‘It’s dangerous to put a number on them’. Media coverage of koalas during the 2019–2020 ‘Black Summer’ bushfires in Australia

Almost half (47%, n = 175) of the articles included an estimate of the numbers of koalas killed by fire in NSW or an estimate of the proportional loss to koala populations due to fire, and 53% (n = 197) of the articles quantified either the numbers of koalas killed by fires or area of koala habitat lost to fire (Table 2). The firm data that we obtained are the number of articles, whereas the koala numbers mentioned in the articles are far from certain and not amenable to calculations.

Estimates of koala deaths from fire published in the media varied greatly among different media articles, often within the same issue. There was no apparent difference in the reported numbers of koala deaths from fire or koala population estimates between publications or groups of publications, such as those owned by News Corp Australia versus those under the Australian Community Media group (formerly Fairfax Media). A more detailed analysis was not possible, as the figures were often vague, and without any reliable source mentioned.

Estimates of koala deaths in NSW were first published on 30 October 2019, following the fires in the Lake Innes Nature Reserve, south of Port Macquarie on the mid-north coast of NSW. Articles from October and November 2019 consistently stated that around 350 koalas died in the Lake Innes Nature Reserve fire, as reported in a social media post by Cheyne Flanagan and Sue Ashton from the Port Macquarie Koala Hospital (Port Macquarie Koala Hospital Facebook newsfeed, 30 and 31 October 2019). By February 2020, at the end of the bushfire season, media articles stated that between 7000 and 14,000 koalas had been killed by the ‘Black Summer’ bushfires in NSW. These estimates were lowered to around 5000 koalas killed by fire in NSW following a report prepared by the ecological consultants Biolink for the International Fund for Animal Welfare in February 2020 (Lane et al. 2020). That report stated that, if their estimates of the numbers of koalas occurring in NSW are correct, it implies that nearly 4000 koalas across NSW were killed by fires between September and mid-December 2019.

With the release of the NSW Upper House Inquiry into koala populations and habitat in New South Wales on 30 June 2020 (NSW Parliament 2020), there was a review of the numbers of koalas killed by fire. Under the heading, ‘The north coast’, in the Upper House Inquiry, there was a statement on numbers: ‘Dr Blanch told the committee that WWF Australia had commissioned Dr Stephen Phillips and Biolink ecologists to survey three sites in northern New South Wales, for which it had pre-existing data. … Dr Blanch stated that it was unclear how many koalas died or how many had moved into gullies or cooler parts of unburnt forest’ (NSW Parliament 2020, item 1.31, p. 9).

In the Chair’s foreword (June 2020) in the report of the Inquiry, Cate Faehrmann opened with the statement: ‘This season’s significant bushfires have resulted in devastating losses to koala numbers across NSW, so it is imperative that remaining populations and habitat are protected’, and then provided detail, stating that ‘at least 5000 koalas [were] lost in the fires, potentially many more’ (NSW Parliament 2020, p. x).

However, the estimates of the koalas killed by fires in NSW published in the media increased again after August 2020 to an average 10,000 and a maximum 15,000 koalas lost in NSW (Table 3). In contrast, comments about impacts on koala habitat from fire were not as variable as the estimates of koala deaths, with most articles suggesting that between 24% and 30% of koala habitat in NSW had been impacted by fire.

Nearly one-quarter (24%, n = 89) of the media articles provided a koala population estimate, either across Australia, in NSW, or an estimate of a local population, and 16% (n = 58) provided an estimate of the NSW koala population (Table 2). Many articles stated that the number of koalas in NSW was unknown or uncertain. The NSW koala population figures published in the media ranged from 1000 koalas (e.g. Sue Arnold, ‘Independent Australia’, 21 February 2020) to 55,000 (Graham Readfearn, ‘The Guardian’, 4 March 2020), with an average of 28,000 koalas (Table 3).

We found that some of the language and imagery used by the media gave the impression that koala habitats within the fire grounds were permanently lost and that 100% of the koalas in those habitats had perished, never to return. The precise location and scale of the fires and the specific koala populations that were impacted were often unclear or gave the impression that koalas across a larger region than the actual fireground, and a higher proportion of the state’s koalas, had been impacted by fires.

Of the 371 media articles we examined, 7% of all articles (n = 26), and 13% of international articles (n = 5) featured clearly misleading or factually incorrect headlines that exaggerated the scale of the impacts of fires on koalas while ignoring other threats. For example, headings such as: ‘Koalas ‘extinct by 2050’ because of fires’ (David Mills and Louise Starkey, ‘The Daily Telegraph’, 24 January 2020); and ‘Nearly 85% of koalas in New South Wales are wiped out after horror bushfires’ (no author, ‘Daily Mail UK’, 4 February 2020), suggest that koalas across all of NSW were equally and severely impacted by fires alone.

