Integrative neural networks model for prediction of sediment rating curve parameters for ungauged basins

被引:39
作者
Atieh, M. [1 ]
Mehltretter, S. L. [1 ]
Gharabaghi, B. [1 ]
Rudra, R. [1 ]
机构
[1] Univ Guelph, Sch Engn, Guelph, ON N1G 2W1, Canada
关键词
Sediment; Rating curve; Artificial neural networks; Ungauged basins; FLOW-DURATION CURVES; SUSPENDED-SEDIMENT; FUZZY-LOGIC; RIVER; DISCHARGE; LOAD; WAVELET; WATER; REGIONALIZATION; TRANSPORT;
D O I
10.1016/j.jhydrol.2015.11.008
中图分类号
TU [建筑科学];
学科分类号
0813 ;
摘要
One of the most uncertain modeling tasks in hydrology is the prediction of ungauged stream sediment load and concentration statistics. This study presents integrated artificial neural networks (ANN) models for prediction of sediment rating curve parameters (rating curve coefficient alpha and rating curve exponent beta) for ungauged basins. The ANN models integrate a comprehensive list of input parameters to improve the accuracy achieved; the input parameters used include: soil, land use, topographic, climatic, and hydrometric data sets. The ANN models were trained on the randomly selected 2/3 of the dataset of 94 gauged streams in Ontario, Canada and validated on the remaining 1/3. The developed models have high correlation coefficients of 0.92 and 0.86 for alpha and beta, respectively. The ANN model for the rating coefficient alpha is directly proportional to rainfall erosivity factor, soil erodibility factor, and apportionment entropy disorder index, whereas it is inversely proportional to vegetation cover and mean annual snowfall. The ANN model for the rating exponent beta is directly proportional to mean annual precipitation, the apportionment entropy disorder index, main channel slope, standard deviation of daily discharge, and inversely proportional to the fraction of basin area covered by wetlands and swamps. Sediment rating curves are essential tools for the calculation of sediment load, concentration-duration curve (CDC), and concentration-duration-frequency (CDF) analysis for more accurate assessment of water quality for ungauged basins. (C) 2015 Elsevier B.V. All rights reserved.
引用
收藏
页码:1095 / 1107
页数:13
相关论文
共 71 条
[1]   Modelling sediment transfer in Malawi: Comparing backpropagation neural network solutions against a multiple linear regression benchmark using small data sets [J].
Abrahart, RJ ;
White, SM .
PHYSICS AND CHEMISTRY OF THE EARTH PART B-HYDROLOGY OCEANS AND ATMOSPHERE, 2001, 26 (01) :19-24
[2]  
Akaike H., 1973, 2 INT S INFORM THEOR, P267
[3]   Suspended sediment load simulation by two artificial neural network methods using hydrometeorological data [J].
Alp, Murat ;
Cigizoglu, H. Kerem .
ENVIRONMENTAL MODELLING & SOFTWARE, 2007, 22 (01) :2-13
[4]  
[Anonymous], 2002, RUSLEFAC REVISED UNI
[5]  
[Anonymous], QUANTITATIVE RISK AN
[6]  
[Anonymous], 2011, WORLD ACAD SCI ENG T
[7]   Fitting and interpretation of sediment rating curves [J].
Asselman, NEM .
JOURNAL OF HYDROLOGY, 2000, 234 (3-4) :228-248
[8]   Entropy-based neural networks model for flow duration curves at ungauged sites [J].
Atieh, Maya ;
Gharabaghi, Bahram ;
Rudra, Ramesh .
JOURNAL OF HYDROLOGY, 2015, 529 :1007-1020
[9]   A genetic programming approach to suspended sediment modelling [J].
Aytek, Ali ;
Kisi, Oezguer .
JOURNAL OF HYDROLOGY, 2008, 351 (3-4) :288-298
[10]   Comparing methods for estimating flow duration curves at ungauged sites [J].
Booker, D. J. ;
Snelder, T. H. .
JOURNAL OF HYDROLOGY, 2012, 434 :78-94