Academic journal article Cartography and Geographic Information Science

Modeling Dynamic Polygon Objects in Space and Time: A New Graph-Based Technique

Academic journal article Cartography and Geographic Information Science

Modeling Dynamic Polygon Objects in Space and Time: A New Graph-Based Technique

Article excerpt

Introduction

The element of time plays a critical role in the analysis of dynamic spatial systems. For example, plant ecologists who study processes controlling landscape heterogeneity through patch dynamics, investigate how size, shape, location, and species composition of individual patches change over time in relation to other patches in a system (Levin et al. 1993). Ecological processes can cause patches to come into existence, merge, split apart, or die off, altering patch topology over time. In dynamic spatial systems, time, locational attributes, and aspatial characteristics combine to form a complex structure that is not readily captured by at conventional map (Peuquet 1994). Moellering (1984) introduced the concept of real and virtual maps to describe this aspect of cartographic data more precisely. Traditional maps that are both directly viewable and have a permanent tangible reality are called real maps, while cartographic data that lack one or both of these properties are classified into one of three types of virtual maps, based on the missing property. The complex spatio-temporal structure of interest here is classified as a Virtual Map-Type 3 because it represents data that are not directly viewable and do not have any permanent tangible reality.

In a spatio-temporal system, the temporal, spatial, and aspatial elements must be managed and analyzed in concert. By treating all three elements together, spatio-temporal analysis can illuminate patterns that are not evident for data at individual points in time or space, or for individual aspatial data attributes (c.f. Kraak 1999). Without the direct incorporation of a spatial component, temporal analysis is restricted to fixed, and in many cases, arbitrary regions. Likewise, without a temporal element, spatial analysis is limited to describing a single point or period in time.

The full integration of temporal analysis with spatial analysis to produce a comprehensive spatio-temporal model significantly increases computer-processing demands over traditional static spatial analysis. Although early spatio-temporal analysis has included dynamic cartography (Moellering 1973) and the measurement of change between maps (Leatherman 1983), it has only been in the last decade that improvements in computer-processing power has facilitated the development of analytic spatio-temporal theory and its application to fundamental research problems involving large data sets (Johnson 1990; Worboys 1996). Progress has also been made in the efficient storage and retrieval of spatio-temporal data (e.g., Langran 1992; Peuquet 1994). Additionally, spatio-temporal models have been developed conceptually and implemented in prototype spatio-temporal systems (Peuquet and Qian 1996). The work on system models has been followed by the development of spatio-temporal semantics that provide a framework for the description of spatial object interaction over time (Claramunt and Theriault 1996; Hornsby and Egenhofer 1997; Cohn et al. 1998). More powerful computers have also provided a fertile platform for the development of interactive dynamic cartographic tools (MacEachren 1995; Peterson 1995), pattern analyzers (Openshaw 1994), and integrated visualization with knowledge discovery systems (MacEachren et al. 1999) for the analysis of large spatio-temporal data sets.

This paper describes a new spatio-temporal model that extends the traditional geographic information system (GIS) model by adding a series of temporal links that represent changes in polygon topology from one spatial snapshot to its temporal neighbors. The model demonstrates how a temporal component can be added to a commercially available GIS system using the standard programming interface. As a practical illustration, we apply the model to study the role of patch size dynamics in determining mortality in an ecological system.

Temporal GIS

A temporal GIS must integrate the fundamentally different element of time within a spatial analysis system. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.