This dataset was compiled to assess exposure of 36 threatened Australian dryland mammals to future extreme heat. The study compares baseline (1981–2010) and projected (2041–2070) thermal envelopes defined using BIO5 (maximum temperature of the warmest month) from CHELSA v2.1 (Karger et al., 2017, 2021), evaluates the thermal suitability of translocation sites under future climate scenarios, and compiles a systematic review of thermal ecology literature for all focal species. Species distributions were compiled from occurrence records, Marsh et al. (2022), and expert consultation; historical distributions were reconstructed to represent pre-European-colonisation ranges. Full details are provided in Appendices S1–S3 of the associated manuscript. Due to their large file size and public availability, raw CHELSA climate rasters are not included; all extracted and processed climate values used in analyses are provided instead.

We calculated the baseline (1981–2010) heat envelopes for each species, defined as the range of maximum temperatures (BIO5: maximum temperature of the warmest month at 2-metres above ground) across both their historical and current geographic distributions. We then assessed future predictions to determine the proportion of each species’ current distribution projected to remain within either its current or historical heat envelope. Both historical and current distributions were included because many species have undergone > 90 % range contractions and the historical distribution may better reflect each species’ climatic tolerance. We defined thermal maxima as the upper limit of each of the current and historical heat envelopes. We compared future climate conditions at existing translocation sites to these current and historical thermal maxima to assess their potential suitability under projected climate scenarios.

The dryland boundaries used in this investigation were defined using the Köppen major climate classifications (Appendix S1; BOM, 2025). We assessed all mammal species listed as threatened under the IUCN Red List or the EPBC Act and included those with 10 % or more of their current or historical distribution within the dryland landscape (see Dickman & Pavey, 2023). For each species, we compiled both current and historical geographic distributions to account for extensive range contractions in Australian mammals since European colonisation. Historical distributions were constructed using occurrence records and expert consultation, while current distributions were adapted from Marsh et al. (2022), with translocation sites removed and boundaries refined using recent records. Full details of data sources, spatial filtering, and distribution construction are provided in Appendix S1.

This resulted in the inclusion of 34 threatened mammal species. Two additional conservation significant population units were included; the Pilbara population of the orange leaf-nosed bat (Rhinonicteris aurantia), which is listed as a vulnerable population (EPBC Act 1999), and the central Australian populations of the brushtail possum (Trichosurus vulpecula), given the restricted extent of these populations and candidature for reintroduction back into the dryland regions (Cooper et al., 2018; McDonald et al., 2015). For these species, historical distributions encompassed the full species range, while current distributions were restricted to the focal population units. Henceforth, unless otherwise stated, conservation status refers to IUCN listings.

We used CHELSA v2.1 climatology data (Karger et al., 2017; Karger et al., 2021) to represent baseline heat exposure from 1981–2010 and compared it to 15 future climate scenarios (five global circulation models × three shared socioeconomic pathways; SSP1-2.6, SSP3-7.0, and SSP5-8.5) projected for 2041–2070. Although many species disappeared from parts of their historical distribution prior to the baseline heat exposure period, we used CHELSA v2.1 consistently to define both current and historical heat envelopes as it uses statistical downscaling, offering more realistic representation of temperature variation in complex or remote terrains compared to alternatives derived from spatial interpolation, especially where weather stations are sparse (Jeffrey et al., 2001). For each species, we defined the baseline heat envelopes as the range of BIO5 values across either its current or historical geographic distribution. Future heat envelopes were defined as projected BIO5 values across each species’ current distribution under each future climate scenario. Analyses were conducted in R (R Core Team, 2025) using the sf (Pebesma, 2018) and terra (Hijmans et al., 2022) packages.

To assess vulnerability to future heat, we placed species into three risk categories utilising the mean across the 15 scenarios. Species were classified as low risk if > 50 % of their future heat envelope remained within both their current and historical heat envelopes. Moderate-risk species retained > 50 % overlap with their historical heat envelope, but < 50 % with their current envelope. High-risk species had < 50 % overlap with either envelope. To be conservative, species were assigned to a higher risk category if the standard error of the mean climate scenario overlapped with these criteria. Using this method, species facing novel heat under future warming compared to both their historical and current heat envelopes are classified as high risk; those exposed to novel heat only relative to their current envelopes are classified as moderate risk; and species with little or no novel heat exposure are classified as low risk (Figure 1).

We assessed predictors of risk by testing relations between risk categories and species-specific traits. Categorical associations, taxonomic order and IUCN status, were tested using Chi-squared tests. For continuous variables, including range size (log-transformed), body size (log-transformed; Jones et al., 2009; Van Dyck et al., 2013), difference in area between historical and current range thermal maxima, historical and current heat envelope width, proportion of distribution in arid zones, and percent of remnant habitat, we used Kruskal-Wallis tests with post-hoc Dunn’s tests to assess differences across risk categories.

We assessed the differences in BIO5 between the historical thermal maxima and their translocation sites. Translocation sites were sourced from Legge et al. (2018) and supplemented by additional information by Woinarski et al. (2023), Crisp et al. (unpublished data), and a screening of the recent literature (Appendix S2). We excluded all failed translocations and any havens where species were remnant, rather than translocated, from analyses. In the case of the brushtail possum, we only included translocations into the dryland areas of its historic range. We treated each translocation site as a point location and assessed the BIO5 value from 1981–2010 and projected the mildest and most extreme future scenarios at the location to compare translocation site temperatures with thermal maxima to assess their potential suitability as thermal refugia.

