Will tropical mountaintop plant species survive climate change? Identifying key knowledge gaps using species distribution modelling in Australia

Biological Conservation 191 (2015) 322–330

Contents lists available at ScienceDirect

Biological Conservation journal homepage: www.elsevier.com/locate/bioc

Will tropical mountaintop plant species survive climate change? Identifying key knowledge gaps using species distribution modelling in Australia Craig M. Costion a,b,c,⁎, Lalita Simpson a, Petina L. Pert d,e, Monica M. Carlsen b, W. John Kress b, Darren Crayn a,e,f a

Australian Tropical Herbarium, E2 Bld. James Cook University, Cairns Campus, PO Box 6811, Cairns, QLD 4870, Australia Department of Botany, National Museum of Natural History, MRC 166, Smithsonian Institution, PO Box 37012, Washington, DC 20013-7012, USA c Centre for Tropical Biodiversity and Climate Change, James Cook University Cairns Campus, PO Box 6811, Cairns, QLD 4870, Australia d CSIRO Ecosystem Sciences, PO Box 12139, Earlville BC, Cairns, QLD 4870, Australia e College of Marine and Environmental Sciences, James Cook University, Cairns, QLD 4870, Australia f Centre for Tropical Environmental Sustainability Science, James Cook University Cairns Campus, PO Box 6811, Cairns, QLD 4870, Australia b

a r t i c l e

i n f o

Article history: Received 29 December 2014 Received in revised form 18 May 2015 Accepted 15 July 2015 Available online xxxx Keywords: Cloud forest Endemics Threatened species Range shifts Acclimation Wet Tropics

a b s t r a c t Species inhabiting tropical mountaintops may be most at risk from the detrimental effects of climate change. Yet few regional assessments have critically assessed the degree of threat to species in these habitats. Here we model under three climate scenarios the current and future suitable climate niche of 19 plant species endemic to tropical mountaintops in northeast Queensland, Australia. The suitable climate niche for each of the 19 species is predicted to decline by a minimum of 17% and maximum of 100% by 2040 (mean for all species of 81%) and minimum of 46% (mean for all species of 95%) by 2080. Seven species are predicted to have some suitable climate niche space reductions (ranging from 1 to 54% of their current suitable area) by 2080 under all three climate scenarios. Three additional species are projected to retain between 0.1 and 9% of their current distribution under one or two of the climate scenarios. In addition to these declines, which are predicted to occur over the next 30 years in northeast Queensland, we discuss and outline pressing research priorities that may be relevant for the conservation of biodiversity on tropical mountaintop environments across the globe. Specifically, further research is needed on thermal tolerances, acclimation potentials, and physiological constraints of tropical mountaintop taxa as current species distributions are primarily determined by climatic factors. © 2015 Elsevier Ltd. All rights reserved.

1. Introduction Large-scale changes in the distributions of species and complete turnover of future ecosystems across the world have been predicted as responses to climate change (Parmesan and Yohe, 2003; Williams et al., 2007). In particular, the montane tropics, although little studied in comparison to documented cases in the temperate zone (Thomas, 2010), are considered highly threatened by even marginal increases in temperature due to steep environmental gradients (Corlett, 2011), narrow thermal tolerances of tropical species (Cunningham and Read, 2003b; Laurance et al., 2011), and by the secondary effects of rising ocean surface temperatures that are expected to increase the average altitude of cloud base formation (Foster, 2001; Still et al., 1999). Evidence for direct impacts of recent climate change on tropical mountaintops includes widespread amphibian extinctions and altering of community structure in Costa Rica (Pounds et al., 1999, 2006), and upslope ⁎ Corresponding author at: Department of Botany, National Museum of Natural History, MRC 166, Smithsonian Institution, P.O. Box 37012, Washington, DC 20013-7012, USA. E-mail address: [email protected] (C.M. Costion).

http://dx.doi.org/10.1016/j.biocon.2015.07.022 0006-3207/© 2015 Elsevier Ltd. All rights reserved.