Another example of misreporting occurred following an interview with the then Federal Environment Minister Sussan Ley on ABC Radio National (27 December 2019). Ley was asked by Kim Landers: ‘How many of these animals have been killed, particularly in that NSW mid-north coast region?’ Ley replied: ‘It may well be 30% of that population in that region because up to 30% of their habitat has been destroyed.’ This statement was heavily misquoted and misrepresented in the media, leading readers to believe that 30% of the entire koala population of NSW had been lost due to the fires, for example: ‘As many as 30% of koalas in New South Wales may have been killed in just weeks during Australia’s recent devastating round of wildfires, the federal minister for the environment warned on Friday.’ (Mary Papenfuss, ‘The Huffington Post’, 28 December 2019).

A notable source of factual error occurred when journalists attempted to convert estimates of the proportion of a koala population impacted by fire into actual numbers of koala deaths using incorrect koala population estimates. For example, in an article published in ‘The Guardian Australia’ on 27 December 2019, journalist Naaman Zhou attempted to convert Sussan Ley’s statement into precise numbers of koala deaths from fire. Zhou wrote that: ‘The state’s mid-north coast is home to a significant number of Australia’s koalas, with an estimated population between 15,000 to 28,000.’ Zhou went on to calculate 30% of the high end of that population estimate to state that: ‘Sussan Ley’s estimate suggests up to 8400 koalas may have perished in the bushfires’. Zhou did not provide a source for the population estimate, but the figure was likely drawn from a report published by WWF in 2019 that stated that: ‘a compilation of the latest survey and estimate data presented here shows there are currently 37 to 38 metapopulations in NSW with a likely total population size of 15–28,000 animals.’ (Paull et al. 2019, p. 9). Thus, Zhou appears to have applied a state-wide koala population estimate to the mid-north coast fires, grossly inflating the number of koalas that had been impacted by fires at that time.

Following the publication of ‘The Guardian’ article, the figure that 8000 or more koalas had been killed by fires in NSW was repeated in 20 subsequent articles from a range of Australian and international publications including ‘The Australian’, SBS News, news.com.au, and the Australian Associated Press, and even in peer-reviewed publications (e.g. Jiang et al. 2019). For example: ‘The majority of the state’s koalas, which are numbered at 15,000 to 28,000, live in the region, meaning that more than 8000 koalas may have perished’ (No author, SBS News, 28 December 2020).

Journalists sourced estimates of koala death rates from statements, interviews, and reports made by policymakers, wildlife carers, advocates, local landholders and less frequently, by scientists and wildlife managers. When scientists who were actively involved in koala research were consulted, they often expressed caution and cast doubt on the figures of koala deaths by fire reported by the media and the accuracy of the koala population estimates.

Dr Christine Hosking of the University of Queensland pre-figured Professor Corey Bradshaw’s caution when asked how many koalas had been killed in fires in NSW and Queensland early in the fire season: ‘In the case of the bushfires, we need to wait a bit longer – the fires are still burning… In ecology it’s always grey – there’s never black and white, because koalas are very hard to track and count’ (Dr Christine Hosking, in Sam Langford, The Feed, SBS News, 25 November 2019).

Dr Stuart Blanch of WWF commented during a committee hearing for the NSW Parliamentary Inquiry into Koala Populations and Habitat in NSW on 18 February 2020: ‘We are relying on data that is probably at least a decade old and nobody really knows how many koalas there are in this State. I would not be surprised, from what I have heard from people doing the surveys, that we might have lost 10,000 koalas from the fires and the drought.’ (NSW Parliament ‘Committee hearing 18 February 2020’ p. 10.).

While Blanch and other scientists estimated the combined impacts of drought and fire on koalas, many journalists and politicians reported that the loss of koalas across the state was caused by the fires alone; e.g. ‘We lost maybe 10,000 koalas in NSW in the “Black Summer” fires’ (Cate Faehrmann, Upper House MP and Chair of the NSW Parliamentary Inquiry [The Greens NSW 2020, reported in Aslan Shand, ‘Echo Net Daily’, 20 October 2020]). We note that this figure of 10,000 is twice the number that Cate Faerhmann had presented in the report of the Parliamentary Inquiry in June of 2020, but it is the same number that Stuart Blanch reported that included both the preceding drought and the ‘Black Summer’ bushfires.

Misinformation about the fire impacts was reported even after post-fire surveys had been completed. Steve Phillips and colleagues, working on behalf of WWF, undertook detailed post-fire surveys at 123 sites across six fire grounds in the north-east of the state near Port Macquarie and Taree (Phillips et al. 2020, 2021). They chose sites where the koala population was known prior to the fires to enable them to assess the impact of the fire, and surveyed preferred koala food tree species for koala faeces as evidence of koala presence. They found that the impacts on koalas varied among the sites depending on the severity of the fire and reported a fall in koala occupancy rates between 34% and 100% of their pre-fire levels, with an average of 71% reduction in pre-fire koala numbers at the surveyed sites (pre-fire here meaning within one koala generation of 6 years) (Phillips et al. 2020). The authors also showed that koalas in low severity fires in the region were five times more likely to have survived compared with forests that burned with high severity.

