Factors that influence the decision to release rehabilitated koalas (Phascolarctos cinereus) to fire‐affected home sites

Abstract

This study appraises recent research to critically evaluate knowledge on the factors that influence decisions around the release of rehabilitated koalas to fire‐affected home sites. Availability of koala release sites could be compromised at times in the future by the increasing incidence of large and intense bushfires due to climate change. In addition, factors such as the fragmentation of koala habitat and increasing urbanisation impose other threats to the long‐term survival of the released koala. Our investigation is conducted as a Critically Appraised Topic (CAT) with eight papers selected to be appraised. We highlight three key areas that affect koala rehabilitation and release to fire‐affected home sites ‐ the resilience of the koala in recovery, the adaptation ability of the koala upon release and the components of the fire‐affected habitat. The evidence shows that the koala is a good candidate for release given that burnt koalas are resilient in rehabilitation and many can recover and then adapt quickly to their release site. Findings from this study indicate that, when the release site is affected by fire, the conditions that are more favourable for koala release are low severity of fire impact, greater distance from urban areas, connection to unfragmented habitat and previous management of predators. These findings can inform wildlife release and rescue as well as government planning and urban development.

In Australia, bushfire is a recognised threat to wildlife, and has had significant impact on many koala (Phascolarctos cinereus) populations. ¹ , ² , ³ Due to climate change, the incidence of large and intense fire is predicted to increase. ¹ , ⁴ During Australia's Black Summer bushfires of 2019–20, more than 15,000 bushfires burned across the country, ⁵ severely impacting or destroying more than 3.5 million hectares of koala habitat. ⁶ In 2020, following these bushfires, the koala populations of Queensland (Qld), New South Wales (NSW) and the Australian Capital Territory (ACT) were classified as one of 810 priority species and ecological communities identified for urgent intervention. ⁷ On 12 February 2022, under the Environment Protection and Biodiversity Conservation Act 1999, the status of these koala populations was changed from “Vulnerable” to “Endangered”. ⁸

It is important for welfare and population management that rehabilitated koalas be released in appropriate locations. With loss of habitat due to urbanisation, ⁹ the fragmentation of koala occupied land ¹⁰ and the increasing number and severity of fires due to climate change, ¹ , ⁴ it is likely that the suitability of release locations for rehabilitated koalas will be an increasing challenge. “No suitable release location” is one of the criteria for euthanasia of rescued koalas under the relevant Code of Practice in NSW. ¹¹ It is understood that rescued wildlife should be released to original home sites where possible, however there is limited published research specifically informing how release decisions should be made and what precise factors indicate readiness for release of rehabilitated wildlife. ⁹

This study will focus on one aspect of habitat, the fire‐affected home site, as a release location for rehabilitated koalas. We will critically evaluate factors that may impact the decision of whether, and when, to release koalas to fire‐affected home sites, identify knowledge gaps and develop any possible recommendations for management or future research. This study is a research synthesis presented as a critically appraised topic (CAT).

Research question

In rehabilitated koalas awaiting release to original but fire‐affected home sites, what factors should influence release?

Methodology

Searches for published peer reviewed research were conducted on the databases PubMed and Scopus (Table 1), and all available years were searched. The search involved combinations of koalas, rescue, rehabilitation, fire, stress and behaviour; the term behaviour was included as changes in this could indicate stress in animals and research on wildlife rehabilitation shows that physiological factors such as stress can have an impact on rehabilitation and release. ⁹ , ¹² Search criteria involving land management and conservation were not included in this study to increase focus on those papers specific to our research question. The search was first performed in December 2022, then again in April 2023 and in February 2024, and emerging literature was added so that this study kept up with the most recent published research. Inclusion and exclusion criteria (Table 2) were then used to filter results according to our research question.

Table 1.

Search terms and databases

Search	Database	Dates searched	Search terms	Number of results
1	PubMed	1900–2024	Koala AND rehabilitation AND (rescue OR fire OR stress OR behaviour)	11
2	Scopus	1960–2024	Koala AND rehabilitation AND (rescue OR fire OR stress OR behaviour)	16
3	Scopus	1960–2024	Koala AND fire	47
4	Pubmed	1900–2024	Koala AND fire	13

Table 2.

