Four fundamental questions to evaluate land change models with an illustration of a Cellular Automata – Markov model

Keywords: cellular automata, land change, Markov, model, simulation, validation, verification
Abstract Type: Virtual Paper Abstract

Abstract

Numerous models exist to simulate land change to facilitate communication with an audience of stakeholders. Our manuscript raises four fundamental questions to help users decide whether to use any particular model: 1) Can the user understand the model? 2) Can the audience understand the model? 3) Can the user control the model? and 4) Does the model address the goals of the specific application? Our manuscript applies these questions to a popular Cellular Automata - Markov model. We compare the software’s documentation to the model’s behavior in 120 runs. Results show that the Cellular Automata feature corrupts the Markov feature in counterintuitive ways that the model’s documentation fails to describe. The model’s counterintuitive behavior is likely to cause users to misinterpret the validation metrics and to miscommunicate with audiences. We encourage users to answer our four questions to identify more clearly the characteristics of any simulation model.