The claim

More than 1 billion animals are said to have died in Australia’s bushfires so far this fire season.

Chris Dickman, an ecology expert from the University of Sydney, initially suggested on January 3 that as many as 480 million animals were likely to have died in the NSW fires.

Less than a week later, he updated that number to 800 million and projected “nationally” that more than 1 billion had died.

“The 480 million estimate was made a couple of weeks ago, and the fires have now burnt over a large area of further country. That means over 800 million mammals, birds and reptiles have been affected by the fires. Australia-wide, it’s probably over a billion,” Dickman told US radio station KOSU.

Is this correct? RMIT ABC Fact Check investigates.

The verdict

Dickman’s claim is a conservative estimate.

In making the calculations, he has consciously understated animal density rates for areas that previously were bountiful habitats for diverse Australian wildlife species.

Using figures drawn from a 2007 research paper on the number of animals per hectare, and multiplying them by the extent of land affected by fire, he has concluded that more than 1 billion animals, birds and reptiles have been lost in NSW alone.

However, he has intentionally understated the animal density figures of the traditionally wetter, richer eastern ranges habitat by using lower density rates that apply in sparser, drier regions.

The calculation does not include estimates for the number of platypuses or bats lost.

As well, his estimate is based on the extent of fire-affected areas in NSW and Victoria only; it does not include areas affected by fire in other states.

How do the numbers stack up?

Dickman’s calculation is based on native animal density estimates in certain habitats in NSW.

Those density figures were published in a 2007 research paper for the World Wildlife Fund co-authored by Dickman, Dr Hal Cogger, Professor Hugh Ford and Professor Chris Johnson.

The 2007 paper was commissioned in response to concerns about how government-approved destruction of native vegetation, and consequent disruptions to habitats, were affecting wildlife.

The researchers drew on a variety of detailed field surveys and specific species studies to assess the impact of land clearing on mammals, birds and reptiles in NSW.

At that time, they concluded that “104 million native animals, birds and reptiles have died or will die as a result of the clearing of native vegetation approved by the NSW government between 1998 and 2005”.

The researchers categorised the state of NSW into two broadly distinct geographical and climatic regions:

  • Coastal and eastern ranges
  • Tablelands, western slopes and plains.

Table 3 of the 2007 paper shows density rates (the number of animals per hectare) for each of those areas across classes of animals: echidnas; koalas; the common wombat; possums and gliders; kangaroos, wallabies and rat-kangaroos; bandicoots; antechinuses, dunnarts and other carnivorous marsupials; native mice and rats.

For example, the researchers arrived at a density of 0.05 koalas per hectare in eastern ranges and coastal areas (compared with 0.08 koalas per hectare in the tablelands, western slopes and plains).

The 0.05 figure is what the researchers agreed is conservative considering the range of less than 0.01 to 4.4 koalas per hectare from various survey sources.

For possums and gliders, they decided to use a density rate of 15.5 per hectare in eastern ranges and coastal areas, being below the mid-point of the sources’ range of 0.5 to 34.0 possums and gliders per hectare.

Underpinning Table 3 are 19 sources (peer-reviewed papers or chapters from edited books) that draw on surveys of specific species in NSW forests, woodlands and scrub. Separate tables and density estimates are included for birds and reptiles in the 2007 paper.

The researchers also drew on animal-density estimates compiled by Cogger, Johnson and Ford for a 2003 paper (also commissioned by the World Wildlife Fund) which examined land-clearing impacts in Queensland.

The 2007 calculation of 104 million animals lost through land-clearing in NSW was achieved by multiplying the animal densities for each of the two general regional areas by the land area (in hectares) the government had approved for clearing in the seven-year period to 2005.

A wildlife rescuer with a koala on Kangaroo Island (Image: AAP/David Mariuz)

The methodology applied today

Dickman told Fact Check his estimate of 1 billion-plus animals was derived by utilising the density rates from the 2007 paper and multiplying them by the number of hectares of land that has been lost to fires this summer.

He says that while the extent of land-clearing may change from year to year, the animal density rates in each of the broad geographical classes would have remained fairly constant.

However, rather than using the animal density rates for the wetter — and therefore more populous — eastern ranges and coastal region, Dickman used rates applicable in the sparser land areas (tablelands, slopes and plains).

“At every point, where there was a conservative option to take, we took it,” he told Fact Check.

“The estimates are for terrestrial mammals — that is, everything except bats and the platypus, as shown in Table 3 of the 2007 report — birds and reptiles. So frogs, fish and invertebrates were all excluded.”