Inclusions and exclusion criteria to filter search results

Inclusions	Peer‐reviewed published papers
	Published research on koala rehabilitation and release
	Published research on fire‐affected habitat and its suitability for koalas
	Studies that inform release after fire
	Studies that look at stress in koalas
Exclusions	Studies not related to the research question
	Studies relating to disease/trauma or chlamydia
	Studies looking at conservation, environment or management of land
	Studies on the release of koalas but which were not related to fire‐affected sites

Search strategy

See Table 1.

Inclusion and exclusion criteria

See Table 2.

Search outcome

See Table 3.

Table 3.

Outcome of search strategy showing total relevant papers, cross‐referenced with search items from Table 1

Search items from Table 1	1	2	3	4
Number of results	11	16	47	13
Excluded—Does not answer research question	5	6	27	8
Excluded—Relating to disease/trauma/chlamydia	1	2	2	2
Excluded—Relating to conservation or management	1	1	7	1
Excluded—Not related to fire‐affected habitat	1	2	3	0
Excluded—Duplicate	0	3	5	2
Total relevant papers	3	2	3	0

Results

Eight peer‐reviewed published papers were identified for appraisal in this study (Table 3). Each paper presents evidence to answer our research question.

Summaries of the evidence

Author	Narayan E, Vanderneut T 13
Year	2019
Population	Rescued wild koalas in care in Adelaide Koala and Wildlife Hospital (AKWH).
Intervention studied	Influences of habitat, types of stressors and clinical condition on rescued wild koalas.
Outcome studied	Physiological stress of koalas in clinical care and how this affects their outcome.
Design and sample size	Faecal glucocorticoid metabolites (FCM), cortisol and corticosterone, were measured in faecal samples collected from 22 koalas at AKWH between 2015 and 2016. Samples were collected at rescue and during clinical care.
	FCM levels were used to give a quantitative measure of physiological stress, and these were analysed against host‐associated variables ‐ their health data, habitat (location of rescue), stressors and disease/trauma diagnosis.
Main findings	Koalas found on the rural–urban fringe had highest mean FCM levels, and koalas rescued from national parks had the lowest.
	Regarding environmental trauma, FCM was highest for koalas with chlamydiosis followed by those koalas affected by bushfire. Bushfire‐injured koalas had significantly greater FCM levels compared with koalas in all other injury categories, as well as with FCM levels related to veterinary checks.
	Elevated FCM concentrations were also associated with clinical management factors such as veterinary care, the length of stay in hospital, their diagnosis and treatment. However, they were not predictive of rehabilitation outcome.
	Koalas with renal failure, arthritis or diabetes were euthanised, while those suffering burns, heat stress and respiratory illness were all ultimately released. Outcomes were variable for koalas with mucosal shedding of Chlamydia pecorum with signs of infection by C. pecorum such as red eyes, wet bottom and swollen genitals.
Limitations	Authors acknowledged that the small sample size in this study meant that there was a very small sample for some disease categories. Although diagnoses with n = 1 were excluded from statistical analysis in this study and assumptions for the ANOVA test used in this study were satisfied, the stability of the test should be considered with caution.
	At the time of writing, there was no known reference interval for FCM levels in wild koala populations, so it is not possible to compare levels to a known “normal” baseline.
	The faecal metabolite used to indicate stress is under question. Newer research has shown that tetrahydrocorticosterone rather than cortisol was the most sensitive metabolite to indicate stress in koalas, and showed differences between males and females, as well as in breeding and non‐breeding seasons. 14

