Parameterizing the spatial structure of torrential rain for flood impact assessments

In response to climate change, torrential rains undergo changes in their magnitude and temporal properties. In many places, it is predicted that these extreme short-duration rainfalls will intensify, become more frequent, and shorten in duration. In climate change impact studies, the spatial aspect of torrential rain is often overlooked, despite many case studies demonstrating its importance. The main reason why changes in rainfall structure are ignored is the lack of information derived from convection-permitting models. They are capable of simulating the rainfall at the fine spatial and temporal scales required to accurately represent torrential rainfall under various warming scenarios, although they are very computationally demanding. Recent studies have shown a relationship between rainfall spatial structure and air temperature but the relationship is not universal and varies by location. In some places, increasing temperature is associated with an increase in rainfall heterogeneity and a decrease in storm area, while in other locations it is associated with rainfall intensification and an increase in storm area. For flood impact assessments, we propose a parameterization method to change the spatial structure of extreme short-duration rainfall. The method is applicable to either simulated data derived from convection-permitting models or observed data obtained from remote sensing devices (such as weather radar) and can be adapted to a variety of global warming scenarios. We will illustrate how the method can be applied to alter the spatial profile of a design storm and demonstrate its implications for assessing changes to flood statistics.