TEN-YEAR REVIEW OF THE NATIONAL SCIENCE FOUNDATION'S LONG-TERM ECOLOGICAL RESEARCH PROGRAM

EXECUTIVE SUMMARY

In 1980, the National Science Foundation began the Long-Term Ecological Research (LTER) program with a budget of $1.2 million and five sites selected on the basis of competitive proposals. Subsequently, the LTER program has grown to $11 million and eighteen sites located throughout North America and Antarctica, now involving more than 200 research scientists and about 300 technical personnel and students. The goals of the LTER program are: (1) long-term analysis of site-specific ecological phenomena; (2) comparison of observations across diverse ecosystems and in terms of general systems theory; and (3) provision of large, secure, ecologically diverse sites with well developed support capabilities.

Each participating LTER research site is expected to address five core research topics:

• pattern and control of primary production;
• spatial and temporal distribution of populations selected to represent trophic structure;
• pattern and control of organic matter accumulation in surface layers and sediment;
• pattern of inorganic inputs and movements of nutrients through soils, groundwater, and surface waters; and
• pattern and frequency of disturbance to the research site.

The NSF Biological Sciences Directorate constituted the Ten-Year LTER Review Committee and charged it with appraising the ten-year performance of the LTER program, including activities at the sites and of the entire network. In addition, the Committee was requested to identify those aspects of the current system that could benefit from change and to provide recommendations for the future of the LTER program.

During the ten-year period in which the program has been fully operational, scientists involved in the LTER program have produced a distinguished and extensive literature describing important ecological processes and numerous ecological characteristics from the sites. These long-term research programs have documented the status and trends of key environmental conditions from the eighteen sites. In addition, the LTER program has established interactions with and provided guidance to related programs in several foreign countries. In many instances, research from the LTER program has been instrumental in guiding policies and providing a scientifically credible basis for prudent management decisions about natural resources. For example, a significant portion of the information used to establish the U.S. Forest Service's "New Perspectives in Forestry" originated from studies at three LTER sites: H. J. Andrews Experimental Forest, Coweeta Hydrological Laboratory, and Hubbard Brook Experimental Forest. Similarly, the current re-orientation of most of the federal agencies toward a more comprehensive "ecosystem management" approach is built upon many studies from the LTER sites and program.

Despite the unquestionable success of the LTER program, there are several issues that need attention if the program is to improve. For example, the program definition as provided by the NSF Calls for Proposals has purposefully remained flexible to permit as much scientific latitude as possible on the part of the investigators proposing LTER studies. This intentional flexibility has led to some ambiguity in the relative importance of site-specific long-term studies, cross-site comparisons, and furthering ecosystem theory through in-depth studies.

A unique strength of the LTER program is that it specifically studies phenomena that occur over longer time periods, and the funding and research support are designed to recognize these long-term dynamics. At intervals of six years, each site is required to submit a proposal to continue the program. These proposals are expected to have new ideas, but the site is then faced with the need to continue the initial studies while also beginning new ones. The current level of funding is inadequate to meet both continuing and new needs of the research programs at the LTER sites. Thus, there is an obvious and serious discrepancy between the current expectations of the LTER sites and actual funding levels (about $500,000 per site per year). An LTER site is expected to develop an infrastructure for promoting, conducting, and preserving research in all five core areas while also conducting innovative work focusing on current theories and problems in ecology. The expectations of each site increase incrementally as the accumulation of required research increases, as more demands are made on the site from other investigators who want to use the data or work on the site, and as the results of the LTER program become more widely known and receive increased attention and demand. Over the duration of the program, funding has increased, but not at the rate of the increasing demands on the science in the LTER program.

The power of the network approach of the LTER program rests in the ability to compare similar processes (e.g., primary production or decomposition of organic matter) under different ecological conditions. As a result, LTER scientists should be able to understand how fundamental ecological processes operate at different rates and in different ways under different environmental conditions. Some intersite comparisons have been conducted, but the power of the network of coordinated research sites has not yet been fully realized. In addition, the criteria that have been used to select the sites that participate in the program have not uniformly included specific expectations for developing a comprehensive, representative network of sites. Some important ecosystem types are not represented in the current constellation of sites, for example, urban, Great Basin, southern swamps and wetlands, large rivers, and additional marine systems.

The five core research topics were designed to provide guidelines to ensure that sites are selected for broad-based ecological research rather then narrowly focused studies. However, these core areas are defined so broadly that they do not provide significant operational assistance in site selection or evaluation. Moreover, because of the LTER organizational structure within NSF, the research interests of the LTER investigators, and the nature of the core areas themselves, there is some perception of a bias toward ecosystem-level research.

The present LTER review process has been fair. The reliance on peer reviewers for site proposals has ensured that the primary criterion for selection and renewal has been the quality of the science. However, because many reviewers of proposals are accustomed to evaluating hypothesis-driven, single-investigator proposals, some of them have tended to emphasize short-term rather than longer-term and transdisciplinary strengths of LTER proposals. Further, some reviewers tend to focus on individual site achievements rather than on contributions sites make to the overall LTER network.

