Regional Analysis of the Maximum Annual River Flow Data in East Azarbaijan Province Rivers Using the L-moments Method

Analysis of extreme values based on a random sample is very important. Accurate forecasting of the extreme values is important in management and scientific scheduling to confront with deleterious aspects of water resources, such as flood, drought, etc. In the recent decades, application of statistical distributions in hydrology to obtain the designing data has been noted by many researchers. Many of the common statistical distributions (including some of the distributions used in flood-frequency analysis) such as the Generalized Extreme Value, Generalized Logistic, Generalized Normal, Pearson Type(III), Generalized Pareto distribution, and so on are the special cases of the four-parameter Kappa distribution. In this study, the regional analysis was used for fitting the above mentioned distributions to the annual peak flow data of the East Azerbaijan province. Using the L-moments technique, the whole province was divided into the four distinct regions (lack of discordant sites) and then regional parameters and quantiles of the mentioned distributions (including the Kappa distribution) were estimated. The results indicated that the Kappa distribution as a regional distribution was fitted well to the annual maximum discharges data.