Estimation of Suspended Sediment Yield in Natural Rivers Using Machine-coded Linear Genetic Programming

Abstract

Estimation of suspended sediment yield is subject to uncertainty and bias. Many methods have been developed for estimating sediment yield but they still lack accuracy and robustness. This paper investigates the use of a machine-coded linear genetic programming (LGP) in daily suspended sediment estimation. The accuracy of LGP is compared with those of the Gene-expression programming (GEP), which is another branch of GP, and artificial neural network (ANN) technique. Daily streamflow and suspended sediment data from two stations on the Tongue River in Montana, USA, are used as case studies. Root mean square error (RMSE) and determination coefficient (R 2 ) statistics are used for evaluating the accuracy of the models. Based on the comparison of the results, it is found that the LGP performs better than the GEP and ANN techniques. The GEP was also found to be better than the ANN. For the upstream and downstream stations, it is found that the LGP models with RMSE = 175 ton/day, R 2 = 0.941 and RMSE = 254 ton/day, R 2 = 0.959 in test period is superior in estimating daily suspended sediments than the best accurate GEP model with RMSE = 231 ton/day, R 2 = 0.941 and RMSE = 331 ton/day, R 2 = 0.934, respectively.

The need for operational reasoning in data‐driven rating curve prediction of suspended sediment

“…They also assert that 'the results suggest that GEP may provide a superior alternative to the sediment rating curve and multiple linear regression techniques' (p. 297). Similar conclusions remain in more recent studies (cf Guven and Kisi, 2010 (1) re-modelled with two new GEP solutions developed on inputs [Q t ] and [Q t , Q tÀ1 ], using the same basic methodology and settings used by Aytek and Kisi (2008) (2) re-modelled with three additional statistical and data-driven approaches that provide a range of linear and non-linear counterparts against which the relative performance of GEP solutions can be assessed.…”

“…They also assert that ‘the results suggest that GEP may provide a superior alternative to the sediment rating curve and multiple linear regression techniques’ (p. 297). Similar conclusions remain in more recent studies ( cf Guven and Kisi, ). However, these conclusions are based on comparisons between rating curve and regression solutions and the best performing GEP solution, which is developed on operationally inapplicable input combinations (i.e.…”

“…The Tongue River has been a focus of several recent papers examining data‐driven modelling approaches in hydrology (Kisi, ; Aytek and Kisi, ; Guven and Kisi, ). Of particular importance to this study is the paper by Aytek and Kisi () in which an explicit formulation of the sediment‐discharge relationship in the Tongue River is developed using GEP.…”

Efficient methods for prediction of velocity fields in open channel junctions based on the artifical neural network

“…In the present study, hyperbolic tangent was used in hidden layers as activation function. To train ANN, Levenberg-Marquardt method was applied (Guven & Kisi, 2011;More, 1978). In this method, backpropagation algorithm is used to find the weights and bias of neural network.…”

Gene-Expression Programming for the Development of a Stage-Discharge Curve of the Pahang River

“…Keijzer and Babovic (2002) derived empirical equations using real-world hydraulic data, Giustolisi (2004) determined the Chezy resistance coefficient in corrugated channels, Kizhisseri et al (2005) explored a better correlation between the temporal pattern of a flow field and sediment transport by utilizing numerical model results and field data, and Guven and Gunal (2008) predicted local scour downstream of gradecontrol structures. In more recent applications, Guven and Kisi (2011) managed to successfully apply GP for the estimation of suspended sediment yields in a natural river using stream flow and sediment data for the Tongue River in Montana (USA).…”