Physics-Based Modelling of Erosion in Catchment Scale

Abstract

The study of erosion is motivated by the prediction and prevention of soil loss from agricultural lands, the rehabilitation of mining and hazardous waste sites and to improve our understanding of the Earth surface response to changing climate and anthropogenic land use. This dissertation contributes to our understanding of erosion by investigating the effects of unsaturated zone hydrologic processes, rainfall distribution, and vegetation on erosion in the catchment scale. Methods used in this study include an integrated surface and subsurface hydrological model augmented with a sediment transport model for uvial sediment transport. The hydrology model solves the diffusive wave equation on the surface and Richards equation in the subsurface domain, with an exchange water flux term that couples the surface and subsurface. Sediment transport is calculated using water depth on the surface which couples sediment transport and hydrological processes. This coupling of sediment transport with hydrology allows investigation of the parameters that govern the hydrologic response of the catchment and their relation to erosion. The surface topography is updated at short regular intervals so that the ow fields accommodate the new topography and the system co-evolves in time. The coupling of erosion to the surface water flux allows an accurate prediction of hill-slope and river channel erosion for the scenarios investigated in this study. The coupled hydrologic and erosion models are used to evaluate three hypotheses. The effects of variations in the precipitation times series with a fixed annual precipitation on the surface erosion are investigated in the first hypothesis. Results indicate a large difference in model predicted erosion rates for variations in precipitation intensity, duration and dry intervals. More specifically, increase of mean intensity by multiple of two can increase bed shear stress by about 10%. The spatial distribution of the erosion is also shown to be dependent on the precipitation patterns. The high intensity and longer duration events produce erosion throughout the catchment. High intensity and short duration events produce large bed shear on the upstream regions where the slopes are steeper. The low intensity, longer duration and shorter dry intervals do not produce a significant amount of bed shear stress anywhere in the catchment as the precipitation mostly evaporates or in ltrates into the subsurface. The second hypothesis investigates the e ects of vegetation cover and duration of the growth season on the erosion and bed-load discharge from the catchment. Simulations with identical climate and hydrological properties were conducted with di erent forest vegetations that have di erent patterns of growth, canopy size and root depth. The forest types studied are categorized as forests containing mostly Deciduous, Coniferous and Hardwood trees. Changes in the canopy density a ect the surface albedo and potential evapotranspiration. Results indicate that the ground water table elevation changes and responds di erently for each vegetation type. At the growth season, the water table in the Deciduous forest is about 0:5m lower than Coniferous water table depth. The transpiration of the plants a ect the subsurface levels and runo generation mechanism change accordingly. It is shown that the Hardwood forests discharge about 5% less sediment than the Deciduous or Coniferous forests. The third hypothesis investigates the e ects of the climate change and anthropogenic induced reduction in the vegetation density on erosion. This hypothesis is evaluated by wild re and logging induced reduction of canopy size. The reduction in canopy changes hydrological response of the catchment and subsequently the erosion and sediment transport are a ected. Results indicate that a reduction of canopy size due to re or logging a ects the subsurface water table, runo generation and bank storage. These hydrologic changes facilitate faster groundwater response in the subsequent rains and increase erosion rates by 12% to 15% for leaf area index changes of 50% in logging scenarios. Furthermore, results demonstrate that the a ects of canopy size reduction are not homogeneous throughout the catchment. Taken together, the results from investigation of these three hypotheses show that variations in the subsurface ow eld due to di erent climate or vegetation change scenarios are su ciently large enough to in uence both runo generation on hill- slopes, and subsurface contributions to river discharge. The combined e ects of canopy and evapotranspiration change, surface and top soil evaporation determine the amount of runo and therefore control the erosion rates in the catchment.