IHDP. Environmental data must be integrated with social datasets already archived by organizations such as the Institute for Cooperative Programs in Survey Research or by researchers who maintain their own distributional mechanisms. Improved access to data via the World Wide Web, along with advances in software and metadata standards, have greatly improved the ability of researchers to search for specific types of data and then manipulate or download them.

Although considerable progress has been made in understanding the human dimensions of global environmental change, there are still many unresolved questions and several important new areas for research. In the committee's review of progress, we identified many areas where knowledge was lacking or research results were inadequate. In this section we attempt to crystallize a research agenda of high-priority questions that might yield valuable information in the next 5 to 10 years, given sufficient attention and resources. These research imperatives have emerged from recent meetings and reports of National Research Council (NRC) committees with an interest in human dimensions research,¹⁵⁰ from a review of other national and international efforts to identify research priorities, and from consideration of the significant intellectual gaps and opportunities identified in the review above.

Some of these research imperatives directly support particular themes of the USGCRP, such as atmospheric chemistry and seasonal to interannual climate prediction (see Table 7.4). Some focus on particular elements of the traditional

framework of causes, consequences, and responses used in human dimensions research (see Table 7.5). Others cut across several of these themes, develop understanding of fundamental social processes that affect human-environment interactions, or suggest broadening of the overall USGCRP agenda.

Previous research has identified changes in the use of land, energy, and materials as priority subjects in understanding the causes of global change. Although the driving forces for the use of these resources include population growth and technological change, in many regions the most important determinant of environmental impacts is the per capita consumption of energy and materials. Debates over the relative roles of "northern" consumption and "southern" population growth, and over the responsibility of different social groups within countries, have confounded international environmental negotiations and domestic policy development. A recent report¹⁵¹ identifies the study of environmentally significant consumption as an important area for research, a point that is echoed by the recent joint statement on consumption of the National Academy of Sciences and the Royal Society of London, as well as by new research initiatives of the Organization for Economic Cooperation and Development and the European Community.

Consumption, that is, the human transformation of energy and materials, is environmentally significant "to the extent that it makes materials or energy less available for future use, moves a biophysical system toward a different state, or,

through its effect on those systems, threatens human health, welfare, or other things people value.''¹⁵² Environmentally significant consumption and direct alterations of biological systems are the two main ways in which humanity affects the global environment.

Currently, the most environmentally significant consumption from a global perspective consists of the major activities that burn fossil fuels (e.g., travel, space heating and cooling, electricity production, industrial process heating) and activities that use chlorofluorocarbons, nitrogen, and certain other materials responsible for stratospheric ozone depletion, pollution of ecosystems, and other global environmental changes. Other activities (wood and water use, meat and fish consumption, toxic chemicals and waste disposal) are considered of greater environmental significance at local levels and by some groups. New information about biophysical processes can improve understanding of the relative environmental importance of consumption activities.

Two trends in consumption are of the greatest importance to global change. One is the rapid growth of consumption associated with the emergence of a global middle class—growing segments of populations, particularly in developing countries, that are able to afford the consumptive amenities of the developed world, such as motor vehicles, refrigeration of food and living space, and air travel. The other, potentially countervailing, trend is toward decreased consumption per unit affluence, particularly in wealthy countries, probably brought about by technological improvements, saturation of demand for some amenities, the increasing effectiveness of environmental movements, and shifts in the structure of economies. At the global level, the central question about consumption is whether the second trend can counteract the first before consumption causes unacceptable environmental changes. Below the global level, however, the question looks quite different from the vantage points of high-income and low-income countries and populations, which differ greatly in how much benefit they perceive from further increases in consumption.

The social determinants of environmentally significant consumption¹⁵³ include changes in human populations, development and diffusion of consumptive technologies and behavior patterns, economic resources available to households and firms, prices of fuels and equipment, human values and preferences, availability and use of information, structural change in economies, and public policies. One of the overall challenges is to understand the links between individual demand for goods and services and the ultimate environmental impacts of meeting those demands. Particular human needs and wants may be satisfied by a variety of products and processes that cause different types and magnitudes of environmental change. Understanding the connections can help in finding ways to decrease the environmental impact of meeting human needs.

