Presented by Brian Kyanjo – Computational Math Science and Engineering emphasis
Overland flooding is a natural phenomenon that can have disastrous consequences, particularly in low-lying areas or regions with inadequate drainage systems. To predict the extent, velocity, bathymetry, and depths of floodwaters, GeoFlood simulates flood flow behavior over a two-dimensional domain, taking into account factors such as topography, surface roughness, and flow discharge. The model is based on fluid dynamics principles and employs finite volume methods to solve the depth-averaged shallow water equations that govern flood flow. Real-world data from previous flood events, such as the 1959 Malpasset floods, is used to calibrate and validate the model, and its performance is compared to that of existing flood models. Preliminary results demonstrate the model’s ability to accurately simulate flood propagation and inundation patterns for the Malpasset dam break scenario. Standard benchmark tests will be used to assess the model’s performance in terms of accuracy, efficiency, and scalability. The model will be accelerated using Compute Unified Device Architecture (CUDA), and its performance will be evaluated on a variety of parallel computing architectures. This model’s development will improve flood forecasting and emergency response planning, lowering the risks and impacts of overland flooding on communities and infrastructure.