The current LTER program has been highly successful in terms of both research output and provision of a structure for long-term ecosystem research in the United States. The current level of high-quality research has been achieved by involving excellent scientists and rigorous review of individual proposals and by subsequent periodic review of all components of the LTER program. This same level of high expectations must continue, and the current safeguards must remain in place.

Several steps will improve the LTER program in the future. First, on the basis of the experience gained over the past ten years, the goals should be restated: the goals of the LTER program should be to conduct long-term, continuous measurements and analyses of ecological patterns and processes at specific sites, to integrate and synthesize results both within those sites and among sites, and to seek ways to generalize these results over broader spatial scales.

The five core research topics should not be used as organizing principles per se; rather, the original objective of ensuring breadth, continuity of measurements, and comparability should be accomplished but in a fashion which overcomes the problems with the core areas. Specifically, each site should investigate and document key selected ecological patterns and processes that determine the spatial and temporal ecological characteristics and behavior of the particular ecological systems under study. Moreover, at each LTER site there must be a defined core of site-appropriate observations and measurements that are consistently taken for the duration of the project. All measurements and samples should be collected and analyzed in a manner to be as standardized and comparable as possible with other LTER sites.

The LTER program has focused primarily at the ecosystem level of ecological organization. Because long-term ecological research is also needed at other levels (e.g., population, community, landscape, global), and because understanding of ecological patterns and processes at one level is vastly enhanced by studies at adjacent levels, LTER sites should include more research at these other levels of ecological organization.

The LTER program has developed from an initial constellation of sites, joined only by a common interest in long-term ecological phenomena. More recently, the program has begun to take advantage of the network dimension and to develop and institute comparative intersite studies. Integration and cross-site studies should be question-driven, encouraged, and facilitated within the LTER program.

The LTER program has not been effective in communicating its strengths. Much of the attention of the program seems to have been devoted to site-specific studies, and relatively little effort has been directed toward the future of a powerful set of integrated research sites. As the program continues to mature, more programmatic attention must be applied to developing an ever-expanding vision of the value of the integrated LTER program and to communicating the successes of the program to a wide scientific and non-scientific audience.

The governance of the LTER program needs to evolve as the program develops and matures. The organizational structure needs to disperse the governance process more widely throughout the program, and the Coordinating Committee (which includes the Principal Investigator from each site and the Network Office), through its Executive Committee, must take additional responsibilities. Also, external advisory groups should be established for the LTER program as well as for NSF as it continues to support the program. These changes will facilitate strategic planning and broader participation in decision-making.

Although the NSF review process has been fair, several steps can be taken to improve the review of initial proposals and of proposals produced by continuing sites. These improvements include clarification for the reviewers of the expectations of the LTER program and some consistency of reviewers over the duration of the project.

In the past few years, it has become abundantly clear that managing the earth's resources is not only important but complex. Complications arise for several reasons: economic demands on both renewable and non-renewable resources often damage the long-term productivity necessary to support both the regeneration of renewable resources and the ecological services upon which all life depends; there are frequently competing demands for the natural resources located on parcels of land; and retaining important natural resources and ecological processes may constrain short-term economic development. These issues are important because: this generation needs to consider the consequences of its actions on the next generation; local actions may affect regional and global processes and vice versa; and management decisions about natural resources affect the political and economic stability of societies.

Decisions concerning the management of natural resources will always be made in the political arena, usually based on the values expressed by many constituencies. These decisions will represent judgments by decision-makers who consider as much evidence as possible. In the past, many such decisions have been made without adequate information about the consequences of the actions as they relate to the near-term and long-term availability and quality of natural resources and vital ecological processes. Frequently, these decisions have resulted in the loss of resources or in the need for expensive recovery actions. Thus, the challenge is to provide the best possible information so that decisions can be based on the most accurate predictions of the consequence of any action.

The most useful information about natural resources for a decision-maker will have

several characteristics:

Any program designed to provide this information must be based on a stable framework so there is both a continuity to the data collections and a convenient methodology for translating the data to the decision-maker in a manner that is most useful.

Currently there is no environmental research program that meets all of these needs. However, the LTER program contains many of the essential elements, particularly a structure for providing long-term measurements, continuity, and an organized framework for collecting, archiving, and making information available.

The LTER program could easily continue its present trajectory as exemplary research, involving excellent scientists who conduct research on important long-term ecological phenomena and questions. In so doing, the program would continue to contribute in very significant ways to the nation's environmental research effort. However, this review has convinced the committee that there is a far broader and more important role for an LTER program. In fact, as just described, many of the challenges of designing and operating a sustainable biosphere can be most effectively and economically confronted with a newly defined LTER program. The present LTER program thus contains the nucleus of a vitally important national effort.

The current constellation of LTER sites and individual research programs can and should be expanded and molded into an operational network including a wider range of ecological organizational levels; a more complete suite of scientific disciplines; research programs designed to address directly topics that are important for the formulation of policies and practices leading to sustainable ecological systems; the development of technologies to mitigate the adverse environmental effects of human activities; and support for broader ecological education. This new program, building on the strengths of the current LTER program, would involve an expanded and much more integrated network of sites, called the Long-Term Ecological Research Network (LTERN). The ultimate success of the LTERN will, of course, depend on the broad success of the nation's ecological and environmental research program and, therefore, should be developed in concert with related efforts.

