Abstract
Changes in streamflow volumes resulting from climatic change can affect the management of water resources. However, the detection of a statistically significant trend can be difficult because of the large interannual variability and the relatively short period of instrumental record. In this paper, five statistical tests are applied to the historical annual streamflow series of 30 unregulated Australian rivers, located in protected natural catchments, in order to detect trends or changes in the means. The Student tβstatistic is also used to describe the relationship between interannual variability and the magnitude of change and length of data required to identify a statistically significant trend. The analysis indicates that, with the current data set, there is no clear evidence to suggest that the greenhouse effect is impacting on Australian streamflow volumes. The change required to identify a statistically significant change in the mean is directly proportional to the interannual variability, and less dependent on the length of data available. The interannual variability of Australian streams is twice that of Northern Hemisphere streams, therefore, a greater change is required in Australian flow volumes before a statistically significant change can be identified. The analysis of available historical records, however, indicates that statistically significant trends can be detected in the future flow volumes if the scenarios predicted by the G'CMs are realized.