Investigating the performance of the empirical models of SEDMODL2, STJ-EROS and WARSEM in estimating sediment yield from forest roads with different standards

Forest roads are recognized as potential source of sediment production in forest watersheds. The aims of this research were to determine the performance of the empirical models in estimating sediment yield from forest roads and investigate the effect of technical standards of road on sediment yield.

In the forest roads of Bahramnia forestry plan, 30 segments which were suitable for sediment trap installation were determined. These segments were classified into low standard, standard and high standard segments based on some technical characteristics Such as average longitudinal slope, percentage of cutslope coverage, distance from runoff source to nearest reputation, traffic density, pavement condition, etc . Sediment yield from segments were measured after each rainfall event (from March 2019 to February 2020) with the use of trapping at the end of each segment. The location of the trap was at the lowest point and adjacent to the reputation Inside each trap, 5 graduated indicators were installed to repeat the measurements. After each rainfall, sediment height (meters) and sedimentation area (square meters) were measured to calculate sediment volume. At the end of the year, three samples were prepared with the help of cylinders from each sediment trap to measure the apparent specific gravity of the sediment.Then estimated sediment yield by SEDMODL2, STJ-EROS and WARSEM was compared with real field values.

Results indicated the amount of clay, distance of runoff to culvert, amount of silt, ditch cover, bare soil on cutslope, road surface quality and traffic volume were respectively effective factors in sediment yield from roads. Annual sediment yield from roads with low standard (0.35 ton per year) was significantly more than standard roads (0.24 ton per year) and high standard roads (0.27 ton per year). The measured sediment yield from road segments (0.27 ton per year) was considerably less than estimated values by SEDMODL2 (1.74 ton per year), STJ-EROS (1.32 ton per year) and WARSEM (1.58 ton per year) models. Statistical analysis showed that only in roads with low standards there wasn’t significant difference between measured values and estimated values by models.

From the results of present study it can be concluded that none of models can estimate the annual sediment yield from forest roads of study area and all of them shows values more than real measured values. It is suggested that other models were evaluated and localized according to local properties.