How to reuse and modify an existing land use change model? Exploring the benefits of language-centered tool support

Abstract

A model-driven approach with domain-specific modeling languages (DSLs) aims to support both, the transparency of modeling processes and the reusability of simulation models in computer-based system analysis. However, in contrast to general-purpose programming languages used as modeling languages, a single DSL is not expressive enough to formalize all necessary aspects of computer-based system analysis, since DSLs are tailored to serve a specific purpose, e.g. process modeling or the description of experimentation and data analysis. Thus, we propose a languagecentered approach to tool support for environmental modeling and simulation (EMS), where all relevant aspects are specified using different coupled tool-independent DSLs. Extending previous work we propose to use functionality of Geographic Information Systems (GIS) for data analysis by inclusion of a DSL that is based on the semantics of GIS. We show how such GIS-based DSLs can be integrated with DSLs for system and experimentation modeling. We apply this language-centered approach to land use change modeling in the Greater Tirana region, using the wellestablished SLEUTH approach. Despite sophisticated data processing, the available data for the study area does not satisfy the requirements of the original SLEUTH model, thus modifications are required. Modifications apply to both, the land use change model defined by a Cellular Automaton and the experimentation model including a calibration routine that is based on data analysis. In conclusion, we show that the language-centered approach to tool support facilitates the transparent modification of an existing well-established scientific model. In contrast to modeling with general-purpose programming languages, this should enhance the tool support for sound transparent scientific modeling processes.