Authors: Jose Castejon, Anzy Lee, Noelle Patterson, Belize Lane, Colin Phillips – Utah State University; Rebecca Manners Diehl — Univ. of Vermont
Title: Quantifying topographic variability as a key indicator of HAND-FIM performance
Presentation Type: Lightning Talk
Abstract: While the Height Above Nearest Drainage (HAND) flexibility and computational efficiency make it a powerful tool for near real-time, continental-scale Flood Inundation Mapping (FIM), it does show performance limitations, especially in channel settings with high topographic variability. HAND, a terrain-based approach adapted for FIM generation, utilizes topographic data as the core element of its configuration. However, to date it remains unclear how and for what range of flow stages terrain complexity affects HAND-FIM performance, and whether it can be quantified solely from topography. In this study we evaluate HAND-FIM performance in different channel and floodplain settings and propose a method to quantify topographic variability as a key indicator of HAND-FIM prediction capabilities. Here we develop and test a topography-based Signal to Noise Ratio (SNR) approach that indicates ranges of stages where HAND-FIM will perform poorly regardless of using an accurate discharge prediction and a calibrated Synthetic Rating Curve. This method uses topographic data to estimate values of SNR, at the transect scale, for different flow stages (signals), and links them to their corresponding terrain variability (noise) and performance metrics (% agreement). Our study shows that HAND-FIM performance decreases for flow stages with low SNR values, which tend to occur at very low flows and around bankfull stage and in more topographically variable landscapes. Highlighting this fundamental li