Remote Sensing and Fully Distributed Modeling for Flood Forecasting

Abstract

The advantage of introducing Landsat‐derived land‐cover information into a flood‐flow forecasting model is examined. Two modeling alternatives, with and without Landsat data, were applied to 10 km by 10 km grid sizes. The landcover information is primarily derived from Landsat imagery and used directly for rainfall‐excess estimation and runoff routing. Runoff is calculated separately for each of six land‐use/land‐cover classifications for each watershed element. Without Landsat data the model is a lumped‐parameter mpdel, while with Landsat data it is a fully distributed model. The advantage of calculating runoff for each landuse/land‐cover class separately is that a watershed element can be substantially larger than a typical homogeneous hydrologic unit. The paper reports that the improvement of predicting flood peaks and total runoff gained by using Landsat data is at the 10% level of significance.