SMOOTH INTEGRATION OF MODELS AND DATA INTO UNCERTAINTY AND RISK ANALYSIS: ADVANTAGES PROVIDED BY REDUCING NUMERICAL DAEMONS

A major goal of cyberinfrastructure advances is to make information more accessible when, where, and to whom it is needed. Information comes from many sources, including models and the data used in their development. The role models play in this process is currently made more difficult by the existence of “numerical daemons”, which make models produce solutions that are more nonlinear and erratic than are the phenomena they are meant to represent. This talk discusses these difficulties and the advantages derived from models designed to avoid them.