The regional nature of global challenges: a need and strategy for integrated regional modeling
- First Online:
- Cite this article as:
- Hibbard, K.A. & Janetos, A.C. Climatic Change (2013) 118: 565. doi:10.1007/s10584-012-0674-3
- 608 Views
In this paper, we explore the regional nature of global environmental challenges. We take a broad approach by examining the scientific foundation that is needed to support policy and decision making and identifying some of the most important barriers to progress that are truly scale-dependent. In so doing, we hope to show that understanding global environmental changes requires understanding a number of intrinsically regional phenomena, and that successful decision making likewise requires an integrated approach that accounts for a variety of regional Earth system processes—which we define to include both human activities and environmental systems that operate or interact primarily at sub-continental scales. Understanding regional processes and phenomena, including regional decision-making processes and information needs, should thus be an integral part of the global change research agenda. To address some of the key issues and challenges, we propose an integrated regional modeling approach that accounts for the dynamic interactions among physical, ecological, biogeochemical, and human processes and provides relevant information to regional decision makers and stakeholders.
Global change research has been enormously and understandably influential in scientific and policy discussions for several decades. Major global changes such as climate change, loss of biodiversity, desertification, degradation of landscapes, and the overharvesting of the world’s oceans—each of which is driven in large part by human activities—have created extraordinarily challenging environmental problems. The scientific community has responded, in part, through global environmental change programs and through scientific assessment processes, most notably the Intergovernmental Panel on Climate Change (IPCC). These programs, along with the availability of powerful new global observations from satellites and the development of increasingly sophisticated models of climate processes, major biogeochemical cycles, and human activities, have led to a much more sophisticated understanding of Earth as a system.
While the scientific and policy communities have made great strides in understanding major global issues, gaps in regional information and process-level understanding have emerged. These gaps represent major barriers to predicting and responding to the large environmental and human system challenges identified above. This paper considers three overarching objectives in terms of key gaps and argue the necessity of (1) interdisciplinary integration, (2) the need for regional analyses; and (3) current capabilities through an overview of some key gaps in three areas—terrestrial sciences, atmospheric sciences, and decision making—including a short summary of the current state of knowledge and critical issues that remain to be addressed to improve our understanding of and ability to respond to climate change at regional scales. It focuses on the factors and issues highly relevant to regional energy and climate-related decision making for land systems. Other environmental challenges or components of the Earth system are obviously important, too, but our focus precludes an encyclopedic review of these.
1.1 Terrestrial sciences
In the terrestrial sciences, the variability of vegetation across landscapes is an intrinsically important part of understanding patterns and forces affecting biological diversity, ecosystem functioning, the delivery of ecosystem services, and exchanges of materials and energy with the atmosphere. The details of regional land-use history also cast a long shadow on processes that affect carbon exchange with the atmosphere; for example, the aging of temperate deciduous forests in North America may account for the “missing” carbon in the terrestrial part of the global carbon cycle (Tans et al. 1990; Houghton 1993). And of course, vegetation types, and therefore the habitats used by all manner of plants and animals, vary according to climate, topography, soils, and myriad other factors of the landscape that vary on regional scales.
1.2 Atmospheric sciences
In the atmospheric sciences, the direct and indirect forcing of aerosols varies significantly on regional scales (e.g., Ramanathan et al. 2001) because the processes that contribute to aerosol precursors, emissions, chemical transformations, and cloud interactions all vary regionally. Similarly, the ability to simulate precipitation depends on accurate representations of cloud and aerosol processes as well as on getting local topography represented correctly at fine spatial scales in rough terrain (see Wang et al. 2004 for review).
Regional simulations of climate variability and change seek to address some of these issues. Such simulations now include variable-resolution global models–with higher resolution in a specific region–as well as nested or limited-area models that use global fields as boundary conditions, or the downscaling of global model output using statistical or dynamical techniques. By focusing on a smaller geographic domain, regional climate models are able to directly simulate or more accurately approximate phenomena such as the influence of topography and other variations in land surface parameters on regional energy fluxes and precipitation. Advances in understanding and the available computing power have led to notable improvements in the resolution, accuracy, and comprehensiveness of regional climate simulations (e.g., Wang et al. 2004).