Author	Charalambous R, Simonato T, Peel M, Narayan E 15
Year	2021
Population	Three wild rescued koalas in rehabilitation at Port Stephens Wildlife Hospital (PSWH).
Intervention studied	Cortisol levels during rehabilitation on days when stressors were and were not recorded.
Outcome studied	Understanding how koalas in rehabilitation respond to stressors.
Design and sample size	Faecal and fur samples were collected sporadically between September and December 2018 from three koalas and cortisol measurements were taken from these samples. Days on which samples were collected were classified as stressor and non‐stressor days. Eight stressors were identified: bellowing (noise made by koalas), campers, maintenance, fire‐crackers, visitors, moving cages, veterinary intervention and weather. Glucocorticoid levels on stressor and non‐stressor days were compared using Single Factor ANOVA. Measurements of individual glucocorticoid levels for each koala were taken between September and December 2018 and these were compared using box and whisker plots.
Main findings	There was a significant difference between all three koalas in their mean faecal cortisol levels however one koala had a mean (218 ng/g) much higher than the other two koalas (56 ng/g and 63 ng/g respectively).
	Fur cortisol levels were also significantly different between all three koalas. There was a significant difference found between stressor and non‐stressor days for two of the three koalas.
	The most common stressors were found to be visitors, maintenance and bellowing, however it must be noted that this study collided with breeding season where bellowing occurs and a natural rise in glucocorticoid levels is seen.
	The study concludes that glucocorticoid levels differ based on individual responses to stressors as a whole, but this was seen individually in two of the three koalas.
Limitations	Authors acknowledge that there has been no known previous research to establish baseline glucocorticoid values in koalas so study results could not be compared. This was a pilot study with a very small sample size of three.
	While the degree of variance seen is useful for future research, with the lack of baseline data it may not be useful here to compare results between koalas, but instead concentrate on changes seen within an individual. Each of the three koalas was of different ages and in rehabilitation with different injuries needing various types of veterinary intervention. These factors could affect cortisol levels in koalas, and this further emphasises the likelihood of heterogeneity impacting comparisons between koalas.

Author	Lunney D, Gresser SM, Mahon PS, Matthews A 16
Year	2004
Population	A group of 16 injured (burnt), rehabilitated and released koalas were compared with a group of 23 uninjured wild koalas, in Port Stephens, Australia, in 1994.
Intervention studied	Comparison of the long‐term survival and reproduction of released rehabilitated injured koalas to uninjured wild koalas in the same habitat after fire.
Outcome studied	Survival rates and reproduction over a three year period and two breeding seasons after fire, and how this contributes to population recovery.
Design and sample size	Immediately after each of three fires in Port Stephens, extensive ground searches located injured or orphaned koalas to be taken into care for rehabilitation. Injured koalas had burns mainly to the skin, hands and feet and the average time in care was 168 days. Of the 17 koalas rescued, 16 were rehabilitated and released, back to the original fire‐affected site and monitored by radiotelemetry.
	Between June 1994 and July 1996, a group of 23 uninjured koalas from the same habitat were captured to install a microchip and then monitored by radiotelemetry alongside the released group. Koalas thus had staggered entry into the study depending on which fire they were found after, but all were tracked for 5 days a week after release, from March 1994 to September 1996, and then once a week until February 1997.
	For all koalas in this study, sex and age were recorded along with the number of days observed after release and their fate.
	Survival was calculated using the Kaplan–Meier estimate using two methods, one of which adjusts for staggered entry into the study, and a log‐ranked test was used to measure the difference in survival between the two groups.
	Reproduction was measured by recording whether the koala had young on its back at the time of observation, and was quantified as the number of young divided by breeding opportunities (number of seasons).
Main findings	This study found no significant difference between survival of rehabilitated compared with uninjured wild koalas, regardless of which method was used to calculate survival.
	While the survival rate was low with 13 koalas left at the end of the study, the main cause of mortality for both groups was predation by dogs, with other causes being age related and unknown causes.
	Reproduction values were calculated as 80% for the rehabilitated injured koalas and 73% for the uninjured wild koalas and no significant difference between groups was found.
	This study concludes that serious injury and rehabilitation do not disadvantage released koalas once returned to the fire‐affected home habitat. Within 5 months of fire (and 3 months after start of vegetative growth), koalas were seen in burnt trees with epicormic growth.
Limitations	Technical failures of radio trackers and disappearance of individuals meant that eight koalas were excluded from the study, decreasing the sample size. Sample sizes were too small to make comparisons based on age and sex.
	Survival and reproduction by the unburnt population, could have been affected by the introduction of the rehabilitated koalas which would increase competition for resources, however this study viewed this as negligible due to the rapid regeneration of burnt forest that takes place after fire.

