Land Processes DAAC Science Advisory Panel Meets
--G. Bryan Bailey
(gbbailey@edcserver1.cr.usgs.gov) USGS, EROS Data Center,
Tel. 605/594-6001; Fax 605/594-6567
The Land Processes DAAC (LPDAAC) Science Advisory Panel held a regular meeting at the USGS EROS Data Center (EDC) on May 6-8, 1996. Attending were a quorum of Panel members, as well as meeting participants and interested observers from EDC, NASA, and EOSDIS Core System (ECS) contractor Hughes Information Technology Systems (HITS).
Panel Co-Chairman, Tony England, opened the meeting with some reflections on certain events and developments of the past year. He noted that the Panel had gone a full year since its last meeting, opting to postpone the meeting scheduled for last fall due to confusion and uncertainties stemming from EOSDIS federation efforts and proposed recompetition of the DAACs. He also challenged the DAAC and the Panel to recognize that there yet remains a great deal to be accomplished in two years to get to the point where the DAAC hopes to be by launch of AM-1.
Following a tour of the new EDC building addition, where the LPDAAC will be located, Panel members reviewed outstanding action items from the April, 1995, meeting. Notably, DAAC Manager, Lyn Oleson, reported on the anticipated role in DAAC development of the ESDIS Project IV&V contractor and on the DAAC's interaction with the EOSDIS Ad Hoc Working Group on Production; Mike Abrams clarified current plans for ASTER data acquisition request (DAR) allocations; Darrel Williams reported that there will be an X-band antenna in Alaska for real-time downlink of Landsat-7 data and that NOAA has embraced the concept of Landsat DARs for individual users; and LPDAAC Project Scientist, Bryan Bailey, reported substantial progress in preparing a DAAC Science Support Requirements and Activities document. Review copies were distributed to Panel members in anticipation of further discussion later in the meeting.
Dixon Butler presented information about, and led a discussion on, NASA's current concepts for creating a federated EOSDIS. He presented his ideas in the context of a hierarchy of data priorities and the relationship between environmental information and Earth system science. He called for the establishment of an environmental information economy, and he noted that a federated EOSDIS would be one step in that direction. He went on to define and describe what an environmental information system federation is and what Earth science information partners (ESIPs) are, and he explained NASA's ideas for establishing three types of ESIPs. Butler concluded by presenting current NASA plans and tentative schedule for implementing their EOSDIS Federation concepts.
The opening day of the meeting concluded with EDC Satellite Systems Branch Chief, R. J. Thompson, clarifying some of the similarities and differences between Landsat-7 ground system development activities and other LPDAAC development activities underway at EDC. The discussion also addressed the role of the international ground stations (IGS) in Landsat-7 data acquisition plans. Panel members expressed the point-of-view that cooperation, and even partnerships, with the IGS is needed to enhance Landsat data access by the international science community, and they expressed concern that charging the IGS large station access fees will hurt enhanced data access.
The second day of the meeting opened with a spirited discussion of EOSDIS IMS status and plans, by Lyn Oleson, who characterized the "IMS of today as the client of tomorrow." He acknowledged that there are not as many users of the X-windows version of the IMS as anticipated, partly because it may be too complex for the casual user. Someone else observed that "it is too slow for the knowledgeable user, so who is it for?" Dave Skole pointed out that part of the problem is that some users are expecting more from the Version 0 IMS than ever was intended.
He also stressed the need for stronger involvement of the user community in various IMS prototyping activities and its development process. It was noted that the ECS Evaluation Package 7 (EP-7) will be released this summer. The Panel assigned the DAAC an action to sponsor a "client workshop" this summer to examine existing IMS implementations, both operational and prototypical including EP-7, and develop a list of recommended ECS system functionalities that would be of particular use to the land science community.
Separate from related reports provided by Panel members associated with ASTER and MODIS Science Teams, DAAC staff summarized certain specific DAAC-ASTER and DAAC-MODIS activities. Jim Lacasse reported on ASTER beta science software integration and test (SSIT) activities at the LPDAAC. He provided an overview of the ASTER standard data product development strategy, identified the higher level ASTER products currently undergoing SSIT at the DAAC, reviewed the recent SSIT activities and their results in some detail, and outlined planned activities.
Bryan Bailey reported on the status of efforts to implement the Japanese ASTER Level 0 to 1 processing software at the LPDAAC, implement an expedited ASTER data processing system at the DAAC, and to implement an ASTER precision processing capability at the DAAC. Glenn Kelly reviewed the ASTER DEM standard data product software procurement status. He discussed details of the requirements analysis that precedes release of any Request for proposals (RFP), as well as the current schedule for procuring the Commercial Off The Shelf (COTS) software for Digital Elevation Model (DEM) generation from ASTER data.