Primary goals for this research program are, therefore, to address effectively the long-term ecological, scientific questions of today; to develop an integrated ecological research program that can serve the nation's need to make defensible environmental decisions and policies; and to implement practices leading to a sustainable biosphere. To achieve these goals, the LTERN must accomplish the following:

• Make long-term ecological measurements in order to help understand the dynamics of ecological systems, provide a baseline against which to measure environmental changes, and evaluate and mitigate the adverse impacts of human activities.
• Enhance the spatial dimension of the program so that it will be possible to determine the conditions under which lessons learned in one location can be applied to other locations.
• Expand the network of sites to represent additional important ecosystems that are absent from the existing program.
• Increase the scope of the research program to include a primary emphasis on various levels of biological organization (e.g., genetic to landscape levels) since research has conclusively demonstrated that multiple levels of investigation are required to understand key ecological processes.
• Expand the range of scientific disciplines to include the physical sciences and the social sciences, since understanding ecological phenomena requires the integration of these disciplines and because practical solutions to environmental problems can only be developed and implemented by involving these additional disciplines.
• Institute programs to develop new technologies for measuring and understanding the environment and, as appropriate, for testing the environmental impacts of new technologies.
• Contribute to the nation's educational effort at all levels, particularly by providing educational opportunities and materials directed toward multidisciplinary, field oriented, and problem-solving education and training.

To meet these needs, the committee recommends that the new LTERN include a broader representation of key types of ecological systems by expanding the number of primary sites to no more than double the current number. Spatial resolution of the network should be increased by adding satellite sites to the primary LTERN sites and by taking advantage of the broad range of shorter-term ecological research sites supported by NSF and other agencies. Research programs should be directed specifically toward providing increased understanding of ecological systems, particularly long-term processes, and also toward using that understanding to provide answers to questions of resource management and environmental protection. The LTERN research program must include a broader range of scientific disciplines, and the biological component must involve all levels of organization. Special emphasis must be placed on technologies, both for developing new techniques for measuring environmental processes and for testing the impact of new technologies in a variety of environmental conditions. Moreover, the LTERN program must have formalized ties to the nation's decisionmaking processes and to its educational programs.

To meet these new challenges, the LTERN annual budget must increase by an order of magnitude. This budget must include funds for research and management of the primary and satellite LTERN sites, for intersite coordinated studies, for studies conducted by other scientists on problems and ideas that are related to the LTERN, for conducting collaborative research with other agencies and organizations, and for increasing the network and communication activities of the program.

This new budget represents a significant investment, especially in these times of fiscal constraint. However, this investment will be repaid many times over as the LTERN provides sound scientific information for policy and management decisions that protect our natural resources and that prevent environmental damage requiring enormously expensive steps to clean up and restore. The LTERN, as described in this report, presents an opportunity for the National Science Foundation, in concert with the scientific community, to meet the sizable and continuing challenge of producing sufficient ecological understanding for sustaining human societies and for providing responsible stewardship of the biosphere.

TEN-YEAR REVIEW OF THE NATIONAL SCIENCE FOUNDATION'S

LONG-TERM ECOLOGICAL RESEARCH PROGRAM

The Long-Term Ecological Research (LTER) program, created in 1980 by the National Science Foundation (NSF) with five initial sites, currently consists of a network of eighteen sites (Appendix A) located across North America and in Antarctica. In 1993, the Biological Sciences Directorate of NSF appointed a ten-member committee (Appendix B) to evaluate the LTER program over the ten-year period of its existence. The committee was charged with appraising the past performance of the LTER program, including activities both at the sites and of the entire network. In addition, the committee was requested to identify those aspects of the current system that need to change as the LTER program continues to evolve and to provide recommendations for the future of the LTER program.

The committee sought information, ideas, and suggestions from a wide variety of sources in evaluating the past performance, current status, and future directions of the LTER program. At the first meeting of the committee, which took place in Washington, D. C., on February 3-4,1993, a questionnaire (Appendix C) was developed to enable each LTER site, and the Network Office, to respond to several key questions. To enrich the views available for evaluation, the committee requested that many participants at each site and the Network Office assist in completing these questionnaires. Individual questions were designed to solicit information and opinions in four broad areas: (1) the effectiveness of the current program for implementing LTER objectives, (2) products and outcomes of the program in terms of scientific insights and impacts on society, (3) future directions the LTER program should take, and (4) any other information the sites chose to include that was not addressed specifically in the questionnaire.

During the review process, committee members discussed the LTER programs with members of the scientific community. Also, to ensure that the program was well understood by the committee, each member was asked to become especially knowledgeable about particular LTER sites by reviewing NSF-provided materials and by visiting the sites if necessary. Representatives from each site and the Network Office were then invited to attend the second meeting held March 16-18, 1993, at Miami University in Oxford, Ohio.

