Minimizing the Risk of Unsuccessful Modeling Exercisesby John J. Warwick, Univ of Texas at Dallas, United States,
Abstract: Monte Carlo simulation techniques were employed to relate input parameter uncertainties to the characteristics of model output accuracy and reliability, and to the probability of successfully applying a calibrated model. A simple Streeter-Phelps dissolved oxygen (DO) model was chosen as an exemplar. Application of these techniques allowed for optimization of the field monitoring (sampling location) and mathematical modeling (selection of calibration data set) exercise. The best average probability of attaining a DO model with an accuracy of ±0.10 mg/L was only 37%. Using the 'wrong' data set for model calibration greatly reduced this probability to only 7%.
Subject Headings: Model accuracy | Mathematical models | Risk management | Simulation models | Dissolved oxygen | Data processing | Model tests | Calibration | Probability | Europe | Monaco | Monte Carlo
Services: Buy this book/Buy this article
Return to search