Time, Representation of

Basic Approaches for Digital Representation of Space–Time Data

There have been recent introductions of GIS products for analyzing the occurrence of discrete point data over space and time (e.g., incidence of disease or crime) and for tracking point objects as they travel through space and time (e.g., delivery vans). The Tracking Analyst from the Environmental Systems Research Institute (ESRI) is one such GIS product. Dealing with areal data, however, is more challenging. Areal objects, such as residential districts, wetlands, and areas of poor air quality, not only move but also can grow, shrink, and change shape. Implementations for dealing with this type of data currently involve improvising with the (static) raster (i.e., grid) and vector data models currently inherent in GIS.

The primary conceptual approach currently used for representation of areal space–time data in GIS is known as the snapshot approach—sequential images or snapshots of values for a given variable over a given area for a known point in time (Figure 1a). These snapshots are conceptually equivalent to slices extracted from the continuous space–time cube. Another conceptual approach for representing space–time data is what Gail Langran called “base state with amendments” (Figure 1b), where only the changes known at specific times are recorded. A third conceptual scheme is the space–time composite, which can be thought of as the result of[Page 488] stacking all successive changes on top of one another as virtual transparencies, flattening out the space–time cube into a single cumulative depiction (Figure 1c).

Figure 1 The Three Space-Based Approaches to Space–Time Representation: (a) Snapshots, (b) Base State With Amendments, and (c) Space–Time Composites

Sequent snapshots usually are implemented as gridded raster images, although vector implementation can also be done. This approach is easy to implement using the standard raster and vector data models available in existing GIS by “tricking” the software. Instead of individual grids representing a set of thematic layers (e.g., elevation, vegetation, soil type), each spatially registered layer represents the state at a given moment in time for one of these themes (e.g., a vegetation sequence). Because modern computer graphics use rasters, the gridded implementation of sequent snapshots allows these to be displayed as a movie similar to the radar map on the Weather Channel. Under software control, the display sequence can be speeded up, slowed down, or reversed. Temporal change can also be calculated by finding the differences between two snapshots.

...