Quick Dips

Forecasting Floods

A new method for assessing the frequency of floods and droughts has been tested and applied by hydrologists at the National Institute of Water and Atmospheric Research's Freshwater Division. The work is the result of co-operation between NIWAR staff and an IBM research centre in the US.

The method is claimed to allow better assessments of the probability of rare events. Its development was prompted by the unsatisfactory flood frequency methods in use in the US and the need to ensure improved assessment of flood risk for expensive structures, such as dams, situated on or near rivers.

In New Zealand, damage attributed to floods is estimated to cost $120 million per year. Large losses were also caused by the energy crisis resulting from an unexpectedly severe drought in the area of the southern hydro lakes last winter.

Hydrologist Charles Pearson states that a new and simpler statistical technique has reduced the inaccuracies in current methods.

"The key simplification is the use of linear combinations of collected river discharge data, rather than more complicated number crunching," he says.

The method also enables data from several locations to be pooled together to provide improved estimates of hazard within a region.

For example, the data for the severe flood that struck Invercargill in 1984 can be included when assessing the flood hazard of other coastal Southland rivers. This is important because limited budgets make it impossible to collect data from all the locations where it is needed.

The new technique has already been used to estimate the frequency of large floods in the Manawatu and for helping design flood protection schemes in Christchurch. Work is now under way to begin applying it to low flows, which are excellent indicators of drought severity.