Estimation of regional and at-site quantiles of extreme winds under flood index procedure

Ishfaq Ahmad, Touqeer Ahmad, Usman Shahzad, Muhammad Athar Ameer, Walid Emam, Yusra Tashkandy, Zanib BadarHeliyon 2023 · DOI

TL;DR

This paper presents a robust method for estimating extreme wind quantiles in Pakistan's Sindh and Balochistan provinces, crucial for infrastructure design and disaster preparedness. By combining L-moments regional frequency analysis with a flood index procedure, the study identifies the Generalized Logistic (GLO) distribution as the most suitable for modeling wind data and provides valuable at-site quantile estimates for various return periods.

What Problem Does This Paper Address?

Extreme wind events are a growing concern globally, posing significant risks to infrastructure, human safety, and economic stability. Accurately predicting the frequency and magnitude of these events is essential for effective disaster risk reduction, urban planning, and the design of resilient infrastructure like buildings and wind turbines. Traditional at-site frequency analysis, which relies on data from a single location, often suffers from sampling variability, especially for long return periods. This limitation is particularly acute in regions with limited historical data. The study addresses this by employing Regional Frequency Analysis (RFA) to leverage data from multiple stations within homogeneous regions, thereby improving the reliability of extreme wind quantile estimates. The "flood index procedure" is then used to translate these regional estimates into site-specific quantiles, making them directly applicable for practical purposes.

Methods

The core of this study lies in the application of Linear Moments (L-moments) for regional frequency analysis. L-moments are statistical measures derived from order statistics that are known to be more efficient than traditional moments, especially with small sample sizes and outliers. The methodology involves several key steps:

  1. Data Cleaning and Screening: L-moments-based discordancy measures were used to identify and remove any inconsistent or outlier stations from the dataset.
  2. Formation of Homogeneous Regions: Cluster analysis (specifically Ward's method) and heterogeneity tests were employed to group stations into homogeneous regions. This ensures that stations within a region share similar statistical characteristics, a prerequisite for RFA.
  3. Selection of Best-Fit Probability Distribution: Five probability distributions (Generalized Normal (GNO), Generalized Logistic (GLO), Pearson Type 3 (P3), Generalized Pareto (GPA), and Generalized Extreme Value (GEV)) were fitted to the data within each homogeneous region. The best-fit distribution was selected based on the Z-statistic criterion.
  4. Parameter Estimation and Quantile Estimation: Once the best-fit distribution was identified (GLO for both regions), its parameters were estimated. The flood index procedure was then applied to convert these regional parameters into at-site quantiles for various return periods. This involves scaling the regional quantiles by the sample mean or median of each individual station.
  5. Robustness Assessment: Monte Carlo simulations were used to assess the robustness of the chosen GLO distribution by evaluating its relative bias (RB) and relative root mean square error (RRMSE) compared to other candidate distributions like GEV and GNO.

Datasets

The study utilized Daily Annual Maximum Wind Speed (DAMWS) data from 21 weather stations located in the Baluchistan and Sindh provinces of Pakistan. The data spans the period from 1990 to 2019 and was downloaded from the NASA website.

Key Insights and Ideas

  • GLO Distribution Reigns Supreme: The study found that the Generalized Logistic (GLO) distribution consistently provided the best fit for extreme wind data in both identified homogeneous regions. This is a significant finding as it offers a reliable statistical model for predicting extreme wind events in the study area.
  • Flood Index Procedure's Practicality: The successful application of the flood index procedure highlights its utility in translating regional statistical insights into actionable, site-specific estimates. This is particularly valuable for regions where direct at-site data might be scarce or unreliable for long-term predictions.
  • Homogeneity is Key: The rigorous process of identifying and validating homogeneous regions through cluster analysis and heterogeneity tests underscores the importance of this step for the reliability of RFA.
  • Lasbella: The Outlier: The identification of Lasbella as a discordant station is a crucial data cleaning step, ensuring that the analysis is not skewed by anomalous data.

Key Results

  • Two Homogeneous Regions Identified: The 21 stations were successfully grouped into two homogeneous regions based on cluster analysis.
  • GLO as Best Fit: The Generalized Logistic (GLO) distribution was identified as the best-fit distribution for both homogeneous regions, outperforming GEV and GNO in terms of relative bias and root mean square error.
  • At-Site Quantile Estimates: The study successfully generated at-site quantile estimates for various return periods (2, 5, 10, 20, 50, 100, 500, and 1000 years) for all 21 stations. These estimates are presented in Tables 9 and 10.
  • Mean vs. Median Scaling: The analysis indicated that using the sample mean as a scaling factor generally resulted in smaller standard errors for at-site quantile estimates compared to using the sample median, suggesting its superiority for this dataset.

Limitations and Future Work

  • Data Scarcity: While the flood index procedure helps mitigate data scarcity, the study acknowledges that the availability of historical data can still be a limitation.
  • Assumptions: The study mentions that the flood index method has some limitations due to statistical assumptions and data availability.
  • Future Enhancements: The authors suggest that future work could incorporate Bayesian methods and spatial phenomena into the regionalization process to further enhance the analysis. Additionally, exploring the impact of different scaling factors beyond mean and median could be beneficial.

Summary updated

Abstract

Extreme winds are becoming more common among environmental events with the most catastrophic societal consequences. A regional frequency analysis of Daily Annual Maximum Wind Speed (DAMWS) is necessary not only for a comprehensive understanding of wind hazards but also for infrastructure design and safety, wind energy potential, disaster risk reduction, insurance and risk assessment in a particular region of study. This study investigated regional frequency analysis of DAMWS of Baluchistan and Sindh provinces of Pakistan. L-moments regionalization techniques along with flood index procedure were applied to DAMWS records of 21 stations from 1990 to 2019 across the study area. We intended to find the regional frequency distribution for maximum winds and predict the returns for extreme winds events in the future. Only one station namely Lasbella was found to be discordant. With the help of cluster analysis, the remaining 20 stations were further divided into two homogeneous. Heterogeneity measures validate that both regions are homogenous with allotted stations. Regional quantiles for both regions are estimated through best-fit probability distribution among Generalized Normal (GNO), Generalized Logistic (GLO), Pearson Type 3 (P3), Generalized Pareto (GPA), and Generalized Extreme Value (GEV). Robustness of GLO distribution compared to GEV distribution is assessed through Monte Carlo simulations of relative bias and relative root mean square error. Findings clearly show that GLO distribution is the best for regional modeling. Furthermore, with the help of index flood procedure we determined at-site quantiles of all stations for various return periods. These estimated quantiles are of valuable information for various sectors, including infrastructure, energy, disaster management, and climate resilience, leading to improved planning, development, and risk reduction in the face of wind-related hazards in Sindh and Balochistan provinces of Pakistan.
BibTeX
@inproceedings{ahmad2023estimation,
  author = {Ishfaq Ahmad and Touqeer Ahmad and Usman Shahzad and Muhammad Athar Ameer and Walid Emam and Yusra Tashkandy and Zanib Badar},
  title = {{Estimation of regional and at-site quantiles of extreme winds under flood index procedure}},
  booktitle = {Heliyon},
  year = {2023},
  doi = {10.1016/j.heliyon.2023.e23388},
}
Figure