Geospatial Analysis and Modelling of Urban Structure and Dynamics

Geographic information research and technologies have experienced over four decades of development, from the mainframe to the workstation to the desktop, and to today's laptop and mobile devices. Every important GIS development is driven by a significant breakthrough of mainstream information technology. For example, the 1980s was characterized by the popularity of personal computers that were increasingly becoming affordable to university departments, governmental agencies and private sectors. Many university GIS programs were established during this time period, and the NSF-funded NCGIA played an important role in coordinating the development of course curriculum and related research activities. The next decade can be named the age of GIScience. GIScience is the science behind GISystens, dealing with fundamental questions raised by the use of GISystems and technologies (Goodchild 1997). It occurred at the time when the Internet and the World Wide Web started to change the way we led our lives and ran our businesses. It was the Internet and the Web that made the GIS community think of a service oriented approach to GIS, namely GIServices (Günther and Müller 1999). Instead of owning a GIS, end users can be served by GIS functionalities from a remote GIService center. GIServices aim to develop distributed or decentralized GIS to serve individuals and communities for spatial planning and decision making, as well as for their daily life. Another perspective of GIS is GIStudies for studying the impacts of geographic information and technologies on society.

The increasing availability of geographic data at high resolution and good quality has improved our ability to investigate a city's social geography thanks to the data's enhancement of our capacity to integrate three fundamental dimensions: residential distributions, the built-up environment's properties, and individuals' perceptions. This paper discusses the benefits and constraints of using such data in current approaches to the investigation and modeling of urban residential segregation. Based on several studies conducted on residential segregation in Jaffa, a mixed area in Tel-Aviv, we conclude that high-resolution data, especially at the house-level, has significant potential to enrich our understanding of the involvement of the built-up environment and individuals' spatial preferences in ethnic residential segregation. In addition, the constraints associated with using such data are discussed. On the methodological level, they refer to privacy considerations and cartographic presentation; on the conceptual level, they touch upon the gap between observed behavior, spatial preferences, and the dynamics of social residential segregation.

In this paper we describe our novel use of space syntax for the design and development of pervasive systems. Pervasive systems are computer systems that are designed to be “invisible” to users because they are designed to blend in with their environment and become part of the fabric of everyday life. Pervasive systems consist of fixed, mobile and embedded components, each of which may entail interactive capabilities and intelligence. Due to their close relationship with the build environment, pervasive systems are an ideal domain for adopting a space syntax methodology. The contribution of this paper is two-fold. First, we present an adaptation of the space syntax methodology aimed at researchers and designers of space and pervasive systems. By developing an adaptation of the space syntax methodology we intend for space syntax to be added the arsenal of tools and theories that researchers use to understand and design pervasive systems. To exemplify how this can be achieved, we present three case studies that demonstrate crucial ways in which space syntax analysis can aid the design of pervasive systems. The case studies presented here show how space syntax can be used in the development of pervasive system as an application development tool, as an exploratory tool, and as a modelling tool.

The aim of this proposed research is to develop an urban simulation model (USM) specifically designed to study the evolution and dynamics of systems of cities. It is innovative in three respects: First, in its structure – the proposed model is built as a superposition of two types of models; the first is an Agent Based Urban Simulation Model (ABUSM) that simulates the movement and interaction of agents in the urban space and the second is a network model that simulates the resultant urban network as it evolves. Secondly, it is innovative in the specific behavior of its agents – urban agents in our model act locally (as usual) but in order to do so, they perceive the city globally, i.e. they “think globally and act locally”. In our model, the local activities and interactions of agents give rise to the global urban structure and network that in turn affects the agents’ cognition, behavior, movement, and action in the city and so on in circular causality. The third aspect of innovation is connected with the specific urban phenomena it simulates – the vast majority of USM simulate the growth and expansion of urban systems but few simulate the reverse process of re-urbanization and gentrification; our model simultaneously captures the two processes and the interplay between them.

Cities are complex systems by their nature. They have originally emerged, and are still developing, out of the interactions between many agents that are located and move in space and time. These interactions result in many links that create complex networks which form the city. In the last two decades CA and AB USM have provided the main approaches to studying the dynamics of cities as complex self-organizing systems. In the last few years, models based on the new science of networks have been introduced as well. This research direction is new in the conjunction it suggests between traditional network analysis (e.g. graph theory) and complexity theory. We introduce a new dynamic model for city development, based on the evolution and structure of urban networks.