As a major part of the agenda for the meeting in March, each site representative was

asked to make a brief presentation responding to the following three questions:

In addition, several panel discussions were conducted which focused on the present status and future directions for the LTER program and on the relationship of the LTER program and its research to societal issues such as biodiversity, global change, and sustainable ecological systems. Dr. James Edwards represented the National Science Foundation and participated during the second half of the meeting.

In its evaluation of the LTER program, the review committee incorporated information provided by discussions with members of the scientific community, visits to selected LTER sites, literature relating to the LTER program (Appendix D), the questionnaires (Appendix C), and the ideas presented during the March meeting. The following report considers issues at the site and network levels from three interrelated standpoints. Part one focuses on the past performance, successes, and shortcomings of the current LTER program; part two suggests short-term steps that might be taken to improve the LTER program given that it continues to operate with some increase in the level of funding and size in the future; and part three provides recommendations for a more ambitious network of sites which builds on current LTER strengths but also capitalizes on new and emerging technologies and expanded research opportunities and national needs.

The Long-Term Ecological Research (LTER) program sponsored by the National Science Foundation has been remarkably successful. From its beginning with just five sites in 1980, the program has grown to include eighteen sites (over the course of the program, two sites have been discontinued) representing a broad array of ecosystems and research emphases. The current network includes participation of more than 200 research scientists and about 300 technical personnel and students.

The mission of the LTER program is to conduct and nurture ecological research by:

• · understanding general ecological phenomena which occur over time periods of a few to many years and over a broad range of spatial scales;
• · creating a legacy of well-designed and documented long-term experiments and observations for use by future generations;
• · conducting major synthesis and theoretical studies; and
• · providing information for the identification and solution of societal problems.

The LTER Network Office, located at the University of Washington in Seattle, has the

responsibility for:

• facilitating communication and data sharing among the LTER sites and between the LTER program and other scientific communities;
• supporting the planning and conduct of collaborative research efforts, including provision of some technical support services;
• leading some intersite scientific activities;
• providing a focal point and collective representation of the LTER program in its external relationships; and
• developing linkages with other relevant long-term research programs, site networks, and science and technology centers.

The Network Office provides direct access to electronic mail on all major networks, online data bases, information bulletin boards, and Internet. An annual personnel directory, a site directory, and a semiannual newsletter as well as specialized reports are also published by the Network Office. The LTER Remote Sensing and GIS Laboratory assists in the acquisition, archiving, and analysis of satellite and GIS data for the network of sites. Among the important products at the Network Office is the Core Dataset Catalog, which provides both LTER and non-LTER scientists with information about data and studies across the entire LTER program.

Although the LTER system was originally conceived as a loose confederation of sites united in their focus on long-term studies, the potential value of coordination of certain data gathering and monitoring activities has become very clear as the program has developed over the past decade. With the establishment of the Network Office in 1988, LTER scientists agreed on protocols for standardization, communication, and analysis of certain basic meteorological data sets referred to as the "minimum standard installation" (MSI). MSI elements include the common hardware and software elements necessary for the computer networking system that links all eighteen sites as well as standardization of other computer applications such as geographic information systems.

During the ten-year period in which the program has been fully operational, scientists involved in the LTER program have produced a distinguished and extensive literature describing important ecological processes and numerous ecological characteristics from the sites. These long-term research programs have documented the status and trends of key environmental conditions from the eighteen sites and have established interactions with several foreign countries. In a number of instances, research from the LTER program has been instrumental in guiding policies and providing a scientifically credible basis for prudent management decisions about natural resources.

The LTER network generates site-based, long-term observations, providing a baseline against which changes can be detected. Such baseline studies will be increasingly important, especially as the scientific community addresses the structure and function of the global system and its interactive physical, chemical, and biological processes in order to understand and predict global change. These comprehensive baseline and process studies have greatly increased the value of long-term ecological research sites. The understanding gained from ecosystem-based studies is one of the essential building blocks for assessments of the contributions of specific ecosystems to global change as well as the effects of global change on ecosystems. Likewise, long-term studies are making significant contributions to our understanding of the importance of biological diversity to ecosystem structure and function, the role of ecosystems in transforming and transporting pollutants, and the effect of human population growth on land cover and use.

The records of atmospheric carbon dioxide concentrations from Mauna Loa, started in 1958, have provided tremendous insight not only into trends over time but also into the role of biological systems in its regulation. The long-term records of soil chemical properties and agricultural yield from Rothamsted Experimental Station in the United Kingdom have provided invaluable information on soil organic dynamics over 150 years of study. In similar ways, the long-term ecological research at these specific LTER sites will prove invaluable, and the importance of these observations will increase rapidly over time.

In the 1960s, several ecosystem projects were initiated which addressed the dynamic behavior of ecosystems and provided the opportunity to study process controls and detailed functions of these ecosystems. The Hubbard Brook Experimental Forest site in New Hampshire is one example where manipulative experiments begun over 30 years ago continue to contribute substantially to the understanding of ecosystem structure and function and to the relationship between ecosystems and human activities. The scientific publication rate from this site has increased rather than decreased over time, showing the power of established experimental sites with long-term commitments. Long-term ecological research will continue to provide an in-depth understanding of the behavior of ecosystems, thereby guiding our use of renewable resources provided by ecosystems and allowing us to assess how our use of non-renewable resources affects ecosystems.

