Modelling the Effects of Soil Properties, Rainfall and Road Geometry to Erosion in Unpaved Roads

Erosion of soils seriously challenges the sustainability and safety in unpaved roads. It leads to faster deterioration of these roads by formation of rills and gullies in the running surface. Many factors related to the soil properties, rainfall parameters, and road geometry affect erodibility of soils at the surface of unpaved roads. However, little is known about the relationships between those factors of erodibility for a single rainfall event. This paper models the contributions of soil properties, intensity and duration of the rainfall, and road’s length and gradient to the quantity of eroded soils from unpaved roads. For a 30-minute duration and two consecutive days; rainfall intensities of 30 mm/hr, 51 mm/hr and 68 mm/hr were used to test the erodibility of soils. The tested bed surfaces were set at slopes of 0% and 6%, in a small (large)-scale testing box of 0.6 m (1.2 m) x 0.3 m x 0.17 m (length x width x height). RapidMiner Studio software was used to predict quantities of eroded soils based on the measured eroded soils under the same influencing factors of erodibility. Six predictive models were developed based on the firstand second-day rainfall events. The predictive models can perform well with the Nash and Sutcliffe’s coefficients of efficiency (ME) ranging from 0.62 to 0.74. Also, clay content and mean particle size of the surface soils, rainfall intensity and slope gradient were the most contributing factors to the quantity of eroded soils from unpaved roads.