A High Resolution Global Pluvial Flood Model for Predicting Physical Risk in a Changing Climate

Pluvial flooding is a phenomenon that poses a danger to life and property worldwide, regardless of distance to a water body. High spatial resolution is a prerequisite for pluvial flood modeling, however, high-fidelity pluvial models are often too computationally expensive to be simulated globally at a high spatial resolution. Common approaches to pluvial flood modeling use dynamical or depression-filling algorithms, resulting in significant development and runtime costs when running simulations on a global scale. We present a model that incorporates a uniquely efficient depression-filling model from Barnes et al. (2021) paired with a fast calculation of standard hydraulic equations. The result is an efficient solution that incorporates both depression-filling and the characteristics of overland flow without needing to run a 2-D hydraulic model. Model uncertainty was incorporated from static input parameters and precipitation amount. To predict physical risk, pluvial flood outputs were produced for six different recurrence intervals under multiple climate change scenarios through the year 2100. The model was run globally at 90-m resolution and regionally at 10-m resolution. We highlight the level of detail achieved at 10-m resolution and comment on the observed relationship between increasing resolution and compute time.

Reference:

Barnes, R., Callaghan, K. L., and Wickert, A. D.: Computing water flow through complex landscapes – Part 3: Fill–Spill–Merge: flow routing in depression hierarchies, Earth Surf. Dynam., 9, 105–121, https://doi.org/10.5194/esurf-9-105-2021, 2021.