displacement and range shifts of montane species in Madagascar (Raxworthy et al., 2008), Borneo (Chen et al., 2009) and Costa Rica (Pounds et al., 1999). A recent study examined data on the browning and greening of vegetation in response to recent climate changes in tropical montane zones in five continental regions and found that the rates of vegetation change were dependent on elevation with warming rates more pronounced at higher elevations (Krishnaswamy et al., 2014). Here we investigate the impacts of climate change on a tropical montane flora in northeastern Australia's Wet Tropics bioregion (Fig. 1). A rise in the cloud base has already been projected to affect the availability of high and consistent moisture in Queensland's coastal mountain habitats (McJannet et al., 2007). In particular, the mountaintop regions of northeast Queensland were identified as “disappearing environments” in a global assessment of projected impacts of climate change on specific vegetation types (Williams et al., 2007). A subsequent study extended this assessment using species composition dissimilarity data and found similar results (Williams et al., 2012). The analysis suggested that the current tropical montane environment which harbors a unique assemblage of plants and animals will simply disappear and is unlikely to occur or be replicated anywhere else in

C.M. Costion et al. / Biological Conservation 191 (2015) 322–330

323

Fig. 1. The Wet Tropics bioregion of northeast Queensland, Australia (a); region overview (b); and centres of species diversity for the mountaintop endemic flora zoomed in for regions north (c) and south (d) of the Black Mountain Corridor.

Australia by 2070. Prior to this alarming prediction, Williams et al. (2003) investigated how climate change projections will affect the bioregion's vertebrate species and reported that increasing temperature is predicted to result in a significant reduction or complete loss of the core habitat of all regionally endemic vertebrates, the majority of which are restricted to upland habitats. The Wet Tropics bioregion of northeastern Queensland in Australia is a World Heritage Area and is renowned for its antiquity and relictual Gondwanan taxa. It is home to 65 regionally endemic vertebrates, 674 endemic vascular plants, including 62 monotypic genera and one monotypic family (Metcalfe and Ford, 2008), and includes rain forest areas that are believed to have served as dynamic refugia for over 40 million years. A total of 5% of this bioregion occurs above 1000 m elevation and 96% of this area occurs within well functioning protected areas. An investigation of this bioregion provides an ideal opportunity for a first approximation of climate change threats to a well-protected environment of high conservation value. Historical climate change has been a primary force shaping biodiversity patterns of this region. Pleistocene contractions of rainforest during glacial periods and subsequent expansions during global warming events left their mark on the extant vertebrate species assemblages (Williams and Pearson, 1997; Winter, 1997). The same process facilitated the migration of Asian plant lineages into tropical Australia (Crayn et al., 2015) creating novel Asian–Gondwanan mixed species assemblages in rain forest re-expansion areas (Costion et al., 2015). Historical climate changes have been a primary driver of extinctions and speciation in the region (Schneider et al., 1998; Williams and Pearson, 1997), and tropical mountaintop environments are widely accepted as highly vulnerable to climate change (Brooks et al., 2006; La Sorte and Jetz, 2010), possibly more so

than temperate mountain environments (Sheldon et al., 2011). For these reasons we tested the prediction that climate change will have direct and significant impacts on the distribution of the endemic montane flora of northeast Queensland. Using environmental niche modelling methods we assessed the change in the geographical extent of the climate niche of 19 endemic plant species under three climate scenarios (a1b, a2, b1) over three time periods (2040, 2060, and 2080). 2. Methods We used herbarium specimen records to identify plant species that are endemic to the Wet Tropics Bioregion and restricted to areas above 1000 m in elevation. Our use of the term “mountaintops” includes some local geographical areas that are more correctly defined as elevated plateaus or tablelands ≥1000 m. Initially, records from the Australian Tropical Herbarium (CNS) were searched to identify species for which all records occurred at ≥ 1000 m within the Wet Tropics bioregion (Metcalfe and Ford, 2008). These records were then refined to current taxonomy using the Australian Plant Name Index (APNI: http://www. cpbr.gov.au/apni/index.html accessed August 2013). A final search was performed using Australia's Virtual Herbarium (AVH: http://avh. ala.org.au accessed August 2013) to obtain all digitised collection records in Australian herbaria for each taxon. Many of the records were collected prior to GPS technology; however, in most cases these collections were made with precise elevation data. All GPS location records for each specimen were manually checked for accuracy using available data on each corresponding herbarium label and validated or corrected using Google Earth where necessary for maximum accuracy of locality data. This process necessitated further filtering of taxa from the final list