1.3 Decision processes: mitigation
Regional scales are also important for policy and decision-making communities. Global agreements are extraordinarily difficult to negotiate, and at least as difficult to monitor and verify. There are both regulatory and market-driven approaches to improving environmental quality—from addressing surface ozone pollution to reducing sulfur emissions—where implementation is intrinsically regional (e.g., U.S. Environmental Protection Agency (EPA) 2011). Responding to environmental challenges also demands consideration of local-to-regional economic and social conditions, tradeoffs, and implementation issues. Long-term global policy and technology goals are not realistic if they cannot be implemented on regional, shorter-term scales.
Policy decisions focused on climate change mitigation, or legislation designed to reduce the effects of climate change must account for changes in energy supplies, demand and technologies. The mix of energy technologies has evolved to satisfy economically-driven changes in energy supply, demand, markets, and regulatory requirements. It will continue to respond to all these forces, including the possible adoption of targets for atmospheric concentrations of greenhouse gases, policies to promote energy independence, and other influences. There is a large literature from integrated assessment models (IAMs) that explores the role of the energy sector in driving—and potentially mitigating—climate change, including such issues as the value of technology development in meeting greenhouse gas concentration (or radiative forcing) targets, the potential role of new technologies, and the interaction of the energy and agricultural systems (e.g., Edmonds et al. 2001, 2007).
Many studies using global IAMs make the simplifying assumption that each region makes decisions simultaneously in an economically rational way—which of course is not how the real world works. The consequences of different countries and simulated regions of the globe making strategic decisions with respect to energy and climate policy at different times and with different rates were explored by Clarke et al. (2009). These differences turn out to have a substantial influence on climate outcomes; for example, if rapidly developing countries delay reducing emissions for several decades, it becomes very costly—and in some cases physically impossible—to meet stringent greenhouse gas concentration targets.
1.4 Decision processes: adaptation
Adaptation decisions, in contrast to many mitigation decisions, are often governed much more by the particular circumstances of potential changes on regional and local scales—including both the specific regional manifestations of climate variability and change and the inherent spatial variability in the structure and function of human and natural systems, which lead to local and regional variations in the sensitivity of those systems to changes in climate. Adaptation strategies also reflect the fact that the management of most human and terrestrial systems, from agriculture to energy production to endangered species, is inherently regional (e.g., National Research Council 2010).
The regional nature of decision making is also evident in responses to global socioeconomic signals. For example, while demand for commodity crops such as corn or soybeans is driven in large part by global markets and trends, the management responses of individual farmers and groups of farmers to meet those demands is conditioned in large part by local climates, soils, availability of agricultural technologies, access to seed varieties, access to capital and labor, and so on. As a different example, adaptation decisions about energy infrastructure will require balancing between the expectations of future demand, the influence of energy policy and regulation, and consideration of climate-driven variability (such as changes in the supply of water for cooling). As an example, the current drought in Texas resulted in some localities prohibiting fracking for natural gas because of limited water availability for cooling and the need to continue agricultural production (Malewitz 2011).
2 Why is a regionally integrated approach needed?
One of the major complications in understanding and responding to global changes is that they are often characterized by multiple interactions, feedbacks, and tradeoffs among different human activities and environmental processes across both temporal and spatial scales. For example, energy supply, urban development, and agricultural production often compete for land and water resources. The extent and productivity of agricultural land depends on water supply, weather, and market demand, among many other factors. From the physical climate system perspective, agricultural land exchanges carbon, nutrients, energy, and water with the atmosphere and surrounding landscapes differently than other land-cover types. From climate policy perspective, the productivity and yields of agricultural land interact with policy choices about carbon pricing to influence the amount of land required to meet global and regional food demand. (e.g., see Wise et al., 2009).
Many of the physical and biogeochemical factors that determine the evolution of landscapes and their interaction with the atmosphere depend on the demand for ecosystem goods and services (Millennium Ecosystem Assessment 2005), some of which are priced in markets, and some of which are not. Therefore, accounting for these interactions, at the relevant spatial and temporal scales, is critical for informing and supporting effective decision-making. Our overarching hypothesis is that to understand and quantify gaps in our knowledge between decision making and process understanding, integration across different regional-scale physical, biogeochemical and human systems is critical. In the context of climate change, and especially climate-related decision making, this means representing regional-scale processes effectively, in a manner that accounts for human activities as well as physical, ecological, and biogeochemical processes. This requires interdisciplinary information from national or global scales that influence regional decision making, as well as high resolution processes that combine at smaller temporal and spatial scales to influence regional characteristics (e.g., Root and Schneider 2002).