Author	Matthews A, Lunney D, Gresser S, Maitz W 17
Year	2016
Population	55 koalas in a remnant forest in Port Stephens.
Intervention studied	Tracking the movement of 55 koalas between March 1994 to February 1997, in a fire‐affected habitat.
Outcome studied	Examining movement patterns and home range of koalas recolonising their burnt home site after fire.
Design and sample size	The sample of 55 koalas tracked consisted of 28 rehabilitated koalas (17 with burn injuries and 11 with other reasons for rehabilitation) and 27 unburnt wild koalas. All koalas were ear tagged, fitted with a radio tracker collar and microchipped. The 28 rehabilitated koalas were released once recovered, the first group in March 1994, and the 27 unburnt koalas were released within a day of capture from June 1994. All koalas were tracked for five days per week till September 1996, and then once a week till February 1997.
	Koala daily movement, defined as the distance koalas moved between consecutive daytime trees, was calculated by recording koala locations (fixes) each day, mostly calculated in the daytime. The monthly displacement was determined by first defining a mean centre, the location where the koala was spending most of its time, and then recording displacement from this centre each month for each koala, with annual displacement calculated for koalas who were able to be tracked for at least 12 months.
	Home ranges were calculated using probability distributions for koalas with more than 50 fixes, while for koalas who had shifted their range, home range was calculated using only the locations after their move.
Main findings	Daily movement calculations showed that koalas were capable of moving long distances. The mean distance moved in one day was 189.3 m and the maximum was 2540.9 m. Males moved longer distances in the day, and overall daily movement increased in breeding season.
	Monthly displacement showed that most koalas kept to their home range site with a mean monthly displacement of 199.3 m, but this increased significantly during breeding season with the maximum recorded monthly displacement of 5980.9 m. However, those with a higher monthly displacement were more likely to die due to dog attacks and vehicle collisions.
	It was found that koalas tended to stay in the same area each year with average annual displacement of 230.9 m, however some travelled for long distances before settling on a home site, the longest being 20.6 km. Fifteen koalas dispersed more than 600 m to another site and these koalas came from all groups showing that rehabilitation was not an important factor. In this shifting of home ranges there was a variety of dispersal methods, some exploratory, some along a direct path and some incrementally, allowing for habitat colonisation.
	The average size of home range was 39.5 ha with males having a significantly larger home range than females. It was found that home range size did not increase when the proportion of burnt bushland was higher, which could indicate that the quality of feed was not less adequate than that from unburnt trees.
	This study showed that the highly mobile nature of koalas allowed recolonisation of regenerating forest and establishment of home ranges within burnt areas. While males seemed to disperse for longer distances there was no difference in movement patterns and home range sizes between rehabilitated koalas and unburnt wild koalas.
Limitations	Due to limited number of night‐time fixes, the daily movement paths that were separated by more than one day were extracted.

Author	Phillips S, Wallis K, Lane A 18
Year	2021
Population	Koala population from six fire grounds in the coastal hinterland of NSW between Forster and Ballina, where pre‐fire occupancy data was known, and which was affected by the 2019‐20 bushfires.
Intervention studied	Comparing intragenerational changes in the use of habitat by koalas, before and after fire.
Outcome studied	To gain insight into the direct impact of bushfire on resident koala populations.
Design and sample size	Pre‐fire koala occupancy data collected between 2013 and 2019 from a previous study was used to identify 123 surveyable sites across six fire grounds, 49 of which had koala occupancy previously recorded. A survey of koala occupation was undertaken after the 2019‐20 bushfires on these sites (post‐fire occupancy) between March and May 2020.
	Occupancy of habitat was determined by identifying koala faecal pellets around Preferred Koala Food Tree (PKFT) species at each site, using Rapid‐SAT sampling. Burnt faecal pellets indicated pre‐fire occupation while unburnt pellets indicated post‐fire occupation.
	Data on fire severity was also collected by examining how severely the tree canopy was affected, using Google Earth Engine Burnt Area Map to compare changes in colour pre and post fire. Habitat was then classified as fully burnt, partially burnt or unburnt.
	Statistical analysis using log‐likelihood ratios was then used to compare pre and post fire occupancy levels at each fire ground. The proportion of pre‐fire occupied sites was calculated using pre‐fire data in the survey area from a previous study. The proportion of post‐fire occupied sites was calculated in the same survey area, and these proportions were compared to see the change in habitat use.
	The relationship between fire severity and koala activity was also analysed using log odds.
Main findings	Post‐fire habitat use by koalas was reduced across all six fire grounds (123 sites) with one fire ground recording no post‐fire occupancy. The reduction in occupancy ranged from 44% to 100% with a median reduction of 71% from the original pre‐fire occupancy rate.
	Out of the 49 sites known to have been previously occupied by koalas, 18 were recorded as having post‐fire occupancy, showing a significant reduction in site use to 50% of unburnt sites, 44% of partially burnt sites and 14% of fully burnt sites.
Limitations	Not all sites in each fire ground could be surveyed due to safety concerns and logistical issues. Of the 227 sites in the six fire grounds, 123 were surveyed.
	The detection of koalas relied on the prior knowledge of their preferred food tree and the Rapid‐SAT sampling technique. While this has been a previously validated research method it has been questioned whether this non‐random technique could introduce some bias and affect the results. 19
	Faecal pellet searches were restricted to a one meter radius around the tree and this could mean that some results were missed if they were not directly under trees or located a bit further away. 19

