access to the NCAR Director's Opportunity Fund would be a plus for the community.

In addition:

• The Policy Committee sees itself as protecting resources for the 2008 proposal
• Unidata's community is not commensurate with UCAR's or NCAR's. It is noteworthy that Unidata includes many smaller undergraduate institutions. 
• In any merger scenario the inevitable block funding process would have a natural tendency to block the community interactivity process.

The Process

• At this time June first is the target for beginning implementation of the NCAR realignment.
• NSF supports the management of UCAR/NCAR in achieving its strategic goals, while recognizing that the goals of each are different and is neutral in its judgment of the proposed structure. The metric quantifying the reorganization's success will influence the agency's position on the reorganization.
• It may be possible and desirable to hold workshops to educate the NCAR community on what these changes auger, how they will benefit the community, and how intereactions between NCAR and its community will be structured.
• Concerns: how will this help collaboration? Implementation groups need to consider this.

Other

Other discussion included the following points/questions

• Concern about the loss of "brands" (HAO and RAP were specifically named), but those two at least would not change.
• DLESE and NSDL are being hosted by UCAR, but may not want to be part of the new structure
• Further development of mathematical science applications within UCAR and university programs will lead to more advanced use of observational data.
• Are increased two-way visitor programs a possible outcome of the reorganization
• Could "service to society" be another theme.
• USWRP: is this an example of a "theme" project?
• Successful models have in common that they have advisory panels.
• In general, the alignment of DLESE, GLOBE (NASA), NSDL, and other groups with funding from outside of ATM could create some difficulties. These other entities may not wish to be so closely aligned with NSF/ATM.

Summary:

Here is a summary of Unidata's concerns about a possible merger of NCAR and UOP:

1. Being swallowed
2. Our community is different from NCAR 's
3. Community accessibility issue: Unidata's smaller size encourages interactions with its community
4. Unidata is service oriented, both from the program center to member institutions
5. The synergy issue: what or how would the proposed reorganization change it?
6. The UPC community could lose its voice.

Finally in wrap up discussion two points were made:

In principle, BOT approval gives the green light to the reorganization, but many implementation issues remain.

During town meetings held by NCAR, there was little response of any kind from UOP staff. The probable reason is that the content of the town meetings focused on NCAR; and UOP staff, while interested, probably did not have problems with the NCAR reorganization. When it became clear that UOP was included in a reorganization, UPC staff's reactions were not positive. Jack clarified: these are two separate processes, and staff opinions on both sides matter.

There are alternatives to simply merging UOP and NCAR. The fact that UOP entities are smaller and more autonomous and can be reviewed strictly on the merits of each small group is critical and needs to be recognized.

However: the major concern is the impact on the Unidata community culture.

GIS Activity at the UPC (Domenico)

Unidata's original versions of WXP, GEMPAK, and McIDAS used "stovepipe" analysis and display applications. Data access occurred by datasets being delivered via the IDD to users and then decoded into local files. The data were in local files on users' desktops.

Data delivery occurred pretty much the same way in the GIS world, but without IDD delivery. Each of these approaches served their users well.

In the thin client approach the stovepipe was connected. Thin clients allowed users to access actual images while creating them was done on the server side. The Community Data Portal at NCAR is an example of this. While powerful, this method represents a loss of control in data visualization and analysis.

Thick clients on the other hand, allow users to move and to manipulate portions of data streams to local systems and to manipulate them.

THREDDS incorporates a client/server approach that enables integrated analysis and display of datasets from distributed servers via user-selected thick client. Among other advantages are that data residing on remote servers can be analyzed with desktop software. In this approach: the user has to know where the data are, and many systems have to work together.

GIS users' data is a collection of static features with geographic features where the features can be stored and manipulated in a database. Atmospheric and oceanic communities on the other hand, use data that are a set of parameters with the behavior of the parameters being governed by equations, i.e., data are approximations to mathematical functions.

The standards approach to data has both positive and negative implications. Among the former, it would provide a single interface for clients while usage and discovery metadata interfaces are part of the specification. Among the latter is the fact that many steps remain incomplete while, at the same time, legacy systems are installed, are working, and have momentum.

At this time it seems clear that no single solution exists, but participants are benefiting from multiple approaches. The payoff point is important to both communities and to evolving collaborations. Open GIS/ISO standards seems the best hope at this time.

