Keywords
hillslope erosion, sediment delivery ratio, travel time
Start Date
1-7-2006 12:00 AM
Abstract
Approximately 90-95% of soil eroded on a hillslope is redeposited and does not enter a stream. However, sediment transport models such as SedNet require an estimate to be made of the actual delivery of sediment from a hillslope into a stream based on predictions of hillslope erosion. In a companion paper in this volume (Post et al. 2006) we have developed a spatially explicit hillslope delivery model based on travel time and applied it to erosion estimates (based on the Revised Universal Soil Loss Equation, RUSLE) for a small subcatchment of the Burdekin River, Weany Creek (13.5km2), at 5m grid size. In this paper we look at the effect of increasing grid spacing on the HSDR model’s predictive capability at Weany Creek. We suggest that the model is only valid in the range of grid sizes up 2500 m2 (50 m grid size). For grid sizes between 50 -100 m (which equate to ¼ and 1 channel threshold area), the model is unstable and should not be applied. For grid sizes beyond 100 m, the model is not applicable and a constant HSDR may be a suitable approximation. Our calibration term for HSDR, β, shows a strong dependency on grid size as it must compensate for the reduction in predicted erosion from the RUSLE with increasing grid size. The term γ, defining the decay rate of HSDR with hillslope travel time, appears to be grid size independent. The choice of channel threshold area is of critical importance in defining the spatial nature of HSDR and places very strong constraints on the range of grid sizes to which the model may be applied.
Evaluation of the scale dependence of a spatially-explicit hillslope sediment delivery ratio model
Approximately 90-95% of soil eroded on a hillslope is redeposited and does not enter a stream. However, sediment transport models such as SedNet require an estimate to be made of the actual delivery of sediment from a hillslope into a stream based on predictions of hillslope erosion. In a companion paper in this volume (Post et al. 2006) we have developed a spatially explicit hillslope delivery model based on travel time and applied it to erosion estimates (based on the Revised Universal Soil Loss Equation, RUSLE) for a small subcatchment of the Burdekin River, Weany Creek (13.5km2), at 5m grid size. In this paper we look at the effect of increasing grid spacing on the HSDR model’s predictive capability at Weany Creek. We suggest that the model is only valid in the range of grid sizes up 2500 m2 (50 m grid size). For grid sizes between 50 -100 m (which equate to ¼ and 1 channel threshold area), the model is unstable and should not be applied. For grid sizes beyond 100 m, the model is not applicable and a constant HSDR may be a suitable approximation. Our calibration term for HSDR, β, shows a strong dependency on grid size as it must compensate for the reduction in predicted erosion from the RUSLE with increasing grid size. The term γ, defining the decay rate of HSDR with hillslope travel time, appears to be grid size independent. The choice of channel threshold area is of critical importance in defining the spatial nature of HSDR and places very strong constraints on the range of grid sizes to which the model may be applied.
