The U.S. Geological Survey (USGS) has a long history in the development of geospatial data, starting with the topographic mapping of the nation that began in the late nineteenth century and continued throughout the twentieth century. The USGS developed topographic maps on several scales that have been used over the past 100 years by professionals and citizens alike. To carry out its mapping mission the USGS gained considerable expertise in cartography and, with the advent of computer technology, was one of the leaders in the development of techniques and standards in the field of digital cartography. Even further advances in technology have pushed the USGS to strive toward a fully electronic implementation of the topographic maps and related geospatial data, which has been named The National Map.
The challenges of developing The National Map differ greatly from those faced by cartographers even 20 years ago. Meanwhile, geographic information system (GIS) technology has become ubiquitous, with digital mapping sources to be found anywhere from specialized government agency sites, to state and local government web pages, to commercial sites that have caught the interest of the general public. Within this fast-changing environment, the USGS realized the need to assess the focus of its research in geographic information science (GIScience) to determine how it could best meet the needs of The National Map, the USGS, and the nation.
THE EMERGENCE AND FOCUS OF CEGIS
The idea of a Center of Excellence for Geospatial Information Science (CEGIS) was first proposed by McMahon et al. (2005) in a report that describes a science strategy for geographic research, including GIScience, at the USGS between 2005 and 2015. CEGIS was initiated by the Associate Director for Geospatial Information in 2006.
CEGIS is housed within the National Geospatial Program Office (NGPO). The NGPO was created in 2004 when the USGS reorganized its geospatial information programs to better invest in technology and partnerships aimed at modernizing its collection, management, processing, updating, and delivery of geospatial information.1 The major elements of USGS’s geospatial programs and services unified under NGPO include The National Map, the National Atlas of the United States of America®, the Federal Geographic Data Committee secretariat, Geospatial One-Stop, and other geospatial program elements (Figure 1.1). Geospatial information is one of five disciplines within USGS, the others being water, geology, geography, and biology (Figure 1.1).
Among the NGPO’s responsibilities are defining the overall GIScience (see Box 1.1) research agenda and championing GIScience research as a component of USGS's science portfolio. CEGIS undertakes these GIScience research responsibilities. The USGS’s vision for CEGIS is to “conduct, lead, and influence the research and innovative solutions required by the National Spatial Data Infrastructure (NSDI)” (CEGIS, 2006).2 CEGIS’s mission is to “conduct, support, and collaborate in research to address critical Geographic Information Science questions of importance to the USGS and to the broader geospatial community” and “as an outgrowth of and complement to this research program, CEGIS will support and collaborate in technological innovations that further the implementation of the NSDI” (CEGIS, 2006).
Definition of Geographic Information Science
The individual who coined the term GIScience defined it as “a multidisciplinary research enterprise that addresses the nature of geographic information and the application of geospatial technologies to basic scientific questions” (Goodchild, 1992).
CEGIS staff recently identified the reach of GIScience as including “the traditional mapping disciplines of surveying, aerial photographic interpretation, photogrammetry, remote sensing, and cartography. It also encompasses a broader scope of issues related to the modeling and representation of geographic phenomena, data, and processes; human cognition of geographic information; the analysis, depiction, and use of uncertainty information; spatial analysis and modeling, including geographic information systems (GIS); scale sensitivities; geographic ontologies; visualization; and other similar topics” (CEGIS, 2006). GIScience relies on expertise from many allied fields and has intimate ties to geospatial technology and applications.
As stated earlier, the starting point for planning CEGIS’s GIScience research activities is a study by McMahon et al. (2005) that describes a science strategy for geographic research, including GIScience, at USGS between 2005 and 2015. The recommendations of the McMahon report were, in fact, quite broad and include the needs of the other USGS disciplines. Furthermore, many of the authors, including McMahon, were from disciplines within USGS other than geography, or were from outside of the USGS. The McMahon report recommended that USGS establish CEGIS to lead USGS GIScience research (for details on the McMahon report’s ideas for CEGIS, see Appendix C). Proposed areas of focus within CEGIS are drawn from goals 8, 9, and 5 in McMahon et al. (2005), namely:
Provide timely, efficient, and intelligent access to new and archived USGS geographic data needed to conduct science and support policy decisions (Goal 8).
Develop innovative methods of modeling and information synthesis, fusion, and visualization to improve our ability to explore geographic data and create new knowledge (Goal 9).
Develop credible and accessible geographic research, tools, and methods to support decision making related to the human and environmental consequences of land change (Goal 5).
While the call to action in the McMahon report is the primary reason that CEGIS is being established now, the agency feels a need to retain key talent with a critical mass of researchers for which a center of excellence would be helpful (Steve Guptill, USGS, personal communication, 2006). The McMahon report suggests that such a center would build, nurture, and maintain a core of GIScience researchers and provide a focal point and sense of identity for these
researchers. Lastly, CEGIS will build a science role within NGPO in addition to its operational and leadership roles.