Over the next decade, research on environmentally significant consumption should address major questions such as the following:

• What are the constituents and determinants of energy use and other environmentally significant consumption in countries and populations at different levels of economic development?
• How is consumption likely to change with increasing affluence in low-income countries and populations and does this change always follow the path that high-income countries and populations have followed?
• What social forces drive the most environmentally significant consumption types, such as travel, the diffusion of electrical appliances, agricultural intensification, water use, and purchases of high-energy-consuming vehicles?
• What are the relative roles of various determinants of consumption in different countries?
• What policies at the national level lead to greater attention in communities to such issues as urban sprawl, reducing the cost of home-to-work commuting, expansion of green spaces, and enhanced recycling of materials?
• Which materials transformations have the greatest environmental significance and what determines related kinds of consumption?
• What interventions can effectively alter the course of the most environmentally significant kinds of consumption?
• What determines public support for effective consumption policies and how do these factors vary across countries?

This research will involve analysis of disaggregate data on particular consumption types in relation to prices, policies, and physical infrastructure within countries; surveys of consumption behavior and related values and beliefs in households and firms; data comparisons across consumption types and countries; and experiments with interventions. It will also attempt to understand how culture, fashion, advertising, and various kinds of opportunities and constraints influence consumption and will investigate the ways in which economic and cultural globalization and corporate and government decisions increase, limit, or expand individual consumer choices.

Over the next 5 to 10 years this research can meet several critical goals:

• Improve understanding of the constituents, determinants, and time paths of energy use in developing countries, thus improving the ability to model and anticipate anthropogenic carbon emissions, use of biomass fuels, and emissions of local and regional air pollutants.
• Improve projections of emissions and pollution in high-income countries, along with understanding of the key behaviors driving those emissions.
• Improve understanding of how changes in water, food, and wood consumption influence land use and vulnerability to climate variability and change.

• Promote more realistic analyses of the policy options for achieving national and international targets (e.g., limitation of greenhouse gas emissions, control of regional air quality, more efficient use of water resources), taking into account knowledge about public acceptability and the requirements of successful implementation.

This research priority will develop basic understanding of consumption and provide insight about its causes, dynamics, and trajectories that will be essential for making informed decisions. It will also provide important support for scientific research in atmospheric chemistry and decadal to centennial climate change, in understanding how water demands create vulnerability to climate change and variation and how ecosystem pollution and food and materials consumption are drivers of land use change.

The rate of technological change is one of the most significant sources of uncertainty in climate models as well as in understanding future uses of land, ecosystems, and water. Moreover, development and adoption of new technologies together constitute one of the most important methods available for achieving national and international commitments to environmental protection and sustainability.

Improving emissions scenarios for greenhouse gases and other globally significant pollutants such as SO₂ and hydrogenated chlorofluorocarbons requires not only more accurate demographic trajectories but also sectoral studies of changes in consumption and technology, particularly in developing countries. The implementation and enforcement of international treaties to control ozone depletion and greenhouse gas emissions and the projection of future changes in atmospheric chemistry all will require a much more realistic and geographically disaggregated assessment of land use, technology substitution, consumer preferences, incentives, and trade than has been undertaken to date. This effort requires research ranging from studies of the industrial ecology of individual sectors, firms, and farms to analyses of corporate strategy, trading relations, and responses to regulations.

Over the next decade, research on technological change and the environment should address such questions as these:

• What factors determine variations among sectors and actors and change over time in the approximately 1 percent per year decrease in national energy intensity generally attributed to technological change?
• What factors determine average rates and variations around the average in the adoption of new production technologies that reduce inputs of energy and virgin materials per unit output?

• What factors determine the rate at which production costs of an environmentally benign technology decline as output increases?
• What have been the effects of prescriptive standards, best-available-technology rules, public recognition, and awards to encourage voluntary technology adoption and other technology-related environmental policy instruments on actual rates of innovation?

This research on technological change will involve econometric modeling of change in energy and materials productivity and case studies of technology adoption in firms and industries, changes in production processes, and responses to regulation and incentives, and case comparisons. Over the next 5 to 10 years this research can achieve the following:

• Account for a significant proportion of variance in ''autonomous" energy efficiency improvements, thus enabling more accurate modeling of improvements in energy intensity and identifying factors responsible for rapid efficiency improvements in some sectors and by certain actors.
• Identify characteristics of technology-related policy instruments associated with more rapid innovation by industries.
• Identify important sources of variation in technology adoption across firms within industries.
• Identify some of the causes of "learning" in production processes.

Research on technological change, particularly research documenting rapid adoption of environmentally beneficial technology, will suggest effective policy options for encouraging beneficial technological change. It is also critical to improving the modeling and anticipation of the climatic impacts of greenhouse gases, to understanding regional changes in atmospheric chemistry, and to examining the role of technology in human impacts on terrestrial and marine ecosystems.