324

C.M. Costion et al. / Biological Conservation 191 (2015) 322–330

where species were found to occur below 1000 m and for taxa deemed to have too few collection records to be accurately modelled (i.e. b8 records). A total of 32 species were identified as endemic to areas N1000 m in the Queensland Wet Tropics bioregion. Eleven of these species were excluded from our modelling analysis due to lack of sufficient occurrence data. These include five fern species, two orchids, and four woody angiosperm species: Pilidiostigma sessile N. Snow, Leionema ellipticum Paul G. Wilson, Ceratopetalum hylandii Rozefelds & R.W. Barnes, Hollandaea porphyrocarpa A.J. Ford & P.H. Weston. Two additional woody angiosperms that meet our criteria for mountaintop endemics, Dracophyllum sayeri F. Muell. and Hypsophila halleyana F. Muell., were excluded from the analysis due to suspected errors in the collection records. The final dataset used for our analysis included a total of 19 species of trees and shrubs. MAXENT ver. 3.3.3, niche modelling was used to infer the potential distribution of suitable climate for 19 species, within the Wet Tropics bioregion, under current and future climate scenarios (Phillips et al., 2006). Suitable climate is defined as an area or areas providing a climatic niche that a species currently occupies. This was used as a surrogate for habitat suitability following VanDerWal et al. (2009) and is referred throughout the text as “suitable habitat” to be consistent with the IUCN Red List terminology. MAXENT was chosen as it has previously been reported to outperform other modelling methods for predicting potential species distributions (Elith et al., 2006). Training of the models was based on 8– 122 (mean 37) occurrence records for each species. Climate layers used in the analysis consisted of bioclimatic variables, mapped at 250 m resolution across the Wet Tropics bioregion, sourced from the James Cook University Tropical data hub (http:// tropicaldatahub.org/, accessed 14 December 2012) which had previously been created using the “climates” package in R (VanDerWal et al., 2011) with baseline climate surfaces from ANUCLIM 6.1 software with a climate baseline of 1975–2005 (Hutchinson et al., 2000). Initially, test models were run using all 19 bioclimatic variables available, and these variables were ranked for significance against each species based on a jack-knife statistical test of variable importance. The 12 bioclimatic variables containing the highest significance for all species were then selected to further model the distribution of suitable habitat for each species. This modelling approach is commonly used however it could bias results specially when the number of variables included in the model is higher than the number of locality records used to train the model. In our case since six species had less than 19 records, including one species with eight records and two species with 13 records, it was deemed necessary to account for this potential bias. We addressed this by running a second set of models by selecting only uncorrelated and species-specific bioclimatic important variables. To account for potential correlations between the bioclimatic variables, we used Pearson's correlation coefficient on the climatic data extracted for each locality occurrence point in the entire dataset. Uncorrelated variables included: Annual Mean Temperature (Bio1), Mean Diurnal Range (Bio2), Isothermality (Bio3), Temperature Seasonality (Bio4), Max Temperature of Warmest Month, (Bio5), Min Temperature of Coldest Month (Bio6), Temperature Annual Range (Bio7), Mean Temperature of Wettest Quarter (Bio8), and Annual Precipitation (Bio12). We then ran test models in MAXENT using all the uncorrelated bioclimatic variables and selected specific variables for each species based on the jack-knife test of variable importance. For our final modelling run the total number of bioclimatic variables used for each species ranged from four to seven (Table 1). Modelling was completed with 10 replicates using the cross validation option with linear, quadratic, product, hinge and the 10th percentile training presence threshold features enabled. To assess the impact of predicted climate change, the distribution of suitable habitat was modelled for the time periods: 2040, 2060 and 2080, using the same 12 environmental variables mentioned above. Climatic conditions were simulated by CSIRO's Mk3.5 climate model (Gordon et al., 2010) based on three future CO2 emission scenarios, A1B, A2, B1 (IPCC special report, 2000).

