Assessing the uncertainty in parameter estimation of Log Pearson type III Distribution

Design flood quantile estimation at critical locations in a river basin is essential for various hydrological applications. Regional flood frequency analysis using the L-moment-based approach offers a robust and efficient method for estimating flood quantiles at ungauged and sparsely gauged sites. The literature suggests that LH moments—higher probability-weighted moments—place greater emphasis on the tail of the distribution. This study explores the performance of the LH-moment-based approach for regions modeled using the Log Pearson Type III (LP-III) distribution, applying techniques such as the method of moments, maximum likelihood estimation, L-moments (a special case of LH-moments), and LH-moment parameter estimation. A Monte Carlo simulation experiment was conducted to assess the accuracy and reliability of these parameter estimation techniques for design flood estimation. The analysis was applied to four river basins in South India to evaluate the ability of the LP-III distribution to model annual maximum series across different climate zones (arid, temperate, and tropical). The findings have significant implications for flood risk management, infrastructure design, and policy-making, especially in regions undergoing rapid environmental changes. This research enhances the understanding of regional flood dynamics and provides a framework for more accurate flood risk assessments and improved management strategies.