Continental-scale high-resolution river geometry and real-time inundation mapping
MetadataShow full item record
Flooding is the most threatening natural disaster worldwide considering the fatalities and property damage it causes. Recent flood disasters have raised concerns for accurate and responsive inundation forecast due to the rapid spread and astonishing destructive power of these events. Although recent development in large scale hydrologic simulation has enabled the real-time streamflow simulation operating on millions of river reaches, a framework for converting the forecast discharge into corresponding water surface elevation and inundation maps at a continental-scale is absent to better support local flood response. To accurately map flood inundation extent, a comprehensive description of the geometry of the channel is indispensable. As such, this dissertation presents an innovative approach for estimating river geometry and conducting inundation mapping at a continental-scale with a high spatial resolution. This approach is based on the concept of Height Above Nearest Drainage (HAND). Advanced hydrologic terrain analysis workflows have been designed to derive channel hydraulic properties, stage-discharge rating curves, and inundation extents using HAND. After the mechanism being presented, the implementation of this approach across the contiguous United States has been demonstrated using the 10-meter National Elevation Dataset. The integrity of the outputs has been validated through the comparison with best available references at multiple test sites. Considering the increasingly availability of high-resolution topographic data derived from lidar technology, the dissertation further presents how advanced geomorphic feature extraction tools are integrated into the proposed approach to overcome the challenges associated with the enrichment of terrain details. At last, this dissertation presents how banklines, an essential piece of river geometry characteristic as the boundary differentiates channel zone from floodplain, is detected with enhanced geomorphic feature extraction tools for improving large-scale hydrologic simulation and inundation mapping accuracy.