One of the great challenges of global change research is to make scientific information, such as the results of climate modeling and analysis and studies of vulnerability and adaptation possibilities, more relevant to decision making at the local level. Regional assessments have been identified as a priority by the IPCC, the USGCRP, and the International Research Institute for Climate Prediction. Regional assessments can be developed for scenarios of global warming, decadal climate shifts, and seasonal forecasts and have the potential to address many issues of concern to local resource managers, corporations, and citizens. This research priority is also highly relevant to scientific efforts to provide more useful

seasonal to interannual climate predictions and to understand decadal to centennial shifts and changes in climate.

An important intellectual shift in climate impact assessment in recent years has been an increased focus on understanding vulnerability and adaptation. Understanding how global warming or ENSO will affect a local region is as much a question of understanding the social and economic characteristics of the region as it is of obtaining the appropriate results from a climate model. For example, drought impacts on crop production are mediated by access to adaptive technologies such as irrigation, fertilizer, and seeds, as well as crop prices and subsidies and coping mechanisms such as environmental information and insurance. Adaptive technologies and coping mechanisms vary considerably by region, sector, and social group. Thus, improved regional assessments require detailed studies of how vulnerability develops and can be reduced. Future models and economic analyses are likely to include vulnerability indicators and findings about processes that affect vulnerability.

Research priorities for understanding the impacts of long-term climate change should address some of the agendas established by the IPCC. These would include studies that take advantage of mesoscale model outputs and downscaling techniques to improve regional projections of climate impacts and detailed analysis and longer-term projections of changing vulnerability and adaptive strategies. As scientific understanding of other decadal shifts in atmospheric and ocean circulation improves, the research might also focus on climate interactions with the management of resources such as ocean fisheries, forests, and water and agricultural systems. For example, the dynamics of fisheries in the context of decadal climate shifts cannot be understood without an understanding of human pressures on marine resources.

In the area of seasonal to interannual climate prediction, an NRC panel has developed a research agenda to increase the social usefulness of such predictions.¹⁵⁴ As described in Chapter 5, there have been significant improvements in our ability to forecast climate 3 to 12 months in advance, especially in relation to changes in sea surface temperature and atmospheric circulation associated with ENSO. Because ENSO appears to be correlated with large impacts on agriculture, health, water resources, and ecosystems, this improved forecast capability has significant implications for people, especially when combined with information on vulnerability and adaptive responses.

The NRC panel considered such issues as measuring and monitoring the social impacts of climate variability, analyzing changes in vulnerability to climatic extremes and variations in vulnerability across social groups, and identifying opportunities and barriers for the beneficial use of seasonal forecasts, including improved understanding of interactions with markets and improved communication of uncertainties in the policy process and to forecast users. Research to make climate predictions more socially useful must be undertaken with close links to researchers and policy makers in affected regions and with frequent communica-

tion with those producing the predictions. As understanding and predictability extend to new regions, lead times, and levels of certainty, human dimensions research can provide important insights into local vulnerability and policy contexts as well as into human needs and responses to improved climate information.

Over the next decade, research to make climate predictions more useful at a regional level should focus on a number of questions, including the following:

• What are the sectoral impacts at the regional level of climate change and seasonal to interannual variations?
• Are there impact and vulnerability indicators that can be useful to detect the extent and severity of the impacts of global change on human populations?
• Can historical data be used to project future human vulnerabilities to climatic variation and change?
• How does climate change interact with other social and ecological changes to influence crop yields, water use, and other impacts?
• Can the mesoscale outputs of climate models be better linked to models predicting the regional impacts of climate change?
• How are the impacts of climate change and variability affected by the coping techniques available to vulnerable groups?
• When science can provide early warnings of possible catastrophes, how can this information be transformed effectively into public understanding and appropriate policy responses?

This research will include case studies of responses to past climate variations; quantitative analyses of the social and economic consequences of such variations, including adaptations and the distribution of impacts across regions and social groups; development and testing of vulnerability indicators against past data; building of models that project future vulnerability; development and linking of models; and analysis of responses to climate forecast information.

Over the next 5 to 10 years this research should be able to develop the following:

• Methods for linking mesoscale and other climate model outputs to models of regional water resources, agricultural production, energy needs, and health conditions.
• Assessments of the vulnerability of many regions to climate change and variation.
• Estimates of the potential regional impacts of future climate change and variability and the value of improved climate information, including seasonal forecasts.
• Improved methods for delivering climate forecast information.