The study by Phillips et al. (2020, 2021) was widely reported in national and state-wide media including the ‘Sydney Morning Herald’, ‘The Daily Telegraph’, and ABC. However, the coverage focused on the reported impacts of the high-intensity fires, and not the lower-intensity fires. Further, journalists did not consider details that are key to understanding the impacts in terms of koala conservation: e.g. what was the mortality rate for koala populations in low and moderate severity fires; how many koalas were present in the region before the fires, and how many koala populations were not impacted by fire? Phillips et al. (2020) found no post-fire evidence of koalas at Kiwarrak, near Taree, and this was widely reported as a local extinction event in the media. However, mid-north coast reporter Julia Driscoll of the ‘Great Lakes Advocate’ (23 September 2020) pointed out that post-fire surveys undertaken by NSW Forestry Corporation prior to the Phillips et al. (2020) survey had found koalas surviving at Kiwarrak, but this was not widely reported in the media.

It seems that the problem of misinformation that arose was not from the study by Phillips et al. (2020, 2021), but from a misreading and misrepresentation of their study by the media. The authors acknowledge the possibility of false negatives using their rapid survey method, saying that ‘it is possible that survey effort was not sufficient to detect koala survival’ (Phillips et al. 2020, p. 18; Phillips et al. 2021, p. 86), however this uncertainty was not communicated by the state-wide media.

In contrast to the media portrayal, there were vast local and regional differences in the extent and severity of the fires (Gibson et al. 2020; DPIE 2021a; Beale et al. 2022), with many locations in NSW where koala populations were not impacted at all by fire. For example, aside from a few isolated locations such as north of the Clarence River and Lake Innes Nature Reserve, the coastal strip on the north coast was largely unaffected by fire (DPIE 2021a). Similarly, the koala populations in Campbelltown (south-west of Sydney) and in Gunnedah (north-west NSW) were not affected (DPIE 2021a).

Further, rescued koalas were returned to regenerating habitats within 3–4 months following intense fire, such as in the Blue Mountains west of Sydney (e.g. B.C. Lewis, ‘Blue Mountains Gazette’, 2 April 2020). Unburnt refuges within fire grounds and unburnt areas with koalas adjacent to high-severity fire grounds were not reported on, even though these are essential areas for koalas to recolonise burnt areas after regeneration.

Much of the focus of the media was on a few high-severity fires in koala habitat, particularly around Port Macquarie and the mid-north coast region (Fig. 2), however even within this region, many areas of koala habitat were affected by only moderate and low severity fires (Gibson et al. 2020). Law et al. (2022), working in north-east NSW, compared their field assessments with the expert elicitation study by Legge et al. (2022) and demonstrated that experts reliably assessed the impacts of high severity fire on koalas, but overestimated the effects of low severity fire, where Law et al. (2022) detected no difference. These authors make a particularly strong point that is relevant to how the media understood the extent of the fires, namely that assessments of fire impacts on koala populations should not rely on simple fire extent as a surrogate for impact, instead, an assessment of severity is also required.

Cristescu et al. (2023) recognised that estimating the number of koalas killed is crucial to assess koala conservation, but that it is difficult because accurate, pre-fire data on koala distribution and density are patchy. In their study, they found that many koalas were still present in fire-affected areas after the 2019–2020 bushfire season, with koala density being 24–71% lower in fire-affected sites compared with control sites. In contrast, they found that one pair of sites they observed had a 32% higher koala density in the fire-affected site. They make the point from their field data that koalas can be present in fire-affected areas. It is a point that was missing from the media.

The 2019–2020 fires were extensive, covering 4,813,000 ha (7% of NSW) (DPIE 2021a), but there were marked differences in their severity across the burnt areas. The class of high, full canopy scorch, with partial consumption, covered 1,158,800 ha; and extreme, full canopy consumption covered 989,400 ha (DPIE 2021a). This gives a total of 2,148,200 ha, and that means that the remaining 2,664,800 ha were in the classes of low burnt understorey, unburnt canopy (1,171,300 ha), and moderate, partial canopy scorch (1,493,500 ha). Understanding the importance of the different fire severity classes is a crucial step in wildlife management and conservation, with Legge et al. (2022) employing fire severity as a key step in their study. Of course, these figures were not available until well after the fire season was over, so these important data were not available to the media. However, it would have been apparent to many observers, such as firefighters, local residents, koala managers and koala rescue teams that there were marked differences in the fire severity and there were many locations with no fires at all. This message was largely missed in the media, and the harrowing accounts where high and extreme severity fires were dominant became the focal point of media attention.

While there is no doubt that the impacts to many NSW koala populations were serious, the true impact of the fires on koalas in NSW was always uncertain, for three reasons: (1) the population numbers of koalas that existed in the affected areas prior to the fires were uncertain, or more likely, completely unknown, as were the numbers of koalas remaining in these forests post-fire; (2) the severity of the fires within the affected areas was patchy, as shown by Gibson et al. (2020), thus not all koalas and habitats were affected evenly; and (3) scientific assessments of the impacts of the fires could not begin until the fires had been extinguished, and even then, long-term impacts of the fires will only be known with monitoring and long-term studies. Despite the uncertainty, almost half of the 371 media articles we examined published quantified estimates of the numbers of koalas killed by fire in NSW, while less than one-quarter published estimates of the koala population that existed before the fires, which is essential in understanding how many koalas may have survived the fires to become the source of the post-fire recovery. In general, estimates of the number of koalas in NSW that were killed by the ‘Black Summer’ bushfires published in the media were not reliable, they were inconsistent among the media sources (as is evident from the data in Table 3), and the reported figures can be considered as misleading or wrong because there was no solid evidence of numbers and only rarely were there well-informed scientific ballpark estimates.

