Comparing Of Standard And Recent Extreme Probable Distributions Corresponding To Iraqi Streams Flow

Abstract

Annual extreme flood distributions were statistically analyzed and studied to determine a best probability distribution that describes the observed annual maximum peak flows for six streams of Tigris River in Iraq. This is done through fitting of three different probability distributions and estimating the goodness of the fit to confirm a hydrologic series of 39 annual maximum peak flows of the six streams. In this study, two standard probability distributions were used as well as a new one model. The two standard distributions are the Extreme Value Type I distribution (the Gumbel's distribution) and the log Pearson type 3 distribution. While the new model is the Demetris distribution which is a recent model developed by Demetris Koutsoyiannis depending on theoretical and empirical grounds for many hydrologic series in the world. The analyses showed that the Demetris distribution is significantly overestimating the actual flows for small return periods of about less than 10 years, where it seems approximately as the two standard distributions for a return period greater than 10 years.