C4. Integration of models with decision-support systems

Many environmental decision support systems require the integration of computational models into highly interactive decision support systems (DSS) for end users. Historically such model integration has been carried out in a manual fashion, by implementing (linking) models directly into DSS code, or by wrapping models into callable services and by calling such wrapped models from the DSS code (service-based model integration). In either case, integrating the models with the DSS is a manual programming process which requires changes in the code if models need to be replaced or modified. Recent requirements (as well as recent advances) in the field of Enviromatics (Environmental Informatics, Eco Informatics) suggest that DSS need to be much more flexible and adaptable to new user needs. Part of this adaptation also effects model integration. It would, for instance, be highly desirable to be able to replace one model by another one, to cope asynch ronously with models which require large computing time, to run similar models in parallel in order to provide comparison, or to schedule large numbers of model runs for the purpose of ensemble modelling. Similarly, it has become very important to couple models in more complex work flows in order to answer complex questions combining multiple domains of expertise. Service-based model integration has been carried out since the early 2000s. Developments in the field of environmental service infrastructures and SDI's (Spatial Data Infrastructures) since approximately the mid 2000s have provided concepts and solutions to execute work flows in distributed environments. In addition, recent advances in virtualisation, for instance Cloud Computing offer new opportunities. But still the way models are integrated has not really changed - it still requires heavy coding and a large amount of intelligence in the interactive clients. This session calls for contributions which demonstr ate concepts for and practical experiences with DSS-model integration. Contributions shall help defining the problem space or shall discuss how to go beyond the traditional model described above. The aim of the session is to foster more flexible and more powerful DSS-model systems which can better satisfy end user needs.