High-resolution topographical data and high-performance-computing tools for the morphodynamical modelling of realistic flood events

The application of depth-averaged two-dimensional morphodynamic models for the simulation of real-scale long-terms flood events has been historically limited not only by the lack of field data of sufficient precision, resolution and frequency for model validation and parameter calibration, but also by the enormous computational effort required by those 2D models when run in medium-to-highly refined spatial meshes. In the last decade, new topographic data acquisition processes have improved the resolution and accuracy of the river bed measurements. Moreover, recent numerical models have been able to perform hydrodynamic flood simulations with large-scale domains in acceptable computational times thanks to computing in Graphics Processing Units (GPUs). Furthermore, in the last 5 years, GPU-accelerated High-Performance-Computing (HPC) tools have been also developed for two-phase, suspended load and bedload surface flows. In this work, both lines of progress are combined with the aim of modeling in detail the topographical changes caused in a natural alluvial river channel, 12 km long section of the Cinca River at the spanish Pyrenees, as a consequence of a highly erosive long-term flood event.