Natural science research has identified and is evaluating several kinds of rapid and discontinuous environmental changes that might overwhelm human adaptive capacities, at least in some localities. They include rapid climate change events (as from a major disturbance of the North Atlantic Oscillation), major outbreaks of diseases in humans or key crop species, and rapid destruction of the reproductive capacity of key ecosystems resulting from chemical releases to the environment. The damage to society should such changes come to pass is obvious; it is less obvious how to deal best with the prospects of such changes. Societies must function and plan for the future in the face of continuing revelations from environmental science and high uncertainty about potential catastrophes. We must also deal with meta-uncertainty—not knowing how uncertain we are.

It is not just rapid environmental changes that may produce global change surprises. Social systems can also change rapidly and discontinuously in ways that may greatly alter environmental systems and human vulnerability to global change.¹⁵⁵ History provides a number of illustrations of such changes, including the environmental impact of European political decisions to colonize (including the rapid spread of diseases and land use changes) and the impact of regional and global warfare on resource consumption and ecosystems. More recent rapid political and economic changes also have major global change implications. For example, the collapse of the Soviet political bloc altered greenhouse gas emissions and land use, restructured trade and property rights, and altered political alliances. The environmental and social implications of such rapid change should be a research priority. Another form of dramatic social change has been the rapid spread of democratization and liberal economic policies in Latin America and Africa. These too alter human-environment relationships in unforeseen ways. As economic liberalization changes the terms of trade, land use and industrial production can change quickly, with pollution patterns shifting along with industrial relocation and deregulation and vulnerability to climatic extremes changing with the restructuring of agriculture and food systems. Democratization can transform public attitudes and open up decision making processes to popular movements, altering the policy process responsive to global change. One of the most challenging questions is to understand how these rapid social changes may interact with rapid environmental change.

Some of the key questions about surprises for the next decade are the following:

• What are the human consequences of rapid climate changes in the past and present?
• What are the global environmental change implications of rapid political and social changes in the past and present?

• How have environmental and social surprises interacted?
• Which human activities (e.g., patterns of land use and management, chemical releases) can significantly alter the potential for major environmental surprises?
• When science can provide early warnings of possible catastrophes, how can this information be transformed effectively into public understanding and appropriate policy responses?
• How can hazard management systems, including insurance strategies, subsidies, technological investment, and warning systems, be organized to increase resilience in the face of major surprises and at what cost?
• How can society deal with the possibility that citizens will become immobilized by warnings of possible, but highly improbable, environmental catastrophes?

Research on environmental surprises should include comparative studies of past environmental catastrophes and hazard management systems, simulations that superimpose plausible hazard events on existing hazard management systems, and experiments to test responses of individuals and organizations to information about possible environmental surprises.

Research in the next 5 to 10 years can achieve the following:

• Document and analyze the environmental implications of recent rapid social changes, such as post-Soviet restructuring, democratization, trade liberalization, and resource privatization.
• In collaborations between social scientists and natural scientists, significantly elaborate the human dimensions of credible but low-probability geophysical catastrophes, ecological collapses, and disease outbreaks.
• Examine the consequences of and responses to catastrophes in the historical and prehistoric records, to identify the characteristics of hazard management systems that have been associated with effective response in the past.
• Evaluate the immobilization hypothesis and, if it is a serious threat to response, suggest ways of presenting information about possible surprises that could overcome such tendencies.
• Develop some practical approaches through which those responsible for hazard management systems can consider the implications of catastrophe scenarios for those systems.

To make effective and well-informed decisions to anticipate the threat of global environmental change, society needs better understanding of how social institutions influence environmentally significant human actions. This need can

be seen from the following observations: international agreements set targets without much consideration to what is feasible; governments often set resource extraction limits at unsustainable levels; national policies often appear to local resource managers to be part of the problem; techniques for estimating the full social costs of natural resource consumption rarely result in either social consensus or policy decisions; institutional change often has unforeseen or unfair distributional impacts; and even when there is widespread agreement about a global change phenomenon among specialists, many people perceive a high level of scientific disagreement. Such difficulties afflict resource management institutions at levels from local to international.

Global and national institutions, which are at the same scale as the problems, must be better coordinated with local institutions, which are often at the same scale as the solutions. Decision makers need more information about how to achieve this coordination. They also need to develop institutional approaches for allocating environmental resources when market prices give incomplete or misleading information. Research on environmental management institutions advances our understanding of the causes, consequences, and responses to global change and should thus be given a high priority.

Over the next decade, research to meet these needs should address such questions as the following:

• What are the characteristics of effective institutions for managing global environmental change?
• What are the correlates of effectiveness for the management of international environmental and natural resources by international regimes and institutions? In particular, what are the conditions favoring effective implementation of commitments to protect biodiversity, forests, oceans, and stratospheric ozone and to prevent climate change?
• What are the implications, applicability, and limits of particular policy instruments, including market-based instruments and alterations in property rights institutions at international, national, and local levels?
• How do declaratory targets, consensus policies, and review processes interact to influence behavior and restructure the power relationships of states and nonstate actors?
• Which characteristics of national institutions are most conducive to sustainable resource management by local institutions?
• How can knowledge about the conditions for successful local resource management be applied to problems at national and international levels?

