Low flow frequency analysis by L-moments method (Case study: Iranian Central Plateau River Basin)

Abstract

Knowledge about low flow statistics is essential for effective water resource planning and management in ungauged or<br />poorly gauged catchment areas, especially in arid and semi-arid regions such as Iran. We employed a data set of 20 river<br />flow time-series from the Iranian Central Plateau River Basin, Iran to evaluate the low-flow series using several frequency<br />analysis methods and compared the result of these methods in their ability to set low flows for the conservation of<br />instream water uses. Theoretical frequency distributions including the log-normal, three-parameter lognormal, Gumbel,<br />Pearson type III, and log Pearson type III were applied with the low-flow series. Goodness-of-fit tests including Lmoment<br />and conventional moment methods for the observed data were applied to identify the best distributions. For each<br />distribution, the calculated values of the residual sum of squares (RSS) was applied to compare between the conventional<br />moment and L-moment methods, and the best method was selected to determine the most appropriate probability<br />distribution. The lowest RSS values were used to select the best distribution for each station. Then, T-year low-flow series<br />was estimated using the best probability distribution. Our results suggested that, for annual low flows, based on the<br />computed RSS, Pearson type III, log Pearson type III, Gumbel distributions, and L-moment method were suitably<br />distinguished for 85%, 10%, and 5% of the stations, respectively. Finally, Compared to the conventional moment method,<br />L-moments method was found to be more adequate to identify low-flow series probability distributions in the Iranian<br />Central Plateau River Basin, while Pearson type III was found to be the best probability distribution for modeling<br />minimum flow series in the study area.