Application of wavelet analysis and a particle swarm-optimized support vector machine to predict the displacement of the Shuping landslide in the Three Gorges, China

Landslides occur frequently in the Three Gorges in China, posing threats to human life and the normal operation of the Three Gorges Dam. A number of preexisting landslides have been reactivated since the initial impoundment of the Three Gorges Reservoir in June 2003. An effective and accurate method of predicting landslide displacement is necessary to mitigate the effects of these disastrous landslides. This study carries out a landslide displacement prediction for the Shuping landslide using 7 years of monitoring data, wavelet analysis, and a particle swarm-optimized support vector machine (PSO-SVM) model. The landslide’s displacement is strongly influenced by periodic precipitation and reservoir level fluctuations, and the cumulative displacement curve versus time indicates a step-like character. Based on the deformation characteristics of this landslide, the total displacement is divided into its trend and periodic components by means of the wavelet analysis. An S-curve estimation is used to predict the trend displacement via the curve fitting of the historical displacement versus time. Five primary factors are used as the input variables for a PSO-SVM model to predict periodic displacement. These factors include cumulative precipitation over the previous month, cumulative precipitation during a two-month period, maximum continuous decrement in the reservoir level during the current month, and cumulative increments and decrements in the reservoir level during the current month. The mean squared error, squared correlation coefficient, and Akaike’s information criterion of the wavelet-PSO-SVM model at GPS monitoring points ZG85 and ZG87 are 2.45, 0.945, and 20.80 and 10.46, 80.981, and 36.38, respectively. This method can be applied to the prediction of displacement in colluvial landslides in the Three Gorges. This study may provide useful information to engineers and planners involved in landslide prevention and reduction.