Moira Zellner

Path dependence and the validation of agent-based spatial models of land use

In this paper, we identify two distinct notions of accuracy of land‐use models and highlight a tension between them. A model can have predictive accuracy: its predicted land‐use pattern can be highly correlated with the actual land‐use pattern. A model can also have process accuracy: the process by which locations or land‐use patterns are determined can be consistent with real world processes. To balance these two potentially conflicting motivations, we introduce the concept of the invariant region, i.e., the area where land‐use type is almost certain, and thus path independent; and the variant region, i.e., the area where land use depends on a particular series of events, and is thus path dependent. We demonstrate our methods using an agent‐based land‐use model and using multi‐temporal land‐use data collected for Washtenaw County, Michigan, USA. The results indicate that, using the methods we describe, researchers can improve their ability to communicate how well their model performs, the situations or instances in which it does not perform well, and the cases in which it is relatively unlikely to predict well because of either path dependence or stochastic uncertainty.

A new framework for urban sustainability assessments: Linking complexity,information and policy

Abstract Urban systems emerge as distinct entities from the complex interactions among social, economic and cultural attributes, and information, energy and material stocks and flows that operate on different temporal and spatial scales. Such complexity poses a challenge to identify the causes of urban environmental problems and how to address them without causing greater deterioration. Planning has traditionally focused on regulating the location and intensity of urban activities to avoid environmental degradation, often based on assumptions that are rarely revisited and producing ambiguous effects. The key intellectual challenge for urban policy-makers is a fuller understanding of the complexity of urban systems and their environment. We address this challenge by developing an assessment framework with two main components: (1) a simple agent-based model of a hypothetical urbanizing area that integrates data on spatial economic and policy decisions, energy and fuel use, air pollution emissions and assimilation, to test how residential and policy decisions affect urban form, consumption and pollution; (2) an information index to define the degree of order and sustainability of the hypothetical urban system in the different scenarios, to determine whether specific policy and individual decisions contribute to the sustainability of the entire urban system or to its collapse.

Embracing complexity and uncertainty: The potential of agent-based modeling for environmental planning and policy

Environmental degradation is often defined as a public goods problem, emerging when property rights are not clearly defined and costs are externalized to other parties. Proposing corrective regulation that enforces technological fixes or market-based approaches is often met with political resistance and doubts about its effectiveness. This is partly due to the complexity of interacting physical and socio-economic components that obscure the impacts of human decision-making on environmental functions. Yet, understanding the complexity of integrated human-environmental systems can help planners and stakeholders frame environmental problems, view their role in them and design effective policies to address them. This article examines the potential and limitations of agent-based models as metaphors that can contribute to the understanding of such complex systems, illustrating the argument with a hypothetical application in groundwater management.

Linking MODFLOW with an agent-based land-use model to support decision making.

The U.S. Geological Survey numerical groundwater flow model, MODFLOW, was integrated with an agent-based land-use model to yield a simulator for environmental planning studies. Ultimately, this integrated simulator will be used as a means to organize information, illustrate potential system responses, and facilitate communication within a participatory modeling framework. Initial results show the potential system response to different zoning policy scenarios in terms of the spatial patterns of development, which is referred to as urban form, and consequent impacts on groundwater levels. These results illustrate how the integrated simulator is capable of representing the complexity of the system. From a groundwater modeling perspective, the most important aspect of the integration is that the simulator generates stresses on the groundwater system within the simulation in contrast to the traditional approach that requires the user to specify the stresses through time.

Generating policies for sustainable water use in complex scenarios: an integrated land-use and water-use model of Monroe County, Michigan

Rapidly declining groundwater levels since the early 1990s have raised serious concern in Monroe County, Michigan. Hydrological studies suggest that land-use changes have caused this decline. The mechanisms linking land-use and groundwater dynamics are not clear, however. In this paper I present WULUM, the Water-Use and Land-Use Model, an agent-based model that serves as an analytical framework to understand how these processes interact to create the observed patterns of resource depletion, and to suggest policies to reverse the process. The land-use component includes the main groundwater extractors in the county—stone quarries, golf courses, farms, and households. The groundwater component includes the glacial deposits and the underlying bedrock acquifer. The behavior of water users is defined by simple rules that determine their location and consumption. The dynamics of groundwater are represented through infiltration and diffusion rules between each cell and its immediate neighbors. Initial explorations with the model showed that land-use patterns contributed significantly to groundwater declines, while eliminating quarry dewatering did not entirely solve the problem. Both low-density and high-density zoning restrictions improved aquifer conditions over medium-density development, suggesting a nonlinear relationship between intensity of residential use and groundwater levels. Moreover, of all the natural and policy variables, zoning had the greatest influence on urban settlement and therefore on resource consumption.

Planning for deep-rooted problems: What can we learn from aligning complex systems and wicked problems?

