Throughout human history, dams have played an important role in the advancement of water security and supply.

In the modern era, the capacity of these reservoirs is immense, and often located in close proximity to high population centres. This means that there are inherent risks which need to be managed effectively and proactively.


Dr. Monte Azmi, Senior Associate Engineer – Water Resources at SMEC, has published a paper in the Journal of Natural Hazards, focusing on enhancing estimations of crucial dam breach parameters such as peak discharge, breach formation time, and average breach width. This research is pivotal for understanding and mitigating the impacts of dam failures.


Bridging the gap in dam safety: Transforming dam break modelling from subjective to statistically grounded precision

Dr. Monte Azmi, SMEC Senior Associate Engineer – Water Resources


There are many causes of dam failures, including structural defects, significant seepage, piping, overtopping, earthquakes, and landslides. Concrete dams are susceptible to instantaneous failures, while earth fill dams typically undergo a gradual breach process. This research has focused on earthfill dams.  The complex relationships between different elements of a dam breach event and limited recorded information means that estimating dam breach parameters have always been challenging.


In Monte’s career, he noticed a range of approaches, preferences and interpretation of guidelines across companies, and even experts working within a company. “The research aims to introduce an objective framework that evaluates the effectiveness of a data fusion approach compared to individual predictors, provide a simple method for quantifying estimation uncertainties, and validate or adjust a 2-dimensional dam break model based on estimated breach parameters” Monte explained.


Monte and his co-author reviewed various methodologies, from lab-scale dam break replication to Computational Fluid Dynamics and 3D modelling, and data-driven techniques, highlighting, the existing limitations due to methodologies itself and/or dependencies on reliable and sufficient data.

The scarcity of reliable recorded information and complex relationships among different elements of dam breach events often result in estimates with high uncertainties. Hence, this research adopts a systematic approach based on data fusion-based approach, presenting superiority over existing empirical equations in reducing uncertainty across all breach parameters. Moreover, the research demonstrates practical applications of these models in real world cases, aiming to improve our ability to predict and mitigate dam failures, thereby enhancing dam safety and risk management.
———— Dr. Monte Azmi, SMEC Senior Associate Engineer – Water Resources

The study’s primary focus lies in identifying key findings, concerns, and recommendations regarding dam breach simulations. It underscores the importance of international efforts to gather accurate datasets and conduct laboratory-based dam breach simulations, which could significantly enhance data-centric models. Additionally, it emphasises the necessity of considering all relevant information, particularly geotechnical data, to ensure the reliability of estimations, especially for large dams with potentially catastrophic consequences.

Practitioners are advised to exercise caution when utilising breaching equations for dams equipped with additional safety components, ensuring a thorough understanding of dam structures to derive reliable outcomes. The study also points out ambiguities and limited information regarding the ‘erodibility’ term in certain datasets, suggesting a reassessment of equations and the collection of more geotechnical data for future research. Furthermore, it stresses the need for a robust framework to filter and select the best predictors, acknowledging the significant impact of input data on data-driven approaches.

Lastly, the study highlights the importance of selecting empirical equations based on comprehensive datasets to ensure accuracy in peak failure discharge estimations, while also recognising the necessity for continuous refinement of methodologies with improvements in historical data and predictor updates.

Dam break modelling is an imprecise science involving subjective parameter selection and substantial variability in outcomes. Monte's analysis provides our industry with a statistically founded method that can provide consistency, eliminating subjectivity and improving our understanding of outcome variability.
———— Tim Rhodes, Technical Principal – Water Resources

Monte is looking forward to seeing how this research can inform the global dam industry, “On a global scale, anyone from any part of the world could replicate and or use this methodology and compare/validate against their local procedure and models that they have.”

Monte, along with his colleague Tim Rhodes will represent SMEC at the “Advances in Dam Breach Assessment webinar hosted by the Australian Water School on 8 May 2024. Here, he will speak more on the use of advanced statistical methodologies to reach the outcome, while leveraging this platform to share his expertise with professionals and researchers worldwide.