The scaling relations indicating fractal nature can be employed to associate a hierarchy and a network of cities. The cascade structure of an urban hierarchy follows the 2 ⁿ rule of size class, which can be formulated as a set of exponential functions. From a pair of exponential laws, we can derive a power law indicating the scaling relation between city number and city size of different classes. The scaling exponent is the fractal dimension of the city-size distribution. Owing to the relationship of mathematical transformation between a hierarchy and a network, the scaling analysis of hierarchical structure can be used to enrich geographical spatial analysis such as spatial interaction studies. The interaction among cities from different classes has locality property, i.e., in theory, cities of one class act merely on cities of the immediate adjacent classes. Not all cities have significant impact on other cities where hierarchy is concerned. Only the interaction between adjacent cities, both in space and size, is strong enough to have a significant effect on the structure of urban systems.

Virtually all major cities around the world face severe parking problems in their centers. While existing models of parking search and choice behavior do provide insight into the basic dynamics of parking in cities, as well as into the phenomenon of drivers cruising for on-street parking, virtually all models discussed in the literature ignore a number of key factors that influence parking behavior and parking dynamics. This paper makes a first step in this direction, by proposing a non-spatial model of parking search and an explicit geosimulation model of the parking process, termed PARKAGENT, which accounts for street network, drivers’ decisions and their destination. We employ both models to analyze the phenomena of cruising for parking and compare the models’ outcomes, focusing on the impact of space on parking dynamics. We estimate the main characteristics of these dynamics, and specify the conditions under which spatial effects are, or are not, important for analyzing parking. In particular, we demonstrate that traffic engineers’ recommendation that about 15% of all on-street parking places should remain vacant to ensure easy ingress and egress and prevent cruising for parking can be decreased to 7% and even less, especially in case of relatively low parking turnover levels. The paper ends with a short discussion, in which we explore the implications of our model study for establishing urban parking policies.

This study investigates the convenience of urban transportation in Beijing municipality area. The analyses reported here consist of three parts − the spatial patterns of urban transportation Level of Service (LOS) perceived as transportation convenience by residents in Beijing, the spatial patterns of urban transportation LOS as reflected by objective transportation and traffic data, and the comparison of these two sets of patterns and measurements. The spatial unit of analyses was Jie Dao (街道), which is the basic civil administration unit in Chinese cities. This is one of the first studies linking subjective assessment of transportation convenience in Beijing with objective measurement; it reveals the displacement between the patterns of transprotation convenience reflected from the two aspects. Further examination are needed to explain the discrepancy between the perceived and measured transportation LOS in different Jie Daos in Beijing. The related investigation shall shed light on better understanding the effectiveness of different practices of urban transportation management, planning, and project investment in different types of neighborhoods in Beijing.

We present an object-oriented data model to represent the multi-modal, indoor/outdoor transportation network of an urban area, that can be used for route planning and navigation and to perform other network analyses. The data model divides the urban transportation network into two main parts, one being the network for modeling movement inside buildings and the other being the network outside buildings using multiple modes. The network inside buildings is intrinsically three-dimensional (3D), thus requiring to model vertical connectivity between floors. Such 3D indoor network is modeled using a 2.5D approach that combines the existing 2D data structures and 3D visualization techniques available in commercial off-the-shelf GIS software. The network outside buildings is modeled as a multi-modal network consisting of modes operating on streets, walkways, and public transit routes. The data model is implemented for the multi-story buildings and transportation network on the campus of the State University of New York at Buffalo, USA. The effectiveness of the data modeling scheme is evaluated by performing path finding analysis in the study area using the data model with standard and customized GIS software tools.

