IDRISI GIS Analysis
Like all GIS software systems, IDRISI handles both vector and raster data. However, its orientation is primarily raster. While vector systems are focused on the management of objects in space, raster systems focus on space itself. Using a fine grid data structure, raster systems have exceptional analytical power. With their powerful implicit topology, raster systems are excellent for modeling geographic processes, particularly those that involve flows over space.
Map Algebra / Modeling
Perhaps the most fundamental property of raster systems is their ability to treat map layers as variables in an equation. TerrSet’s IDRISI GIS provides an extremely rich set of logical and mathematical operators such as one would find on a scientific calculator. In this manner, mathematical models such as the Universal Soil Loss equation can be solved. The basic tools such as arithmetic operators, exponentiation, trigonometric functions and logarithmic operators can be performed either through the use of simple dialogs, or written as equations using the Image Calculator. In addition, TerrSet provides a very powerful graphical modeling tool complete with feedback loops and iterations.
The foundation for TerrSet is the IDRISI GIS suite of analytical tools. This figure illustrates some of the many GIS analysis tools provided. These include a risk mapping based on a multi-criteria evaluation, a modeling of surface runoff incorporating information on precipitation and soil infiltration and the use of the Image Calculator for basic map algebra tasks. The bottom right shows a view of the FlyThrough interactive 3-D flight viewer.
Distance and Context
Operators
The IDRISI GIS component in TerrSet provides a very rich set of tools for the assessment of distance across space including Euclidian distance, non- Euclidian cost distance where the effect of frictions to movement are accommodated, to anisotropic cost distance where the frictions are different in various directions. Tools are also provided for optimal path analysis and spatial allocation based on distance. Context operators (also known as neighborhood or focal operators) derive values based on individual cells and their neighbors. The IDRISI GIS includes a wide selection for filtering, pattern analysis and determining rates and directionality of change (such as slope and aspect). The latter can be used to describe force vectors and tools are provided for their combination to derive resultant forces.
Surface Analysis
Surface interpolation procedures in TerrSet’s IDRISI GIS include inverse distance weighting, triangulated irregular network (TIN) modeling, Thiessen polygons, trend surface mapping and Kriging. Tools are provided to derive topographic features such as slope gradients, aspect, illumination (hillshading), and curvature and facilities to delineate watersheds and viewsheds, determine surface runoff and flow patterns, evaluate sedimentation and model soil erosion.
Spatial Statistics
IDRISI provides a wide range of tools for statistical analysis of map layers including descriptive statistics, point distribution measures, autocorrelation analysis, pattern and texture measures, polynomial trend surface analysis, linear and multiple linear regression, logistic regression and multinomial logistic regression.
Decision Support and
Uncertainty Management
A hallmark of TerrSet’s IDRISI GIS has been the development of tools for multi-criteria/multi-objective decision support and uncertainty management. This development continues, and now TerrSet incorporates a major graphical modeling tool for multi-criteria and multi-objective decision support – the Spatial Decision Modeler (SDM). SDM incorporates tools for use of fuzzy sets to convert variables into comparable factors, the Analytical Hierarchy process for the derivation of factor weights, Ordered Weighted Averaging for multi-criteria evaluation and a newly designed procedure for multi-objective land allocation (MOLA). The new MOLA allows one to set targets for allocations based on either area or accumulated cost of land acquisition. In addition, it allows one to control contiguity, the number of contiguous allocations and the relative degree of compactness (specified as the minimum spanning distance). This is a first-of-its-kind implementation in the GIS industry and a major breakthrough in decision support.
TerrSet’s IDRISI GIS also contains a wealth of additional procedures for uncertainty management including error propagation tools through Monte Carlo Simulation, the evaluation of decision risk as a result of propagated error, calculation and aggregation of Fuzzy Sets, and the aggregation of indirect evidence to support a weight-ofevidence conclusion using both Bayesian and Dempster-Shafer approaches.