The fervour for constant news updates during the crisis may have meant there was little time, and perhaps little public appetite, to cross-check information, seek supporting evidence or alert readers to the uncertainties of the impacts of the fires on koalas. However, the misleading numbers continued to be published for most of the remainder of 2020 (Table 3). In a study published in February 2020, but describing data to 10 December 2019, Steve Phillips reported that 4000 koalas may have been killed (Lane et al. 2020). Around the same time, Phillips was cited in the media by Peter Hannam (‘Sydney Morning Herald’, February 18, 2020) thus: ‘… his team used surveys of known koala sites in northern NSW to estimate more than 6000 koalas died in the fires’. Other than the research by Phillips et al. (2021) in the year following the fires (although these results were in reports by Lane et al. (2020) and Phillips et al. (2020); i.e. in the year of the fires), there was no reporting of the refereed papers that had studied koala populations following the major fires of 1994 in Port Stephens, and published before the 2019–2020 fires (Lunney et al. 2004, 2007; Matthews et al. 2007, 2016). As a consequence, the rapid recovery rates of the koala population reported in these studies was not reported. Taken together, the media articles built a general impression that the fires may lead to the extinction of koalas across NSW and, more generally, a fear that the fires herald a new age of extreme environmental disasters caused by climate change (Thomson 2021).

In their detailed study, Reich and Barnoy (2023) commented that ‘evidence does not surface as a purely distilled epistemic element labelled as such … Rather, evidence is part of a versatile, malleable, dynamic and chaotic flow of information. It is the journalists who select the relevant parts in this flow.’ Given that the fires we report on burnt from September 2019 to March 2020, there was ample time during the fires, and more reflectively in the subsequent 7 months of our study period, for evidence-based journalism to emerge and to transcend the chaos of the many months of continuing fires. It strikes us that the images of blazing forests and dead koalas were taken as adequate evidence for journalists, and their media managers, to report on koala numbers without any reliable numbers of koalas killed or surviving in the wild. Thus, we conclude that the public and decision makers were ill-served by the media during and after this unprecedented ‘Black Summer’ bushfire season.

In what they see as a journalistic problem, Reich and Barnoy (2023) point out that, ‘even if news reporters were ‘evidence averse’ when Ettema and Glasser (1984, 1998) made their highly influential observations, today one may expect them to rely broadly on evidence’. Further, they emphasise that ‘Reliance on evidence is necessary to restore public trust in the media’. We contend that trust in the media fell given the hundreds of media pieces that reported on details of koala numbers killed or surviving without providing the evidence. This implies that Reich and Barnoy’s view that journalists now rely on evidence is not supported in this case.

In their study of ‘Scientific research in news media: a case study of misrepresentation, sensationalism and harmful recommendations’, Dempster et al. (2022) make what should be an uncontested point that ‘accurate representation of scientific research in news media is important.’ They add that ‘media both shape and reflect public opinion (Caulfield et al. 2014)’ and that the ‘public receive a significant amount of their health information from the media.’ One only has to replace the word ‘health’ with ‘koala’ and their thesis can be applied to the reporting of koalas in the media during the 2019–2020 fires. Their closing point is that ‘Scientists, science communicators and journalists have an obligation to frame science as interesting and newsworthy without jeopardising the truth.’ We agree, and koalas and fire is an interesting and newsworthy subject, and the sheer number of articles we obtained (371 articles in 12.5 months) demonstrates that it was a matter of continuing interest. However, our concerns are not only whether the truth was jeopardised by reporting numbers that were unsupported by evidence, but that the scientists who did speak, such as Corey Bradshaw’s advice, ‘It’s dangerous to put a number on them’ given at the middle of the fire season, were ignored, or just as bad, not understood and not even checked as to what they might have meant. Dempster et al. (2022) remarked that journalists need to regain some of the scientific expertise and resourcing that has been lost as newsrooms have declined in overall staff, including science journalists. We agree, koala conservation depends upon it.