This chapter explores the utilities of remote sensing and GIS integration for urban growth characterization by using a loose-coupling approach. Central to this approach is a time series of satellite imagery for detecting urban land changes. The change detection output is combined with other biophysical and socio-economic data to examine some causal factors leading to the observed changes. The historical urban extent data derived by remote sensing are used to calibrate a dynamic model, which is further applied to assess the spatial consequences of future urban growth under different scenarios. This approach has been applied to a case study site, Atlanta, a typical postmodern American metropolis having undergone rapid demographic and economic growth during the past several decades. Remote sensing-based change detection reveals far-reaching suburbanization in Atlanta during 1973–1999, as indicated by large urban land area growth that substantially outpaced the rate of population growth, as well as more scattered urban spatial patterns. The growth and change revealed by remote sensing are found to be highly correlated with population and economic growth and accessibility conditions. Future urban growth simulations suggest that numerous edge cities would emerge and smaller ones would coalesce together to form larger urban clusters. If current growth patterns do not alter, the process of suburbanization would deplete vegetation and open space by around 2030. The research approach reported here contrasts with many other urban growth studies that focus on the remote sensor data analysis without enabling GIS-based spatial analysis and modeling to go beyond data description and exploration into areas of simulation and prediction.

Urbanization is a critical factor affecting the ecological and environmental balance of the Greater Toronto Area (GTA), the most populous metropolitan area in Canada, in the past three decades. The purpose of this chapter is to examine the relationship between landscape change and population increase patterns as well as to evaluate ecological impacts of urbanization in the GTA. Multi-temporal remotely sensed data have been used to derive vegetation changes from 1992 to 2003. Land use change is derived from historical land use maps. Five landscape fragmentation indices are calculated for different periods using FRAGSTATS software. Population change is compared with land use and vegetation changes. The landscape fragmentation rate is then compared with the population growth rate. Our results show that the mean normalized difference vegetation index (NDVI) is negatively correlated with the percentage of the urban settlement land and population density at the census tract (CT) level. Changes in the percentage of urban land use show relatively weak correlations with the five fragmentation indices. It was found that shape and fractal dimension indices are better at characterizing the urbanization process than the indices of diversity, contagion, and percentage of like adjacencies.

Precipitation is an important component of the global water cycle and a proxy for changing climate. Proper understanding and quantification of spatio-temporal precipitation variability is critical for a range of meteorological, hydrological, and climate processes. Past and current literature has presented theories and observational studies on how urbanization affects precipitation. Assessment of the urban environment’s (land use, aerosols, thermal properties) impact on precipitation will be increasingly important in ongoing climate diagnostics and prediction, global water and energy cycle (GWEC) analysis and modeling, weather forecasting, freshwater resource management, urban planning-design, and land-atmosphere-ocean interface processes. This chapter presents a review of findings and methods related to “urban rainfall effect” studies with an emphasis on numerical modeling strategies. Numerical modeling of atmosphere-land interactions enables controlled experimentation to address fundamental research questions.

Measures of population pressure, referring in general to the stress upon the environment by human consumption of resources, are imperative for environmental sustainability studies and management. Development based on resource consumption is the predominant factor of population pressure. This paper presents a spatial model of population pressure by linking consumption associated with regional urbanism and ecosystem services. Maps representing relative geographic degree and extent of natural resource consumption and degree and extent of impacts on surrounding areas are new, and this research represents the theoretical research toward this goal. With development, such maps offer a visualization tool for planners of various services, amenities for people, and conservation planning for ecologist. Urbanization is commonly generalized by census numbers or impervious surface area. The potential geographical extent of urbanism encompasses the environmental resources of the surrounding region that sustain cities. This extent is interpolated using kriging of a variable based on population wealth data from the U.S. Census Bureau. When overlayed with land-use/land-cover data, the results indicate that the greatest estimates of population pressure fall within mixed forest areas. Mixed forest areas result from the spread of cedar woods in previously disturbed areas where further disturbance is then suppressed. Low density areas, such as suburbanization and abandoned farmland are characteristic of mixed forest areas.

A place exists in a space (footprint) which evolves over a temporal extent (lifeline). This study focuses on those places such as cities or component places within cities (hereafter referred to as city places or simply places). It concerns the modeling and analysis of places from the integrative spatio-temporal perspective. Drawing on recent research development related to spatio-temporal ontology, this study discusses a spatio-temporal ontology specific to city places. The ontology distinguishes between the static view of places (Static-Place) and the dynamic view of spatial-temporal regions led by spatio-temporal processes (ST-Place). The study considers places as spatio-temporal constructs. It proposes the concept of primitive ST region for the construction of spatio-temporal regions and ultimately spatio-temporal places. The study then presents a spatio-temporal data model to link spatial features and the spatio-temporal processes. In this way, the study models an ST-place from a spatio-temporal perspective by forming high dimensional spatio-temporal regions. Based on this conceptual data model, the study further discusses the potential use of this model in spatio-temporal analysis in general and in reasoning spatio-temporal topological relations in particular.