The CEGIS budget covers three full-time equivalent (FTE) employees who are considered CEGIS staff—one in Reston, Virginia, and two at Rolla, Missouri. In addition, CEGIS funds GIScience-related projects conducted by two to three other USGS FTEs and a support staff of four to six FTEs.3 The USGS also funds a CEGIS-affiliated postdoctoral research position managed by the University Consortium for Geographic Information Science (UCGIS) (CEGIS, 2006). In late 2007, three additional postdoctoral positions in GIScience will be added through USGS’s participation in a National Research Council (NRC)-administered postdoctoral program.
The FY 2007 budget has two components. The first is $1.2 million to cover the activities listed above, which are located not only at Reston, but also at the Center for Earth Resources Observations and Science (EROS) in Sioux Falls, South Dakota, and in Rolla, Missouri (to support all of the FTEs mentioned above). The second component of the CEGIS budget is an FY 2007 call for proposals, which resulted in approximately $1 million being awarded to seven interdisciplinary projects that involve people across USGS in collaboration with non-USGS partners.
CEGIS’s research portfolio is thus a mix of (1) preexisting GIScience-related projects already funded within USGS and placed under the management of CEGIS when it was formed in 2006 and (2) the seven new projects funded under the FY 2007 call for proposals.
FOCUS OF THIS REPORT
Recognizing the need to develop a set of research goals and priorities for CEGIS that will best meet its needs for future capabilities in GIScience, the USGS approached the NRC through its Mapping Science Committee (MSC) and asked the MSC to form a committee to develop these research goals and priorities (Appendix A). Using knowledge of the current state of the art in GIScience and information from USGS on current and future needs and capabilities, the committee was asked to determine which areas of research would be most effective for the CEGIS to pursue. The three primary tasks follow:
Identify current and future USGS needs for GIScience capabilities.
Assess current capabilities in GIScience research at the USGS and recommend strategies for strengthening these capabilities and for collaborating with others to maximize research productivity.
Using knowledge of the current state of the art in GIScience, make recommendations regarding the most effective research areas for the CEGIS to pursue.
To complete its task, the committee met three times in person—twice in Washington, D.C., and once in Irvine, California—and numerous times by phone. The committee benefited from input from a range of experts (Appendix B) and drew from a broad range of documents listed in the references.
Given the short time frame and potentially broad scope of its task, the committee chair and study director met with the USGS Associate Director for Geospatial Information, the CEGIS Director, and the Associate Director’s Chief Scientist (Karen Siderelis, Steve Guptill, and Anne Frondorf, respectively) prior to the first full committee meeting to discuss the sponsor’s expectations from the study and to gain insights into their priorities among the items in the committee’s task. As the primary audience for the report, these senior USGS staff indicated that they were most interested in the committee’s insights on item 3 and the second half of item 2 of its task. These tasks were summarized as, What should CEGIS focus on and how can this be achieved? The committee’s discussion of and recommendations on these tasks are covered in Chapters 3 and 4, respectively.
In addition to emphasizing a desire for the committee to focus on tasks 2 and 3, USGS steered the committee away from primary research on the first task and the first half of the second task. Instead, the USGS urged the committee to draw on recent reports such as McMahon et al. (2005) and on responses to the call for proposals for this information.4 These tasks are covered in Chapters 1 and 2.
In addition to the USGS’s guidance on emphasis among the committee’s three tasks, it encouraged the committee to shape the CEGIS research portfolio based on its need for GIScience research across USGS (i.e., not solely within NGPO). However, the committee is keenly aware of the importance of The National Map to the mission of NGPO and to the USGS as a whole, not to mention
the USGS’s topographic mapping mission responsibilities. In 2002, the NRC stated that “developing The National Map is the most important single initiative in the Geography Discipline at the USGS” (NRC, 2002), while recognizing that “The National Map as a database product and an information base is an attainable goal by 2010, but some of the basic knowledge needed to create it (and other spatial data products) is not yet available,” and that “present knowledge, methods, and tools are inadequate to create The National Map … ”. In the committee chair’s meeting with Karen Siderelis, Steve Guptill, and Anne Frondorf, they confirmed that supporting The National Map is of highest priority and that there were critical research needs for accomplishing that objective.
The USGS urged the committee not to constrain the scope of research based on current CEGIS resources and to think in terms of the next decade of research. Even with this guidance, the committee concurred with the USGS sponsors that the range of needs for GIScience research (as described in USGS, 2001; NRC, 2002; McMahon et al., 2005, and responses to the FY 2007 call for proposals, for example) would readily exceed even the most optimistic expectations of resource availability for CEGIS. Consequently, the committee focused its view of CEGIS’s role on applied, technical aspects of GIScience and away from software engineering, product development, and nontechnical aspects (e.g., institutional issues, digital rights management challenges) of supporting the National Spatial Data Infrastructure (while recognizing the need for these roles elsewhere in USGS). The committee’s process of prioritization of research tasks is explained in Chapter 3.
GISCIENCE CAPABILITIES AT USGS
The USGS employs a small cadre of GIScience professionals. In addition, it uses several mechanisms that bridge to external GIScience expertise.