Jeff Eidenshink reviewed for the Panel various coordination activities ongoing between the LPDAAC and the MODIS Project. He characterized coordination responsibilities as: representing requirements and concerns of the LPDAAC to the MODIS Science Data Support Team (SDST), ECS, and the ESDIS Project; providing various and specific V0 experiences to ECS development; and representing traditional user requirements to the MODIS Science Team. Coordination is accomplished through regular participation at meetings (also telecons) of the SDST, MODIS Science Team, and MODIS Land Science Team (MODLAND), and through participation in major activities, such as developing plans and identifying requirements.
Eidenshink also summarized various MODIS-relevant "lessons learned" from the Global Land 1-km AVHRR Project. In the area of data processing, standards and documentation are critical for production staff and data users, and benchmarks are necessary to validate algorithm performance and quality assurance. Benchmarks should be run on a routine schedule. As relates to data distribution, tape media are the overwhelming choice for product delivery; networks are not sufficient for large data volume transfers to a majority of users; and spatial, spectral, and temporal subsetting are necessary. Other general lessons learned include the fact that tools are needed by users for such things as facilitating use of multiple map projections, and reprocessing is difficult because of the data volume of the total data set.
Jay Feuquay reported on the DAAC's high-speed communications activities with MODIS. These include an experiment to investigate high-speed networks for large-volume data transfers where 60 GB of AVHRR data are transferred to the MODIS Science Team. Ten gigabytes of data have been sent to the team, so far. Feuquay reviewed the physical and logical connections involved in the data transfer, and he reported results of 1.25 Mbytes/sec average throughput with a peak throughput of about 3 Mbytes/sec. Plans for future work also were noted.
During the reports from Instrument Team members, Chris Justice identified the new members on the MODIS Team and briefly explained their roles. He also provided updates or commented on the MODIS validation plan, QA system design, common processing requirements, inter-instrument data dependencies, MODIS processing design, and various DAAC-related issues. In addition, Barbara Putney drew attention to recent increases in MODIS data processing and archiving loads, and she noted certain efforts being undertaken to deal with problems created by such increases. Putney also reviewed certain MODIS-related issues from the recent ECS critical design review (CDR).
Mike Abrams provided an ASTER activity report, covering developments in the areas of systems engineering, algorithm and software development, operations and mission planning, future plans, and management. Highlights from ASTER include the fact that Moshe Pniel has taken over as JPL ASTER Project Manager, replacing Dave Nichols, who accepted a Radarsat Project management position; the beta version of science processing software was successfully implemented at the LPDAAC; first-year funding for a new airborne ASTER/MODIS simulator (MASTER) was approved; and a major cal/val field campaign will be held in Nevada in June.
Diane Evans provided an update on SAR activities. She reported that the National Research Council (NRC) SAR review stressed the importance of frequency and polarization diversity, endorsed the proposed Shuttle Topographic Radar Mission (SRTM), and generally came to many of the same conclusions as the Interagency Ad Hoc Working Group on SAR. Evans also noted that NASA appears to be moving toward a potential new start for a LightSAR in FY 1997, and she encouraged the DAAC to remain proactive in assessing potential involvement with current and future SAR missions.
Landsat Project Scientist, Darrel Williams, provided a crisp status report on Landsat Project activities. Topics covered included status of Landsat-7 image assessment system development, addition of Level 1 processing for Landsat-7, approval of a second X-band receiving station, plans to synchronize the orbits of Landsat-7 and EOS AM-1, activities related to the definition of and selection by NASA's New Millennium Program (NMP) of an advanced land imager as the first instrument to be developed by the NMP, and the process of selecting a new Landsat Science Team, the members of which were announced in June.
The balance of the meeting's second day was spent reviewing status and plans of LPDAAC activities, with different DAAC staff presenting information and responding to Panel member questions. Mike Benson covered DAAC User Services and data distribution topics. He noted that the workload for the User Services staff has increased dramatically since the beginning of the 1996 calendar year, and, if the workload continues to increase, an additional full-time person will have to be added in FY 1997. He also reviewed selected user profile statistics and summarized DAAC outreach activities. In addition, Benson reviewed LPDAAC data distribution statistics by product, as well as by media, and he presented information that helped to characterize the users of various LPDAAC products.