Capturing the complex interactions between the physical and biogeochemical processes in the climate and environmental systems as well as engineered systems (such as energy infrastructure), will require appropriately resolved modeling frameworks. In addition, observations that provide insight into the processes governing decision making, and regional environmental feedbacks to the climate system as well as constraining and evaluating these models will be needed. Regional modeling in general, and integrated regional modeling in particular is challenging for a number of reasons. Characterizing the myriad interactions and feedbacks among energy, water, climate, and ecosystems—to name just a few of the relevant systems—at regional scales requires not only accurate representations of each individual system, but also a detailed understanding of the scale-dependent interactions among them. In addition, addressing the questions that regional decision makers are asking will require the development of models capable of evaluating different adaptation strategies, testing different mitigation options, and accounting for the tradeoffs, co-benefits, and uncertainties associated with these actions or combinations of actions—such as how technology cost, performance, and availability will impact results.
There are at least three reasons an integrated and regional approach is needed. First, a regional framework provides specificity and an opportunity to explore the unique characteristics for that place. Within a regional framework, it is important to ensure consistency with global boundary conditions, but less so to provide feedback to the global system. Secondly, consistent global data on energy systems and economies at the required level of specificity for credible results within a global framework are not available. Global climate models address these gaps through global observational systems and physical constraints that lend to interpolations around missing data and, often with quite good spatial resolution. Current global observational capabilities do not capture the required economic activity and energy data, which are not constrained by physical or biophysical processes. Third, part of what a regional framework should do is investigate the interactions between human systems and physical systems in a way that can illuminate decision-making and the processes that frame these decisions. Representation of regionally specific human and environmental systems within the global models will not be available in the near term, and, even so, will not necessarily represent those processes that are important to regional decision makers, but will account for the overarching objectives of the global models. Quantifying and providing regional policy and decision making relevant integrated analyses at these scales is urgently needed in the near term.
3 Regional impacts of climate change
Many studies have evaluated the current and anticipated future impacts of human-induced climate change on various sectors in different regions based on downscaled climate information. However, the vast majority of these studies neglect to account for regional-scale processes and interactions that could fundamentally alter the nature of the response to climate. For example, many studies have projected the future impacts of climate change on agriculture in different regions based on changes in global atmospheric carbon dioxide concentrations and regional temperature, precipitation, and other climatic factors (IPCC 2007a, b; CCSP 2008), but far fewer studies have accounted for other relevant factors such as changes in agricultural demand, competition over land and water resources for other uses (such as bioenergy production), and the availability and cost of new agricultural technologies (e.g., see Bhardwaj et al. 2010; Loarie et al. 2011). Even less well appreciated are the potential confounding influences of other global changes such as changes in dietary patterns, the availability of other food sources (e.g., the effects of overfishing, ocean acidification, and pollution on seafood production), the influence of invasive species, or changes in regional trade and agricultural incentives made for other (non-climate change related) reasons (Nelson et al. 2005). While some of these processes and interactions are global in scope, many are either regionally specific or will require regional resolution to effectively understand one or more of the relevant systems. A more robust and accurate understanding of the regional impacts of climate change on agriculture—or any other sector—over the next several decades will require an integrated approach that is capable of accounting for the complex, interacting processes and factors that influence outcomes on regional scales.
4 Implementation of greenhouse gas mitigation measures and adaptation strategies
5 Interactions among land, water, and energy
In addition to improving projections of the impacts of climate change on sectors and understanding the implications and limitations associated with adaptation and mitigation options, integrated regional approaches are needed to address questions and issues that span multiple sectors. One notable example, which has increasingly been recognized as a key issue in many regions and may be particularly amenable to an integrated modeling approach (see, e.g., Janetos et al. 2009), is the nexus of land, energy, and water (e.g., Skaggs and Hibbard 2012). Analyses of bi-sectoral issues in energy-water date back to the late 1970s (e.g., Harte and El-Gasseir 1978); more recently, several proposed power plants have recently been turned down in the permitting process over concerns related to the availability of adequate water supplies. Competition for land between the energy sector and agriculture—especially related to biofuel production—is a more recent concern (see, e.g., Reilly and Paltsev 2009, Wise et al. 2009).