Table 1 Model parameters used in the second set of MAXENT modelling analyses (i.e. species-specific models). Numbers of environmental variables used for each species correspond to standard BioClim variables. Species

BioClim environmental variables used

Austromuellera valida Cinnamomum propinquum Cryptocarya bellendenkerana Diospyros sp. Mt Spurgeon Elaeocarpus linsmithii Elaeocarpus sp. Mt Misery Eucryphia wilkiei Micromyrtus delicata Phaleria biflora Planchonella sp. Mt Lewis Polyosma sp. Mt Lewis Solanum eminens Symplocos graniticola Symplocos orbesia Symplocos bullata Syzygium fratris Tasmannia sp. Mt Bellenden Ker Uromyrtus metrosideros Zieria alata

5,6,8,12 4,5,6,7,8,12 4,5,6,8,12 4,5,6,8,12 1,4,5,6,7,8,12 4,5,8,12 1,5,6,7,12 3,4,6,8,12 1,4,5,6,8,12 2,4,5,6,8,12 2,4,5,6,7,8,12 2,7,8,12 2,3,4,5,6,8,12 1,3,4,5,7,8,12 1,4,5,7,8,12 1,5,7,12 2,4,6,8,12 2,4,5,6,7,8, 1,4,5,6,7,8,12

To create a binary distribution map of suitable and unsuitable habitat the 10th percentile training presence threshold was applied. This threshold predicts unsuitable habitat for 10% of occurrence records where environmental variables are considered the most extreme. Thresholds for each species were applied to the averages of all 10 runs for current and each projected date 2040, 2060, 2080 and climatic scenario, and displayed as binary distribution data (suitable or unsuitable habitat). The results of this second set of more species-specific models were used to test the accuracy of the overall trend identified in the first run using the same 12 variables for all species. Output ascii files from the MAXENT analysis that were converted to binary grid files (where 1 = suitable habitat, and 0 = unsuitable habitat) were projected using ESRI ArcGIS ver. 10.2. Species richness maps were calculated showing the sum total of suitable habitat for all species under each climate scenario using the functions of Spatial Analyst extension tools. The total suitable habitat was then calculated in hectares for each species under each climate scenario and date, and then recalculated for each species using only suitable habitat known within each species current known range. The proportion of the current suitable habitat remaining for each time and climate scenario was then calculated and presented as “percent decline” for each species where a 99% decline represents a 1% of the current suitable habitat. 3. Results 3.1. Inferred habitat declines The 19 mountaintop species analysed here are most abundant at seven locations in the region (Table 2, Fig. 1), each of which maintains some suitable habitat in at least one of the climate scenarios by 2080. Three of these mountaintops, Carbine Uplands, Mount Bartle Frere, Table 2 Centres of diversity for the 19 mountaintop regional and local endemic plant species modelled for the Wet Tropics of northeast Queensland, Australia. No.