In addition, there needs to be greater effort for what is termed ‘boundary spanning’, a term to describe linking an organisation’s internal networks with external sources of information. As Safford et al. (2017) explain, ‘Boundary spanners endeavour to achieve an appropriate balance of salience (relevance of the science to users’ needs), credibility (the perceived reliability and adequacy of the science employed), and legitimacy of information (the perception that the production process has been respectful of stakeholders’ values and beliefs) through three main functions: communication, translation, and mediation…’ In the concluding section of their paper, Safford et al. (2017) present what is a most salient example: ‘In 2009, the Joint Fire Science Program, a federal interagency fire science research and applications partnership, commissioned a study to investigate the dissemination and use of fire science information by US fire and resource managers (Kocher et al. 2012).’ Unsurprisingly, Safford et al. (2017) conclude that, ‘By inhabiting the boundary between science and application, translational ecologists are uniquely positioned to wrestle with the scale, uncertainty, risk, and entrenched interests that make many current environmental problems seem intractable (Balint et al. 2011).’ We note that Safford et al. (2017) are discussing the outreach from within an agency that manages wildfire. We concur, and it also applies to NSW, and as always, the case can be made that more communications are needed, especially in crises such as the ‘Black Summer’ bushfires. Conversely, we also argue that there is a responsibility of professional journalists and the media organisations to seek the relevant scientific information, make contact with the boundary spanners in the relevant government departments and universities and employ boundary spanners themselves to ensure salience, credibility and legitimacy. From our analyses of the reporting in the media on the ‘Black Summer’ bushfires in relation to koalas, boundary spanning was well below the level that could be considered acceptable.

Ewart and Cook (2024) examined the ‘mainstream news media coverage of Australia’s 2019–2020 ‘Black Summer’ bushfires to explore how that coverage characterised the leadership of the country’s Prime Minister Scott Morrison during that disaster.’ What is different about their study is that their exclusion criteria included articles that focused on animals, plants and climate. As in our study, Ewart and Cook drew on news media coverage as the disaster unfolded and in the months that followed. They found ‘a significant gap between the Australian Prime Minister’s leadership and community expectations of that leadership.’ They concluded that ‘the news media remains a key source of information for those affected by disasters and those watching at a distance.’ Their findings are consistent with our general conclusions concerning koalas during these same fires. Further, they demonstrate that analyses of the media’s reporting on a subject as harrowing as koalas killed in bushfires needs to become part of how we re-imagine how to manage koalas. The media has not shown the way in this crucial endeavour for the koala as an endangered species.

The highly variable and fluctuating NSW koala population and koala mortality estimates in the media from the start of the fires in October 2019 continued well after the end of the fire season (Table 3). This reflects the absence of accurate population figures for koalas in NSW, as discussed above (e.g. Cristescu et al. 2023). Media seized on ‘guesstimates’ – from any source, however dubious, and resorted to calculating numbers from habitat statistics and general population estimates without including the caveats that scientists included with these estimates. The question that arises is what authoritative sources of such information were available to the media from government, scientists, koala care groups, academics etc? We can also consider whether there is now, at the time of concluding writing this study 4 years after the fires, a better information base with which to evaluate the media reports.

As shown in Table 3, the average of the media’s estimates of koala numbers in NSW was 28,000 (range 1000–55,000) from the 57 articles that posited the number of koalas in the State, which was also the upper estimate in the more authoritative study by Paull et al. (2019). What is even more remarkable is that when we averaged the estimates in the 126 articles reporting koala deaths in NSW, the number was 5755 killed by the fires (the range was 200–14,000). This is very close to the estimate of 5000 in the 2022 NSW Koala Strategy (DPE 2022, p. 7). It would seem that the scramble for numbers has fed figures backwards and forwards between authorities, media and the public to not only arrive at what might be considered as official figures, as presented in the 2022 NSW Koala Strategy, for the number of koalas in NSW at the time of the fires, but also the number killed by the ‘Black Summer’ bushfires. There is no evidence that anyone has carried out the analyses that we have presented from our media analysis, but it is certainly possible that the constant media drive, with its numbers, had been unwittingly absorbed at a level below formal recognition.

What is also remarkable is that the numbers have become stuck in the media. Under the heading, ‘If you want your kids to see koalas, we have to change’ (‘Sydney Morning Herald’, 18 February 2023), Mike Foley states, ‘It is estimated as many as a third of NSW’s koalas – about 10,000 animals – perished in the 2019−2020 Black Summer bushfires and preceding drought…’ This puts the NSW koala population at the start of the fires in late 2019 as being 30,000. What is important to note here is that the estimated 10,000 koalas that perished did so as a result of drought as well as fire. Therefore, 3 years post-fire, there are the same sets of figures, and in this case, combined with the drought that led into the fires. What is also pertinent to note in Foley’s writing is his use of the passive voice, ‘It is estimated…’. What the passive voice obscures is the source of these figures. We postulate that it is simply a repeat of what was in the media at the time of the fires. There have been no scientific studies that have established the figures, so the media holds sway. Given the headline is sensational – ‘If you want your kids to see koalas…’ – it signals that it is the media that is driving opinions and imperatives on the future of koala populations in NSW, not science.

Lunney et al. (2020) showed that the readily visible koala population of Warrumbungle National Park in the 1990s had dwindled to rarity during the Millennium Drought before the severe fire of January 2013 had left the Park without any koalas. While the fire was widely reported, and was the subject of a coronial enquiry, few recognised that the drought caused far more losses of koalas than the fire, as droughts are too gradual to be considered news until our agricultural lands are withering. Thus, to focus on fire alone over the summer of 2019–2020 is to make an error in ecological understanding of what causes koala populations to decline. This goes to the point that trends are important to understanding population changes. Of course, if counts could be regular, such as the human census of the Australian population, or kangaroos in NSW, which is one of the best long-term wildlife datasets in Australia (Lunney et al. 2018), then counts are most valuable. However, humans are the only species where a complete enumeration of the population is attempted. It is more effective to sample to detect occupancy, trends and status of wildlife populations.