Research on these questions will demand a variety of methods, including systematic empirical study of existing regimes and institutions for managing global change issues at levels from the international to the local; conceptual studies of proposed institutional policy instruments, focusing on bargaining prob-

lems and links among international, national, and local levels; theoretical studies of the bargaining, contracting, and principal-agent aspects of implementing commitments at higher levels by delegating substantial authority to lower-level agents (e.g., tradable permit systems, joint implementation, federal systems); institutional and political study of the applicability of institutions at lower levels of organization to the design of national and international policy instruments; quasi-experimental studies, case comparisons, and simulation studies of the effects of major changes in institutions and rules, based in part on data from archival records and the recollections of participants; and small-scale simulations and experiments.

Over the next 5 to 10 years, research on these issues can be expected to lead to a number of achievements:

• Identification of conditions, potential contributions, and pitfalls associated with specific policy instruments, such as tradable permits, and with specific designs of environmental institutions.
• Development of a larger and more consistent body of data on international institutions and regimes and on regional and local property rights and other institutions with which to conduct comparative studies of their formation, evolution, and influence.
• Identification of conditions under which particular national policies assist or impede the efforts of local resource management institutions to sustain their resources and identification of insights from the experience of local resource management institutions that are transferable or adaptable to national and international institutions.
• Identification of the contributions of process-based international review mechanisms to changed behavior.

Considerable progress is already being made in understanding land use/land cover change and changes in human population processes. All land use is local, but the forces influencing the dynamics of land use and land cover come not only from individuals, households, and communities but also from processes at regional, national, and global levels. To understand land use and land cover change requires knowledge of how forces within and beyond the individual actor combine to affect decisions, particularly the conditions conducive to land use decisions that are either destructive or restorative to the environment. We do not yet fully understand how individual perceptions, attitudes, and socioeconomic situations affect land use choices or precisely how various external conditions, such as trade and international political economy, in addition to local rules for access to resources, insurance regulations, distance to markets, infrastructure development,

and other factors, interact in the calculus that people use in making decisions about resource use.

One of the current challenges in understanding land use change, as well as changes in the use of water, marine ecosystems, and other resources, is characterizing the role of population in environmental change and degradation. As noted, considerable progress has been made in understanding population-environment relationships and in explaining more basic population dynamics. There is some agreement that, in the future, migration, rather than changes in human fertility and mortality, will be the key demographic link between the two dynamic processes of land use and land cover change. Causation and feedback will probably move in both directions: environmental changes will likely cause migration, and migration will likely change the environment. Careful research into the relationships between population mobility and environmental change is also needed because of the growing popular concern with environmental refugees, the environmental impacts of immigration, and the role of population in environmental conflict and security.¹⁵⁶ There is very little empirical documentation of the relationships between migration and environment.

Population migrations in the United States, however, illustrate the process. There were large migrations into the Midwest from the middle of the nineteenth century until about 1920, after which time the population of the country became increasingly urban—until the past decade, that is, when, across the country, a growing number of households have moved outside of cities and either commute to work or work at home part of the time. This shift, if it continues, may have significant environmental consequences in terms of the consumption of fossil fuel and other resources and for land use and land cover.

To understand the interaction of migration patterns and land use/land cover change requires improved data and data analysis both on prior migrations and intended future migrations. Data are needed for individual and household levels, as well as for more aggregated levels. Data on migration and other social variables must be linked with biophysical data from remote and land-based sources on soils, climate, and other biophysical factors. The data must be collected and coded in such a manner that they can be geo-linked at spatial and temporal scales with resolutions appropriate for the theoretical issues addressed. The necessary temporal depth can be achieved through prospective and retrospective techniques. Retrospective approaches allow temporal depth in a cross-sectional survey; prospective approaches permit the inclusion of intentions and attitudes, which cannot be obtained retrospectively. The social science community is now in a position to collect and analyze the requisite migration data. In the past 15 years, considerable progress has been made in improving the quality of retrospective migration data by embedding its collection in a broader life history approach. More recently, advances have been made in collecting prospective migration data by incorporating the insights of social network analyses into the data collection

process. Methodological aspects of geo-links to biophysical data are currently being worked out.