Author	Cristescu RH, Gardiner R, Terraube J, McDonald K, Powell D, Levengood AL, Frère CH. 20
Year	2023
Population	Koala populations in eight sites in NSW and Qld, four of which were fire‐affected, and four unaffected. Each state contained two pairs of fire‐affected and unaffected sites.
Intervention studied	The impact of fire on koalas by comparing population density in fire‐affected habitat with unaffected habitat.
Outcome studied	Density of koala populations in fire‐affected sites.
Design and sample size	The four fire‐affected sites were paired with four unaffected (control) sites. All sites selected were previously known to be occupied by koalas and site pairing was based on geography and vegetation with the aim to match sites environmentally, with paired sites located in the same state.
	Koalas were detected by two methods. In NSW, koalas were detected by drone thermal technology and the area surveyed was recorded by flight control applications, and then confirmed by visually checking for the presence of scats. In Qld, detection dogs were used to locate koala scats, which were then genetically fingerprinted to determine the number of individual koalas. The area surveyed was mapped using data from GPS collars worn by detection dogs. In NSW, detection started 30 days after fire, while in Qld detection started 109 days after fire.
	Koala density was calculated for each area as the number of koalas detected per area surveyed each day. For each state, density was then compared between fire‐affected and control sites.
Main findings	Koalas were present in all fire‐affected sites in relatively high numbers when compared with paired control sites.
	In three of the four fire‐affected sites, koala density was lower compared with their matching control sites. The two NSW sites showed a 24% and 26% decrease respectively, and a Qld site showed a 71% decrease.
	Koala density at one Qld fire‐affected site increased by 317% compared with its control site. This different result was attributed to multiple factors including site selection given the unaffected site was close to urban areas while the fire‐affected site was connected to unfragmented habitat and had been managed for predators.
Limitations	Authors acknowledge the following limitations:
	Some search areas overlapped from previous days so there could be some overcounting if koalas moved between surveyed areas, however this did not affect density count as this was calculated per day.
	Different detection methods were used in each state and therefore koala density between states could not be compared.
	Using scats to determine density could potentially introduce a correlation between fire severity and scat density as scats are known to persist in the environment and unburnt sites would have years of accumulation compared with burnt sites.
	The small number of sites meant that koala density compared with fire severity could not be studied.
	The choice of optimal control sites was challenging and may have introduced too many variables for an accurate comparison, such as proximity to urban areas, degree of fragmentation, whether there had been predator control previously or landscape differences, given that some sites were 70 km apart. These factors could explain why one Qld site unexpectedly had greater density in a fire‐affected site.