The NSF vision for LTER has proven to be an overall success not only in this country but elsewhere. It has provided internationally recognized leadership in developing long-term ecological studies, and other countries (e.g., the United Kingdom's Environmental Change Network [ECN] and the Chinese Ecological Research Network [CERN] of the Chinese Academy of Sciences) now use the LTER program as a model. Despite these successes, however, the future contributions that the LTER program makes nationally and internationally will depend in part on how it meets the challenges described in this report.

In its original call for LTER proposals, NSF stressed that "environmental biological research may require more time than allowed by the usual two-to-three-year grant period," and support was made available for long-term ecological research projects. This was a visionary approach taken by NSF that recognized an important set of conditions in ecological research. Many ecological processes are slow and require long-term studies to elucidate their rates, environmental controls, natural fluctuations, and long-term trends. Thus, the LTER program was designed to match the research funding process with these dynamics of ecological processes.

As the LTER program has matured, its value as a network of sites has begun to increase. This constellation of eighteen sites represents different types of ecosystems and habitats, each with distinctive ecological conditions. With this diversity of ecological conditions, the network has successfully begun to facilitate the conduct of comparative studies between and among sites. The power of this network approach rests in the ability to compare similar processes (for example, primary production or decomposition of organic matter) under different ecological conditions. As a result, LTER scientists have been able to understand how fundamental ecological processes operate at different speeds and in different ways under different environmental conditions.

An example of the power of comparative approaches is provided by studies of landform effects on ecosystem pattern and processes involving four separate LTER sites, the H.J. Andrews Experimental Forest, the Niwot Ridge, the Central Plains Experimental Range, and the North Temperate Lakes. Landforms define the spatial context of specific locations within ecosystems and represent the physical "stage" on which biotic processes are played. Comparisons of forest, lake, alpine, stream, and grassland ecosystems have identified landform controls on energy, nutrient, and sediment budgets as well as controls on disturbance regimes. Because landforms and landscapes are easier to observe than many ecosystem processes, an understanding of these connections provides great power for predicting ecosystem behavior. For example, the position of lakes and streams in drainage basins greatly affects many aspects of water chemistry and biology. LTER research has demonstrated that sensitivity of lakes and streams to acid deposition is highly correlated with landscape position. In terrestrial ecosystems, landforms affect the behavior of disturbances as well as the patterns of post-disturbance ecosystem recovery. The LTER program is now extending these landscape studies to include fifteen LTER sites and, as a consequence, will further increase the power of the predictions between land form and ecosystem behavior.

In addition to the long-term nature of these comparative studies, LTER research has provided a critical context and perspective for intensive short-term studies. The decades-long environmental records at such sites as the Hubbard Brook Experimental Forest, Coweeta Hydrological Laboratory, and the H.J. Andrews Experimental Forest provide ample proof of the importance of sometimes rare episodic climatologic events on ecosystem processes and energetic and elemental budgets. Studies at the Cedar Creek LTER demonstrate that biological diversity is often most critical to ecosystem function during extreme events such as droughts (see Box 1). The occurrence of Hurricane Hugo provided the Luquillo Experimental Forest LTER with an opportunity to understand the ecological impacts of major natural disturbances on Caribbean tropical forests. Studies such as this would have virtually no meaning if it were not for the long-term records and coordinated research that are the hallmarks of the LTER program.

In addition to research funded through the LTER program, LTER sites and research have served as a platform for the development of a wide range of research programs supported by other funding sources. These collaborative programs enhance the value of LTER research and the importance of the LTER sites themselves, e.g., the extensive collections of arthropods from the H.J. Andrews Experimental Forest LTER provided the basis for the development of the Western Forest Insect Biodiversity Center at Oregon State University (see Box 2).

In numerous instances, research from the LTER program has been instrumental in guiding policies and providing a basis for prudent management decisions about natural resources. For example, research at the Central Plains Experimental Range and Konza Prairie LTERs has focused on the potential impacts of livestock grazing and fire on the long-term productivity and diversity of public lands (see Box 3) and, as a result, has provided answers to questions about how grasslands should be managed. Population, community, and ecosystem level studies at the Coweeta Hydrological Laboratory, Hubbard Brook Experimental Forest, and H.J. Andrews Experimental Forest LTERs have greatly influenced forestry practices nationwide and have provided the scientific basis for detailed planning and management guidelines. In fact, much of the information base for the U.S. Forest Service's "New Perspectives in Forestry" originated from these three LTER sites. Moreover, studies from these and other LTER sites have been crucial in developing the more comprehensive "ecosystem management" approach now being adopted by all the land-management federal agencies.

During the past decade, anthropogenic impacts on our global environment have become a matter of great concern. Although such issues were not part of the original LTER objectives, LTER research has proven extremely valuable in the study of global processes and the interrelationships between these processes and human activities (see Box 4). Some data for examining changes in ecological systems within the U.S. Global Change Research Program has been collected on the LTER sites. For example, the large NASA-FIFE program designed to examine energy and gas exchange between the biosphere and atmosphere using satellite data was conducted at the Konza Prairie LTER, using the spatial and temporal ecological data bases provided by the LTER program for validation. The Hubbard Brook Experimental Forest LTER site was the source of much of the data and many of the ideas that led to our current understanding of acid precipitation.