The misguided obsession with numbers that emerged during the sustained media blitz of numbers of koalas killed by the ‘Black Summer’ bushfires has had an impact in policy statements. One could reasonably conjecture that the timing of the uplisting of the koala to endangered by the Commonwealth in February 2022 was a consequence of the media focus on the numbers of koalas killed followed by an increasing emphasis on governance in the aftermath of the fires. In the five reasons provided for the uplisting, the statement within the first point, ‘weather conditions which promote bushfires’ gives an insight as to the importance of bushfires in the decision to uplist the koala (DCCEEW 2024).

A numerical emphasis is also evident in the Chair’s foreword to the NSW Parliamentary Inquiry on Koalas: ‘With at least 5000 koalas lost in the fires, potentially many more, it was deeply distressing but extremely important for committee members to agree to the finding that koalas will become extinct in NSW before 2050 without urgent government intervention.’ (Cate Faehrmann, Chair, June 2020, p. xi in NSW Parliament 2020). We note that the term ‘extinct’ is in effect simply putting a number of zero on koala numbers by 2050. We further note that the number of koalas killed by the fires, as stated in the foreword, does not appear in the text of the report. However, ‘Finding 2’ in the report did state extinct, as follows: ‘That, given the scale of loss to koala populations across New South Wales as a result of the 2019–2020 bushfires and without urgent government intervention to protect habitat and address all other threats, the koala will become extinct in New South Wales before 2050’. (NSW Parliament 2020, p. 12). The loss of koalas to the bushfires was canvassed in the report of the enquiry, but the emphatic linking of scale of loss of koalas and the bushfires is not readily apparent. We think that the media had driven this emphasis on numbers, especially as the impact of the fires, which had stopped only 4 months earlier, was still a prominent media theme. Given the importance of the enquiry, one cannot be surprised at any flow on effects. Consider the subtitle of the NSW Koala Strategy (DPE 2022): ‘Towards doubling the number of koalas in New South Wales by 2050.’ After witnessing the torrent of misinformation in the media on koala numbers in relation to the ‘Black Summer’ bushfires, the public goal of doubling the numbers makes sense in a media context. We contend that the bandying about of koala numbers in the media during the fires, and in the months that followed, exposed koala policy to an over-emphasis on numbers. Any target needs to make ecological sense and it could be argued that a primary focus on one number on one date overlooks such features as trends in relative numbers and the normal volatility of populations depending on preceding conditions, changes in distribution, the differences among populations in different parts of NSW, the importance of private land conservation, the long-term value of national parks and other protected areas, as well as what we shall learn from continuing research. Given the importance of the media in communicating science, the jeopardy created by misinformation in the media has emerged as a problem for koala conservation, as well as generating confusion as to what the real problems are and how to tackle them.

In ‘The Conversation’, under the subheading ‘Harnessing our grief’, Rumpff et al. (2023) observed that, ‘The Black Summer fires showed people care. The disaster triggered an outpouring of grief from Australia and around the world.’ That grief, as noted by these leading scientists, and the acknowledgement that people cared, reflects the media coverage. Fielding et al. (2023), in their study of predicting intentions to protect koalas, found very positive attitudes towards koalas, and commented that the 2019–2020 bushfires doubtless increased positive sentiments because of the preponderance of imagery of bushfire-affected koalas in the media.

The koala has an unusual, almost unique, position in the public’s consciousness. No other animal can be used to such great effect to encourage public donations to wildlife conservation and welfare, or galvanise politicians to pledge conservation actions. The frequent and emotive portrayal of koalas in the media during and immediately following the ‘Black Summer’ bushfires doubtless helped raise awareness and concern for environmental issues and helped drive fundraising efforts. Between September 2019 and May 2020, a total of A$80 million had been donated to wildlife rescue and care groups in NSW (DPIE 2021b). Such figures demonstrate the power that the media exerted in profiling a tragedy and inspiring public participation in conservation efforts. Caldwell and Henry (2020) describe the role of the media and media ‘marketing’ during the fires in motivating the public to exert pressure on government for koala protection and help drive policy change for conservation, including the rationalising of the NSW State Environmental Planning Policy on koala habitat protection on private lands. However, despite any benefits of raising awareness of the plight of koalas, we posit that the inaccuracies in the media coverage of fires and koalas have the potential to ultimately hinder rather than help koala conservation efforts by misdiagnosing and oversimplifying the problems koalas face, eroding trust in science and media reporting, and sowing confusion and despair about the value of conservation action.

There is also an ethical dimension to the way that the popular media portrays animals, an issue covered at length by Molloy (2011, p. 2). She notes Rollin’s pithy observation that ‘Animals sell papers’ (Rollin 2008, in Molloy, p. 2). Molloy (2011, p. 1) summarises her views at the outset with her statement: ‘…animal narratives are economically significant for popular media industries, which in turn play an essential role in shaping the limits and norms of public discourse on animals and animal issues and so constitute a key source of information, definitions and images.’ Molloy does not include koalas nor any Australian animals, but the concepts discussed are applicable in the Australian context.