Author	Law BS, Gonsalves L, Burgar J, Brassil T, Kerr I, O'Loughlin C 21
Year	2022
Population	Male koala populations in state forests of north‐east NSW that were impacted by the 2019–20 mega‐fires.
Intervention studied	Impact of mega fires on the koala.
Outcome studied	Density of male koala population and its relationship with fire severity.
Design and sample size	Three areas, each with different severity of fire impact (categorised as low, moderate and high), and with previously known koala density estimates, were surveyed using acoustic arrays. This study followed a before‐after‐control‐impact (BCAI) design where the initial survey in the fire grounds was conducted prior to the 2019–20 fires, a second survey was repeated immediately after the fires and a third survey was undertaken one year after the fires. In addition, control areas in nearby unburnt national parks were also surveyed.
	Surveys were conducted using acoustic arrays which deployed sensors strategically placed to record koala bellows from sunset to sunrise, and then recordings were scanned using koala recognisers, and spectrograms were manually checked for false positives.
	Spatial count models were used to model spatial density and estimate the number of locations (home range) of koala bellows and mathematical modelling was used to estimate density.
	Fire severity and charcoal height of trees were also recorded at each sensor. Spatial density models were overlaid with fire severity mapping to investigate how fire severity influenced changes in density at a local level.
	The mean koala density was calculated for each of the three fire severity categories in each array. The extent of fires was also recorded using a pyrodiversity index model which indicated fire severity within an array.
Main findings	Across all three areas koala density decreased post‐fire, with significant positive association found between the level of decline and fire severity, but not with the extent of any fire. Variation in pyrodiversity also did not change possibly due to epicormic growth masking initial fire impacts.
	In the low‐severity fire ground, pre‐fire koala density was estimated at 0.05 males/ha with 17 bellows/night heard. Immediately after fire, 5 bellows/night were recorded and, one year later, density was estimated at 0.4 males/ha.
	In contrast, on the moderate‐severity fire ground, estimated pre‐fire density of 0.08 males/ha decreased by 50% to 0.04 males/ha one year after fire. The high‐severity fire ground recorded pre‐fire density as 0.04 males/ha with no bellows heard immediately after fire and then 0.016 males/ha one year later.
	While overall density decreased in the moderate and high‐severity fire grounds, it was not sufficiently different pre and post‐fire in the low‐severity fire ground. Density in the unburnt control areas showed negligible change. The spatial variation in density when compared with fire severity showed that in moderate or high fire severity areas, focal density was decreased across all areas, while in the low‐severity fire ground only the focal area of high‐severity had decreased density. Given the results from the high‐severity fire ground, it was concluded that there is evidence of some localised recovery one year after fire however in focal high‐severity areas koalas did not remain.
Limitations	Authors acknowledge that only males were recorded in this study as there is no evidence that females regularly bellow. However, it is unclear whether calls were scrutinised for belonging to males or whether attempts were made to distinguish female bellows. 22 Although it seems likely that the number of female bellows would be small, these can be distinguished from male bellows 23 and this would add further clarity to the composition of population accounted for in each fire ground.
	The initial pre‐fire survey was conducted in spring, a time that is known for heightened koala activity and increased bellowing due to breeding season. The second survey immediately after fire was not conducted in spring but in summer ‐ while this is still within the breeding season, there could have been a naturally decreased incidence of bellows at this time.

Author	Lane MR, Lowe A, Vukcevic J, Clark RG, Madani G, Higgins DP, Silver L, Belov K, Hogg CJ, Marsh KJ 24
Year	2023
Population	56 koalas in the Snowy Monaro Shire of NSW impacted by the 2019–20 bushfires, 31 of which were rescued, rehabilitated and released, 9 of which were non‐rescued and living in the burnt environment and 16 of which were non‐rescued and living in the non‐burnt environment.
Intervention studied	Health of rescued and released koalas after bushfire compared with non‐rescued koalas in burnt and unburnt habitat.
Outcome studied	Health data and long‐term health risks for rehabilitated koalas who have been released into burnt habitats.
Design and sample size	After the 2019–20 bushfires, 31 koalas in the Snowy Monaro region of NSW were rescued and rehabilitated between January and March 2020. Initial health data was collected from all individuals however 12 of them were fitted with trackers when released and their health data was compared with non‐rescued koalas also fitted with trackers in the same area at pre‐release (May–December 2020) and recapture (February–May 2021). Due to the mosaic of fire burn, the non‐rescued group was divided into nine individuals from burnt habitat and 16 individuals from unburnt habitat.
	Three health checks were conducted ‐ the initial health check on the 31 rescued koalas, the pre‐release health check on the 12 rehabilitated koalas fitted with trackers and the non‐rescue groups, and the recapture health check on individuals still fitted with trackers ‐ by this stage, the rehabilitated group contained four individuals and the non‐rescue groups contained eight individuals from burnt habitat and six from unburnt habitat.
	Health checks were conducted under sedation and koalas were checked for age (from tooth wear), sex, presence of burns, young in pouch, physical condition, abdominal fill and aural health. In addition, blood collection was performed for haematology, biochemistry and genetic analysis, and ocular swabs, genital swabs and ear tissue biopsy were taken, and Chlamydia status was determined.
Main findings	There was a statistically significant difference in body condition score (BCS) in the rehabilitated koala group between rescue, pre‐release and recapture health checks. The BCS at the initial check was very low (1.3 out of 5), but by recapture, there was no significant difference between BCS of koalas in the rehabilitated and the nonrescue groups. This indicates that within four months of fire, there is sufficient quantity and quality of food to support healthy koalas.
	Three of the rehabilitated koalas had burn injuries at rescue, one of which required regular bandaging, and all made a full recovery.
	Ocular chlamydiosis was detected in three individuals in the rehabilitation group and none in the non‐rescue groups. Urogenital chlamydiosis was detected in 58% of the rehabilitated group, 44% of the non‐rescue from burnt habitat group and 63% of the non‐rescue from non‐burnt habitat group. It was found that rates of chlamydial shedding did not differ between groups indicating rehabilitation or fire did not have a major impact on Chlamydia status.
	Blood biochemistry parameters were mostly within reference intervals while the following differences are relevant to koala rehabilitation: Urea concentrations were higher for rehabilitated and non‐rescue koala groups in the burnt habitat compared with non‐rescue koalas in the unburnt habitat, but by the recapture health check urea was not significantly different between groups, and this could indicate improvement in nutritional status. The liver enzyme AST was elevated at the initial health check in rehabilitated koalas and this could indicate stress.
	Haematology results were also within the reference range for most individuals with no significant differences found between koala groups or health checks for PCV, TPP, haemoglobin, MCHC, platelet count, leukocytes, monocytes, eosinophils or basophils, and while differences were found between lymphocyte and neutrophil values, they were within reference range. There was no significant difference found between haematology parameters and Chlamydia status or survival rates.
Limitations	Of the 56 koalas in the original survey, only a small subset ultimately received health checks, with the final health check involving only four individuals from the rescued group and eight individuals from the nonrescued groups. Given the small sample size, individual variation and that some differences noted were within reference ranges, it is questionable whether conclusions can be made about haematology and biochemistry parameters.

