DDDAS Predictions for Water Spills

Craig C. Douglas1,4, Paul Dostert2, Yalchin Efendiev3, Richard E. Ewing3, Deng Li1, and Robert A. Lodder1

1University of Kentucky, Lexington, KY 40506-0046 

2University of Arizona, Tucson, AZ 85721-0089  

3Texas A & M University, College Station, TX 77843-3368  

4Yale University, New Haven, CT 06520-8285, USA
douglas-craig@cs.yale.edu

Abstract. Time based observations are the linchpin of improving predictions in any dynamic data driven application systems. Our predictions are based on solutions to differential equation models with unknown initial conditions and source terms. In this paper we want to simulate a waste spill by a water body, such as near an aquifer or in a river or bay. We employ sensors that can determine the contaminant spill location, where it is at a given time, and where it will go. We estimate initial conditions and source terms using better and new techniques, which improves predictions for a variety of data-driven models.