An Event-Based Downscaling Approach to Modeling Extreme Cloudburst Precipitation Events

Cloudburst events are defined as high intensity, short duration rainfall episodes that cause flash floods endangering life and property. The environments that initiate these storms are relatively complex but can be simplified to three basic conditions: moisture availability, atmospheric instability, and either a triggering mechanism or lift. These conditions vary spatially at the cloud-resolving model scale, which leaves a large gap between the spatial and temporal scales of convection-permitting models and global climate models.

Two popular strategies to close the gap across scales are empirical statistical downscaling (ESD) and regional climate modeling (RCM). These approaches simulate the local changes that global climate models do not explicitly represent, but each has some disadvantages. ESD can create large ensembles of spatially resolved precipitation projections, but it assumes a stationary relationship in time between local and large climatic scales. RCM can create dynamically consistent evolutions of atmospheric fields, but it is computationally expensive at the convection-permitting scale, which limits the number of scenarios that can be generated. Another strategy is an event-based downscaling approach termed Pseudo-Global Warming, hereafter PGW, where observed extreme storms are selected and changes in the intensity, total amount, and spatial distribution of precipitation within those storms are simulated under various scenarios. PGW retains the dynamical advantages of RCM but is efficient enough to perform broader ensembles than RCM. PGW case studies can provide insight into how historical cloudburst events may behave in future climates, which could inform urban planners.

In this study, the PGW approach is explored using the Weather Research and Forecasting (WRF) for specific observed cloudburst events. A heterogeneous vertical warming profile is calculated from analogous projected environments. This method is compared against a homogeneous +4K delta approach and an ensemble mean of vertical warming at the end-of-century from the CMIP5 RCP 8.5 scenario. Analysis of the sensitivity to PGW profile is provided to inform best practices in the event-based downscaling approach to future flood risk decision making.