Discussion

This study evaluated current evidence around three key areas of koala rehabilitation and release for their relevance to decisions on release of rehabilitated koalas to fire‐affected home sites: the koala's resilience during recovery; its ability to adapt upon release; and the attributes of the fire‐affected environment.

We found this field of research was characterised by a small number of papers to be appraised, namely eight, and typical of many studies of wildlife, conclusions were limited by small sample sizes, and the different methods of measurement and survey techniques used. With only two papers suited to inclusion, evidence around koala rehabilitation was more limited than that relating to koala release and fire‐affected sites. Similarly, an additional perspective by Burton and Tribe (2016) showed the need for more research around koala rehabilitation, highlighting the requirement for consistency in treatment regimens of wildlife hospitals to give clarity for outcome decisions. Despite the limitations of individual studies, the consistency in findings allowed us to effectively appraise three key determinants of success: the resilience of the koala in recovery; the adaptation ability of the koala upon release; and the characteristics of the fire‐affected habitat.

While significant numbers of fire‐affected koalas can be treated and released successfully, and one study suggests koalas can tolerate physiological stress during this process, ¹³ there is limited evidence on criteria for success in the clinical setting. Severity of koalas' burn injuries vary significantly ²⁵ and, not only are euthanasia and mortality rates before and at admission likely to be highly variable depending on fire intensity, there is also limited evidence on how triage, treatment, husbandry and release criteria impact recovery and post‐release survival. ²⁴ One study supports the ability of fire‐affected koalas to tolerate physiological stress during rehabilitation, and potentially more so than those affected by vehicle collisions, predator attacks or disease trauma, ¹³ but the significant variation in koalas' responses to typical stressors faced in care such as visitors and noise ¹⁵ suggests that individual personality ¹⁵ , ²⁶ and the characteristics of the clinical and rehabilitation setting may significantly influence outcomes. Research papers evaluated in this study did not detail triage, treatment and rehabilitation protocols or criteria for release of rehabilitated koalas, and there is currently no standardised method for this. Different criteria could produce different outcomes, and this area has already been identified as one where more research is needed. ⁹ This is demanding as it requires clear standardised criteria, replication across multiple settings, and control for the impact of habitat on post‐release survival. Incorporation of this research into bushfire response preparedness planning would be beneficial.

