Multi-objective optimization for construction of prediction interval of hydrological models based on ensemble simulations

Practice experience reveals that prediction interval is more reliable and informative compared to single simulation, as it indicates the precision of the forecast. However, traditional ways to implement the construction of prediction interval is very difficult. This paper proposed a novel method for constructing prediction interval based on a hydrological model ensemble. The excellent multi-objective shuffled complex differential evolution algorithm was introduced to calibrate the parameters of hydrological models so as to construct an ensemble of hydrological models, which ensures a maximum of the observed data to fall within the estimated prediction interval, and whose width is also minimized simultaneously. Meanwhile, the mean of the hydrological model ensemble can be used as single simulation. The proposed method was applied to a real world case study in order to identify the effectiveness of the construction of prediction interval for the Leaf River Watershed. The results showed that the proposed method is able to construct prediction interval appropriately and efficiently. Meanwhile, the ensemble mean can be used as single simulation because it maintains comparative forecasting accuracy as the traditional single hydrological model.

[1]  Hoshin Vijai Gupta,et al.  Do Nash values have value? , 2007 .

[2]  A. Rango,et al.  MERITS OF STATISTICAL CRITERIA FOR THE PERFORMANCE OF HYDROLOGICAL MODELS1 , 1989 .

[3]  Hoshin Vijai Gupta,et al.  Decomposition of the mean squared error and NSE performance criteria: Implications for improving hydrological modelling , 2009 .

[4]  M. Franchini,et al.  Global optimization techniques for the calibration of conceptual rainfall-runoff models , 1998 .

[5]  George Chryssolouris,et al.  Confidence interval prediction for neural network models , 1996, IEEE Trans. Neural Networks.

[6]  S. Jain,et al.  Fitting of Hydrologic Models: A Close Look at the Nash–Sutcliffe Index , 2008 .

[7]  R. McCuen,et al.  Evaluation of the Nash-Sutcliffe Efficiency Index , 2006 .

[8]  Yi Liu,et al.  A Novel Multi-Objective Shuffled Complex Differential Evolution Algorithm with Application to Hydrological Model Parameter Optimization , 2013, Water Resources Management.

[9]  Henrik Madsen,et al.  Parameter estimation in distributed hydrological modelling: comparison of global and local optimisation techniques , 2007 .

[10]  João Corte-Real,et al.  Automatic Calibration of the SHETRAN Hydrological Modelling System Using MSCE , 2013, Water Resources Management.

[11]  John W. Nicklow,et al.  Multi-objective automatic calibration of SWAT using NSGA-II , 2007 .

[12]  R. Moore.  The probability-distributed principle and runoff production at point and basin scales , 1985 .

[13]  K. Sudheer,et al.  Quantification of the predictive uncertainty of artificial neural network based river flow forecast models , 2012, Stochastic Environmental Research and Risk Assessment.

[14]  Soroosh Sorooshian,et al.  Toward improved calibration of hydrologic models: Combining the strengths of manual and automatic methods , 2000 .

[15]  Yuqiong Liu,et al.  Uncertainty in hydrologic modeling: Toward an integrated data assimilation framework , 2007 .

[16]  P. Krause,et al.  COMPARISON OF DIFFERENT EFFICIENCY CRITERIA FOR HYDROLOGICAL MODEL ASSESSMENT , 2005 .

[17]  Richard H. McCuen,et al.  A proposed index for comparing hydrographs , 1975 .

[18]  S. Sorooshian,et al.  Effective and efficient algorithm for multiobjective optimization of hydrologic models , 2003 .

[19]  S. Sorooshian,et al.  Calibration of a semi-distributed hydrologic model for streamflow estimation along a river system , 2004, Journal of Hydrology.

[20]  Aidong Adam Ding,et al.  Backpropagation of pseudo-errors: neural networks that are adaptive to heterogeneous noise , 2003, IEEE Trans. Neural Networks.

[21]  Soroosh Sorooshian,et al.  A framework for development and application of hydrological models , 2001, Hydrology and Earth System Sciences.

[22]  Henrik Madsen,et al.  Automatic calibration of a conceptual rainfall-runoff model using multiple objectives. , 2000 .

[23]  Jeffrey G. Arnold,et al.  Automatic calibration of a distributed catchment model , 2001 .

[24]  V. Singh,et al.  Evaluation of the subjective factors of the GLUE method and comparison with the formal Bayesian method in uncertainty assessment of hydrological models , 2010 .

[25]  Henrik Madsen,et al.  Uncertainty assessment of integrated distributed hydrological models using GLUE with Markov chain Monte Carlo sampling , 2006 .

[26]  S. Sorooshian,et al.  Effective and efficient global optimization for conceptual rainfall‐runoff models , 1992 .

[27]  Chandranath Chatterjee,et al.  Development of an accurate and reliable hourly flood forecasting model using wavelet–bootstrap–ANN (WBANN) hybrid approach , 2010 .

[28]  John A. Nelder,et al.  A Simplex Method for Function Minimization , 1965, Comput. J..

[29]  T. Gan,et al.  Automatic Calibration of Conceptual Rainfall-Runoff Models: Optimization Algorithms, Catchment Conditions, and Model Structure , 1996 .

[30]  H. Riedwyl.  Goodness of Fit , 1967 .

[31]  Jasper A. Vrugt,et al.  Comparison of point forecast accuracy of model averaging methods in hydrologic applications , 2010 .

[32]  Henrik Madsen,et al.  Generalized likelihood uncertainty estimation (GLUE) using adaptive Markov Chain Monte Carlo sampling , 2008 .

[33]  Abbas Khosravi,et al.  Particle swarm optimization for construction of neural network-based prediction intervals , 2014, Neurocomputing.

[34]  Paul Goodwin,et al.  Do forecasts expressed as prediction intervals improve production planning decisions? , 2010, Eur. J. Oper. Res..

[35]  R. Tibshirani,et al.  An introduction to the bootstrap , 1993 .

[36]  D. Legates,et al.  Evaluating the use of “goodness‐of‐fit” Measures in hydrologic and hydroclimatic model validation , 1999 .

[37]  Henrik Madsen,et al.  Parameter estimation in distributed hydrological catchment modelling using automatic calibration with multiple objectives , 2003 .

[38]  Soroosh Sorooshian,et al.  Optimal use of the SCE-UA global optimization method for calibrating watershed models , 1994 .

[39]  Peter A. Vanrolleghem,et al.  Improved design and control of industrial and municipal nutrient removal plants using dynamic models , 1997 .

[40]  G. Kuczera.  Efficient subspace probabilistic parameter optimization for catchment models , 1997 .

[41]  K. Sudheer,et al.  Constructing prediction interval for artificial neural network rainfall runoff models based on ensemble simulations , 2013 .

[42]  J. Nicell,et al.  Evaluation of global optimization methods for conceptual rainfall-runoff model calibration , 1997 .

[43]  W. J. Shuttleworth,et al.  Parameter estimation of a land surface scheme using multicriteria methods , 1999 .

[44]  Soroosh Sorooshian,et al.  Toward improved calibration of hydrologic models: Multiple and noncommensurable measures of information , 1998 .