Multiobjective sensitivity analysis and optimization of distributed hydrologic model MOBIDIC

J Yang,Y Chen,F Castelli

doi:10.5194/hess-18-4101-2014

Abstract

Abstract. Calibration of distributed hydrologic models usually involves how to deal with the large number of distributed parameters and optimization problems with multiple but often conflicting objectives that arise in a natural fashion. This study presents a multiobjective sensitivity and optimization approach to handle these problems for the MOBIDIC (MOdello di Bilancio Idrologico DIstribuito e Continuo) distributed hydrologic model, which combines two sensitivity analysis techniques (the Morris method and the state-dependent parameter (SDP) method) with multiobjective optimization (MOO) approach ε-NSGAII (Non-dominated Sorting Genetic Algorithm-II). This approach was implemented to calibrate MOBIDIC with its application to the Davidson watershed, North Carolina, with three objective functions, i.e., the standardized root mean square error (SRMSE) of logarithmic transformed discharge, the water balance index, and the mean absolute error of the logarithmic transformed flow duration curve, and its results were compared with those of a single objective optimization (SOO) with the traditional Nelder–Mead simplex algorithm used in MOBIDIC by taking the objective function as the Euclidean norm of these three objectives. Results show that (1) the two sensitivity analysis techniques are effective and efficient for determining the sensitive processes and insensitive parameters: surface runoff and evaporation are very sensitive processes to all three objective functions, while groundwater recession and soil hydraulic conductivity are not sensitive and were excluded in the optimization. (2) Both MOO and SOO lead to acceptable simulations; e.g., for MOO, the average Nash–Sutcliffe value is 0.75 in the calibration period and 0.70 in the validation period. (3) Evaporation and surface runoff show similar importance for watershed water balance, while the contribution of baseflow can be ignored. (4) Compared to SOO, which was dependent on the initial starting location, MOO provides more insight into parameter sensitivity and the conflicting characteristics of these objective functions. Multiobjective sensitivity analysis and optimization provide an alternative way for future MOBIDIC modeling.

Highlights

With the development of information technology, there has been a prolific development of integrated, distributed and physically based watershed models (e.g., MIKE-SHE, Refsgaard and Storm, 1995) over the past two decades, which are increasingly being used to support decisions about alternative management strategies in the areas of land use change, climate change, water allocation, and pollution control
(2) Both multiobjective optimization (MOO) and single objective optimization (SOO) lead to acceptable simulations; e.g., for MOO, the average Nash–Sutcliffe value is 0.75 in the calibration period and 0.70 in the validation period
The procedure applied here consists of two-step analyses, i.e., a multiobjective sensitivity analysis generally characterizing the basic hydrologic processes and singling out the most insensitive factors, and a multiobjective calibration aiming at trade-offs between different objective functions

Summary

Introduction

With the development of information technology (e.g., highperformance computing cluster and remote sensing technology), there has been a prolific development of integrated, distributed and physically based watershed models (e.g., MIKE-SHE, Refsgaard and Storm, 1995) over the past two decades, which are increasingly being used to support decisions about alternative management strategies in the areas of land use change, climate change, water allocation, and pollution control. In principle, parameters of distributed and physically based models should be assessable from catchment data (in traditional conceptual rainfall– runoff models, parameters are obtained through a calibration process), these models still need a parameter calibration process in practice due to scaling problems, experimental constraints, etc. Problems arising from calibrating distributed hydrologic models include how to handle a large number of distributed parameters and optimization problems with multiple but often conflicting objectives.

Methods

Findings

Discussion

Conclusion