Locally Relevant High‐Resolution Hydrodynamic Modeling of River Floods at the Regional Scale

Abstract

This paper deals with the simulation of inundated areas for a region of 84,000 km 2 from estimated flood discharges at a resolution of 2 m. We develop a modeling framework that enables efficient parallel processing of the project region by splitting it into simulation tiles. For each simulation tile, the framework automatically calculates all input data and boundary conditions required for the hydraulic simulation on‐the‐fly. A novel method is proposed that ensures regionally consistent flood peak probabilities. Instead of simulating individual events, the framework simulates effective hydrographs consistent with the flood quantiles by adjusting streamflow at river nodes. The model accounts for local effects from buildings, culverts, levees, and retention basins. The two‐dimensional full shallow water equations are solved by a second‐order accurate scheme for all river reaches in Austria with catchment sizes over 10 km 2 , totaling 33,380 km. Using graphics processing units (GPUs), a single NVIDIA Titan RTX simulates a period of 3 days for a tile with 50 million wet cells in less than 3 days. We find good agreement between simulated and measured stage–discharge relationships at gauges. The simulated flood hazard maps also compare well with local high‐quality flood maps, achieving critical success index scores of 0.6–0.79.

HORA 3D: Personalized Flood Risk Visualization as an Interactive Web Service

“…For hora 3D we utilize flood simulation results already calculated in the country‐wide project hora 3.0 [BKWC*22] as well as georeferenced data acquired from open data services and authorities. Figure 2 gives an overview of the data used for our visualization.…”

“…Our web service provides detailed information about different risk factors for a whole country, involving data larger than 6 TB. The data basis is derived from the hora 3.0 project [BKWC*22, BWBK*22], where a high‐resolution digital twin of the whole of Austria with an area of about 84000 km 2 and a river network of 33880 km length was created for hydrodynamic simulations. In this project, four river flood scenarios have been simulated, from which we derive individual risk factors and visualize them via hora 3D.…”

GPU‐Accelerated Urban Flood Modeling Using a Nonuniform Structured Grid and a Super Grid Scale River Channel

“…The growing availability of new terrestrial and remote survey techniques, coupled with advancements in computing capabilities, facilitates the development of high resolution urban flood models. These models exhibit the capacity to resolve individual buildings, with an enhanced resolution in characterizing channel and floodplain hydraulics that is consistent with known processes (Bates et al, 2005;Buttinger-Kreuzhuber et al, 2022;Jung et al, 2012). In particular, increases in the abundance of data collected through Light Detection and Ranging imagery (LiDAR) provide city-scale Digital Elevation Models (DEMs) of 1 m resolution or finer, with current vertical errors of approximately 10-20 cm (Chen et al, 2012;Dottori et al, 2013;Fewtrell, Duncan, et al, 2011).…”

A Simple Model of Flood Peak Attenuation

“…PAIVA AND LIMA 10.1029/2023WR034692 2 of 22 using simplified and limited flow routing Muskingum and Muskingum-Cunge-Dooge methods (Cunge, 1969;Dooge, 1973;Koussis, 2009;McCarthy, 1938), which are commonly introduced in hydrology textbooks (e.g., Chow et al, 1988;Dingman, 2009;Ponce, 2014) and used in distributed hydrological models (e.g., Arnold et al, 1998;David et al, 2016;Thober et al, 2019). A great effort has been placed on computational hydraulics since the 50s (Cunge & Hager, 2015) and current hydrodynamic models (Teng et al, 2017) can solve full Saint-Venant or 2D shallow water equations providing a detailed representation of floods at local (USACE, 2020), regional/continental (Buttinger-Kreuzhuber et al, 2022;Hodges, 2013;Paiva et al, 2013;Siqueira et al, 2018), and global scales (Sampson et al, 2015;Wood et al, 2011;Yamazaki et al, 2011) including complex terrain, land use and hydraulic infrastructure complexities. For instance, a popular representative of current flood modeling practice is HEC-RAS software (USACE, 2020).…”