--Bryan Bailey (gbbailey@edcmail.cr.usgs.gov), EROS Data Center
The Land Processes Distributed Active Archive Center (DAAC) Science Advisory Panel held a regular meeting at the Raytheon EOSDIS Core System (ECS) Offices in Landover, MD, on February 4-5, 1998. This two-day meeting of the Panel was preceded by a one-day meeting devoted to addressing topics and issues of importance to the DAAC's support of and interaction with the ASTER, MODIS, and Landsat-7 Instrument and Science Teams. That meeting consisted of a short plenary session where "across instrument" topics were discussed, followed by concurrent sessions where instrument-specific topics were addressed. The results of "Instrument Team Day" discussions were presented during the meeting of the full Panel.
The first item of business was to review the status of action items from the October 1997 Panel meeting. John Dwyer presented information on map projection preferences indicated by users ordering Landsat data. History shows UTM to be the most popular map projection, with SOM a distant second. Alan Strahler pointed out that many MODIS land products will be in the Intergerized Sinusoidal Projection (ISP), so it would be desirable for the DAAC to provide tools by which ancillary or other related data not in the ISP could be transformed to ISP. The DAAC will investigate the possibility of adding ISP to the General Cartographic Transform Package (GCTP). Bryan Bailey reviewed the strategy and plans for making sample ASTER, MODIS, and Landsat 7 data and products available to the general user community in the near-post launch time frame. Panel members were mixed on the issue of whether or not implementing a prototype sample data and product capability was important to its success.
John Daucsavage provided more detailed information on a previous action concerning the archive and management of EOS data and products not yet ready for release to the general user community. ECS has mechanisms to limit distribution of data to users on a designated access list, and these tools and their implementation schedule were described in the presentation. Tom Kalvelage reported on the status of network testing by describing the testing that already has occurred and presenting a schedule for planned testing. Panel members expressed lingering concerns about the status of and plans for network testing. They indicated their belief that network testing must be comprehensive and have metrics to assess performance. Actual blocks of data need to be moved across the networks to adequately test them. It was pointed out that the DAACs really are the customers of the networks, and the responsibility for network development, implementation, and testing lies with the ESDIS Project.
In response to a Panel request, Bill North presented a table that compared dollars budgeted in FY 1998 by the different DAACs for data distribution and user services. The EDC DAAC budgeted amount is substantially more than that of any of the other DAACs. Primary reasons for the difference include the large size of the data sets distributed by the EDC DAAC and the large number of products distributed. Finally, the DAAC provided an overview of the Vegetation Canopy Lidar (VCL) mission and some preliminary information about its potential impact on DAAC resources should the DAAC archive and distribute VCL data and products. Vanessa Griffin reported on the status of a variety of NASA projects, activities, and other happenings, particularly as they relate to the Earth Science Systems Program Office. She noted that both AM-1 and Landsat-7 have experienced delays that likely will result in launch slips. She said that the Office of Earth Science fared well in the President's budget, and she mentioned some personnel changes occurring in the Office. Griffin reported that 24 Type 2 & 3 Earth Science Information Partners (ESIPs) had been selected as working prototype Federation partners, and cooperative agreements are being negotiated with each selectee. There will be an NRC Workshop on the Federated EOSDIS in late February to explore possible approaches to establishing a "Federated" structure for managing EOSDIS. The DAACs have been invited to participate in the workshop. Griffin reviewed the status of DAAC certification, which should be concluded by early summer. Eleven companies were selected to participate in Phase 1 of the Scientific Data Purchase Program, but Griffin expressed some disappointment that only one company proposed to fly a new space mission. One company proposed to use the Land Processes DAAC at EDC for archive and distribution of their products.
H. K. Ramapriyan presented an EOSDIS Overview that addressed overall status, strategy for launch-readiness, emergency back-up plans, and the new paradigm adaptive approach to the PM-era and beyond. Rama noted that the DAACs support a broad and growing user community, and receive orders from about 2300 distinct users each month. The EOS Review Group is meeting this week, and after that meeting a decision will be forthcoming about which system (ECS or emergency back-up) will be used to support generation of standard data products at launch. Rama reviewed the ECS Version 2.0 development strategy whereby enhancements are being delivered in a series of 5 incremental "drops." Activities are being prioritized to ensure "launch critical" functions are operational before launch. Finally, Rama described the New Paradigm-Adaptive Approach, the objective of which is to assure that EOSDIS is responsive to the NASA Earth Science Program and science needs by adapting to new ways of doing business, including more collaborative approaches with PIs, and by evolving from predetermined implementation approaches to hybrid implementation based on cost, technical, and programmatic factors.
