EcoRisk is developed in response to the growing demand for computer support of modeling, acquisition and utilization of the expanding environmental knowledge. EcoRisk is the basis for development of software applications for decision making in emergency situations, contingency planning and study of long-term effects on the environment. The purpose is to create an intelligent environment for spill modeling and decision support for environmental emergencies. The system integrates: substance characteristics database, geographical information subsystem, numerical models of environmental processes (propagation of the polluted substances in ground, air and water, weather forecasting, etc.), knowledge-based decision support. Knowledge processing techniques are used for applying expert knowledge for evaluation situations and giving recommendations for actions, as well as for control of the user interface and of the whole system. Special knowledge-based tool - "prompting engine" - leads the inexperienced user when preparing data for environmental modeling and choosing the most appropriate model.
EcoRisk has two layers: The Architectural Layer integrates conceptual framework, general design principles, process models, and communication protocols. The Software Components Layer consists of various reusable building blocks for application development.
Applications of EcoRisk can be utilized in:
EcoRisk allows seamless integration of various physical models for substance behavior and other physical processes. The system design permits integration of external models. The space-time kernel manages events and parameter passing between the models, the user interface and the inferencing processes.
The event-driven user interface components provide visualization of geographical information, spill location and distribution, concentration diagrams. The emergency and novice users are supported by a prompting engine that provides step-by-step instructions.
The intelligent agents represent environmental objects and the knowledge about them. They directly communicate with the space-time kernel. An example of an intelligent agent is an animal population and the knowledge about the reaction of that population to various substances. Intelligent agents have visual representation. In the example given, this is the area on the map, in which the animal population lives.
The knowledge bases for the intelligent agents and the prompting engine are represented in InfoMaps notation and are fully accessible for modifications. Inferencing mechanisms support production rules and multidirectional inferencing, fuzzy logic, neural nets, and pattern matching. Graphical editors allow domain experts to access, maintain and verify knowledge bases without expensive professional help. Additional inferencing mechanisms and knowledge representations can be integrated if required.
Applications, based on EcoRisk, are designed as open, scaleable and multiplatform systems. Various GIS systems, DBMS, CD-ROM drives, LAN and WAN, data acquisition, communications, and popular file formats can be integrated.