Tim Davies: Why floods came, and will come again

The devastation in the Bay of Plenty community of Matata was caused by erosion that was inevitable and could have been foreseen - although its timing could not be predicted - through a hazard assessment of the site.

A number of events of this type have claimed lives, including the Peel Forest flash flood of 1975, the Arthurs Pass debris flow of 1979, the Coromandel events of 1985 and 2002 and the Rees Valley tragedy of 2002. It was extremely fortunate that no lives were lost at Matata.

Many developments and dwellings are at risk from such events in all parts of the country.

Unless adequate provision is made for the identification and mitigation of these situations, more such events will destroy dwellings and more people will die.

The science of these disasters is well known. Wherever a stream flows from a steep, erosion-prone catchment onto flatter land, it deposits sediment that builds an alluvial fan. Any development on the surface of a fan, such as housing and infrastructure, faces a direct risk from flooding.

Smaller alluvial fans pose a particularly significant risk. This is because small, steep catchments with erodible rock are prone to a quite different erosion phenomenon known as a debris flow. A series of debris flows devastated parts of Matata.

These flows occur when a large amount of fine sediment enters a stream, often from a major slip on a hillside. In sufficient volume, this fine material can become a thick muddy slurry, capable of picking up and transporting rocks and boulders of virtually any size.

The flow builds into the consistency of wet concrete, gaining more debris and more destructive power as it moves down the valley.

A debris flow can also be generated by a landslide blocking a stream temporarily, then washing away.

Debris flows differ from normal flood flows in several ways: They are not steady; they tend to flow in violent surges and move quicker than water in a flood.

Unlike floodwaters, debris flows tend to ignore the course of the stream or river, especially on bends. They are more powerful, with enough force to move houses and boulders and scour a flood channel to bedrock. The debris flow carries large rocks at the front of the flow, forming a huge battering ram. It can fill a fanhead channel quickly, and travel to any part of the fan.

Debris flows occur in steep, erodible catchments of usually less than a few square kilometres and they form fans with usually more than a 5deg slope. It is logical to expect they can happen anywhere there is evidence they have occurred before - at Matata there were large boulders in gardens before the recent devastation - but often the evidence has been wiped away.

Any intense rainstorm can start a debris flow, but since the intensity of a storm is difficult to predict, so, too, is the likelihood of a debris flow.

The destruction at Matata is a reminder that any housing development on a small, steep alluvial fan could be a disaster waiting to happen.

Preventing damage from a debris flow is usually not economically feasible. Normal flood protection works are completely inadequate. Small towns cannot afford the strong protective works needed for an event that might not happen for several hundred years. Indeed, the only place this has been done in New Zealand is at the Hermitage Hotel in Mt Cook National Park.

Managing the risks from debris flows is difficult. They come without warning and travel quickly through small catchments, so any kind of warning-evacuation system becomes impractical. And, besides, it is better to be caught in a house than out in the open after evacuating too late.

Our greatest protection, therefore, is good risk assessment. To do this we must first decide what level of risk is acceptable. International experience tells us the acceptable risk of 100 deaths from a dam burst is of the order of one in a million a year.

In a situation such as Matata, the number of deaths during a debris flow is more likely to be of the order of one to 10, so the acceptable risk is of the order of one in 10,000 a year.

If Matata has a major debris flow on average at least once every 100 to 1000 years, which is highly likely, the risk is about one in 100 to one in 1000, or 10 to 100 times greater than considered internationally to be acceptable.

Compare this with the codes for the 1991 Building Act, which recommend that no dwelling should be exposed to structural damage by earthquake or wind more often than once every 475 years. If this rule is applied to debris flows, it means that any fan that experiences a flow more often than once in about 475 years on average is not suitable for housing.

At Matata there is evidence of two similar but smaller events last century before today's fanheads developed.

There has been no estimate of how many homes are built on high-risk sites, but we have all seen examples of houses and buildings located on or near a small alluvial fanhead. Many new rural developments deliberately choose fanheads because they are well-drained, above valley flood levels, have nice views and often have idyllic little streams from catchments clad in native bush bubbling past large boulders. The problem is thus increasing as rural housing development increases.

The only way to reduce the unacceptable and unmanageable risks of debris flow is to avoid development on these sites. This conclusion is inescapable, given that the science of these hazards is well known, the risk levels can be calculated with an acceptable level of accuracy, and the risks cannot be mitigated at affordable cost with protection systems.

The 2002 Civil Defence and Emergency Management Act requires civil defence groups to carry out regional hazard assessments on behalf of their councils. The intention is to identify natural hazards, such as housing on a short alluvial fan at Matata. But many local authorities have neither the expertise nor the resources to hire it, and some may not yet realise they need it.

Yet unless urgent efforts are made to identify similarly vulnerable situations, many similar and worse disasters will occur.

Future development on debris-flow fans must be prevented. Regional councils, which are responsible for preventing such disasters under the Resource Management Act, must be given the skills and resources to make competent assessments. If they are not, avoidable deaths will come.

* Dr Tim Davies is a senior lecturer in Canterbury University's department of geological sciences. He runs postgraduate courses in hazard and disaster management.