Introduction to Unit 9: Environmental Spatial Decision Support

The previous units have looked at some of the ways in which GIS can be used to support specific tasks in environmental management such as species reintroductions, timber-felling operations, and the construction of new recreational facilities. Having looked at these specific examples, we now return to consider the broader role of GIS in helping managers to make decisions about the environment.

Environmental GIS as spatial decision support systems?

Of particular relevance is the concept of decision support systems, first introduced by Simon (1960), but later adapted for spatial data by Densham (originally in 1991 but see his later paper from 2002 below). Simon identified three types of problem:

• structured problems: These are problems which can best be tackled by software. As an example, many home insurance quotations take into account a property’s risk of flood damage. This is often based on previous history of flooding at that property, although this is also an active area of GIS research. The cost of the insurance is automatically calculated by computer and the agent selling the insurance has no opportunity to override the quotation provided by the computer.
• unstructured problems: These are problems that defy computer-based solution. The objectives involved may be too nebulous and any criteria for making a decision are likely to be qualitative and difficult to capture in a computer. Recruiting staff for an environmental organisation, for example, would still largely be regarded as too complex a task for a computer-based system.
• semi-structured problems: These are problems that fall somewhere between the other two groups. It may be difficult to define the objectives of a semi-structured problem precisely and there may be certain elements of the problem that are qualitative rather than quantitative. Computers may be used to assist in solving semi-structured problems, but the final decision rests with individual or groups rather than software. Examples might include deciding on the site of a new national park or deciding when and where would be the best time to fell stands of trees.

The distinction between these three classes of problem have changed over time. For example, some problems which were once thought of as semi-structured or unstructured (e.g. whether or not to provide a bank loan) are now completely automated.

Decision Support Systems have evolved to help resolve semi-structured problems and there has been some use of GIS as Spatial Decision Support Systems – in other words, as a specialised form of Decision Support System designed for decisions with a spatial dimension. Examples of such decisions would be delineating the area in which farmers will receive compensation for crops lost to drought or identifying the area suitable for replanting a native tree species. Typically, a decision support system can be used to explore alternative solutions, such as the relative benefits of several different sites for reforestation. Very often, instead of running the software and identifying a solution, a user (or group of users) will explore many different scenarios through a decision support system in an iterative way.

Do environmental GIS really constitute a spatial decision support system or are most decisions not actually taken on the basis of analysis and careful sifting through information at all? Or is the reverse true? Are there signs – as with estimating insurance costs for flood damage – that more and more problems are becoming amenable to solution with software packages?

---

---