We produced species-specific heat exposure plots for all threatened dryland mammals in Australia using ggplot2 (Wickham, 2016). These plots present the historical and current heat envelopes of the species and compare these to the mildest and most extreme future climate scenarios. We then compared these heat envelopes to all the current conservation translocation sites for the species, under current, mild, and extreme future conditions. We generated species-specific maps to visualise projected heat exposure across their distributions. For each species, we used the historical thermal maximum as a benchmark to identify areas in their historical range projected to exceed it under all, some, or no future scenarios. These maps can be used to identify potential thermal refugia and priority areas for conservation or translocation. An annotated example of a heat exposure plot is presented in Figure 2 for the endangered numbat.

We conducted a systematic review for all 36 species to assess the literature directly relevant to their thermal ecology. We compiled common names, binomial names, and recent synonyms as listed by Baker and Gynther (2023). A Boolean search string was constructed where species names were combined with primary search terms, such as “heat”, “warming”, and “climate change” (Appendix S3). We conducted literature searches using Web of Science and Scopus in June 2025 and adopted a PRISMA reporting procedure to screen the literature (see Appendix S3; Page et al., 2021). We obtained additional resources by applying a snowball technique using the reference section of publications as a source for other relevant papers (Greenhalgh & Peacock, 2005; Rubenstein et al., 2023). Review papers were excluded from the final step but were used to find additional papers from their cited references.

We included any study that described a species’ response to temperature, functional traits, or processes related to heat exposure or thermoregulation. A Kruskal–Wallis test was then used to investigate biases in the average number of papers per species across the heat risk categories described previously. Papers were grouped into three broad themes: (1) individual responses, the observed behavioural or physiological adaptations of an individual animal in response to an increase in temperature and/or consequences of these events, (2) population responses, the population level responses of a species to an increase in temperature, observed or modelled, and finally (3) innate traits or processes, which includes traits, behaviours, or processes which will have a bearing on individual persistence and survival in extreme conditions but are not explored in this context, such as shelter site use, basal metabolic rate, or pelage properties. These themes were further subdivided into nine key categories: behavioural response, physiological response, consequences, population response, morphological adaptation, climate predictions, thermal refuges, functional traits, and functional processes (Table 1).

Baker, A., & Gynther, I. (Eds.). (2023). Strahan's Mammals of Australia (4th ed.). Reed New Holland.

Greenhalgh, T., & Peacock, R. (2005). Effectiveness and efficiency of search methods in systematic reviews of complex evidence: audit of primary sources. BMJ, 331(7524), 1064-1065.

Hijmans, R. J., Bivand, R., Forner, K., Ooms, J., Pebesma, E., & Sumner, M. D. (2022). Package 'terra'. CRAN.

Jeffrey, S. J., Carter, J. O., Moodie, K. B., & Beswick, A. R. (2001). Using spatial interpolation to construct a comprehensive archive of Australian climate data. Environmental Modelling & Software, 16(4), 309-330.

Jones, K. E., Bielby, J., Cardillo, M., Fritz, S. A., O'Dell, J., Orme, C. D. L., Safi, K., Sechrest, W., Boakes, E. H., Carbone, C., Connolly, C., Cutts, M. J., Foster, J. K., Grenyer, R., Habib, M., Plaster, C. A., Price, S. A., Rigby, E. A., Rist, J., Teacher, A., Bininda-Emonds, O. R. P., Gittleman, J. L., Mace, G. M., & Purvis, A. (2009). PanTHERIA: a species-level database of life history, ecology, and geography of extant and recently extinct mammals. Ecology, 90(9), 2648.

Karger, D. N., Conrad, O., Böhner, J., Kawohl, T., Kreft, H., Soria-Auza, R. W., Zimmermann, N. E., Linder, H. P., & Kessler, M. (2017). Climatologies at high resolution for the earth's land surface areas. Scientific Data, 4(1), 1-20.

Karger, D. N., Conrad, O., Böhner, J., Kawohl, T., Kreft, H., Soria-Auza, R. W., Zimmermann, N. E., Linder, H. P., & Kessler, M. (2021). Climatologies at high resolution for the earth's land surface areas [Raster dataset]. EnviDat. https://doi.org/10.16904/envidat.228

Legge, S. et al. (2018). Havens for threatened Australian mammals: the contributions of fenced areas and offshore islands to the protection of mammal species susceptible to introduced predators. Wildlife Research. 45: 627-644.

Marsh, C. J., Sica, Y. V., Burgin, C. J., Dorman, W. A., Anderson, R. C., del Toro Mijares, I., Vigneron, J. G., Barve, V., Dombrowik, V. L., & Duong, M. (2022). Expert range maps of global mammal distributions harmonised to three taxonomic authorities. Journal of Biogeography, 49(5), 979-992.

Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., Gluud, C., Mayo-Wilson, E., McDonald, S., McGuinness, L. A., Stewart, L. A., Thomas, J., Tricco, A. C., Welch, V. A., Whiting, P., & Moher, D. (2021). The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ, 372, n71.

Pebesma, E. (2018). Simple features for R: standardized support for spatial vector data. The R Journal, 10(1), 439-446.

R Core Team. (2025). R: A Language and Environment for Statistical Computing (Version 4.4.1). R Foundation for Statistical Computing. https://www.r-project.org/

Rubenstein, M. A., Kampe, T. U., & Steele, M. K. (2023). Snowball sampling for systematic mapping and review: an application to urban ecology. Frontiers in Ecology and Evolution, 11, 1106568.

Van Dyck, S., Gynther, I., & Baker, A. (Eds.). (2013). Field Companion to the Mammals of Australia. New Holland Publishers.

Wickham, H. (2016). ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York.