Locality

Elevation (m)

Endemics

1 2 3 4 5 6 7

Mt. Finnigan Thornton Peak Windsor Tableland Carbine Uplands Lamb Ranges Mt. Bellenden Ker Mt. Bartle Frere

1148 1338 1356 1383 1308 1590 1622

4 4 2 13 (6) 2 6 (1) 7 (2)

C.M. Costion et al. / Biological Conservation 191 (2015) 322–330

and Mount Bellenden Ker, have local endemics – species which are restricted to that upland locality only – with the Carbine Uplands showing the highest richness (13 species, six of which are locally endemic). MAXENT model was high for all species, receiving AUC scores between 0.89 and 1.0. Model predictions of habitat suitability under contemporary climatic conditions correspond well to the known distributions. Prediction of suitable habitat for some species included areas outside a species current range, reflecting the potential distribution for a species. Realised distributions often don't meet the potential distribution due to factors such as dispersal limitations or competition. Therefore, results for total suitable habitat were compared within (i.e. realised) and outside (i.e. potential) each species' known current range. When the total declines of suitable habitat per species were compared with the declines of suitable habitat only within the current known range of each species an average difference of only 1% across all three scenarios and dates was found. Results in Figs. 2–3 and Table 3 are derived from the data within each species' current known range to give a better estimate of total projected declines. Results in Fig. 4 are derived from the total suitable habitat data, including areas outside each species current range, to identify the most likely mountaintop areas that in the future could maintain suitable habitat or act as sinks for range shifts and/or dispersal events. A rapid decline in extent of suitable habitat was inferred for all species by 2040 (Table 3, Fig. 2). The mean inferred decline of habitat per species across all climate scenarios by 2040 is 81%. Overall, the general pattern of decline is similar for all three scenarios, in terms of total loss of suitable habitat (Fig. 2) with marginal differences between each of the climate scenarios (Fig. 3). The pattern shown in Fig. 3 indicates that the most significant loss of habitat is predicted to occur between 2040 and 2060 at which time 8–12 species will be at risk of complete extinction. By 2080, suitable habitat for all species declined

325

20 Total spp. with suitable/ unsuitable (red) habitat

18 16 14 12 10

A1B

8

A2

6

B1

4

Potential extinctions

2 0

Fig. 3. Total number of species with/without suitable habitat projected as remaining for each climate scenario and time period. ‘Potential extinctions’ or species without suitable habitat (shown in red) indicates the number of species for which 100% of habitat has been lost in at least one of the three climate scenarios. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

by a minimum of 46%, with a mean decline of 95% across all species and all scenarios (Table 3). All three scenarios predict the loss of all suitable habitat for nine out of 19 species by 2080, with the remaining ten species having lost 46–99.9% of suitable habitat. Seven species are predicted to have some suitable habitat remaining under all three climate scenarios by 2080, and an additional three are predicted to persist under one or two of the three scenarios (Table 3). Declines projected by the first round of analyses using the same 12 climatic variables for all species predicted faster and more extreme declines for many of the species (Supporting Information). The average decline across all species for all three scenarios was 12% higher at 93%, the

Fig. 2. Decline in the total suitable habitat area (ha) per species for each climate scenario (A1B, A2 and B1) over three time periods.

326

C.M. Costion et al. / Biological Conservation 191 (2015) 322–330

Table 3 Summary of the decline in area of suitable habitat for all 19 endemic mountaintop tropical plant species of northeast Queensland for three climate change scenarios (A2, A1B, and B1) for the time periods 2040, 2060, and 2080. Totals per time period show minimum and maximum decline across all three scenarios. Boxes are highlighted in red where decline is 100% for all three scenarios. Species

Austromuellera valida Cinnamomum propinquum Cryptocarya bellendenkerana Diospyros sp. Mt Spurgeon Elaeocarpus linsmithii Elaeocarpus sp. Mt Misery Eucryphia wilkiei Micromyrtus delicata Phaleria biflora Planchonella sp. Mt Lewis Polyosma sp. Mt Lewis Solanum eminens Symplocos graniticola Symplocos orbesia Symplocos bullata Syzygium fratris Tasmannia sp. Mt Bellenden Ker Uromyrtus metrosideros Zieria alata

Total suitable habitat min–max % decline

Suitable habitat (2080) outside current range

Current area (ha)

2040% Decline

2060% Decline

2080% Decline

A2

A1B

B1

24,050 5431.25 50,950 41,662.5 32,281.25 52,425 3350 18,981.25 14,693.75 16,481.25 22,118.75 881.25 20,056.25 37,893.75 18,443.75 2343.75 3487.5 78,281.25 19,850

