Temporal GIS

Temporal geographic information systems (GIS) (also known as spatiotemporal GIS) is a concept and an active area of research within geographic information science focusing on representation, management, manipulation, and analysis of spatiotemporal data. Time presents an important component of spatial analysis. However, conventional GIS approaches adopt a static view to model time-dependent processes; thus, they are incapable of representing the changes and dynamics of geographic phenomena. Research efforts to extend GIS with the capabilities of handling spatiotemporal data have been ongoing since the 1980s.

The efforts of augmenting GIS with the capabilities of handling spatiotemporal data have been directed primarily at developing data models for time-dependent processes and phenomena. Similar to the object-based and field-based views of space, a discrete or a continuous view of time can be adopted in a temporal GIS design. The discrete time view uses two types of time objects to describe individually distinguishable events. A time point labels the existence of an event or the occurrence of a change, while a time interval denotes the duration of a certain status of a geographic entity. On the other hand, the continuous view treats time as a stage where the evolvements of various geographic events occur. This view is particularly important for tracing and studying the process of a geographic phenomenon in which changes take place gradually rather than abruptly.

Time-stamping presents a straightforward approach to adding temporal information to geographic entities. It has been widely used in various temporal GIS applications to integrate the geographic and temporal information of real-world entities. Depending on how a timestamp is applied to an entity, temporal GIS can represent its temporal information at different levels. A timestamp can be attached to a geographic layer (usually stored as a table in a relational database), and all geographic entities in the layer will share the same temporal information. A timestamp can be attached to every single geographic entity in a layer, and each entity will have its own temporal information. Also, multiple timestamps can be applied to different attributes of a geographic entity, and the entity will have attributes with different temporal characteristics. These time-stamping methods have been used in various temporal GIS models to represent changes of geographic entities and support relevant to spatiotemporal queries and analyses.

Another straightforward approach, the snapshot model, uses a collection of snapshots to represent the status change of geographic entities in an area. Each snapshot is taken at a specific time point; thus, all geographic entities in one snapshot share the same timestamp. The sequence of the snapshots then presents the change history of geographic entities in a particular area. In the late 1980s, Langran and Chrisman proposed a space-time composite model to record accumulated [Page 2797]changes in one single composite layer. A composite layer consists of a set of space-time composites that are derived from combining the geographic entities in multiple snapshots and represent homogeneous features sharing the same history and attributes. Each composite unit in the layer has its own timestamp. A historical state of a geographic entity can be constructed by combining several composite units associated to the entity according to their temporal sequence.

...