Methods to Estimate Optimal Parameters

Model, data, and parameter estimation are three fundamental elements in hydrologic process modeling and forecasting. Recent progresses in hydrologic modeling have been made toward more efficient and effective estimation of model parameters. In this chapter, classical and recently developed parameter optimization methods and their applications in hydrological model calibration are reviewed. Those methods include gradient-based optimization methods, direct search methods, and recently developed stochastic global optimization methods. A recently developed surrogate model approach, with the purpose to reduce computational burden of model which runs through replacing the hydrologic process model with a cheaper-to-run surrogate model, is also discussed. Extending from a single objective function parameter optimization, multiobjective optimization methods and their core concept in deriving trade-offs are also summarized. Examples are provided to demonstrate the strengths and limitations of optimization algorithms summarized in this chapter.