Scholarlink

This study presents a flood frequency analysis of the Williams River at Tillegra (Station 210011), located in the Hunter River basin of New South Wales, Australia. This region of Australia is affected by frequent floods. A 93-year record (1932–2024) of annual maximum flood (AMF) data was used in the frequency analysis, with peak flows ranging from 14 m3/s (2014) to 1349 m3/s (1946). Probability distributions including Gumbel, log-normal (LN), log-Pearson Type III (LP3), Generalized Extreme Value (GEV), and Generalized Pareto (GPA) were fitted using FLIKE software, and results were compared with ARR RFFE estimates. The result shows that LP3 and GPA provide the most reliable quantile estimates across return periods, with the LP3 estimate for the 100-year flood (1568 m3/s) aligning closely with GPA and RFFE values. In contrast, GEV and LN overestimated design floods, while Gumbel underestimated them. Trend analysis using regression and non-parametric tests indicated no significant long-term change in AMF data. This is the first study for this river location which compares five probability distributions and RFFE model. The LP3 distribution is recommended for infrastructure design applications at this site.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Flood Frequency and Trend Analysis for Williams River at Tillegra in Hunter Basin New South Wales Australia

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Arfan M. Zaman,
  • Sumya Rahman,
  • Amira Rahman,
  • Ataur Rahman

摘要

This study presents a flood frequency analysis of the Williams River at Tillegra (Station 210011), located in the Hunter River basin of New South Wales, Australia. This region of Australia is affected by frequent floods. A 93-year record (1932–2024) of annual maximum flood (AMF) data was used in the frequency analysis, with peak flows ranging from 14 m3/s (2014) to 1349 m3/s (1946). Probability distributions including Gumbel, log-normal (LN), log-Pearson Type III (LP3), Generalized Extreme Value (GEV), and Generalized Pareto (GPA) were fitted using FLIKE software, and results were compared with ARR RFFE estimates. The result shows that LP3 and GPA provide the most reliable quantile estimates across return periods, with the LP3 estimate for the 100-year flood (1568 m3/s) aligning closely with GPA and RFFE values. In contrast, GEV and LN overestimated design floods, while Gumbel underestimated them. Trend analysis using regression and non-parametric tests indicated no significant long-term change in AMF data. This is the first study for this river location which compares five probability distributions and RFFE model. The LP3 distribution is recommended for infrastructure design applications at this site.