Taking that into account, Dickman’s figures are based on 17.5 mammals, 20.7 birds and 129.5 reptiles per hectare (about 167.7 animals per hectare).

Since September 2019, the official start of the Australian fire season, bushfires have affected mostly one type of land area, namely the coastal and eastern ranges of the continent, including Victoria’s east.

It should be noted, though, that fires have also devastated parts of South Australia, Western Australia, Queensland and Tasmania.

Dickman told Fact Check, his estimate of 1 billion animals is based on projected losses only in NSW and Victoria; his figures do not take into account losses of wildlife in areas such as Kangaroo Island in South Australia, or in the other states.

The overall national losses are likely to be much higher than 1 billion animals.

Dickman says his calculations are confined to NSW and Victoria because the original density estimates (2007) were based on surveys pertinent to NSW conditions.

His justification for applying the NSW density rates to habitats lost by fire in eastern Victoria is that the Gippsland forests and woodlands are very similar to those in the NSW eastern ranges and coastal region.

Dickman says that methodology also is conservative because rainfall rates in Victoria’s eastern forests and woodlands tend to be higher than in NSW: higher rainfall rates tend to support more wildlife, so animal-density rates are likely to be even higher in east Gippsland than in NSW.

How much has been burnt?

The NSW Rural Fire Service told Fact Check that 5.22 million hectares of land had been burnt in that state since the start of the fire season through to January 16, 2020.

The RFS said the calculation was based on a range of sources, including on-ground reports from individual incidents, aerial surveys and satellite views.

It said the information was available internally to the RFS and issued on request, but was not published on a public website.

In Victoria, the State Control Centre (Emergency Management Victoria) estimates 1.487 million hectares had been burnt between July 2019 and January 18, the overwhelming majority of which was the result of fires that flared on and after the first ‘code red’ day of November 21, and the major fires of December and January.

This estimate, which is not published publicly, is the sum of land area lost by fire in state and national parks plus fires that affected both public and private landholdings.

The centre said the area burnt is sometimes overestimated initially; more accurate analysis is obtained through on-ground situation reports and aerial observation, including infrared scanning by helicopters and high-level linescanning by aircraft.

A rough calculation of fires outlined on the Victorian government’s emergency website on the morning of January 20 included: the Orbost complex, which had burnt 601,704 hectares; the Bairnsdale complex (310,263 hectares); Walwa (216,786); Abbeyard (101,503); Beloka region (68,283); south of Hotham Heights (38,250); and Shannonvale (42,163).

That does not give a complete picture of land area affected because the site registers only those fires that are still ‘going’, not those that have been largely extinguished.

The total area burnt in eastern ranges and coastal areas in NSW and Victoria by January 20 was approximately 6.7 million hectares.

Using Dickman’s estimate of 167.7 animals per hectare, more than 1.1 billion mammals, birds and reptiles would have died in NSW and Victoria alone.

What other ecologists say

Michael McCarthy, a professor at the University of Melbourne’s School of BioSciences, noted that Dickman’s calculation rested largely on the density of reptiles and seemed to assume that all affected animals in the fire zones would die.

He said the level of mortality would depend on the species and whether they tend to burrow underground or hide under leaves (which would burn), or whether there were rocks and other potential forms of shelter, as well as the intensity of the fire.

“But many of the impacted individuals will die,” he said.

So in conclusion, the figure of 1 billion animals (mostly reptiles, but also including birds and mammals other than bats) being impacted by the fires seems reasonable to me as a rough order-of-magnitude number.

“And a large fraction of those affected individuals would either die during the fires or in the days after the fires, given the size and intensity of the fires.”

He said the number of invertebrates killed would “greatly exceed” the 1 billion estimate, and the figure also excludes fish and amphibians. McCarthy said the impact on fish and amphibians is “likely to be large, and will extend into areas downstream of the fires”.

He suggested Dickman’s calculation “might be better described as an educated guess rather than an estimate”.

But as a rough measure of the scale of the impact on reptiles, mammals and birds, the figure of 1 billion seems reasonable because it conveys the message that the impact of the fires on animals has been large. And I don’t think anyone would argue that the impact has been anything less than large.

Mark Eldridge, principal research scientist at Australian Museum in Sydney, said Dickman’s estimate is probably an underestimate, especially considering it does not include losses of bats, frogs or, notably, invertebrates which are foundation elements of the food chain.

“The approach is an entirely appropriate way to estimate the numbers of a limited range of vertebrate animals that would have been present in the areas burnt,” Eldridge said.

Given the habitat differences, it is likely to be a cautious underestimate of the actual numbers of these groups of vertebrate animals present in the wetter and more diverse forests that actually mostly burnt. If invertebrate animals … were also included, then the estimate of animals present would increase by orders of magnitude.