Analyses that account for smaller scale human processes in larger scale contexts include Rausch and Rutherford (2010) that examine local and global convergences of market economies by household and O'Neill et al. (2012) that discuss urbanization trends in India and China and the subsequent energy choices. In general, the international integrated assessment modeling community has explored the consequences of simplistic assumptions around global cooperation or singular decision making results (e.g., Clarke et al. 2009). Their results clearly show that even within global frameworks with very broad definitions of “regional”, heterogeneity in the interests of the regional actors, can have extremely large consequences for the outcomes of global policy. These results demonstrate, even within existing frameworks, the importance of understanding regional heterogeneity in decision-making, just as we already understand that regional heterogeneity in natural resources and climate matters.
6 Designing an integrated regional modeling framework
Comprehensive, integrated modeling frameworks capable of simulating human systems, climate processes, and their interactions with the environment on multiple scales do not currently exist. Several recent modeling activities have, however, begun to couple human processes with detailed representations of climate system dynamics at global scales. For instance, Voldoire et al. (2007) coupled an integrated assessment model to the French ocean–atmosphere general circulation model to evaluate land use and land cover change interactions under a future climate change scenario that included significant changes in land use. Their results suggest that demographic and agricultural practices will dominate climate-change feedbacks on land-use decision making in the near term (less than 100 years). However, their global modeling framework was not capable of accounting for regional challenges.
Modeling frameworks that include multiple components of the Earth (e.g., atmosphere, land, coastal, marine) and human (e.g., energy infrastructure and technology, economy) systems on regional scales will require an integrated suite of model components that are collectively capable of simulating the processes and interactions for regionally specific multi-sectoral questions and issues and also relevant to climate and other global change issues. The framework should provide mechanisms to ensure that the regional models remain consistent with global constraints and boundary conditions, such as the output from global climate and integrated assessment models, and that fluxes of water, carbon, energy, and other common components are treated consistently. Finally, to ensure accuracy, consistency, and robustness, an integrated regional modeling framework should either include or readily support the application of sophisticated tools for model evaluation and uncertainty characterization, both for individual model components and for the modeling framework as a whole.
Not every question of regional importance will require all components of the framework to provide insight into the questions that decision makers might have. For water management questions over a time frame of a few decades, one might reasonably decide that model components that describe the slow response of vegetation functional types to changes in climate need not be invoked. The regional modeling framework, therefore, should be flexible such that every module need not be incorporated into an analyses or simulation but rather, for any given regionally-specific question, an appropriate coupling strategy should be employable. Finally, a regional modeling framework should be generalizable to any region in the world, with the primary limitation as available data to parameterize and initialize the models. A framework that is developed and calibrated to only one specific region may be useful for that region, but provides limited benefit outside a given region.
6.1 Incorporating decision-making and stakeholder involvement
One of the main rationales for developing an integrated regional Earth system modeling capability is to provide comprehensive, state-of-the-art scientific input to regional decision-making. It would be impossible to capture the insight needed without engagement with stakeholder and decision making communities (e.g., see National Research Council 2010). In the first US National Assessment of the Impacts of Climate Change and Variability, for example, a substantial effort was expended in identifying issues that regional stakeholders around the country were concerned about (National Assessment Synthesis Team (NAST) 2000; Morgan et al. 2005). Many stakeholders identified land-use and water management issues as being of particular importance, and in most cases as being of greater importance than just understanding the potential for the impacts of climate change alone. But resource management issues are environmental concerns that require a consideration not only of variability in the climate system, but also an understanding of how policy constraints, resource supply and demand, and the availability of technologies interact with each other over time. An integrated regional modeling framework as proposed could illustrate a family of possible solutions to stakeholder needs and decision criteria. As experience with the needs and requirements of these groups evolves, insight into which processes in the models must be improved, as well as managing expectations on what such a framework can do can be provided.
A more complete characterization of uncertainty should not only encompass an understanding of how the model behaves with different parameters and data sets, but also an understanding of the topology of decision-making to which it is being applied. As the modeling framework is exercised to address stakeholder questions, such as how will the competition for land and water influence the need to meet energy demand in the future, new thinking in uncertainty analyses and model evaluation criteria will be needed. Incorporating stakeholder decision support needs in an integrated framework will provide insight into the appropriate questions that an integrated framework should address (Rice et al. 2012). Whereas uncertainty quantification that identifies errors that arise from missing, or limited data, or model error is clearly needed, qualitative uncertainty introduced by the criteria that decision makers actually use (e.g., price and distance of existing infrastructure for new energy development), by the risks they are willing to take, and by their ability to explore alternative scenarios of the future will also be required.
