THE USE OF HISTORICAL FLOOD INFORMATION IN THE IOWA RIVER BASIN TO IMPROVE RISK ASSESSMENT

摘要

Iowa River, a tributary of the Mississippi River, is one of the main sources of flood in the Iowa State. The timescale of gauging record without regulation at Iowa City is 56 years. The 1918 flood is the largest flood event in the gauged record. Flood frequency analysis, using gauged records only with Maximum likelihood estimation (MLE) and L-moment estimation methods (LME), place the 1918 event at a return period of around 100 years. However, a review of historical (pre-gauged) floods gives a different perspective. This paper demonstrates the value of augmenting a gauged river flow record with historical data for flood frequency analysis. The L-moment estimation method, considering historical floods is applied to the parameters of Generalized Extreme Value (GEV) distribution and Generalized Pareto (GPD) distribution. The result shows that incorporation of historical data into flood frequency analysis increases the reliability of risk assessment and ultimately provides a better basis for planning decisions. Meanwhile, comparison between Maximum likelihood estimation method and L-moment estimation method is also carried out. The results suggest these two methods are comparable to the time series in Iowa City.