The recent flooding in New South Wales serves as a timely reminder of the devastating impacts of floods.

A recent review of the international flood guidance in response to climate change shows there has been significant progress in incorporating the impacts of climate change into engineering design.

But what is the best way to consider the additional uncertainty due to climate change when estimating floods for planning and design purposes? And how do we design infrastructure to mitigate the risk of flooding?

How 'critical' the infrastructure is defines the level of protection it needs.

Minor infrastructure, like kerb‑side guttering, may be designed to accommodate floods that might occur (on average) once every five years. Whereas floodplain development for housing is often designed to be protected for a one in 100-year event.

Critical infrastructure, like major bridges, are designed to withstand floods that may be 10 to 20 times rarer than the one in 100-year event.

But this all assumes that flooding is a 'stationary' process. That is, the probability that a one in 100-year event occurs, can be reliably estimated from historical data, and is the same from one year to the next.

So, design is based on our current understanding of the system. However, the risk of flooding is not stationary, and probably has never been.

This is an underlying assumption in flood guidance that all the natural climatic oscillations – like El Niño and La Niña – average out in the long run.

What this is means is the likelihood of a '100 year-flood' in any one year is the same as playing on a roulette wheel that has 100 pockets; each passing of a calendar year is the same as spinning the wheel, where the chance of a '100 year-flood' is the same is the likelihood that a ball will land in the zero pocket of our roulette wheel.

The problem is, as our climate gets warmer, the risks of extreme floods are changing.

Research has shown that human influence on climate has affected the magnitude of floods on a global scale, and it appears that while small floods are getting smaller, large floods are getting bigger. In other words, we are faced with a future where the number of pockets on our extreme-flood roulette wheel are changing with time.

For example, what we now call a one in 100-year event may in the future be closer to the one in 50-year event – which means we are likely to see it more often. In other words, our roulette wheel might have 100 pockets now, but in the future it may have only 50 pockets – so the chances of landing on a big flood will be greater.

So, how do we respond to increasing flood risks?

There has been a significant shift to incorporating 'non-stationarity' in flood guidance.

Based on the best possible science, many guidelines around the world are recommending factoring up rainfall intensities for the expected future rainfall increases.

Australia is no different, already factoring in an increase to future rainfall intensities into infrastructure design.

Although this is a significant advancement, there is one key problem with this approach. It continues to assume a fixed level of design risk based on our current understanding, whereas the science is advancing incredibly quickly.

Approximately every five years, the Intergovernmental Panel on Climate Change (IPCC) releases an assessment report on the state of the climate.

Does that mean we should revisit our flood guidance every five years? Clearly this is impractical.

Engineering design is regularly revisited (retrospectively) – but it's a reactive approach, as our knowledge improves we improve and retrofit our existing design. But could we be proactive?

Adaptive design commits fully to the uncertainty that climate changes brings with it.

This type of design recognises our limited ability to project climate change into the future and allows for design and planning decisions to be updated as knowledge improves. Adaptive design implements a flexible, dynamic plan which is updated and reviewed as understanding of any future changes improves.

We know the future will be different, but how different is a question that is continually being investigated and refined. Adaptive design adopts a design solution that can be updated as the projected (climate) path we are following becomes clearer.

Banner: Getty Images