From a social science perspective, Leimbach and Palmer (2022) note: ‘the iconic and severely threatened koala was a highly visible non-human species directly harmed alongside thousands of species in the order of individual billions.’ They maintain that the ‘social media cast a spotlight on – and propelled into a mediatised, virtual space – the suffering of humans and other species.’ (Using a slightly different spelling, Ampuja et al. (2014) state that ‘advocates of mediatisation suggest that the media are at the core of social change. Such ideas, potentially at least, increase the importance of media research in the social sciences and social theory at large.’). In particular, Leimbach and Palmer (2022) contend that ‘the catastrophic bushfires prompted a heightened multispecies awareness in the greater population.’ They argue that ‘the disaster produced a transversal event through social media communications, one that de-centred the human, allowing for novel connections between the human and non-human, prompting new questions and creating new responsibilities.’ This perspective pushes out the boundaries of the influence of the media beyond such analyses as misreporting, errors and hype into social dimensions that reach beyond science, koala management or policy. Hence, our contribution adds a scientific voice to the public discourse by presenting a critical appraisal of what, in the media, was misleading, exaggerated or misinterpreted with respect to the impact of the fires on koala populations.

Scientists are often warned that journalists will have the final word when reporting on science, because journalists and scientists have different aims, priorities and disciplinary interests, i.e. journalists aim to entertain and provoke readers rather than to maintain scientific rigour (Gregory and Miller 1998; Lunney 2012; Lunney and Moon 2012a). Unfortunately, the media cannot wait for this slow, academic fact-checking and must instead rely on opinion and comments for material, and these comments may often originate from the loudest voices. Journalists prefer to report clear stories and consequences and often avoid reporting scientific uncertainty (Kitzinger 1999; Due et al. 2014). They are also less likely to cite sources who express tentative opinions compared with sources who have definite opinions, even if those opinions are unfounded or unsupported (Kitzinger 1999; Stocking 1999; Lunney and Moon 2012a; Maier et al. 2016).

Implicit in the title of this paper is the gap between how science is conducted and communicated and what the public wants to know. The media occupies much of this gap, and they will report both what they are given and what they think the readership wants to know. During the ‘Black Summer’ bushfires, koala researchers, or their organisations’ media staff, were either not invited or largely unwilling to speculate, perhaps because of the lack of scientific evidence or a fear of being critical, about the impacts of the fires on koalas and the numbers killed. As a result, some wildlife conservation advocates and some wildlife carers dominated the coverage, not scientists. That is not to attribute base motives, but to simply point out that with the forests ablaze, heartfelt views, even if guesses, were welcome and reported by the media.

Licensed wildlife rescue and rehabilitation groups, especially koala rehabilitation groups, have been serving the community in NSW since 1973 as volunteers, and they have developed expertise in koala rescue and rehabilitation, yet it is only recently that their contributions have begun to be documented (Haering et al. 2020, 2021; Lunney et al. 2022c, 2023a, 2023b, 2023c). Thus, when asked, a koala rehabilitation specialist is responding to the tragedy of the fire-affected koalas rescued and brought in for rehabilitation, knowing full well that the fires would be killing far more than were brought into rehabilitation. The response is visceral, real, and readily photographed. Thus, it is not surprising that this is both reported and accepted as genuine. The state-wide reporting of koala rescue and rehabilitation in NSW during the ‘Black Summer’ bushfires illustrates the widespread public concern for koalas affected by the fires, yet only 209 were rescued, although with a good rehabilitation rate of about 50% (Lunney et al. 2022c). The issue here is not that koalas suffered or perished in high numbers – they did – but what the effect was on populations in the wild in the short- and long-term. These are different questions, and here the issue of the effect on a number of individual koalas that were rescued becomes conflated in the media with the effect on a wild population. Reporters for the various media outlets need to be more diligent in distinguishing between them, and there is a responsibility to do so if there is a genuine concern in the media for the ultimate well-being of populations of koalas as well as individual koalas.

The rise of social media platforms, digital science publications, blogs, and podcasts has led some to suggest that we may be entering into a ‘golden age’ in scientific journalism (Hayden and Check Hayden 2018; Willis 2020). Scientists need to take a more active role in communicating with the mainstream media and on social media platforms to avoid being left behind during critical events such as natural disasters. Professor Chris Dickman, a co-author of the WWF report into the impacts of the bushfires on wildlife, ‘Australia’s 2019–2020 bushfires: The wildlife toll’ (van Eeden et al. 2020) demonstrated a keen understanding of media communication and the need to take control of the media conversation: ‘I think there is a feeling among environmental scientists and ecologists in Australia that we’ve been frozen out of the debate, certainly out of policymaking. I think it’s now time to bring the scientists back into the tent’ (‘University of Sydney News’, 8 January 2020).

