Evaluating the reliability of synthetic rating curves for continental scale flood mapping

Abstract

Empirical approaches such as the Height Above Nearest Drainage method in conjunction with Synthetic Rating Curves (HAND-SRC) have emerged as particularly appealing alternatives to the traditional flood mapping techniques due to their lower complexity and fewer data requirements. However, SRCs use Digital Elevation Model (DEM) derived reach averaged hydraulic properties and assume one dimensional steady state flow condition with normal depth. These implicit model assumptions may introduce errors in flood stage and extent estimates using the HAND-SRC approach. This study investigates the reliability of SRC across continental United States (CONUS) by comparing them to the United States Geological Survey’s (USGS) gauge rating curves. Results from this comparison show that the implicit model assumptions used in the SRC-HAND approach add significant error in the SRC derivation. The accuracy of the SRC is found to be related to the stream characteristics, including the bathymetry area, slope of the main channel two-year flow and drainage area. Results also show that SRCs in coastal areas, characterized by low slopes and large drainage areas, have higher error and tend to overpredict the stage height in comparison to the USGS rating curves; whereas they tend to underpredict stage height in the mountainous regions. The SRCs are most reliable for the midwestern plains of Ohio, Mid Atlantic, Tennessee and Upper Mississippi regions, and least reliable (higher error) for the Rocky Mountains. Further, the study finds that Deep Neural Network models can be effectively used to judge the performance of SRC for ungauged river reaches.