The Advantages of Cellular Automata Modeling in the Geographic Sciences

Cellular automata have many advantages for geographic modeling. They are capable of supporting very large parameter spaces for simulation. A one dimensional CA with a binary state set and 13 cells has 213 possible configurations. A two dimensional version of the same CA has 2169 possible configurations, and a three-dimensional CA with the same specification has 22197! If we were to divide the world’s land mass into 1 km2 rasters and run a two dimensional CA over it with 50 states (this is not unlikely, for example, if we were to develop a simple global climate model), the space of possible state switches would be 50153, 295, 000. This is a large number.

The representation of space and time in cellular automata lends an inherent spatiality to the concept. Traditional modeling techniques in geography abstract from spatial detail. Cellular automata, on the other hand, make an implicit use of space and spatial complexity. Cells, neighborhoods, and lattices are inherently spatial. In addition, cellular automata are capable of supporting separate notions of space and time, as well as combined space–time relationships. The basic cellular concept also has a natural affinity with raster based data structures common to GI systems and image processing in remote sensing. The stratification of state variables in cellular automata is also synonymous with raster layering in GI systems. Similarly, the serial treatment of temporal relationships allows for the introduction of a formal hierarchy to dynamics that complements Markov like processes already popular in geographic modeling. Moreover, cellular automata are capable of representing form and function, pattern and process.

Cellular Automata and Complexity

Cellular automata have further advantages stemming from their relationship with complex adaptive systems. Complexity studies focus on the grassroots of system dynamics, emphasizing the local interaction among elements that give rise to phenomena at synoptic scales. In modeling, the complexity approach also focuses on issues such as the importance of historical (seed) conditions, feedback between subsystems, interaction, dynamics, phase transition, noise and perturbations, and so forth. Cellular automata are among the best and most widely used tools for complexity modeling.