Improving event-based methods for modelling flood risk in a variable and non-stationary climate.

While it may be easy to simulate the rise and fall of a flood hydrograph, it is surprisingly challenging to reproduce the magnitude-frequency relationship of a selected flood attribute (i.e. a flood frequency curve) under historic or current conditions, particularly over the range of spatial and temporal scales relevant to flood behaviour at different locations within a single catchment. Under climate change, these challenges only get harder. This article discusses the key aspects of flood behaviour that need to be considered when deriving magnitude-frequency relationships in a non-stationary climate, as commonly required by the industry for planning and design purposes. Specific attention is given to sources of (irreducible) uncertainty that are a function of natural variability, and those sources of (reducible) uncertainty that are due to data limitations and imperfect knowledge. Discussion is provided on the design information and tools that are required to shift practice from a reliance on deterministic models to stochastic frameworks, and the hydroclimatic factors needed to accommodate the non-stationary effects of global warming. The article concludes with an outline of additional investigations needed to support the development and application of enhanced methods, and includes commentary on the disconnect between hydrologic practice and academic research.This article is part of the discussion meeting issue 'Hydrology in the 21st century: challenges in science, to policy and practice'.