Internal GIScience Resources
Much of USGS’s GIScience expertise has already been identified and linked to CEGIS. As one indication of the GIScience capabilities already tapped by CEGIS, the projects funded in FY 2007 (excluding those funded through the bureau-wide call for proposals) focus on the following:
Automated data integration ($280,000)
Generalization for The National Map ($190,000)
Building an ontology for The National Map ($250,000)
Multiresolution raster data for The National Map ($200,000)
LIDAR-derived elevation technology assessment ($80,000)
Elevation feature extraction ($127,000)
Fractal and variogram analysis of scale and resolution effects in geospatial data ($100,000)
Another indication of GIScience capabilities across the bureau arises from the topical focus of the proposals received in response to CEGIS’s FY 2007 USGS-wide call for proposals. This call by CEGIS leadership had a goal to identify “hidden” GIScience talent within the disciplines at USGS that might be a valuable resource for the center (Steve Guptill, USGS, personal communication, 2006). Although each research team must be led by a USGS researcher, the team is encouraged to be multidisciplinary and include non-USGS expertise. Consequently, an additional result from this call is that it reveals the broader network of GIScience capabilities to which USGS experts are already connected.
Of the 69 proposals received in response to the call, 23 were submitted from the water discipline, 20 from geography, 15 from biology, 10 from geology, and 1 from NGPO. In all disciplines there was a broad range of proposed topics that spanned monitoring and data capture through data integration, analysis, and error propagation to modeling and decision support. CEGIS leadership now has a better sense of the distribution and range of GIScience and related capabilities across USGS. The seven projects funded through this call for proposals are:
Scaling, Extrapolation, and Uncertainty of Vegetation, Topographic, and Ecologic Properties in the Mojave Desert ($73,000);
A Landscape Indicator Approach to the Identification and Articulation of the Ecological Consequences of Land Cover Change in the Chesapeake Bay Watershed, 1970 – 2000 ($132,000);
Assessing Local Uncertainty in Non-Stationary Scale-Variant Geospatial Data ($117,000);
Methods to Quantify Error Propagation and Prediction Uncertainty for USGS Raster Processing ($135,000);
The Geoscience of Harmful Invasive Species: Integrating LANDFIRE (Landscape Fire and Resource Management Planning Tools Project) and Invasive Species Data for Dynamic and Seamless Integration of Raster and Vector Data to Meet Management Needs at Multiple Scales ($150,000);
Mapping Inundation at USGS Stream Gage Sites: A Proof of Concept Investigation ($150,000); and
GEOLEM:5 Improving the Integration of Geographic Information in Environmental Modeling through Semantic Interoperability ($150,000).
See Appendix D for more information on these projects.
GIScience research activities funded within and by CEGIS are occurring in the midst of USGS’s shift in emphasis of staff expertise away from paper map production and toward integration of digital data (Figure 1.2). An assessment of resources against the current mission shows an excess of staff engaged in data production and a shortfall of staff skilled in data integration—a function that will be important for assimilating data from other sources as the USGS moves out of internal production and into working with partners.
Leveraging External Resources
USGS uses Cooperative Research and Development Act (CRADA) agreements as one means of connecting with external GIScience expertise. An example of such an agreement for geospatial activities is with Microsoft Corporation on the development of Terraserver.6 In addition to CRADAs, USGS has leveraged external GIScience expertise by arranging a series of visiting academic GIScientists who have been based at Reston, Virginia during sabbatical leave; participating with the National Geospatial-Intelligence Agency (NGA) in the solicitation and review of NGA University Research Initiative proposals; conducting a graduate school training program in which more than a dozen USGS employees pursued GIScience studies at universities including Ohio State University, State University of New York Buffalo, University of California at Santa Barbara, and University of South Carolina; organizing research meetings (e.g., the Public Health Colloquiums) with expert participants invited through UCGIS; funding postdoctoral positions in GIScience at USGS facilities; and sponsoring GIScience professional meetings (Steve Guptill, USGS, personal communication, 2007).
FUTURE USGS RESEARCH NEEDS AND CEGIS
Geography, GIScience, and mapping will be increasingly important to the USGS’s water, geology, and biology disciplines. With the new roles of map and information integration, CEGIS will face increasing demands for solutions to complex geospatial data processing challenges as well as automation of those functions so that USGS researchers can handle large amounts of dissimilar and nonconforming data with frequent updates. In addition, USGS’s major role in analyzing land change over time will require new GIScience-derived methods.
In 2007, the USGS released its report outlining a 10-year science strategy for the agency. Facing Tomorrow’s Challenges: USGS Science in the Coming Decade (USGS, 2007) sets the bar high early in the introduction: “The USGS is the Nation’s and the world’s leading natural science and information agency … [whose efforts] … allow the USGS to map and understand land use/land change trends across the Nation.” The 2007 report highlights the GIScience needs that could be fulfilled by CEGIS. These needs are described in Chapter 2.