The potential now exists to significantly improve understanding of the inter-relationships between human spatial movements and land use/land cover change. Assuming that sufficient migration data are collected to cover both points of origin and receiving areas, we can move beyond such simple statements as ''large migrations of individuals and families into a given area affect the land use/land cover patterns in that area." Such research accomplishments will involve disaggregating both migration streams and land use/land cover change patterns so that specific attributes of migrants can be related to specific aspects of land cover and land use. For example, does the migration of young adults or the elderly have a larger impact on patterns of forest regrowth in rural areas or on water use?

Research priorities for land use studies have been established internationally through the IHDP/IGBP core project on land use/land cover change.¹⁵⁷ The U.S. research community should maintain collaborative ties with IHDP/IGBP, and the USGCRP should work to ensure that this collaboration is maintained. Research over the next decade should address such questions as the following:

• What are the links among land use change, migration, political and economic changes, cultural factors, and household decision making?
• What are the interrelations between migration and environmental change?
• What comparative case studies of land use and land cover change are useful for understanding and modeling land use change at regional and global scales?

This research will include efforts to map land use and land cover and will document changes over time, develop and validate classifications of land use and land cover, develop algorithms for making the classifications accurately from remotely sensed data, undertake comparative and statistical analyses of past relationships between changes in social driving forces and land use and between land use and land cover, and develop and test regional and global models of land use and land cover change. Research progress will depend on remotely sensed data, which can provide key information on land cover and are needed at appropriate spatial and temporal resolutions.

To obtain data on past population migrations and other social driving forces, continued and improved access to earlier generations of remotely sensed data is imperative. Data collected for military and/or intelligence purposes need to be increasingly declassified and made available to the research community. This need applies to data sources of both the United States and the former Soviet Union. For future remote sensing instruments, fine-grain spatial resolution is critical, as is the ability to determine the height of buildings in urban areas. This research should take advantage of the development of enhanced, multilevel, multiscale, and comparative methods in the study of human communities across

the planet;¹⁵⁸ it can also make use of Earth-observing satellites that offer 1-to 3-m resolution and that facilitate observation, archiving, and analysis of human impacts at that scale. Success will depend on collaboration between social scientists and physical scientists in developing better algorithms for analyzing the large datasets provided by fine-resolution satellites to address behavioral questions.¹⁵⁹ Use of remotely sensed data at this fine scale will require attention to confidentiality in archiving and can benefit from past experience with social data.¹⁶⁰

Over the next 5 to 10 years research on land use issues can be expected to meet a number of goals:

• Development of datasets and comparative empirical studies on the social causes and consequences of land use and land cover change in different regions that will permit improved understanding of the relative roles of population dynamics, economics, and other factors in driving environmental change.
• An improved capability to include detailed land use and land cover information in regional-and global-scale models and the development of prototype land use models that can be validated and used to identify gaps in knowledge.
• Use of a wider range of satellite data to study human-environment interactions.
• Improved understanding of the relationship of population mobility to land use change, including the dynamics and environmental impacts of migration.

The link from science to policy is a major weakness in human response to global change. Although science-based understanding is essential for making informed decisions, it is not always obvious to scientists which information would be considered useful and relevant by participants in environmental decisions. It is also difficult for international, national, and local decision makers to make sense of available scientific information on complex environmental systems, much of which is uncertain or disputed and all of which is subject to change. Well-informed choices are even harder to make because they must be acceptable to decision participants who do not share common understandings, interests, concerns, or values. Research should pursue three related aims: improving methods for valuing nonmarket goods; improving analytical methods for integrating multiple types of decision-relevant information (e.g., integrated assessment models, cost-benefit analyses); and developing decision processes that effectively combine analytical, deliberative, and participatory approaches to

understanding environmental choices and thus guide scientists toward generating decision-relevant information.

Over the next decade this research field should address a number of questions:

• Are there ways to improve economic assessments of the costs, benefits, and distributional effects of forecasted climate changes and variations, taking adaptive capacity into account?
• What are the best ways of communicating uncertainty, providing early warnings of food and health problems, and introducing climate information in the policy process?
• How can environmental quality be incorporated into national accounting systems, so that it can more easily be considered in the policy-making process?
• How can information about the nonmarket values of environmental resources be incorporated effectively into decision making about resource use?
• How can we better represent, propagate, analyze, and describe uncertainties and surprises in integrated assessment (e.g., integrating quantitatively specified uncertainty with subjective probability distributions, clarifying the relationship between uncertainty and disagreement)?
• What are the characteristics of institutional processes that ensure that scientific analyses are organized so as to meet the needs of the full range of decision making participants for information and involvement?
• How can the knowledge and concerns of those participating in or affected by environmental decisions be used to inform scientists about how to make environmental information more decision relevant?
• How do expert advice and assessment influence policy, decision making, and collective knowledge of global change issues and how do policy makers interpret information about scientific uncertainty as they frame global change issues?
• How can decision making procedures be structured to bring the quantitative and formal information embedded in assessment models together with scientific judgment and the judgments, values, preferences, and beliefs of elite and nonelite citizens in decision making processes that meet the informational needs of the participants and are appropriate to the decision at hand?