Upon release, koalas appear able to quickly and successfully adapt to a fire‐affected habitat due to their dispersal methods and use of epicormic growth on burnt trees. Their various methods of dispersal and ability to travel long distances allow for effective colonisation in the environment. ¹⁷ Different dispersal methods, which include exploratory, direct or incremental travel, mean that koalas establish their home range at different times post‐release, with some travelling for a longer period before settling, and in different locations with some closer to urban development than others. ¹⁷ Koalas can travel long distances before setting up their home ranges within burnt areas, also helping with dispersal and colonisation. Rehabilitation does not appear to affect the released koala's ability to adapt to a fire‐affected environment, given that studies provide no evidence that movement patterns or reproductive ability of these koalas differ from wild koalas in the same habitat. ¹⁶ , ¹⁷ Furthermore, koalas start to use burnt trees sprouting growth within five months of fire and after three months of start of growth, ¹⁶ and it appears the quality of this feed is adequate given that the proportion of burnt area within a home range does not affect the home range size. ¹⁷ These observations are further supported by Lane et al (2023) who found that the health and progress of rehabilitated koalas released into fire‐affected habitat in the Snowy Monaro region of NSW were not different from that of wild koalas in the same habitat. ²⁴

The success of the koala in adapting to a fire‐affected area upon release appears to be most clearly dependent on how the fire has damaged that environment. Two features of the fire‐affected habitat that affected the outcome following koala release are severity of the fire and location of the habitat. Low‐severity fire‐affected environments are more advantageous for koala colonisation as tree canopy is mostly unaffected, providing more opportunity for koalas to use the site. ¹⁸ There is evidence that it is the severity of fire rather than the extent of fire which indicates whether habitat is suitable for koala occupation, with a decrease in koala post‐ fire occupancy corresponding to an increase in fire severity. ²¹ In addition, while low‐severity fire grounds usually contained sites that were severely burnt, the impact of these only extended to a localised area and therefore did not affect the overall density of koala numbers in that fire ground. ²¹ The location of the fire ground, however, could have a bearing on the survival of the koala. Koala survival is impacted by the proximity of the site to urban areas that expose koalas, especially those that travel long distances when dispersing, to unfavourable fates such as dog attacks and vehicle collisions. ¹⁶ , ¹⁷ In comparison, Cristescu et al (2023) described a fire ground that had been managed for predators and was connected to unfragmented habitat, having an increase in koala occupancy post‐fire. The evidence indicates that the features of the fire‐affected habitat that provide more favourable conditions for koala release are low severity of fire impact, ¹⁸ , ²¹ increased distance from urban areas, connection to unfragmented habitat and previous management of predators. ¹⁶ , ¹⁷ , ²⁰

It may not be possible to find all these features in fire‐affected home sites given that the incidence of large and intense fires is predicted to be increasing ¹ , ⁴ and koala habitat continues to be affected by urbanisation ⁹ and fragmentation. ¹⁰ When original fire‐affected home sites are unsuitable for release, translocation of koalas may need to be considered, and this is an area where further research is warranted. A recently published study on Kangaroo Island koalas supports our assertion on the resilience of the koala and their ability to disperse, indicating that translocation is possible when there is adequate and unfragmented habitat available. ²⁷ Translocation, however, is a complex issue and when translocating due to unavailable fire‐affected release sites, consideration must be given to the animals being translocated as well as the recipient population. The success of translocation is affected by many factors, which include the increased risk of disease, ²⁸ the proximity to original habitat, the choice of location which could cause overcrowding problems, ²⁹ the habitat quality, ³⁰ the potential effects on the koala's gut microbiome ³¹ and the genetic recovery of an affected population. ³² This is a large and complex field considered beyond the scope of this paper but where further study needs to be undertaken, with more research needed on circumstances and processes that could support the successful translocation of koalas.

This study has implications for veterinary professionals who must make decisions about the prognosis of burnt koalas that come into veterinary care. The resilience of the koala in recovery and its ability to adapt to fire‐impacted sites upon release must be considered when triaging such patients. However, having adequate release sites is a third consideration in the triage process and one that involves government planning and development. It would be prudent to understand the consequences of development near koala habitat, and the importance of maintaining wildlife corridors and managing predators for successful koala release and survival. Further research looking into links between fire‐affected release sites, unfragmented habitat and predator management is warranted.

Conflicts of interest and sources of funding

The authors declare no conflicts of interest or sources of funding for the work presented here.

Hill, D. and Higgins, DP. , Factors that influence the decision to release rehabilitated koalas (Phascolarctos cinereus) to fire‐affected home sites. Aust Vet J. 2025;103:464–474. 10.1111/avj.13413

Data availability statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.