A tool for construction of stochastic spatio-temporal models assimilated with observational data

Abstract

A framework and a software tool has been developed to execute two important steps in the model development cycle of temporal numerical models simulating geographic change using rules of cause and effect. The first step supported by the tool is the software implementation of spatiotemporal models. This can be done with the tool by combining generic operations on 2D map and 3D block data in an iterative script structure included in the standard Python programming language. The second step supported by the tool is the optimisation of models using the methods of genetic algorithms and particle filters. A number of case study models is presented to illustrate how the tool works.