But Eldridge noted that there is “no good data” on what to expect in terms of proportion of animals that might have perished outright in the fires or that could have survived. Nor is there any data on what proportion of animals might have succumbed to injuries, starvation or predators in the aftermath.

He suggested that, considering the intensity and extensive nature of the fires, the number killed “is nevertheless likely to be a significant proportion of the animals present”. And he noted that there would have been “significant” additional mortality in fires, especially in the Canberra region, South Australia and Western Australia.

So taking into account these sources of over-and-under-estimation, I think that [more than 1 billion] is an entirely defensible and reasonable number that probably errs on the side of caution. If you wanted to include all animals (vertebrate and invertebrate) across all impacted states and territories, then it would be a gross underestimate.

Chris Johnson of the University of Tasmania, was a joint author of the 2007 paper. He told Fact Check that, for the 2007 paper, he compiled all the animal density figures he could find for Australia and supplied the research team with his own work relating to density figures in Queensland, which he originally published in a paper in 2003. Those figures were then adjusted to account for NSW conditions.

Johnson said the work he did at the time enabled him to create a database for species densities in states other than NSW.

That database is freely available for fellow ecologists and scientists to use, though it is not publicly available online.

He said the method of calculation used by Dickman was valid.

“But I think everyone would accept that it is a very limited calculation and we could not estimate with confidence just now because the density varies over place to place,” he said.

“It’s looking at it in the most general statistical measure we can use, so it gives a very approximate estimate.”

Johnson said the animals killed by fires were inevitably dominated by common species — those that are widespread and which occur at high densities, so have lots of individuals in the path of the fires.

“For many of those species, the large numbers killed in fires represents a great deal of animal suffering. But it does not necessarily amount to a serious threat of extinction, because there are still lots of live animals in the wild in unburnt habitats [that are] able to keep the species going and ensure re-occupation of habitat when it recovers.

Another co-author of the 2007 paper, the University of New England’s Professor Hugh Ford, said he had discussed Dickman’s latest estimates with colleagues and checked other research papers.

He noted that while there were many more reptiles than birds in forests, and “somewhat fewer” mammals than birds, “Chris’ estimate of 800 million animals killed by the bushfires in NSW and 1 billion in Australia are reasonable estimates and definitely in the right ballpark”.

“I can only comment on birds,” Ford told Fact Check.

“On average, there are about 20 birds per hectare in woodland and forest. I’ve seen figures of 6 million hectares burnt, though it is probably more now. This would give 120 million birds.

Of course, there are many qualifiers: birds can escape fire by flying; also, inevitably small and occasionally large patches of vegetation escape being damaged by fire and act as refuges.

He added: “We will not be able to obtain a good estimate of the number of animals killed by fire until we undertake broad surveys of the burnt areas — probably not for many months. More importantly, [at that time] we shall be able to find which species have survived the fire in reasonable numbers and which have not.”

The latter will include some threatened species such as rufous scrub-birds, which are weak fliers, and glossy black-cockatoos, whose food (seeds from sheoaks) will take many years to recover.

Ford also pointed out that the effects of the extreme drought that preceded the fires — and which continues in many regions — would already have put some animal populations under pressure.

“Food — nectar, seeds, fruit, insects, etc — would have been scarce, and animals would have been stressed even before the fires,” he said.

Kangaroo Island

Dr Rosie Hohnen, of the University of Tasmania, has previously conducted biodiversity surveys of the western areas of Kangaroo Island. She told Fact Check that she had surveyed 42 sites in the Flinders Chase National Park, Ravine des Casoars wildlife protection area and Kelly Hill conservation area in 2017 and 2018.

Her focus was mostly on determining sample sizes of small mammals, reptiles and amphibians.

“The traps we had weren’t built to capture kangaroos, large possums, wallabies or koalas, or any birds,” Hohnen said.

“So the data I have is probably a significant underestimate of the number of animals per site.”

The data indicated that, on average, 11.7 individual animals were captured on the sites, which each measured about 3,600 square metres.

Extrapolating these numbers suggests a density rate of 32.5 animals per hectare, which Hohnen says “is a significant underestimate”.

Hohnen estimated that about half of the approximately 210,000 hectares burnt on Kangaroo Island appears to be bushland, indicating as many as 3.4 million animals are likely to have died there.

She noted, however, that even if animals had survived the fire, their survival would be threatened by post-fire shortages of food and shelter.

Principal researcher: Leonie Wood

factcheck@rmit.edu.au

Sources

© RMIT University 2019