Severe weather often becomes high impact weather when certain “tipping points” are reached. Rivers burst their banks, houses lose their roofs, and bushfires exceed suppression capacity as thresholds are crossed. The high adverse impact events tend to be rare, because society and the environment naturally tend to adapt to more frequent events. They are often small scale, or a relatively small part of a larger system. And they are often subject to considerable forecast uncertainty.

Managing the impacts of severe weather is therefore about managing risk. The results of exceeding one of these thresholds are profoundly different to merely approaching it, but the differences in the meteorology may fall within current forecast uncertainty. Balancing the costs of overpreparation and underpreparation in the presence of such uncertainty is a formidable task.

Managing risk when information is uncertain requires that we move from considering a single, “deterministic” forecast that is our best estimate of what will happen, to consideration of a range of possibilities that reliably reflects the forecast uncertainty. We need to consider not just multiple scenarios, but also their relative likelihood. This talk will review progress from around the world into objectively providing such information. Making effective use of this richer but more complex information stream is a challenge, and we will also consider progress in this area.