Optimizing DEM resolution inputs and number of stream gauges in GIS predictions of flood inundation

Abstract

Spatial analysis using Geographic Information Systems (GIS) is evaluated for its ability to predict the potential hazard of a flood event in the Illinois River region in the state of Illinois, USA. The data employed in the analysis are available to the public from trusted organizations such as Illinois State Geological Survey and the US Geological Survey (USGS). Since available GIS data may be limited for flood risk modeling in some parts of the world, the purposes of this study are to examine the use of spatial analysis in a GIS to determine flood inundation risk and to produce an accurate flood inundation vulnerability map employing the least amount of data. This study concentrates on areas that have stream gauge data with definable flood stage(s) and utilizes the inverse distance weighted (IDW) interpolation method on different digital elevation models (DEM) with different spatial resolutions (1 m, 10 m, and 30 m) to determine the extent of flooding over the study area. Resulting maps created for the Illinois River region yielded about 80% agreement to the effects of an actual flood event on the Illinois River near Peoria, IL on April 23, 2013. A four-gauge distribution scenario using a 10 m DEM produced the most accurate results, but all scenarios generated reasonable flood simulation. Thus, we speculate that it is possible to create a flood prediction map with a reasonable amount of accuracy using only two initial input data layers: stream gauges and a DEM.