To increase the importance of long-term studies, research at individual LTER sites has now begun to expand to larger spatial domains. Such "regionalization and globalization" expansions have become central features at several LTER sites. At the Central Plains Experimental Range, LTER data gathered on the scale of tens to hundreds of meters have been used to develop models of the relationships between ecosystem production and nutrient cycling with gradients of land-use topography, soil characteristics, and moisture. Using a combination of remotely sensed data, soil survey data, climatological data, and topography layered into a geographic information system, these models have been extended to predict patterns of ecosystem response over a four-state region. At the Arctic Lakes and Tundra and the North Temperate Lakes LTER sites, models are being developed to facilitate regional and global extrapolation of trophic interactions and nutrient spiraling in rivers and lakes as well as the importance of these ecosystems in boreal and temperate carbon budgets. Intensive studies at Palmer Station, Antarctica, to understand the importance of spatial and temporal patterns of pack ice development on the dynamics of pelagic marine ecosystems are coupled to a basic 200 x 900 km sampling grid with sample points spaced at 20-km intervals. This combination of intensive and extensive studies will permit generalization of results for the entire margin of the Antarctic continent.

The LTER network has provided valuable educational resources across the curriculum and experiences for many students. At the Harvard Forest LTER, for example, more than 2,000 children, representing elementary schools and clubs of all kinds, view educational slide shows and hike along nature trails at the research site. Field demonstrations, exhibits, and audio-visual presentations of LTER projects are an integral part of this program, which emphasizes the importance of long-term research and scientific understanding of ecosystems. At the undergraduate college level, approximately 25 students work at the Harvard Forest each summer on LTER projects. Many of these activities lead to senior thesis projects and decisions to pursue careers in ecology. Also, Ph.D. students from such varied programs as biology, earth and planetary science, and landscape architecture utilize the Harvard Forest for thesis research. Currently there are eleven students pursuing doctoral research on LTER topics because the accumulated data base and long-term mission of the LTER sites are ideal for doctoral studies. Although these examples are from the Harvard Forest, similar programs are conducted at most LTER sites.

In summary, the LTER program has been successful in bringing together highly competent groups of scientists with broad backgrounds to focus on research that would not have been possible under more conventional research programs. There has been a demonstration of the potential value of a network of sites representing a variety of ecosystem types and dedicated to long-term studies. The program has made major contributions to our understanding of important ecosystem management issues nationally and worldwide. The LTER network has also succeeded in educating many students who have become trained in a multidisciplinary, site-oriented environment with interaction and collaboration among many disciplines.

NSF Level: The NSF official description of the LTER system is contained in the four Calls for Proposals (proposal deadlines in 1980, 1981, 1986, and 1988) and has in principle remained constant over time.

The goals of the LTER network were originally to "(1) initiate the collection of comparative data at a network of sites representing major biotic regions of North America and (2) evaluate the scientific, technical, and managerial problems associated with such long-term comparative research" (Call for Proposals 1980, 1981). In the 1986 and 1988 Call for Proposals, these goals were changed to: "(l) long-term analysis of site-specific ecological phenomena, (2) comparison of observations across diverse ecosystems and in terms of general systems theory, and (3) provision of large, secure, ecologically diverse sites with well developed support capabilities." These announcements were intended to provide flexibility in the proposal solicitation process since this was not a directed research program. The first two goals from 1986 and 1988 can be used as guidance in a general way, while the third can be described more accurately as a criterion for proposal review than a goal for the LTER program. Many of the programmatic and administrative recommendations in this report are designed to provide advice on the relative importance of (1) site-specific, long-term studies, (2) cross-site comparisons, and (3) furthering ecological theory through in-depth studies. In some instances, reviewers have emphasized one or more of these components, which has resulted in inconsistent recommendations from the LTER panel reviewing proposals, site review committees for individual sites, and ad hoc proposal reviewers.

Site Level: Although the three goals stated by NSF are interconnected, a more proper evaluation at the site level can initially be made by considering them separately.

Many of the sites have initiated important observational and experimental studies that could not have been performed without the assurance of long-term support. Long-term studies depend not only on long-term funding but also on a stable infrastructure, secured site management, and a base-line data collection system. It is clear that long-term analysis of ecological processes has been the guiding theme for the research teams submitting proposals for LTER funding. Potential LTER scientists have been offered opportunities for conducting long-term research, and they have been eager to accept the challenges. The first goal stated by NSF in the Calls for Proposals has thus been met by the proposals and funded projects at the LTER sites.