Research to improve analytical techniques will use such methods as model development, with particular attention to the modeling and propagation of uncertainties through complex systems, dialogue among modelers using different methods for analyzing the same issues, experimental studies of methods for quantifying the nonmonetary values of environmental resources, and surveys to identify

those values. Research to improve decision making processes will use case studies and comparisons of existing systems that inform management decisions and will conduct experiments and simulations to test alternative processes, particularly methods that involve broadly based deliberative processes, for informing scientists about decision participants' information needs and for informing policy debates.

Over the next 5 to 10 years research on these issues can be expected to yield the following:

• Improved methods for describing uncertainty, scientific disagreement, and the potential for surprise in environmental systems (e.g., subjective probability distributions based on expert elicitation, discursive methods).
• A theoretically grounded understanding of the sources of apparent anomalies in expressed-preference methods of estimating nonmarket values of environmental goods and services.
• Clarification of the nature of conflicts over cost-benefit analyses and other techniques of integrating information in support of environmental policy decisions.
• Improved understanding of the conditions under which particular analytical approaches meet the needs of decision making participants for information and involvement and the conditions under which these approaches need to be supplemented with other techniques.
• Improved ability to incorporate scientific information within deliberative processes that involve scientists, policy makers, and interested and affected parties in informing and making environmental policy decisions. An experimental effort should be undertaken to use dialogue among scientists, policy decision makers, and interested publics to identify promising research areas that would lead to information directly usable for policy. The current national assessment effort might be studied as an experiment in this sort of dialogue and used to identify some new and important research directions.

As outlined in the USGCRP's (1997) Our Changing Planet, human dimensions research should be a component of each science theme as well as a crosscutting issue. What is needed now is to organize the USGCRP so as to make this a reality. For each of the program's four major research themes, key human dimensions research activities relevant to that theme must be identified and supported. For example, atmospheric chemistry would include research on the consumption patterns and technologies that drive emissions-altering atmospheric chemistry, on the impacts of UV changes, and on the institutions and decision-

making processes that result in the control of these emissions (e.g., the Montreal Protocol and its implementation). The theme area of seasonal to interannual climate prediction would include support for research on vulnerability to climate variations and the social implications of seasonal predictions. Decadal to centennial climate change research would incorporate research on consumption and land use changes that alter the global carbon cycle; the driving forces behind these changes; the vulnerability of water resources, agriculture, and fisheries to decadal shifts in ocean-atmosphere circulation; and the social and environmental effects of policies to limit greenhouse gas emissions. Terrestrial and marine ecosystem studies would encompass work on human causes, consequences, and responses to ecosystem changes, including an increased emphasis on ways in which institutions (in the broadest sense of property rights, laws, and markets) promote and prevent land use and ecosystem changes; integrative assessments of the interactions between natural variations and human exploitation of fisheries, grasslands, and forests; and studies of the human impacts of ecosystem changes resulting from multiple stresses (e.g., ecosystem changes and climatic changes). Some steps are currently being taken toward such integration (e.g., NOAA's effort to develop a research agenda on the human dimensions of seasonal to interannual climate prediction). Such efforts need to be encouraged and their research recommendations implemented.

Structuring support for human dimensions research only around themes defined by natural science is inadequate because certain human dimensions issues cut across all of the research themes and require crosscutting and independent research initiatives. These initiatives include those on valuing environmental quality, the problem of developing improved methods for environmental decision making, and some questions about the human driving forces of environmental change. The challenge of organizing research on these crosscutting topics is confounded by multiagency responsibilities for funding. The research priorities identified in this chapter cannot be addressed without focused and coordinated funding. NSF, the agency responsible for the largest share of designated human dimensions research funding within the USGCRP, is the agency with the most experience in engaging basic social, behavioral, and economic science expertise and in providing a strong peer review system for proposals. However, NSF funds primarily investigator-initiated and disciplinary, rather than problem-oriented and interdisciplinary, social science research. Many of the other agencies currently have very small budgets ($1 million to $3 million) devoted to human dimensions research, which restricts them to supporting research focused on the particular agency's responsibilities. The danger exists that certain critical research areas will be perceived as too crosscutting for funding by mission agencies and too interdisciplinary for funding by NSF.

