Effect of Digital Elevation Model Resolution on Cross Section Generation for Flood Plain Modeling -- Seneca Nation of Indian, Cattaraugus Indian Territories

With the increase of extreme weather events in recent years, quick and precise responses are needed to save both lives and property. Flood is one of the most frequently natural hazard causing casualties and millions of dollars in damages. Although the Federal Emergency Management Agency (FEMA) provides flood maps for most streams in Erie County, most of those maps were generated in 1980s and due to the changes in weather, climate and land cover over time, the flood map generated more than 30 years ago may not represents the flood conditions nowadays. Considering updating and enhancing the accuracy of the floodplain maps, remapping the floodplain using up-to-date topographic data is needed. With the development of technology in remote sensing, gathering high resolution and accurate data become possible nowadays. In contrast to the low resolution 30-m Digital Elevation Model (DEM) generated flood maps by FEMA, the high-resolution (5-m, 2-m and 1-m DEMs) based flood maps enable to use flood modeling at much greater detail in particular for relatively flat floodplains or confluence areas of creeks and streams. In previous studies, high-resolution DEM data generated from LiDAR data source has been considered as error free data, however, that's not necessarily true; as it is shown in this study the uncertainties within the DEMs interpolated from LiDAR dataset indeed have an affect on the flood model result. Cross section analysis and uncertainties at different resolutions have a significant effect on flood modeling results. In this study of the Seneca Nation of Indian Cattaraugus Indian Territory, DEMs based on LiDAR at various resolutions resulted in variations of delineated streams. The analysis of cross section areas and water volumes showed significant differences for a fixed flood level. The relationships between DEM resolution, uncertainties and water volume results are discussed and both statistical and visualization results are presented at each analysis step.