Lyn Oleson reviewed DAAC data ingest status and plans. Data and products currently available from the DAAC include: TIMS, NS-001, and TMS aircraft scanner data; Landsat Pathfinder (MSS & TM) data; 1-km AVHRR orbital segments; global 10-day AVHRR and normalized difference vegetation index (NDVI) composites; 1-km DEMs for most continents; SIR-C Survey and Precision data; and SIR-C Educational CD-ROMs. Currently, only the aircraft scanner and some Landsat Pathfinder data are searchable and orderable through the system-wide Version 0 IMS.
However, all of these data sets are orderable through the EDC DAAC WWW home page. Efforts will continue to populate the Version 0 IMS with the additional Landsat Pathfinder data and SIR-C Precision data planned for release to the IMS in July and August. Additional data sets planned for DAAC ingest over the next year or being considered for ingest, include AVIRIS, ASAS, SEASAT, AirSAR, MASTER, and others.
John Dwyer summarized status and plans of mission support activities, which include development of the global 1-km AVHRR, global 1-km topographic, aircraft scanner, SAR, and Landsat Pathfinder data sets. Notably, orbital stitching of the global AVHRR 1-km data has been completed for the April, 1992 through December, 1993 acquisition period, and the global 10-day compo- sites are on schedule for completion by the end of FY 96. Global data acquisition has been extended to October, 1996. The 1-km digital topographic data sets for Africa, North America, Europe, Asia, and Antarctica have been completed and are staged for FTP access; the South America data are expected to be completed in June.
Demand for the SIR-C Survey and Precision data remain high, and the Survey data have been loaded onto digital linear tape for FTP access from the DAAC data distribution system (DDDS). Planning is underway to transfer the SIR-C processor from JPL to the LPDAAC in September. Generation of the Humid Tropical Forest Project (HTFP) Landsat Pathfinder products continues, and previous problems with the Instituto Nacional de Pesquisas Espaciais (INPE) TM data have been resolved. DAAC Systems Engineer, Tom Kalvelage, briefed the Panel on engineering and development activities, specifically Version 0 sustaining engineering, design, and development of the DAAC's new information server, DDDS status and upgrade plans, and DAAC support of ECS development.
Notably, recent sustaining engineering activities have included installation and test of multiple versions of V0 IMS client and communications software and developing software to upgrade DAAC servers to support new client versions. The DAAC is moving from a server shared with USGS functions to a dedicated DAAC IMS server. A prototype has been built with WWW access. The DDDS has experienced some minor problems, but solutions have been effected. DDDS short-term plans include increasing storage capacity to 2.6 TB, and long-term plans call for increasing capacity to more than 9 TB. DAAC support of ECS development has been extensive through participation in many reviews and other meetings, working groups, workshops, and telecons.
The final day of the LPDAAC Science Advisory Panel meeting started with the DAAC's ECS Science Liaison, Saud Amer, presenting an overview of ECS prototyping activities. Primary purposes of ECS prototype activities are risk mitigation, enhancement of system functions, and technology assessment. Amer identified and described numerous specific prototype efforts that ECS is participating in with academic institutions and private organizations. He also discussed the relationship between various prototype efforts and the incremental inclusion of new capabilities in successive ECS Evaluation Package. Amer noted that the prototyping activities are the mechanisms through which: 1) the user community has the opportunity to impact the "look and feel" of the ECS system significantly in advance of implementation, and 2) the ECS contractor pursues the latest technologies and addresses key subsystem features that are important to the user community.
ECS contractor representative, Tom Dopplick, addressed the topic of V0-to-ECS data migration, as well as other related transition topics. Dopplick noted that the rationale for migrating selected data from V0 to ECS includes both preservation of scientifically important historical data and enhanced services to users wishing to access those data. Furthermore, ECS is contractually bound to effect data migration. Dopplick also explained that data format, specifically HDF-EOS, and metadata are related and important considerations in formulating a data migration plan.
The DAAC, with advice from its science constituency, must decide not only which V0 data to migrate, but also whether or not to convert native formats to HDF-EOS and whether or not to enhance metadata attributes or provide only the minimal, mandatory attributes. Also, migrated data must be validated to ensure that the integrity of the original data is preserved during the migration process. Finally, Dopplick pointed out that it is important for ECS to "learn" about the migrated data and metadata before services can be provided to users. Simply placing the migrated data and metadata into an ECS archive at the DAAC results in NO services. The ECS "learning" process requires that migrated data and metadata flow through both the ECS ingest and data server subsystems.
The Panel concluded its meeting with summary discussions that included a review of action items and Panel recommendations. The next meeting of the LPDAAC Science Advisory Panel will be held in the September/October time frame.