2024 – Quantifying Variability in Dam Outflows and Coincident Flows using Stochastic Transposition of Storms

Matthew Scorah

When undertaking flood modelling for dam failure consequence assessment there is a tension between the need to model natural complexity and variability in flood producing storms and pragmatism informed by computational and data limitations. In Australian Rainfall and Runoff 2019 there has been a shift in hydrologic modelling practice towards greater detail when representing the temporal characteristics of storms. However, as a result of computational complexity in hydraulic models this detail is rarely propagated into consequence assessment.

A key input into estimating potential loss of life (PLL) is downstream natural flooding coincident with design storms upstream of the dam (“coincident flooding”). Consequence estimation is undertaken with and without dam breach in order to determine the loss of life attributable specifically to dam failure. Incremental PLL can be both sensitive to and non-linearly related to coincident flooding. Consistent with the limited temporal variability of rainfall typically represented in hydraulic models, coincident flows are often represented by a hypothetical, average value rather than flows reflective of an actual storm.

In this study, a detailed representation of rainfall spatial and temporal variability was applied in order to quantify the associated variability in coincident flows for a large dam in Queensland. Regional storms were identified and a stochastic transposition technique was used to relate the regional rainfalls to the catchment of interest. The modelled events included floods generated both by storms occurring directly over the catchment and storms where the centre is offset from the dam and the rainfall only partially covers the dam catchment. The resulting coincident flows were examined against the corresponding dam outflows. It was found that dam outflows were not a powerful predictor of coincident flows, and for a given dam outflow Annual Exceedance Probability (AEP) the coincident flows could vary significantly. This highlights the challenges associated with making pragmatic but robust simplifications of rainfall variability when undertaking hydraulic modelling and consequence assessment.

Buy this resource

$15.00

IMPORTANT NOTICE

System maintenance will be performed on the servers that host the ANCOLD Digital Guidelines, on Saturday, December 21st @ 10:00 PM AEDT which will require the servers to reboot.

Total downtime during that event could be up to 4 hours and will commence sharply at 10 pm AEDT.  During that time end users will not be able to view or access their secured content.