Bryan Bailey led a discussion on DAAC interactions with the new Federation ESIPs and the new Data Buy Program selectees. Bailey told Panel members the DAAC is in the process of developing appropriate strategies for interacting with the ESIP and Data Buy selectees, and he said those strategies need to include general criteria against which potential DAAC involvement should be determined. John Dwyer described, in some detail, recent interactions the DAAC and EDC have had with the Earth Satellite Corporation (EarthSat) in support of EarthSat's Phase 1 activities as a successful proposer to the NASA Scientific Data Buy Program. Cooperation includes both provision of data from EDC Landsat archives and distribution of data and products from the DAAC.
In response to a DAAC action to assess the pros and cons of charging for data and data products and to assess the actual feasibility of doing so, John Dwyer led a discussion on possible data charging scenarios. The DAAC's study of data charging issues included identifying some basic principles around which a sound data pricing policy should be established, such as basing charges on true costs of data reproduction and delivery but allowing prices actually charged to be less, or "such as" prices charged should be allowed to vary depending on the media or delivery mechanism used. The reasons identified for charging for data and products included cost recovery and reducing the number of "frivolous" requests. Reasons for not charging include the desire to promote, not inhibit, meaningful use of DAAC data and products. As a result of its study, the DAAC recommended to the Panel that the DAAC start charging for its NALC and GTOPO30 products delivered on hard media. After some discussion, the Panel endorsed the recommendation, and the DAAC took an action to implement a data charging policy and track the impact of the new policy on the number of data and products distributed by the DAAC.
In response to another DAAC action to address potential DAAC-unique extensions, Saud Amer reviewed some of the issues and difficulties in selecting and implementing capabilities not available from the ECS, but which are needed by users. It seems this issue has been one that the DAAC and Panel have been talking about for quite some time without making meaningful progress, for whatever reason. Panel members expressed their frustration that little progress has been made. Most agree that high priority, potentially DAAC-unique extensions include spatial subsetting, an EOS-HDF conversion tool, and additional map projection options. The DAAC took a new action to characterize the data products it will produce in terms of the types and levels of service required to get "appropriate versions" of the products distributed expediently to users. The requirements will be compared to capabilities available from ECS to determine which additional (DAAC-unique) capabilities should be offered by the DAAC. All recognize that available funding may not be sufficient to implement many, or even any, DAAC-unique extensions.
Instrument Team representatives to the Panel presented brief status overviews related to development of their instrument or other issues, and they summarized discussions held during the previous day's instrument team splinter sessions. Jim Irons reported that Landsat-7 has been in thermal vac twice and experienced power supply problems. Power supply units have been returned to Santa Barbara Remote Sensing, and launch has been slipped from July to the December time frame. Irons also discussed some billing and accounting issues that have to be worked out in order to be able to charge for ETM+ data from the beginning of the mission. Level 0 data will cost no more than $475 per scene, and Level 1 data no more than $600 per scene.
Mike Abrams briefly reviewed the ASTER splinter meeting by mentioning the topics and issues discussed in the meeting, including status of topographic derivative product generation, expedited data service, and development of the ASTER DEM standard data product system. Abrams also summarized highlights of the Joint U.S./Japan ASTER Science Team meeting held in Tokyo in December, including discussion about how non-EOS investigators can acquire ASTER data. This information will be available soon on the ASTER homepage.
Chris Justice reported that there are no major problems with the MODIS instrument, and Science Software Integration and Test (SSI&T) is pretty much on schedule, as well. A major issue exists regarding ECS tiling for Level-3 products, which currently is not on the schedule. Discussions for a work around are being held. Other issues noted by Justice include the uncertainty about production capacity at the DAAC due to the fact that hardware resources will be shared among instruments. It currently appears more hardware is required if MODIS is to meet its 2-year production capacity. Also, at-launch global product availability is in jeopardy. Finally, Justice stressed the need for an end-to-end testing schedule that is developed with MODIS Team input and includes network testing of real data.
Ken McDonald of the Earth Science Data and Information System (ESDIS) Project presented an overview of the EOSDIS client. McDonald summarized the development approach for the client, including the facts that it is being built using an incremental track, EOSDIS client tools can include non-ECS components, tools will evolve over the life of the mission, and Version 0 capabilities are deemed sufficient at launch. Under the current release plan, the Version 0 WWW Information Management System (IMS) will provide user access to ECS data at launch. Integration of JEST in a test environment will be completed in May, and the first operational release of JEST is scheduled for September (1998). McDonald presented and discussed a viewgraph that illustrated user options at various stages of client implementation and one that illustrated the overall architecture of JEST. He cited actions that are being taken to improve the Version 0 WWW IMS and the visibility of JEST development. McDonald concluded that broad access to EOS data will be provided at launch of AM-1 through complementary capabilities offered by the V0 IMS and JEST clients.
