Qualitative Modeling for Environmental Decision Support

For the support of analysis and decision making in the environmental domain, we propose a knowledge-based system with explicit representation of expertise in the domain and the specific ecosystem in the form of qualitative dynamic models. The choice of qualitative modeling techniques is motivated by the very nature of the processes examined, the available information and style of reasoning employed by human experts. Two other challenges typical for modeling environmental systems are encountered, namely the representation of spatial distribution processes and the interaction of widely different time-scales. We will present the design of the overall system and then illustrate the chosen modeling approach in its application to the phenomenon of algal bloom in a specific river. Model-based Environmental Decision Support In a collaboration between researchers from Brazil, Germany and France, we started the development of an integrated model-based environmental decision support system called SIGMA (see [Guerrin et al. 96]). The architecture is targeted to provide assistance for the tasks of model building, situation assessment, continuous monitoring, diagnosis of problem causes and therapy planning. These tasks are closely related by sharing the knowledge base and by providing information and results to each other. Figure 1 shows the basic architecture with the modules and their dependencies. Taking data from observations and measurements from the actual ecosystem, SIGMA provides support to the user, who interactively controls the accumulation of knowledge and the search for solutions. Finally, plans derived and tested in the computational model can be applied in reality. 1 Dept. of Computer Science, Technical University of Munich, Orleansstr. 34, 81667 München, {heller, struss}@informatik.tu-muenchen.de