Ahmed et al. (2026) Comparison of Best and Poor Performing Catchments in Regional Flood Frequency Analysis in Southeastern Australia

Identification

• Journal: Lecture notes in civil engineering
• Year: 2026
• Date: 2026-01-01
• Authors: Ali Ahmed, Zaved Ahmed Khan, Ataur Rahman
• DOI: 10.1007/978-3-032-18708-6_7

Research Groups

• Charles Sturt University, Orange, NSW, Australia
• School of Engineering, Design and Built Environment, Western Sydney University, Sydney, New South Wales, Australia
• CSIRO Environment, Canberra, Australia

Short Summary

This study investigates factors contributing to substantial inaccuracies in Regional Flood Frequency Analysis (RFFA) in southeastern Australia. It finds that variability in predictors and hydrological coherence are more critical for model performance than regional homogeneity, with the Quantile Regression Technique (QRT) demonstrating superior robustness over the Index Flood Method (IFM) in heterogeneous catchments.

Objective

• To investigate the factors contributing to poor performance (relative errors exceeding 50%) in Regional Flood Frequency Analysis (RFFA) for flood quantile estimates in southeastern Australia.

Study Configuration

• Spatial Scale: 201 gauged catchments across southeastern Australia, specifically in New South Wales and Victoria.
• Temporal Scale: Analysis for 1-in-20 annual exceedance probability (Q20), representing long-term flood frequency estimation.

Methodology and Data

• Models used: Log-log linear model, Quantile Regression Technique (QRT), Index Flood Method (IFM). Model performance was evaluated using a leave-one-out (LOO) validation approach.
• Data sources: Streamflow data from 201 gauged catchments, provided by WaterNSW and the Australian Rainfall and Runoff Revision Project 5 team.

Main Results

• Catchments were classified into best-performing (lowest 25% absolute relative error, ARE) and poor-performing (highest 25% ARE) groups.
• Variability in predictors and hydrological coherence were found to exert a stronger influence on model performance than regional homogeneity alone.
• The Quantile Regression Technique (QRT) demonstrated superior robustness compared to the Index Flood Method (IFM) in heterogeneous catchments.
• Reliable flood quantile estimation in southeastern Australia requires attention to both regional homogeneity and careful consideration of predictors’ coherence.

Contributions

• Identifies specific hydrological and predictor characteristics (variability in predictors, hydrological coherence) as more influential than traditional regional homogeneity in determining RFFA performance.
• Provides a comparative evaluation of QRT and IFM, establishing QRT's superior robustness in heterogeneous catchment conditions.
• Offers practical guidance for improving flood quantile estimation in southeastern Australia by emphasizing predictor coherence alongside regional homogeneity.

Funding

• Australian Rainfall and Runoff Revision Project 5 team (provided data).
• FLIKE sales team (granted free access to TUFLOW FLIKE software).
• WaterNSW (supplied streamflow data).

Citation

@article{Ahmed2026Comparison,
  author = {Ahmed, Ali and Khan, Zaved Ahmed and Rahman, Ataur},
  title = {Comparison of Best and Poor Performing Catchments in Regional Flood Frequency Analysis in Southeastern Australia},
  journal = {Lecture notes in civil engineering},
  year = {2026},
  doi = {10.1007/978-3-032-18708-6_7},
  url = {https://doi.org/10.1007/978-3-032-18708-6_7}
}