Resource

Comparison of Flood Hazard Estimation Methods for Dam Safety - Phase 1 Task 1

Resource Type
Reports
Reference Title
Comparison of Flood Hazard Estimation Methods for Dam Safety - Phase 1 Task 1
Author/Presenter
Sayers, P.
Bowles, D.
Nathan, R.
Rodda, H.
Tomlinson, E.
Gippel, C.
Little, M.
Organization/Agency
CEATI
Sayers and Partners
Publisher Name
CEATI
Year
2014
Date
05/2014
Document Number
T112700 0225
Abstract/Additional Information

This project provides a structured review of the available and evolving methods for estimating flood flows, where they are used and, importantly, the assumptions and limitations they contain. The project outputs are provided in two task reports.

The Task 1 report provides a review of the regulatory frameworks in place around the world and their impact on the estimation of flood flows, and is covered here. A second, Task 2 report focuses on the estimation approaches themselves and is reported separately.The Task 1 review has highlighted that no explicit relationship exists between formal regulations and the methods for estimating the flood hazard. This finding is consistent with the conclusions of the ICOLD Committee on Dam Safety’s draft report on Dam Safety Regulation. The review also confirms that standards-based approaches continue to dominate formal regulations. In those jurisdictions where the supporting guidance promotes the use of risk approaches to help decide whether or not the legislation or regulations have been met, this does have a number of implications for flood hazard estimation. The most relevant of these include the need to:

(i) Account for aleatory uncertainty (arising from natural variability in climate and hydrologic inputs) and estimate a full distribution of flood flows (rather than a single design flow event characteristic of a traditional standards-based approach).

(ii) Explicitly recognize the epistemic uncertainty (arising from errors in data, models and model structure).

(iii) Recognize the stochastic nature of the system response and accept that some uncertainties are irreducible and cannot be resolved through ever more detailed data and modeling.