Degree Name

Master of Engineering (Research)


Civil, Mining and Environmental Engineering - Faculty of Engineering


Due to their low infiltration rates, impervious surfaces generate large amounts of runoff. This runoff usually has high concentrations of pollutants. The impervious surface cover on a catchment therefore has a significant effect on catchment hydrology and water quality. This makes estimates of impervious surfaces critical when preparing hydrologic models of urban catchments. It could be expected that such measurements would be relatively straightforward since, unlike many other parameters used in hydrologic modeling, impervious surfaces are both directly observable and easily interpreted. However, to date measurements have relied on simple human based techniques which have limitations. Engineering research on improvement in impervious cover estimation has been limited, since the extent of impervious cover is best measured through use of a suite of spatial analysis tools that has not to date formed part of an engineer's core skill-set. In addition, remotely sensed imagery, as well as computing power, suitable for such measurement has only recently become readily available. This thesis investigates both the accuracy of impervious surface cover estimation, and the resulting effect on hydrologic predictions. For this study seven methods for impervious surface cover measurement were assessed. These include a range of human based methods and automated computer based methods applied using LiDAR, high resolution aerial photography, and multi-spectral satellite imagery. The accuracy of these methods was validated for a 925 ha semi-urban catchment located near the city of Shellharbour, NSW, Australia. The study found that the most reliable method for impervious cover estimation was Method 7 involving the use of high resolution satellite imagery and vegetation index analysis. Furthermore, it was found that the accuracy of most methods is dependent on the size of the sample considered and the type of dominant land cover. Whilst Method 7 was found to be overall the most accurate method, ultimately method selection will depend on other factors including available data, software and user skill level. As a second part of the study, the hydrologic sensitivity to accurate impervious surface measurement was then assessed for the study catchment. This involved a sensitivity analysis using two hydrologic models, an event based hydrology model (WBNM) and a continuous simulation water quality model (MUSIC). Hydrologic sensitivity was found to be significant where the modeled conditions involved low rainfall events on catchments with high infiltration. Under these conditions it was found that hydrologic estimates have a typical 25% error due to error in underlying impervious cover estimates. This is in addition to other sources of error and highlights the need for accurate impervious cover estimation.

02Whole.pdf (1015 kB)