83–99% 62–72% 74–83% 85–87% 54–75% 97–99% 45–49% 100% 98–99% 99–100% 99–99.9% 70–90% 53–82% 97–98% 96–99% 50–54% 17–58% 76–79% 99–99.9%

100% 76–88% 93–96% 99–100% 78–91% 100% 56–68% 100% 100% 100% 99.9–100% 91–100% 97–100% 99–100% 99–100% 62–75% 46–72% 92–94% 100%

100% 85–99% 92–96% 100% 85–98% 100% 67–81% 100% 100% 100% 100% 91–100% 99.6–100% 99.9–100% 100% 75–91% 46–87% 97–98% 100%

– No No – No – No – – – – – Yes – – No No No –

– No No – No – No – – – – – – – – No No No –

– No No – No – No – – – – No – No – No No No –

minimum decline for a species by 2080 was 97%, and the number of species with projected habitat remaining by 2080 for all three climate scenarios was only 3. The overall total decline of suitable habitat averaged across all species and scenarios by 2080 however was only higher by 4% at 99%. Spatial patterns in the decline of suitable habitat show parallel trends between all scenarios, predicting a decline in 5–7 montane habitats that are highly reduced by 2080 (Fig. 4). Suitable habitat is retained at Bellenden Ker, Bartle Frere, Carbine Uplands, Thornton Peak, and Mount Finnigan for all climate scenarios, while the Windsor Tableland

habitat is only retained in scenario A2, and patches of the Lamb Ranges and Black Mountain are retained in scenario B1. Most striking is the loss of habitat in the Carbine Uplands, the largest, most species rich area (Table 1) for mountain endemic plants in the study area. Suitable habitat on the Carbine Uplands reduces to tiny patches or completely vanishes in all three climate scenarios (Fig. 4). These predictions assume that climate is a limiting factor for all 19 species and do not consider dispersal probabilities between the current populations and other habitats deemed climatically suitable

Fig. 4. Total number of species (species richness) with suitable habitat vs. unsuitable habitat (presence/absence) projected using 10% training presence threshold rule in MAXENT for climate scenario A1B.

C.M. Costion et al. / Biological Conservation 191 (2015) 322–330

in the area. If the model predictions are accurate, at least one of the six species is predicted to have suitable habitat in 2080 (in one or more climate scenarios), will require either natural dispersal to another area or assisted relocation, because this taxon is not currently known from the locality deemed climatically suitable in the future (Table 3). 4. Discussion

327

here. Dated molecular phylogenies can give insight into the origins of the mountaintop endemic flora, which could then provide more information on what types of climates and historical changes these species or their progenitors have endured over time. More detailed population genetics work should identify refugial versus more recently expanded populations. If climatic factors are indeed critical to survival, these species must either acclimate to persist within their current ranges or disperse to new areas where climatic conditions are similar to those in their current distribution.