An earlier generation of planners turned to Rittel & Webber’s 1973 conception of “wicked problems” to explain why conventional scientific approaches failed to solve problems of pluralistic urban societies. More recently, “complex systems” analysis has attracted planners as an innovative approach to understanding metropolitan dynamics and its social and environmental impacts. Given the renewed scholarly interest in wicked problems, we asked: how can planners use the complex systems approach to tackle wicked problems? We re-evaluate Rittel and Webber’s arguments through the lens of complex systems, which provide a novel way to redefine wicked problems and engage their otherwise intractable, zero-sum impasses. The complex systems framework acknowledges and builds an understanding around the factors that give rise to wicked problems: interaction, heterogeneity, feedback, neighbourhood effects, and collective interest traps. This affinity allows complex systems tools to engage wicked problems more explicitly and identify local or distributed interventions. This strategy aligns more closely with the nature of urban crises and social problems than the post-war scientific methodologies about which Rittel and Webber had grown increasingly sceptical. Despite this potential, planners have only belatedly and hesitantly engaged in complex systems analysis. The barriers are both methodological and theoretical, requiring creative, iterative problem framing. Complex systems thinking cannot “solve” or “tame” wicked problems. Instead, complex systems first characterize the nature of the wicked problems and explore plausible pathways that cannot always be anticipated and visualized without simulations. The intersection of wicked problems and complex systems presents a fertile domain to rethink our understanding of persistent social and environmental problems, to mediate the manifold conflicts over land and natural resources, and thus to restructure our planning approaches to such problems.

Exploring the effects of green infrastructure placement on neighborhood-level flooding via spatially explicit simulations

Abstract State and local governments are increasingly considering the adoption of legislation to promote green infrastructure (e.g., bioswales, green roofs) for stormwater management. This interest emerges from higher frequencies of combined sewer outflows, floods and exposure of residents and habitat to polluted water resulting from growing urbanization and related pressure on stormwater management facilities. While this approach is promising, there are many unknowns about the effects of specific implementation aspects (e.g., scale, layout), particularly as urban settlements and climate conditions change over time. If green infrastructure is to be required by law, these aspects need to be better understood. We developed a spatially-explicit process-based model (the Landscape Green Infrastructure Design model, L-GriD) developed to understand how the design of green infrastructure may affect performance at a neighborhood scale, taking into consideration the magnitude of stormevents, and the spatial layout of different kinds of land cover. We inform the mechanisms in our model with established hydrological models. In contrast with watershed data-intensive models in one extreme and site level cost-savings calculators in the other, our model allows us to generalize principles for green infrastructure design and implementation at a neighborhood scale, to inform policy-making. Simulation results show that with as little as 10% surface coverage, green infrastructure can greatly contribute to runoff capture in small storms, but that the amount would need to be doubled or tripled to deal with larger storms in a similar way. When placement options are limited, layouts in which green infrastructure is dispersed across the landscape—particularly vegetated curb cuts—are more effective in reducing flooding in all storm types than clustered arrangements. As opportunities for green infrastructure placement increase and as precipitation increases, however, patterns that follow the flow-path and accumulation of water become more effective, which can be built on an underlying curb-cut layout. If space constraints prevented any of these layouts, random placement would still provide benefits over clustered layouts.

Exploring the Influence of Urban Form on Work Travel Behavior with Agent-Based Modeling

This paper examines the effect of land use regulations on travel behavior by using agent-based modeling. A simulation model for a hypothetical urban area loosely based on the Chicago, Illinois, metropolitan area was used to study the impact of six land use regulation scenarios on transit use and urban form. The key features and techniques of the model development and the scenarios tested are described. The results from the simulations showed that although the land use regulations that were designed to increase the density near the transit station or in and near the urban core were able to achieve the intended land use patterns, they did not increase the transit mode share for the region in a significant manner. More detailed examination of the output revealed that as long as the rules for mode choice, the distribution of employment, and the transit network remained unchanged, land use regulations that affect residential locations produced limited effects on transit use.

Big Data and Urban Informatics: Innovations and Challenges to Urban Planning and Knowledge Discovery

Big Data is the term being used to describe a wide spectrum of observational or “naturally-occurring” data generated through transactional, operational, planning and social activities that are not specifically designed for research. Due to the structure and access conditions associated with such data, their use for research and analysis becomes significantly complicated. New sources of Big Data are rapidly emerging as a result of technological, institutional, social, and business innovations. The objective of this background paper is to describe emerging sources of Big Data, their use in urban research, and the challenges that arise with their use. To a certain extent, Big Data in the urban context has become narrowly associated with sensor (e.g., Internet of Things) or socially generated (e.g., social media or citizen science) data. However, there are many other sources of observational data that are meaningful to different groups of urban researchers and user communities. Examples include privately held transactions data, confidential administrative micro-data, data from arts and humanities collections, and hybrid data consisting of synthetic or linked data.

Evaluating the embodiment benefits of a paper-based tui for educational simulations

Many claims have been made regarding the potential benefits of Tangible User Interfaces (TUIs). Presented here is an experiment assessing the usability, problem solving, and collaboration benefits of a TUI for direct placement tasks in spatially-explicit simulations for environmental science education. To create a low-cost deployment for single-computer classrooms, the TUI uses a webcam and computer vision to recognize the placement of paper symbols on a map. An authentic green infrastructure urban planning problem was used as the task for a within-subjects with rotation experiment with 20 pairs of participants. Because no prior experimental study has isolated the influence of the embodied nature of the TUI on usability, problem solving, and collaboration, a control condition was designed to highlight the impact of embodiment. While this study did not establish the usability benefits suggested by prior research, certain problem solving and collaboration advantages were measured.