A New Model for Optimal Watershed System Regulation and Its Application

郭怀成,伊璇,周丰,陈国贤,高伟
DOI: https://doi.org/10.13671/j.hjkxxb.2012.12.027
2012-01-01
Abstract:Watershed systems are complex and nonlinear,and stochastic and cognitive uncertainties exist in the numerical description of their scheme.In this context,optimizing watershed systems for improving water quality and ecological health is essential for integrated watershed management.To support watershed system regulation and its risk-based decision-making,Bayesian Recursive Regression Tree model (BRRT) and risk-based decision-making analysis under interval-uncertainty were proposed,which coupled with Enhanced-Interval Linear Programming (EILP) formed a new indirect coupled "simulation-optimization" model.The proposed approach was applied to the Swift Creek Reservoir's nutrient TMDL allocation (Chesterfield County,VA) to identify the minimum nutrient load allocations required from eight sub-watersheds to ensure compliance with user-specified chlorophyll-a criteria.The results indicated that the proposed model is able to represent watershed system characteristics;more importantly,it improves the prediction accuracy,computational efficiency of process simulation as well as global optimality and absolute feasibility of optimization algorithm.This model could be applied for forecast and optimal controls of the analogous nonlinear systems.
What problem does this paper attempt to address?