7 Priority challenges
Developing and testing an integrated regional Earth system modeling framework will require considerable effort and coordination across a variety of disciplines including the natural and social sciences as well as engineering communities. While there are many challenges to address, we would like to highlight three particularly vexing ones: mis-matched time and space scales of model components, interdisciplinary communication, the development of appropriate metrics for model evaluation, and the associated availability of regionally specific data.
The temporal and spatial scales of natural and human systems vary tremendously. For instance, dynamically downscaled models of the climate system typically resolve topography and explicitly simulate climate processes at spatial scales down to 10 km using half-minute (e.g., 30 s) time steps (e.g., Skamarock et al. 2008). The current generation of integrated assessment models, in contrast, are global in scope but typically resolve the economies of a relatively small number (14–22) geopolitical regions (see Van vuuren et al. 2011 and references therein). Other key systems—such as agriculture, land use, and precipitation/runoff—can be considered from the individual farm, ecosystem, or sub-basin level, respectively, but are more commonly represented on scales of tens to hundreds of kilometers in climate models. In the framework identified in Fig. 3, we anticipate not only needing to span spatial scales in the atmosphere from a few kilometers to the scale of synoptic weather patterns, but for terrestrial and energy systems, we anticipate analyses at 0.5°-resolved global data all the way down to a kilometer or less on the landscape itself.
Characterizing and quantifying the nature of uncertainty across and within natural and human system modeling components will be critically important to evaluate not only model skill, but consequences of different decision making criteria (e.g., cost to implement a strategy, or minimizing the risk of failure of a strategy). While several examples exist for quantifying predictive uncertainty across climate, weather or hydrologic systems (e.g., McMillan and Brasington 2008; Pappenberger et al. 2005), there are no analyses to date that quantify uncertainty across modeling components that incorporate regional climate, energy and agricultural systems, socio-economics and technology. Understanding how model skill, incorporation of scale-appropriate processes, decision criteria in the context of mitigation and adaptation will be critical to evaluation of such a framework.
There will also be formidable communication challenges across the natural, social and engineering communities. These communities all embrace different notions of model evaluation and model implementation from research mode to application. Models of the climate and ecosystems are generally considered predictive–they simulate processes based on initial conditions and parameterizations to test hypotheses about interactions within those systems, and secondarily are used to explore potential futures, whereas integrated assessment models are primarily intended to explore scenarios of potential futures under different policy options.
A third challenge will be developing the metrics and appropriate datasets for evaluating integrated model performance. Physical or biogeochemical modeling groups typically evaluate their models by initializing with historical climate or ecological information and comparing results to observations. Human system modelers, in contrast, often do not have the requisite data to evaluate their models in this manner—their models typically use the relatively sparse historical record for initialization, and there has not yet been a concerted community effort to develop independent data sets for model evaluation. Integrated assessment model intercomparisons of similar experiments have attempted to serve this purpose (e.g., Clarke et al. 2009). Accounting for different types and modes of economic variability is a related complication, one that is only somewhat analogous to the differences between validating a long-term climate simulation versus a short-term weather forecast. Finally, the single biggest limiting factor for an integrated regional approach may be the availability of regionally specific data—specifically, the quality, resolution, and sheer amount of data needed for model development, initialization, and evaluation.
Is mitigation more regionally constrained than currently anticipated? For example, are there regional bio-physical or economic constraints that make the implementation of new energy technologies or mitigation options more difficult?
How do changes in both mean climate and climate variability affect scale-dependent policy, resource management decisions and adaptation strategies and how might such decisions interact with mitigation strategies?
Are there tipping points in the environment, either from management decisions about mitigation or adaptation, or from changes in the physical environment, that only become noticeable or understandable when regional scales are considered?
Are regional adaptation strategies for the next 25 to 50 years independent of particular mitigation/stabilization regimes, or do they interact in significant ways?
The framework we propose in this paper will allow these, and other similar questions to be addressed. Successfully implemented, it will provide additional insight into the regional manifestations of global change issues and response strategies.