Discussion

• THREDDS development currently is working to achieve integration with GIS. Information about data slow in coming to digital libraries. NCAR is adopting the THREDDS approach to metadata. The technology transfer in this case is from Unidata to NCAR.

Friday 6 February

Equipment Awards: Analysis and Adjustments (Graham)

The equipment awards RFP was posted on the Unidata web site 7 January, and proposals will be accepted until end of business on the 27th of February.

Poor proposal quality is a concern. Speculation on the reasons for that included the possibility that the grant size may eliminate some of the larger sites, who, because of the unfavorable effort to payoff ratio might not apply, which would in turn mean that smaller, less experienced sites constitute the larger number of respondents to the RFPs. This year the RFP announcement on the web site linked to two well-written proposals from 2003. The Users Committee may need to determine why poorly written proposals are submitted.

The review panel will convene on March 10th and will include User Committee members (who are not submitting proposals), Unidata staff, and two other representatives to be determined (most likely from elsewhere in UOP and the user community). In the unlikely event that an unsuccessful proposer protests the decision of the panel, the UOP director will resolve the disagreement. All proposers receive a written panel summary for their proposal.

AMS CONDUIT/CRAFT meeting update (Miller)

CRAFT and CONDUIT stakeholders met during the annual AMS meeting. Kelvin Droegemeier and Bob Gall led a discussion of the history of CONDUIT and CRAFT. (Both use Unidata's LDM for data distribution.)

The two groups will merge, and the merged group will be loosely organized and meet twice a year, one of which will be at the AMS annual meeting. These will be open meetings.

Jim Steenburgh, representing the Unidata Users Committee, presented the idea of providing the National Digital Forecast Database (NDFD) in the CONDUIT data stream. Discussion revealed that the NDFD includes about 50 MB/hr (about 255 files), but plans are in place to continue expansion of the data stream. The data products are released in one lump (about 26-50 MB/hr) with the process taking about 30-90 seconds. This would increase the queuing delay of products (such as the GFS) already seen by CONDUIT.

Level II Distribution (earlier agenda item) requires 24x7 support. The University of Oklahoma will continue providing the data via CRAFT until the NWS announces the selection of four sites for data distribution.

Additionally, NOAA is investigating the idea of providing a Family of Service solution for distribution of the Level II data. This service would be provided through a cost recovery method for the NWS.

Peak Weather Presentation and Discussion (Russ Peterman, Peak president)

Peak Weather Resources Inc. is the for-profit wholly-owned subsidiary of the University Corporation for Atmospheric Research (UCAR) Foundation. The UCAR Foundation (UCARF) was created in 1986 to commercialize UCAR knowledge and technology where appropriate.

UCARF formed WITI Corporation, a for-profit subsidiary, in 1989 and sold it 11 years later. Peak Weather was formed in June 2003. Its mission is to develop a company to further technology transfer and to create an outlet for developed products to sustain R&D funding. The three entities UCAR, UCARF, and Peak support common goals. While Peak is a for-profit entity it is not to be sold.

Peak received seed funding from UCARF is presently in UCAR space. It fill one to two staff positions in 2004 according to the business plan.

Initial business activities include Weather Support for Deicing Decision Making (WSDDM), Dynamic Integrated foreCast (DICast), and Total Instantaneous Precipitation Sensor (TIPS). In selecting these activities the Peak board exercised a great deal of caution.

Wrap up

The committee will take under consideration reducing the number of meetings per year from three to two. There is some thought that a result of greater reliance on e-mail and access grid interaction the need for three meetings could be reduced to two. The Unidata Users Committee, for example, relies on e-mail to transact some portion of its business while holding only two meetings per year.

Some suggestions:

• Two and a half day meetings
• Meet two times per year with the exception of proposal years
• Form an ad hoc committee to review agendas for the past several years

The point was made that the committee needs to have ample information to make informed decisions.

The committee will hold the two already-scheduled meetings of this year, but will make a decision on reducing the number of annual meetings at the May meeting.

Maidment Presentation (David Maidment)

Following the meeting wrap-up, and end of the scheduled agenda, Policy Committee member, David Maidment made a presentation entitled: From netCDF to GIS. An impromptu invitation to ESIG, JOSS, and RAP staff members was issued, and together with those who were able to attend and Unidata staff members the conference room was filled.

Send comments or suggestions to Jo Hansen Program Center Liaison to the Policy Committee.