Streambank Erosion Prediction

Abstract

Bank instability as a result of flow fluctuations may lead to massive bank erosions and subsequent damage of adjacent properties. Continuous erosion process promotes change in the river morphology, sedimentation problems due to the presence of secondary currents and local scouring at piers downstream of the erosion point. Knowledge on the extent of erosion should facilitate river engineers to resolve issues on river training works and river sedimentation problems. A study has been carried out in the field to quantify the amount of eroded materials using erosion pins that were driven into the ground normal to the bank surface. The erosion pins consist of 6 mm diameter metal rods and 60 cm – 80 cm in length. A spatial variation profile for rates of erosion has been identified with units expressed in unit cm per day. The defining parameters for bank erosion rates have included near-bank velocity, Ub, water depth, Y, stream bank geometry and soil bearing capacity. Development of empirical equations had used multiple linear and nonlinear regression techniques to determine the significant erosion predictors. It takes into consideration the coefficient of determination (r-squared) and Root-mean square error (RSME) as determinants for best predictors. Accuracy of developed equations is measured using the discrepancy ratio, D.R. This is the ratio of predicted to measured erosion rate. Analysis suggest that the equation derived using polynomial function (order-2) gave better accuracy compared to the equation derived using linear and power functions. An accuracy of 75% has been obtained. Scatter plots of the predicted to the measured erosion rates have shown to be between 0.5 – 2.0 within the line of good agreement.