5 How is spatial data modeled in a GIS?
5.1 Once the geographic phenomena to be portrayed have been decided upon it is necessary to determine what format to use to store and manipulate the phenomena in a GIS. There are essentially two models commonly used in a GIS today. These are the ‘Entity’ or vector model and the ‘Continuous Surface’ or raster model (Burrough and McDonnell, 1998, pp20-29 and DeMers, 1997, pp97-119). The models are not mutually exclusive, as many modern GIS are able to manipulate and analyse spatial data in either form, provided that care is taken and the potential for error is considered.
5.2 The two models treat space in different ways (Burrough and McDonnell, 1998, pp20-29 and DeMers, 1997, pp97-119):
5.2.1 Vector Model The vector model treats space as being an infinite two or three-dimensional region that is populated with distinct entities that are described by their attributes. Entities are located within the space by utilising a coordinate system. Typically entities are modeled as a set of vector primitives, i.e. points, lines and polygons.
5.2.2 Raster Model The raster model on the other hand treats space as a surface of continuous variation with no distinct boundaries. The continuous surface may be thought of as a grid, with each grid cell representing a limited, but defined portion of space that is also located by utilising a coordinate system.
5.3 The vector model is the more commonly used model in a GIS. It has been used historically due to its more precise way of modeling features and for its economy of computing resources, in that only the features of interest need be recorded. The raster model on the other hand must have a value for each grid cell (even if that value is ‘null’) and as such tends to utilise considerably more in the way of computing resources. The resolution, or size, of a single grid cell determines the precision of the raster model. The raster model is an excellent model to use when comparing and analysing continuous surfaces as it allows for the use of efficient algorithms[1].
5.4 There are variations in the vector functionality supported by GIS vendors. In addition to the vector primitives of point, line and polygon, some vendors allow for the support of higher level features, for example a cluster of points or network of lines or a region of polygons. High-end GIS applications typically use a specialised vector model called a topological model. The topological data model is an efficient way to store information about the connectedness of features; e.g. a polygon knows which lines and points are used in its construction. Using this information it is relatively easy for the GIS to determine which polygons are another’s neighbours.
5.5 Spatial data is usually grouped into themes of like geographic phenomena, with the entire theme manipulated as an entity.
5.5.1 Examples of vector themes are:
5.5.1.1 Points representing sample locations.
5.5.1.2 Lines representing creeks.
5.5.1.3 Polygons representing geological boundaries.
5.5.2 Examples of raster themes are:
5.5.2.1 Landsat image recording the reflectance values of electromagnetic radiation from surface features.
5.5.2.2 Continuous statistical surface representing the distribution of soil types derived from point samples.
5.6 In situations where there is insufficient or unsuitable vector data for use, it is becoming common practice to use a surrogate data source. An example is the use of a scanned topographic map that has been georectified and used as a backdrop to aid in the determination of location. A scanned map is in fact a raster data set.
5.7 Historically themes of spatial data have been collected for a discrete area, for example a single map area or a project area. Many organisations have a requirement to manage spatial data over a large regional or maybe continental area. The magnitude of the data to be managed, has led to many of these organisations adopting the ‘continuous map base’ approach, where themes of similar spatial entities are merged together to form a single regional dataset for that particular theme.
5.8 An emerging trend for organisations managing such large regional collections of spatial data is to move the data from a proprietary GIS data format into a proprietary format within a relational database. The data is still modeled as vector and raster data sets, however the tools available for the management of such large datasets are improved.
[1] Algorithm: ‘A set of rules for solving a problem. An algorithm must be specified before the rules can be written in a computer (programming) language’ (Burrough and McDonnell, 1998, p298).
To make suggestions or corrections to
this web page, please contact posacc@geoinnovations.com.au.
The information on this web page is © Copyright
1999 B.S.Bannerman.
All commercial rights reserved.
B.S.Bannerman materials published on the Internet are protected by copyright law. Apart from fair dealing for the purposes of private study, research, criticism or review, as permitted under the Copyright Act 1968, no part may be reproduced or reused for any commercial purposes whatsoever.