Management and visualization of spatiotemporal information in GIS

Abstract

Although Geographic Information Systems (GIS) have been recognised as the most advanced technology for the management of geospatial information, they are still unable to efficiently manage the temporal dimension. Originally this problem affeeted only the study and analysis of highly dynamic phenomena. Today's expansion of GIS technology, the ease to acquire and store geospatial data and the increased capacity of computing technologies to managc large amounl of data have contributed lo the propagation of this problem across the whole geospatial seclor. The extended use of GIS in decision-making processes is increasing the demand for tools able to manage and 10 analyse dynamic geospatial phenomena where the temporal dimension is crucial. The only temporal model available in commercial GIS packages is based on discretisation of temporal data. Changes are represented as a succession of snapshots. The dynamics and what happens between those stages are not registered. In addition, this approach presents severe problems due to unavoidable multiplication of data volume, abundant redundancies, loss in query efficiency and the impossibility of knowing when the exact timing of changes occurs. Since the late 1980s and particularly in the 1990s, researching the temporal changes and the conceptual and technological options available has been undertaken by the GIS and DBMS sectors. The primary objective of the research presented in this paper is the development of a model for the integration of temporal data with GIS. The method adopted to achieve this objective is based on the combination of Time Geography principies, its graphic language and dynamic segmentation techniques used in GIS. Past research has demonstrated that the difficulty to integrate time with GIS has its origin in the continuous nature of time. Dynamic segmentation in GIS network analysis has the potential to provide the means for a time-GIS integration in a continuous manner. Lifelines, one of the main Time geography's graphic language elements, has been modelled as a set of network segments where the dynamics in attribute information has been attached to different time segments rather than distance segments (for exampIe Euclidean or cost-based) as normally occurs in dynamic segmentation. This paper summarises initial findings of the project. These outcomes have the potential to improve the way the geospatial sector currently handles temporal information. However, the static nature of current GIS technology impedes an appropriate visualisation of dynamic temporal phenomena. To this effect, the paper also explores the possibilities offered by multimedia techniques as a complement to GIS capabilities.