Estimating future changes in flood risk: Case study of the Brisbane River, Australia

• Extreme rainfall events and the mitigating capacity of dams can affect flood risk.
• Projections favour an increase in extreme events but a decrease in annual rainfall.
• These competing effects can complicate estimates of changes to flood risk.
• A synthesis of climate model indicates a range of outcomes for the Brisbane River.
• There is no clear tendency for either an increase or a decrease in flood risk.

Estimates of potential changes to flood risk due to climate change can be of great value but are difficult to estimate for various reasons including uncertain rainfall projections and problems associated with transforming model rainfall values into runoff and inflows at relevant catchment scales. Here we attempt to estimate changes to flood risk for the Brisbane River region of south-east Queensland which has a long history of serious flood events but which now benefits from the mitigating effect of the upstream Wivenhoe Dam. In this specific case study, the existence of good quality long-term records of rainfall, a relatively large number of climate model projections and the fact that the storage levels within the dam can be reasonably simulated as a function of annual rainfall totals provides a basis for estimating possible changes to flood risk. Changes to the risk of more serious floods is assumed to depend on changes to either the magnitude or frequency of extreme rainfall events combined with changes to the amount of water actually stored in the dam. An increase in extreme rainfall events could be offset by lower annual rainfall totals that effectively increase the mitigation capacity of the Dam. We analyse the results from climate models which simulate the effect of increased greenhouse gas emissions and note that they tend to favour an increase in the former and a decrease in the latter. As a consequence, the model results indicate a range of possible outcomes with no clear tendency one way or another. This outcome reflects the fundamental nature of the climate model results for rainfall for this region and will, most likely, dominate all attempts to reduce uncertainty.