In an update of the numbers of animals affected by the 2019–2020 fires, van Eeden and Dickman (2023) produced a figure of 2.8 billion, outlined how they carried out their estimates and made valuable recommendations, such as the need for long-term species monitoring. Although they stayed on message of counts of animals, they included an enlightening insight on the media in Box 12.1 of their chapter. Dickman recounted that on 27 December 2019 he produced a press release where he estimated 480 million mammals, birds and reptiles would have been in the path of the fires. On 8 January 2020, he updated the press release with a new figure of 800 million animals affected in NSW, and more than a billion if the burnt areas of Victoria were also included. By the end of January 2020, the numerical toll on Australia’s wildlife had been noted in at least 6495 media articles in 115 countries. By the end of 2020, the figures were at least 7539 media articles, in 36 languages, reaching an audience of an estimated 25.6 billion people, which the authors note as being more than three times the global population. van Eeden and Dickman (2023) commented that it was hard to maintain a clear message that their estimates were not a death toll but rather represented the number of animals that were within the fire impact area. They formed the view that the media’s black-and-white interpretation of their estimates of a death toll probably contributed to the incredible extent of the media interest.

In his review of the book in which the van Eeden and Dickman chapter appears, Recher (2023) captured many points that are germane to our consideration of the media-fire-koala nexus. Recher made a point critical of fellow ecologists: ‘Dealing with the media and keeping reports factual as opposed to being dramatic and emotional requires a level of skill and experience that few Australian ecologists possess…’ ‘This was a lesson learned by Chris Dickman during the fires when he estimated that the fires had killed ‘more than a billion animals’ nationally … generating considerable emotional, and little factual media coverage.’ Recher (2023) compliments Dickman by the comment: ‘Ignoring his hard knocks, in Chapter 12 Dickman does a service by describing his involvement with the media in commenting on the loss of animals in the fires (pp. 156−157); an account that should be mandatory reading for all ecologists who choose to communicate with the community through the media.’ In our view, Recher’s comments have captured why so few ecologists or conservation biologists directly address the media, why there are so few analyses by scientists of the impact of the media on wildlife reporting or policy consequences, and it helps explain why we choose to tackle the subject of koalas, fire and the media. We had already grasped Recher’s message, for example, Recher (2013), under the heading, ‘What makes this old scientist grumpy’, declared, ‘Scientists are too conservative and lack the skills to communicate effectively with the community.’ Whereas Dickman was commenting on the media responses to his press releases, and Recher underlined what Dickman had learned as a result, our study looked at the press reports themselves, which makes a clear distinction between our focus and that of van Eeden and Dickman, and of Recher. We can add that our review of the media has had its own issues, both methodologically and factually, which we trust are apparent in our text. We encourage others to undertake a similar study with a species with which they are familiar. For the efficient management of koalas, it is important to determine population trends in both distribution and size rather than trying to agree on definitive pre- and post-fire koala population numbers to satisfy media questions. In evidence to the Parliamentary Inquiry on 18 October 2019, that is, before the fires, Steve Phillips added to the qualification about estimates of numbers: ‘Dr Stephen Phillips, Managing Director and Principal Research Scientist of Biolink and author of several koala plans of management, stated that he too was ‘always cautious about the numbers game … because there is always a bit of uncertainty with it’.’ (NSW Parliament 2020, p. 3). What also matters here is that Phillips added, ‘the number of koalas itself does not matter and it is more important to examine the trends’ (NSW Parliament 2020, p. 3). In an attempt to move the focus away from the numbers of koala deaths towards a more ecologically relevant measure of fire impacts, habitat impacts, Dr Alistair Melzer of the Central Queensland University released a media statement during the fires saying: ‘it’s more important to look at the impact on the extent of the habitat, rather than koala numbers’ (‘Central Queensland University News’, 7 January 2020). We note that these sensible, but cautious, statements, were not part of the media obsession in dealing with numbers.

The media landscape has become much more fast-paced, and this is matched by changes in how science is published, communicated, and accessed. However, the practice of science, including experimental design, consistently collected datasets and independent peer review, is still essential for scientists to maintain standards of integrity, transparency, impartiality, and trust. As a result, the long-term conservation of our fauna still ultimately depends on careful scientific assessment, not media hype in a 24-h cycle. It is hard enough to see a koala in the wild in a daily search, let alone being able to determine the outcome of a wildlife disaster for an entire population before going to press for the next day’s news.

In their study of the portrayal of human-wildlife interactions in the print media, Lunney and Moon (2008) found that the journalism was informative, readable, and entertaining. The authors concluded that the media plays a powerful role that will either further the conservation of wildlife or leave it as a neglected element of our heritage. Consequently, they argued, that ‘scientists and the media can be more profitably engaged, but ultimately, the conservation of our fauna will depend on well-supported and diverse teams of scientists and wildlife managers that operate on sound ecological principles, not media precepts’ (Lunney and Moon 2008). That conclusion remains standing from our analysis of the media, the 2019–2020 bushfires and koalas.

To avoid losing control of the story, it is important that researchers and wildlife managers take a more active role in communicating conservation issues with the media, particularly during the fast-paced and emotional reporting of a natural disaster. However, the need for scientists to stay actively involved in the media must not undermine the need to maintain standards of scientific integrity and trust.