Regional frequency analysis of extreme wind in Pakistan using robust estimation methods

Area of Science:

• Climatology and Meteorology
• Statistical Hydrology
• Extreme Value Theory Applications

Background:

• Accurate quantile estimates of extreme wind speed are essential for structural design and risk assessment.
• Regional Frequency Analysis (RFA) and Extreme Value Theory (EVT) are key methodologies for analyzing extreme events.
• Wind speed data, measured at 10 meters, is critical for engineering and safety standards.

Purpose of the Study:

• To perform Regional Frequency Analysis (RFA) of Annual Maximum Wind Speed (AMWS) in Pakistan's Khyber Pakhtunkhwa province.
• To identify the most appropriate probability distributions for estimating regional quantiles of extreme wind speeds.
• To provide reliable quantile estimates for codified structural designs and policy implications.

Main Methods:

• Utilized L-moments for RFA of AMWS data from sixteen sites.
• Employed the wards method to create two homogeneous regions, confirmed by heterogeneity tests.
• Selected the best-fit probability distributions (GEV for Cluster-I, GLO for Cluster-II) using L-moments diagram and Z-statistics.

Main Results:

• Two homogeneous regions were successfully constructed from the sixteen sites.
• The Generalized Extreme Value (GEV) distribution was identified as the best fit for Cluster-I, and Generalized Logistic (GLO) for Cluster-II.
• Robustness of the clusters was validated using relative bias (RB) and relative root mean square error (RRMSE).

Conclusions:

• The study successfully applied RFA and EVT to estimate extreme wind speed quantiles in Khyber Pakhtunkhwa.
• The identified GEV and GLO distributions provide a robust basis for regional quantile estimation.
• The derived quantile estimates have direct applicability in structural design codes and policy development.