Lyn Oleson reviewed the status of the DAAC's 1998 budget and activities. He noted that the budget was not impacted in a major way by recent cuts; in fact, the DAAC received some non-DAAC contingency funds to complete generation of Digital Elevation Model (DEM) and derivative products called for by the Science Working Group for the AM Platform (SWAMP) DEM Working Group. Oleson also reported that recommendations made by the Panel at the last meeting regarding certain relative priorities of activities had been implemented. Oleson said there are no big changes or issues since the last meeting. Emphasis at the DAAC is on launch readiness, testing, and defining resource allocation. Finally, Oleson reported that the visit by the National Research Council (NRC) DAAC Review Committee went well. A great deal of information was exchanged, and the committee made some positive suggestions to the DAAC, including one sions on Wednesday. Knoble reviewed the networking requirements baseline, which reflects a cut of $8 million compared to the previous (February 1996) baseline. The new baseline requirements also assume a production ramp-up (25-50-75-100) for all Quality Assurance (QA)/Science Computing Facility (SCF) flows. Knoble also discussed networking implementation, including existing connectivity and other connectivity options that are being explored. Substantial discussion about networking took place among Panel members, Knoble, and others in attendance. Frustrations were expressed by Panel members about network performance, as were concerns about an apparent lack of requirements coordination and about the apparent lack of a standard engineering approach in addressing networking problems. The topic was left with general agreement that further work in fully addressing networking issues is needed.
Lyn Oleson reviewed the status of DAAC readiness for the AM-1 and Landsat-7 launches, including the status of ECS hardware installation, software installation, and staffing at the DAAC. He also reviewed ASTER and MODIS SSI&T activities and plans, as well as various other plans related to overall system implementation. Oleson noted major challenges facing the DAAC related to system performance, user expectations, systems maintenance, and developing user community relationships. He also pointed out areas where the DAAC particularly will seek guidance and assistance from the Panel in the future. Jim Irons suggested forming a tiger team to address some of the major challenges facing the DAAC, and he expressed his concern that the DAAC has yet to bring together the fundamental information it needs to effectively operate the DAAC. He suggested that by the next meeting the DAAC needs to produce data flow diagrams for each product it will produce. Furthermore, each PGE the DAAC will receive should have a defined delivery date and acceptance date.
On behalf of the Panel, Chris Justice summarized the discussions the Panel had on Wednesday in executive session. The Panel believes focus now must be on launchthe DAAC needs to be prepared to do its part to make AM-1 and Landsat-7 work. Yet, the Panel believes the DAAC also needs to look to the future. It needs to define where it wants to be 5 years from now. To that end, the Panel encourages the DAAC to engage in formal strategic planning. The Panel also discussed to set realistic and measurable success criteria.
The Panel adjourned around 5 p.m. and then met in executive session for about an hour.
Chris Justice presented information about EOS validation activities, particularly as relates to needs by validation scientists for data sets covering a number of designated test sites. Justice proposes that validation data sets covering certain sites be routinely compiled and made readily accessible to maximize efficiency in meeting multiple demands for the same data over validation sites. The Oak Ridge DAAC is involved in compiling and making available in situ data, and he would like the EDC DAAC to step forward and implement a process to "package" needed remotely- sensed data for the validation test sites. Concern was expressed by the DAAC and some Panel members about the impact on DAAC resources that such a major effort would have. The DAAC agreed to further investigate, in cooperation with the Oak Ridge DAAC, a range of potential options for meeting requirements for validation data sets.
Joe Senftle, Deputy Manager of ECS, presented a status overview of ECS development. Senftle briefly discussed various schedules they are trying to meet and the many different types of interactions they have with the Instrument Teams, DAACs, and others to ensure that ECS development occurs expediently and appropriately. Senftle listed the baseline at-launch Product Generation Executives (PGEs) for ASTER and MODIS. There was general concurrence as regards the ASTER PGEs, but there was disagreement between what Senftle showed for MODIS at-launch PGEs and what the MODIS Science Team believes will be ready at launch. There was substantial discussion among the group about MODIS land-product PGE development, delivery, and certification schedules. It was not clear how much of the issue is a matter of semantics and how much warrants the concern expressed by the Panel that certain MODIS products potentially may not be ready at launch. The issue will be investigated further by MODIS, ECS, ESDIS, and the DAAC.
Gordon Knoble, ESDIS network manager, attended the meeting to address Panel concerns about network status and development that surfaced during discus its membership. Chris Justice will step down from the Co-Chair position and phase out as the MODIS representative to the Panel. Alan Strahler will take Justice's place. In addition, other changes will be made to the membership, and the Panel believes that new members invited to serve on the Panel should be in tune with the future direction of the DAAC. Also, the Panel believes the DAAC needs to think about what it wants from its Advisory Panel and strive to make effective use of the talents and expertise of its members.
As a final activity, the Panel reviewed and listed the action items that had been recorded during the meeting. The next meeting of the Panel will be held at the EROS Data Center on June 2-4, 1998. The meeting adjourned at 12:50 p.m.