Catchment-scale water erosion and deposition modelling : a physically-based approach

Abstract

This thesis aims at developing a catchment-scale distributed physically-based erosion model that can be a component of an erosion productivity model or surface water quality assessment model for the sustainable management of land and water resources. The erosion model, referred to as GUEST (Griffith University Erosion System Template) was chosen as it satisfies model selection criteria for the research problem this thesis is aiming to address. The study identifies limitations of the plot- scale version of the selected erosion model, and modifies its components in order to address these limitations. As the result of the modifications made, it was shown that the temporal variation of a key model parameter (soil erodibility) was not as drastic as it was before the modifications were made. This model parameter needs to be stable, if it is to be used for predictions on a site.Rainfall, runoff, surface cover, and sediment loss data used in this thesis were previously collected from a 10 ha grazing catchment at Springvale in the semi-arid tropics of Queensland, Australia, by the Queensland Department of Primary Industries. Data were collected both at the catchment outlet and from experimental plots within the catchment, ranging in size from 30 to 640 m2. Rainfall and runoff data from the experimental plots were measured at three or one minute time intervals during a storm event using tipping buckets and recorded using data loggers. Runoff at the catchment outlet was measured using a Parshall Flume and recorded using a Rustrack Chart Recorder or a Frequency Capacitance Height Recorder. Sediment loss from experimental plots was collected using Gerlach Troughs, oven-dried, and measured.Exploratory analysis was conducted to test the integrity of data collected in the catchment by comparing runoff data collected and processed in different ways. The exploratory analysis showed that the runoff data from the different measuring devices were similar both in magnitude and distribution, which boosted confidence in the quality of the data.The modified erosion model was calibrated using data collected from experimental plots selected to represent the different soil types and cover levels in the catchment. Sensitivity analyses were conducted to evaluate uncertainties in model predictions due to uncertainties in input values. It was shown that the model was more sensitive to changes in some model parameters than others.A major limitation of the existing model is the fact that it can only be applied at the plot scale. With the ever-increasing availability of tools to deal with spatial variation of topography, soils, and vegetation, such as geographic information systems (GIS), it has been recognised that the erosion model could be more useful for the sustainable management of land and water resources if it has the capacity to be applied at the catchment scale. In this thesis a linkage between the erosion model and GIS was established using what is called a qloose couplingq technique, where the model and GIS interact only through input files.A raster-based GIS was used to derive spatial data for erosion modelling. For this study, a 5m cell size has been found to be adequate on the basis of agreement between original contours and contours derived from a DEM of this cell size. Spatial input data of topographic attributes of slope, flow direction, and flow accumulation were derived from a DEM of the Springvale catchment produced by digitising contour maps of the catchment field-surveyed at a vertical interval of lm. Spatial data on soil characteristics were generated from a 1:1000 scale soil map of the catchment published in previous studies. Field-surveyed cover maps of the catchment were also digitised and used to generate spatial cover data.To enable the modified GUEST to be applied at the catchment-scale, algorithms to route water and sediment were developed. A model-GIS interface was also developed to:dnread spatial inputs of slope, flow direction, flow accumulation, soil type, and surface contact cover (which are outputs of the GIS);dncall the erosion module to carry out calculations for every cell;dnroute water and sediment from cell to cell;dnwrite the output of the model to a file in a format that can be read by the GIS without the need for further modification for mapping, and/or further spatial analyses using map algebra, as needed.The resulting distributed physically-based erosion model, referred to as SWEPS (Spatial Water Erosion Prediction System), is then used to model erosion and deposition at the Springvale catchment using parameter values calibrated using data from representative plots within the catchment. This upscaling of parameters from plot to catchment scale has implications for similar spatial modelling activities.n n n n n n n