This essay outlines how to incorporate morphological rules within the exigencies of our technological age. We propose using the current evolution of Geographical Information Systems (GIS) technologies beyond their original representational domain, towards predictive and dynamic spatial models that help in constructing the new discipline of “urban seeding”. We condemn the high-rise tower block as an unsuitable typology for a living city, and propose to re-establish human-scale urban fabric that resembles the traditional city. Pedestrian presence, density, and movement all reveal that open space between modernist buildings is not urban at all, but neither is the open space found in today’s sprawling suburbs. True urban space contains and encourages pedestrian interactions, and has to be designed and built according to specific rules. The opposition between traditional self-organized versus modernist planned cities challenges the very core of the urban planning discipline. Planning has to be re-framed from being a tool creating a fixed future to become a visionary adaptive tool of dynamic states in evolution.

Breast cancer has ranked highest in cancer incidence in Illinois for years. Detection at an early stage helps cancer patients live longer and maintain a better quality of life. Previous research identified two groups of potential risk factors for late-stage breast cancer diagnosis: spatial factors including access to healthcare, and nonspatial factors including socioeconomic and demographic characteristics. Literature suggests that risk factors for late-stage diagnosis behave differently between rural and urban areas, and research needs to separate study areas into various geographic settings. This chapter focuses on an urban area of six counties in Chicago region, and examines possible associations between several risk factors and late-stage breast cancer diagnosis. Based on the data at the zip code level, the study uses the modified scale-space clustering (MSSC) method to form various geographic areas. The MSSC considers both attribute similarity and spatial adjacency while minimizing the loss of information in the clustering process. Therefore, area units defined by the method are more coherent in terms of attribute and spatial closeness for research than geopolitical units. For instance, health literature often suggests the need to separate a study area into urban, suburban and rural areas or even finer-grained area classifications. The method can be used to generate more meaningful geographic divisions than traditional schemes. The analysis results are generally consistent across multiple area units, demonstrating the effectiveness of the method in mitigating the modifiable areal unit problem (MAUP).

The impact of job accessibility on housing prices is examined in the Buffalo and Seattle metropolitan areas using a hedonic regression modeling framework. Global hedonic regression results show that job accessibility is positively associated with housing prices in the two study areas. Local hedonic regression modeling is also conducted to test whether the response of the housing market to job accessibility is spatially stationary. The statistical analysis reveals that the role of job accessibility in the house price-setting process varies locally in each metropolitan area. Empirical challenges with unraveling relationship between transportation and land use, and the policy implications of our findings, are discussed.

Spatial analytic research has explored the issue of where to best site hurricane relief distribution facilities, but it has largely concentrated on the efficient provision of these services. However, equity considerations may also impact decisions on where to locate facilities. Questions of efficiency vs. equity become all the more acute when more detailed assessments of peoples’ socioeconomic characteristics are made as a part of these decisions. This paper examines the issue of siting hurricane disaster relief facilities based on equity vs. efficiency objectives, in light of populations’ socioeconomic differences. Population differences are measured in terms of a household income variable. p-median and vertex p-center problems are applied to find relief center locations in a Southeastern U.S. city. Results show that income differences interact with the location strategies employed to produce variation in people’s accessibility to relief goods.

Visualizing and interpreting regression coefficients from spatially varying coefficient models, such as geographically weighted regression (GWR), can be challenging, given the amount of information the models provide the spatial analyst. Adding to the visualization dilemma are various diagnostic tools for checking the coherence of the model. One such diagnostic tool is based on decomposing the regression coefficient variance matrix to test for collinearity, which is relevant because previous research has shown that collinearity in GWR can lead to estimated regression coefficients for multiple regression terms that are strongly correlated with each other. In this paper, visualization tools, such as linked scatter plots, parallel coordinate plots, and maps, are presented for diagnosing correlation in estimated regression coefficients that can be problematic for statistical inference of relationships between variables. These tools help explain patterns of dependence between regression terms apparent in scatter plots of estimated coefficients and link the structure in the scatter plots to locational information in a map. Visualization of this information should aid in the typical spatially varying coefficient model estimation process. A case study of census undercount in Franklin County, Ohio is presented as an illustrative example of applying the visual diagnostic approach in a GWR analysis.