Competing pressures to test new and challenging hypotheses while also providing consistent long-term science are cited as problems by investigators at nearly all LTER sites. At current levels of funding, these pressures cannot be reconciled because there are insufficient funds for accomplishing both responsibilities. Also, there is the expectation that the research program at each site will grow and evolve, but at the same time, sites are expected to continue research programs presented in the initial proposal. Long-term studies will inevitably be based on the scientific questions or ideas identified when the work began, but such research questions and ideas are likely to change or be modified over the life of a ten-twenty year project, although these long time periods may be necessary to answer the original questions. Thus, the current funding levels are inadequate for research expectations for both new ideas and for the continuity required to answer the questions posed earlier in the beginning of the LTER studies. However, the concept of the LTER program was never intended to support all the studies on a given site. Rather, the LTER program was developed to provide an infrastructure and to fund selected investigations of key ecological conditions and processes that are long-term in nature.

During the first ten years, the LTER program has not been as successful in advancing ecological knowledge through cross-site comparisons as compared to site-specific research. Nevertheless, several initiatives for cross-site comparisons have been developed by individual investigators as well as the LTER Network Office (e.g., comparative decomposition rates, landform control of ecological processes). The increased attention by NSF toward cross-site comparisons has encouraged and facilitated this development of a stronger intersite comparison component of the LTER program. Reviewers of the LTER program, however, have not always placed consistent emphasis or value on cross-site comparisons. As a result, mixed messages have been received by the LTER scientists when the programs are evaluated and when proposals are reviewed. In the future, it will need to be clear that in addition to site-specific scientific questions, proposals may also address critical questions and hypotheses that can best be addressed through multi-site LTER comparisons.

In its original 1980 Call for Proposals, representation of "major biotic regions of North America" was a stated LTER priority. This explicit specification was dropped from the subsequent 1981 Call for Proposals. Little guidance was provided relative to the meaning of comparisons "across diverse ecosystems," and the selection of proposals was made based on site-specific scientific criteria and the quality of the proposal, without any major attention toward representation of major ecosystem types. In the third Call for Proposals in 1986, NSF expressed special interest in proposals including the following ecosystem types: (1) high latitude (i.e., Arctic or subarctic), (2) tropical moist forest, and (3) agroecosystems (especially those combining conventional and innovative practices). In the latest Call for Proposals in 1988, emphasis was given to (1) tropical forest ecosystems, and (2) land margin ecosystems representing the interface of continental land masses and coastal oceans. With the exception of these suggestions, NSF has provided limited direction to the type of sites to be selected, leaving the selection process largely to the decision of the scientific community. Future site selection should retain an emphasis on the quality of science; but now that much of the network is in place, increased attention should be given to potential sites representing important gaps in the program.

Although the rationale for site selection has not always been entirely clear, the LTER sites successfully represent a diverse array of ecosystems. It is also necessary to emphasize the value of including non-LTER sites in comparative studies, and this has occurred in some instances. The power of spatial resolution of ecological phenomena would be severely limited if the current LTER network were considered to contain all the sites necessary for insightful cross-ecosystem comparisons.

The range of size of LTER sites is large (100 to 100,000 ha). However, even the smallest sites appear to be sufficiently large to address the long-term questions being addressed by the LTER program studies.

Security of the sites has been achieved largely by establishing them on federal land (approximately 75% of the sites are on federal land), and all sites seem to be relatively secure from the perspective of the land being available for a long-term commitment of resources and research investment. LTER funding has in many instances provided more secure support for the site and its research program than was available from the institution responsible for managing the site. However, in several instances, the institution has reallocated the administrative overhead reimbursement of the LTER grant back to the site and thus provided further support for long-term stability.

In addition to attracting considerable support from the host institution, the LTER projects have been successful in establishing groups of scientists who have a corporate responsibility for site management and scientific guidance. The sites have thus avoided the danger of relying on one Principal Investigator whose departure could jeopardize the strength and continuation of the site. Thus, NSF and the LTER program have been successful in providing "secure" sites, and over time the value of these sites and their associated data will markedly increase to potential users; this will also increase the security of the sites because of the large user constituency.

The goal to have well-developed support capabilities has in general been successful. Most sites have adequate support, partly provided by the NSF LTER grants and partly from other sources. However, investigators from many sites have expressed concern that site management now demands an increasing share of NSF funding, and the sites have been successful in meeting this need only by attracting considerable external funding (frequently external support is six to seven times greater than the basic LTER support). The demands on limited funds for site support and for research projects will become an increasing problem. Successful renewal proposals to NSF must include new and exciting science, but at the current funding level, much of the support must be used for maintaining the site. It should also be noted that, although sites have different costs for general maintenance (e.g., plot management or travel costs to) from the site), with the exception of the sites in Alaska and Puerto Rico, equal funding is provided to all sites within one cohort of awards.

Network level: The Network Office has been important in fostering cross-site comparisons. A number of strategies have been used, for example, including calling separate meetings to address specific issues, developing an efficient communication network, implementing a minimum standard computer configuration at all sites, and organizing all-scientists meetings. The Network Office and these activities are essential in order to achieve the goal of cross-site comparisons. A central coordinating function within the LTER program is required to achieve the value added through cross-system comparisons. Such cross-system comparisons will increase the scientific value of the program as a whole and of the site-specific studies and will contribute to the development of ecosystem theory.

