Estimation of sediment yield in roads by combining survey sampling design and WARSEM

Excavation and embankment machineries removes vegetation cover during construction of roads, which is resulted in accelerated erosion particularly in mountainous areas. This issue has received little attention in Iran so far. In this study, WARSEM model was used in order to quantify sediment production from 12 km road in Vard-Avarad’s upstream mountainous basin, Tehran providence. To achieve an accurate estimate in any study, collecting adequate data from all population (in this case the road) is prerequisite which resulting in an increase in both time and cost of the project. Survey sampling techniques can be used to reduce the time and cost by studying a few elements (selected road segments) of the population (whole length of the road). As the first step, each 100 meters of road segments was considered as a sampling unit followed by determining the sample size equal to 17 using Cochran equation. In the next step, the selected road segments were located on the road by systematic sampling design and all model factors such as road width, gradient, lithology, surface cover of both cutslope and fillslope were measured. Sediment production was then estimated in each road segment using WARSEM model. The results showed that total annual sediment yield originated from roads is about 3384 ton (989 and 2137 ton from cutslope and fillslope respectively) and specific average load is 157 ton/ha/yr. The findings demonstrate the importance of three factors including surface cover, height of slope, erodibility of geologic formations compared to the other factors. The validation of results was tested based on standard error of estimates, comparison to the volume of debris in the foot of cutslopes and the amount of sediment excavation conducted by Tehran Roads and Urban Development Organization. The relatively low standard error of estimations and the lower cost and time spent as well suggest the privileged performance of sampling design in estimating of sediment yield despite studying only about 14% of the road length. The remained differences between the observed and estimated sediment can be explained by considering the short period of observations compared to the long-term average of sediment yield calculated by the model.