A rapid, out-of-the-box, regional flood modelling framework using FastFlood for Canadian case studies

Developing a scalable, out-of-the-box flood-modelling framework that performs quickly and reliably across diverse hydrological and geomorphological contexts remains a major challenge in large-scale flood risk assessment. Global flood models increasingly aim to reduce dependence on locally available datasets, yet the limited availability of high-resolution data constrains the reproducibility and transferability of existing modelling approaches.  

In this study, we explore and evaluate an out-of-the-box flood-modelling approach using the FastFlood tool. FastFlood is designed for rapid fluvial and pluvial flood assessment and is supported by globally available datasets for topography, land cover, and soil parameters, enabling flood simulations to be initialized even where local inputs are sparse or absent. We implement a structured, multi-scenario framework that investigates performance across low-detail, globally parametrized runs to high-detail, calibrated configurations.  

We will present this new approach through a series of case study applications illustrating its performance across varying levels of detail and contrasting hydrological conditions. These cases demonstrate the method’s potential for deployment in flood-prone regions facing data limitations, supporting advances in global early warning, rapid impact assessment, and anticipatory action. A detailed case study is presented with focus on the Ontario province in Canada, with validation of the model at several levels of detail with historic and reference simulations using HEC-RAS, obtaining extent-wise similarity with the FastFlood output of 98.5 percent over the entire Trent river course.