In addition to the aforementioned goals, the NSF Calls for Proposals have consistently

stated that the research at the sites should address five core research topics:

1. pattern and control of primary production;
2. spatial and temporal distribution of populations selected to represent trophic structure;
3. pattern and control of organic matter accumulation in surface layers and sediments;
4. pattern of inorganic inputs and movements of nutrients through soils, groundwater, and surface waters; and
5. pattern and frequency of disturbance to the research site.

The core areas were originally designed to promote some degree of topical consistency, while at the same time allowing flexibility and choice at the project level. These core topics have provided guidelines to ensure that sites are selected for broad-based ecological research rather than narrowly focused studies. The five core areas are defined broadly, and no standard methodology is stipulated. Thus, these core areas have not provided strict operational assistance in site selection or evaluation.

No single site investigates all five topics in depth, but all of the sites devote attention to some important aspects of all five topics. The breadth of ecological studies conducted at a site has been maintained, and the research groups have not been constrained by undue pressure regarding how the balance is divided among specific core topics.

At the organizational level of the network, the five research topics could have been important for developing a minimum data set. However, standardization of measurements and agreement of a minimum data set have not been fully developed, and there are legitimate questions about just how such an effort should be addressed. A minimum meteorological data set exists, but otherwise the collection of data is not standardized. Although strict standardization of methodology is difficult and perhaps undesirable to achieve, the possibility for cross-site comparisons addressing the five topics has been limited by the lack of comparable data sets.

Network Governance

The LTER program has developed a Coordinating Committee that represents each of the sites. There is an Executive Committee that conducts the more routine activities of the program and works closely with the Network Office and its Principal Investigator, who also chairs the Executive Committee. The Network Office and the Coordinating Committee interact with NSF concerning general program directions. All LTER sites are led by one or more Principal Investigator(s), frequently with an advisory committee composed of investigators at the site. Each site also has an external advisory committee that meets occasionally with the site investigators and provides a review of the research.

As described previously in the report, the Network Office assists in the achievement of the overall goals of the LTER program, provides leadership, and coordinates and facilitates communication within the program and with organizations outside the LTER program. The assessment from the sites of the role of the Network Office has varied from enthusiastic to critical. Part of the necessary support capabilities mentioned in the third goal of the LTER program is provided by the Network Office. For some functions, such as electronic networking, the Network Office has performed an essential function as these systems have been developed, but the need for a continued high level of involvement in this network is not clear. The program should decide whether the Network Office should continue as a central manager or acquirer of remotely sensed data.

The Foundation has put emphasis on the need for appropriate data management at both site and network levels. Many sites have effectively addressed this issue, but significant financial resources are needed to develop and maintain an effective site data management protocol. The participating scientists must recognize the value of a site data base and the necessity of timely inclusion of all relevant data in a manner that makes it accessible to scientists at the site, in the network, and outside of LTER. The Network Office has proven effective in providing a focal point for site data managers to consider data issues and develop common approaches to data problems.

It is also important that the Network Office represents the view and values of the LTER network to the outside world and encourages external links. Such outreach activities should involve the collective abilities and contributions of the entire LTER program scientific community.

The present LTER review process has been fair, although as noted previously, some of the directives to panels and review teams have been less than precise. Because some reviewers are frequently accustomed to evaluating hypothesis-driven, single-investigator proposals, they may tend to emphasize short-term rather than long-term perspectives and disciplinary rather than transdisciplinary strengths of proposals and LTER research programs. Further, a few reviewers tend to focus on individual site achievements rather than on contributions of the sites to the overall LTER network. Thus, there is a need for better, written briefing of reviewers on the unique nature of LTER expectations and on the appropriate criteria for proposal evaluation.

Continuity of research expectations is of particular importance in the evaluation of long-term research; as a consequence, each evaluation must draw on knowledge of the original hypothesis that motivated the initial collection of long-term data sets. Although the LTER program will adjust the specifics of its program as research issues evolve, the review process must not apply shifting demands on a site without due consideration of the site history and of the long-term research goals of the site studies. The necessary institutional memory can be instilled in the review process by overlapping memberships in review panels and by briefing new reviewers with full site histories and providing copies of previous reviews. As discussed previously, one unique characteristic of the LTER program is the requirement that each site conduct long-term research activities in all five core areas. Unfortunately, in the proposal review process, the required focus on core areas is often lost on reviewers, so better briefing of the evaluators is necessary.

In many cases, the reviewers best acquainted with the unique objectives and constraints of the LTER program are the LTER scientists themselves. To reduce the possibility of conflicts of interest, LTER scientists are sometimes appointed to review panels but never constitute a majority. This practice is entirely appropriate.

A procedure exists for giving currently funded sites that do not meet NSF standards during the renewal process a probationary trial period to address the Foundation's concerns. Some sites have been successful in obtaining continued funding after such a trial period whereas others have not, but the criteria used for discontinuing sites are not clear to all program participants. It is critical that this process be well explained and that the sites which continue to be funded have a clear understanding of what is expected from them.

The number of sites actually funded has increased from five in 1980 to eighteen in 1993 (Table 1). The current average grant per site is slightly less than $500,000. However, the total funding for LTER (primarily from the NSF Division of Environmental Biology, but, since 1990, also from the Division of Polar Programs) has grown more rapidly due to increased allocation to network coordination and other supplementary grants.