Inundation maps are a key component of Emergency Action Plans (EAP) for dams. Flood assessments using two-dimensional (2D) models used to support creation of inundation maps were until recently forced to use relatively coarse resolution due to limitations in topographic data, and exceedingly low performance of the sequential computer codes. However, often there is a need to use high-resolution models in floodplains where small scale terrain characteristics, and urban features can have significant impact on the flood behavior such as wave arrival times, and inundation patterns. The increased availability of LiDAR topography has brought the opportunity to use models with cells small enough to ensure capturing the complex urban environment. Still, many 2D modeling tools sometimes require days to run typical dam-break flooding events with this kind of resolution. GPU-based 2D hydraulic models have proven able to accelerate simulations more than 100 times with respect to conventional models, opening opportunities for detailed flood evaluations. This work describes high-resolution (<1m) flood simulations in several municipalities of Puerto Rico with the RiverFlow2D GPU model. The application involved the hydraulic modeling conducted to evaluate the flooding caused by a partial loss of the dam structures of the Carraizo Dam. The project aimed at supporting Emergency Actions Plans for the cities of Trujillo Alto, Carolina, Canovanas, Loiza and San Juan included dam break in a sunny day, dam-break during a 100-yr storm, and dam break during a Probable Maximum Precipitation (PMP) storm scenarios. The high-resolution flood maps that resulted from the analysis of each scenario can resolve flow at high resolution streets, buildings, culverts, levees, and bridges in the mesh and show the importance of mesh resolution to accurately calculate flood progression, and inundation extent. Model results also proved to be very useful to describe, in a very easy and understandable way, the flood progression during table top drills with all concerning agencies.