Basic social science research on human dimensions administered by disciplinary programs at NSF in response to investigator-initiated proposals is very important. But support is also needed in the form of interdisciplinary review

panels, interagency collaborations, and research driven by specific science plans and organized in centers of excellence to advance human dimensions research. There are good models provided by the now-defunct human dimensions review panel at NSF, the recent Human Dimensions Centers and Methods and Models in Integrated Assessment initiatives, and the joint NSF-EPA partnerships in environmental research. The last collaboration supports research on decision making and valuation in environmental policy and on water and watersheds. Similar partnerships could address the research priorities identified in this document on environmentally significant consumption (NSF, EPA, and DOE), land use change (NSF, NASA, DOI), and regional climate assessment (NSF, NOAA, NASA).

Recent developments in international human dimensions infrastructure and planning make it increasingly important for the USGCRP to support U.S. participation in new international research projects and networks. This participation will allow the USGCRP to leverage funds and research contributions from other countries and to strengthen scientific capability in the United States and in developing countries. The USGCRP should support both core activities and U.S. participation in those IHDP programs that are well planned and truly international, as well as other regional and international networks such as the Inter American Institute, Asia Pacific Network, START, and the International Research Institute for Climate Prediction as they organize high-quality human dimensions research. We are at a critical juncture in the development of many of these international programs and networks, and there are opportunities to participate in new and important research collaborations and to assist in defining international research agendas that should not be missed. Many countries are about to organize new human dimensions research initiatives and establish national advisory committees. There are important opportunities for regional and bilateral collaboration with institutions such as the European Commission (1995) through NAFTA, and with Japan and other countries making a renewed commitment to human dimensions research.

With few notable exceptions, social science data have been collected without concern for research questions about the human dimensions of global environmental change. The data collection efforts have mainly been driven by other needs and paid for by agencies that are not part of the USGCRP. As a result, the overwhelming majority of the publicly available social science datasets, although potentially relevant to global change research, are not as well suited as they might be to this purpose. Several steps could be taken to improve this situation. First, when links from social science data to other observational sources are necessary, the USGCRP could make a marginal investment in the ongoing data collection to ensure that sufficient geographical location information is collected to permit linking with data from relevant observational platforms.

Assuming the requisite geographical linkage data are obtained, the issue of protecting the confidentiality of respondents arises. Below the country or perhaps provincial scale, virtually all social science data are obtained with the promise of protecting the confidentiality of individuals, households, organizations, and often communities. For a variety of ethical reasons it is essential that this confidentiality be maintained. Yet doing so makes it impossible to have geo-linked public use files for the research community. Efforts need to be made to facilitate building a secure system that can link social science data to biophysical data to meet legitimate research needs while simultaneously protecting the confidentiality of respondents.

Solutions might include the establishment of physical places where researchers could go to do their analyses under appropriately supervised conditions or a system that involved appropriate legal safeguards backed up by enforceable penalties. The involved federal agencies need to establish a mechanism to study the problem and put an effective solution in place. Without effective solutions, scientific progress will be severely constrained.

Linkages between human health and ecosystem information can combine environmental monitoring with consequences and impact monitoring. For example, NSF's long-term ecological research site in New Mexico now traps rodents for hantaviruses, in collaboration with the Centers for Disease Control and Prevention. Data systems that integrate health outcomes with remote sensing/geographic information system mapping can help researchers evaluate climate and land use impacts on food sources, predators, and habitats for rodents and other ecosystem changes with human consequences.

Linking social and biophysical data presumes stable funding for archiving and disseminating human dimensions data and sufficient financial resources to permit upgrading as the storage and dissemination technology changes. Both the Social and Economic Data and Applications Center and the Institute for Cooperative Programs in Survey Research have experienced substantial uncertainties and interannual fluctuations in their funding, which in turn has created problems for the research community. When the dissemination mechanism involves individual projects, mechanisms need to be in place to continue the archiving and dissemination of these datasets after the project is complete.

In addition, the time is right to carefully examine the extent to which existing and planned social science data serve the science needs of the global research agenda. This issue is discussed more extensively in Chapter 9, but brief reference is needed here. The bulk of social, economic, and health data used in human dimensions of global research is collected for other purposes, at scales well below the global level. To date, the principal federal agencies involved in collecting social science data in the United States and abroad (Census Bureau, Department of Labor, Department of Health and Human Services, and Agency for International Development) have not been part of the USGCRP. Furthermore, as discussed earlier in this chapter, human observations raise confidentiality issues that