After two recent bouts of flooding already this year in Far North Queensland and more in the south-east of the state and in the Northern Rivers of New South Wales, the perennial matter of our use of floodplains is in the news once more. Specifically, the issue of building houses on flood-liable land has come to public attention again.

As part of the discussion, it has become clear that the Queensland Government has, for several years, been selling land for residential development on the floodplain of Townsville’s Ross River, while in Sydney the government of New South Wales has called in proposals for floodplain accommodation which had been rejected by councils and opposed for public safety reasons by the State Emergency Service. The objections were overturned, perhaps indicating the power of the push towards urban consolidation as a partial remedy for further development on the metropolitan fringe. One flood-liable site, next to a creek which flows to the Parramatta River, will now see the construction, on a raised pad, of a residential high-rise block within which residents will be trapped when access is lost during floods. It must be hoped that fire — a not infrequent occurrence in times of flood — does not break out in this intended building when a flood occurs.

These decisions by state governments have been made at a time of heightened concerns in eastern Australia about residential development on floodplains. For those who oppose floodplain development there have been some cases of welcome news: the Minns Government, in one of its first decisions in 2022, curtailed some planned residential and commercial development on the floodplain of the Hawkesbury River in Sydney’s north-west.

Clearly the matter remains very much alive in the community mind, especially after severe flooding in Brisbane and in Queensland’s Lockyer Valley in 2011, 2022 and again this year, Townsville in 2019 and in separate floods in February and March just gone and in Lismore and the valley of the Richmond River in 2022.

New South Wales took the lead in curtailing urban development on floodplains after disastrous floods in many parts of the state during the 1950s. After recurring political debate and considerable to-ing and fro-ing on the details of regulatory controls, the 1% Annual Exceedence Probability (so-called one-in-100 years on average) flood became the basis for the granting of residential floor-level permission on the state’s rivers. Below that estimated level, the floors of new houses could not be built.

Eventually, the 1% level became the basis for the standard on floodplains throughout Australia.

That does not mean, of course, that new residential development is fully excluded on floodplains. The Probable Maximum Flood, the highest flood thought possible at a location, is often much higher than the 1% AEP event there. On the Hawkesbury River at Windsor, for example, the estimated PMF would reach about 26 metres, or nearly seven metres above the estimated level of the 1% AEP flood. At Lismore, the equivalent difference is about four metres; at Maitland it is probably well over two. On the state’s vast western floodplains, where demand for new housing development is small, it is generally much less than a metre.

Some people argue that to end the long-term tendency for the costs of floods to go on increasing, no new housing should be permitted below the PMF level which, after all, defines the limit of floodplain land. This, if we simultaneously removed all prior development (including retailing and manufacturing), would solve the problem. But it is a solution that is worse than the problem it seeks to solve: it would, in effect, sterilise floodplains from all development apart from agricultural and recreational uses. The difference between the 1% AEP level and the level likely to be reached in a PMF at Windsor gives an indication as to how much land would be lost to almost any worthwhile purposes.

Note here that floods approaching the PMF at any location are exceedingly rare: even Lismore’s great flood in 2022 peaked nearly two metres below the assessed PMF level. The term represents a limiting case: there is no Average Recurrence Interval between PMFs measured in years, but if one were applied it would be somewhere between one in 10,000 and 1 in 1,000,000 years. It is simply not practicable in reality to apply such a stringent standard. It is futile to seek to eradicate all risk.

We will never fully overcome the problems of flooding. The best we can do is to reduce the severity of the consequences floods bring. In times gone past, of course, whole towns were moved from the lower parts of floodplains: thus in New South Wales Moama (in 1851), Gundagai (1852) and Bega and Terara were relocated to higher sites, and the same thing happened at Clermont (Queensland) in 1916. But these places were tiny, with little fixed investment beyond basic roads and houses, and at least three of them (Gundagai, Bega and Clermont) had had huge death tolls from flooding. Maitland resisted being relocated in 1830, and the Central Business District did so again after the great flood of 1955.

A rare exception in recent times was Grantham, in the Lockyer Valley, when about two thirds of the dwellings that were caught in the flood in 2011 were relocated to high ground overlooking the original town site. But Grantham, with only a few hundred residents in 2011, was small and on a per capita basis the relocation was very expensive. Moreover, about 50 dwellings still remain in the original town: their owners, by and large, would not move. At Lismore, some people who were affected in 2022 have likewise refused to relocate. Some homeless people, indeed, have moved into buildings purchased by the state government and intended for demolition.

Relocation is as hard to mandate as is evacuation during an actual flood.

We will not fully sterilise floodplains from urban growth, and wholesale or even partial relocation of towns will be rare. What might be possible is a re-evaluation of the standards of physical flood mitigation measures. The Productivity Commission, long a proponent of spending more on disaster mitigation to reduce the enormous dollar costs of relief and recovery after hazardous events, might usefully ask whether the standard for building town-protecting levees is still fit for purpose. Could some existing levees be raised to offer protection against higher floods than those of 1% AEP proportions? And should floor levels of houses in some areas (for example those in which floods substantially higher than the 1% level can occur) be built even higher than current standards require?