Calibrating a watershed simulation model involving human interference - An application of multi-objective genetic algorithms

We calibrate a storm-event distributed hydrologic model to a watershed, in which runoff is significantly affected by reservoir storage and release, using a multi-objective genetic algorithm (NSGA-II). Two specific questions are addressed: What forms of the objective (fitness) function used in the optimization model will result in a better calibration? How does the error in reservoir release caused by neglected human interference or the imprecise storage-release function affect the calibration? Four statistical measures are used to formulate a multi-objective framework to achieve a good fit of the calibration of a watershed streamflow simulation model. The framework can be utilized as a tool to evaluate the suitability of commonly used statistical measures in the calibration of hydrologic models. Reservoir release is studied as a specific (and popular) form of human interference. Two procedures for handling reservoir releases are tested and compared: Treating reservoir releases to be solely determined by the hydraulic structure (predefined storage or stage-discharge relations) as if perfect, a procedure usually adopted in watershed model calibration; adding reservoir releases that are determined by the storage-discharge relation to an error term, which encompasses a timevariant human interference and a discharge function error, and finding the error term through an optimization-based calibration procedure. It is found that the calibration procedure with consideration of human interference not only results in a better match of modeled and observed hydrograph, but also more reasonable parameters in terms of their spatial distribution and the robustness of the parameter values. The calibration analysis is conducted with the Dynamic Watershed Simulation Model.