The effects of bushfires on hydrological processes using a paired-catchment analysis

Summary Several problems associated with capturing the effects of bushfires on hydrological processes in the Goulburn River catchment (6810km²), Hunter Valley, Australia are investigated using a paired-catchment analysis. It is suggested, first, that the within-year-missing data, as defined in this paper, need to be examined carefully when using the double-mass curve of annual total discharge in a paired-catchment analysis. Second, in order to provide an accurate precipitation background, which is one of the most important prerequisites for a paired-catchment analysis, the use of annual average precipitation is strongly recommended together with annual total precipitation when there are large amounts of within-year-missing data. Third, caution is needed in comparing multi-year average precipitation and streamflow data before and after the fire when the data series is not statistically long enough, since the average values for precipitation and streamflow over a different number of years may produce completely contrasting results. Fourth, the analysis of the flow duration curve, which is another useful technique in paired-catchment analysis, needs to be interpreted from the precipitation duration curve. This is because the change of flow duration curve can be caused either from the change of precipitation or the fire. Fifth, the change in streamflow, calculated by subtracting the average streamflow for the non-fire years from the observed streamflow for the years in which fires occurred, is not an efficient way of capturing the fire effects. The problem associated with this approach is not just that the streamflow is strongly dependent on rainfall, as reported elsewhere, but also, it can lead to misinterpretation using the hydrograph when the average streamflow in non-fire years is close to the average streamflow in fire years.By taking into account the above problems there was no effect of fires found on streamflow in the Goulburn River catchment. This result contrasts with the conclusions reported from other studies that have reported an immediate increase in streamflow by experimental analyses, paired-catchment analyses or modelling studies. Instead, it is shown that the spatial pattern of precipitation over the Goulburn River catchment is more important in shaping the hydrograph than the effects of bushfires. The ratio of fire extent to catchment area is approximately 4% in this study, which we suggest is a minimum area required to identify a hydrological response to the fires. The fact that other studies have focused more on capturing the generally expected effect of an immediate increase in streamflow after fire, than on why this effect occurs, makes it highly desirable to undertake micro-meteorological experiments to obtain observed evapotranspiration data before and after fire. Also, it is important to develop a coupled soil-vegetation-atmosphere-transfer dynamic mechanism and high resolution numerical weather prediction model with a distributed hydrological model in order to simulate more realistically the effect of fire on hydrological processes.