4.1. How accurate are the predictions? Our analysis presents an alarming scenario for northeast Queensland's mountaintop endemic flora. These results must be interpreted carefully, however, because they assume that the current climate occupied by each species is essential for its survival. Because this assumption remains uncertain, our results highlight an important research gap that should be considered a high conservation priority. Other factors may explain the current distribution of these 19 mountaintop endemic species, such as adaptation to exposed basalt outcrops and/or soil types. Identifying appropriate biological links between climatic factors and species distributions are essential to determine if these projected declines are realistic. Our results give a high level of certainty that the current climate conditions on the mountaintop environments of northeast Queensland will be significantly altered by 2080. However, not all species occupying this generalized climate niche will respond in a similar manner. Indeed, using different sets of climatic variables for the analyses demonstrated that a one size fits all approach with a higher number of environmental variables predicted faster and more drastic declines. Doing a high throughput modelling run of a large number of species using the same set of climatic variables for all of the species can identify general trends accurately and may be appropriate for a first look or preliminary assessment of a region. However, this approach may overpredict the severity of declines. In our case, the general pattern of decline across all species did not change, but more habitat was retained for more species by 2080 when fewer variables are selected and the analysis run separately for each species. For species with very narrow ranges, such as found in mountaintop environments, a small amount of habitat can be significant for species survival. If it is assumed that climate is a limiting factor for these 19 species, then declines in the area of suitable habitat observed in our data for all species in each of the three projected future climate scenarios are very abrupt, with the majority of declines occurring between the present and 2040. By 2080 no suitable habitat is predicted within the Wet Tropics bioregion for 84% of species, under all emission scenarios. Although further work is required to link climate tolerances to the distribution of these species, these results are consistent with projections that have been previously conducted for other mountaintops worldwide (Williams et al., 2007) and Australia-wide (Williams et al., 2012). Current levels of climate change are considered unprecedented for observed events over the last 2000 years. These events resemble a similar transition from the humid Holocene to late Holocene drier period, which caused abrupt changes in global vegetation patterns (Thompson et al., 2006). Ice core evidence spanning the last glacial cycle indicates that historical climate changes were abrupt, with shifts of temperature and climate up to half as large as the entire difference between the ice age and modern conditions occurring across the globe in years to decades (Alley, 2000). Because the species in our study pre-date abrupt changes in past climate periods in the geological time-scale, they have the potential to persist through climatic shifts predicted for the coming decades. The abrupt drying and cooling changes during the past glacial periods are likely to have favoured expansion of habitat for these mountaintop taxa allowing their ranges to expand to lower elevations. Therefore, their current distributions may actually be in refugial states. Little work to date has focused on dating the origins of the taxa studied

4.2. Physiological tolerances and acclimation potential It is generally accepted that species range shifts are relative to the physiological tolerances of the taxa (Scheffers et al., 2013; Sunday et al., 2012). Very few documented assessments of thermal tolerances of tropical trees have been conducted to date and none of these have included assessments of mature trees (Corlett, 2011). Our results rest upon a primary assumption that the ranges of species included in this study are determined by climate and their physiological tolerances fall within the climatic niche that they currently occupy. Although ample adjacent habitat below 1000 m exists, without any geographic barriers (other than elevation), no occurrences are recorded in these areas. The lack of occurrences below 1000 m might be explained by their inability to compete in habitats at lower altitudes due to variable physiological tolerances to increases in temperature and/or decreases in moisture at lower altitudes. Since the northeast Queensland flora is regarded as being largely derived from a cooler, more temperate Gondwanan rainforest flora and Earth temperatures over the last 800,000 years were generally cooler than today (Jouzel et al., 2007), it is more likely that evolution in these plant species involved stronger selection for cold tolerance than for heat tolerance (Colwell and Rangel, 2010). Acclimation and/or rapid evolutionary change may be an essential factor for the survival of tropical mountaintop species. Evidence for acclimation potential has been documented for deciduous trees (Gunderson et al., 2010) and eucalypts (Ghannoum et al., 2010) in temperate environments, and evidence from another study (Cunningham and Read, 2003b) has indicated that temperate species have a higher acclimation potential than tropical species. The latter study is of particular relevance here because it compared temperate and tropical rainforest trees of Australia. One of the temperate species in this study, Eucryphia lucida, is closely related to one of our mountaintop species, Eucryphia wilkiei. A followup to this temperate/tropical acclimation work further showed that temperate rainforest species have maximum growth at temperatures higher than the average summer temperatures of their native climates (Cunningham and Read, 2003a) suggesting the likelihood that temperate species have a stronger capacity to acclimate to the increased temperatures projected by climate change than tropical species. Acclimation was also observed for the temperate southwestern European plant Umbilicus rupestris (Woodward, 1990), which evolved rapidly to new low temperature responses after being experimentally transplanted outside its natural geographical range. However, not all species will show adaptive responses to climate change. Some plants simply have predicted rates of evolutionary response that are much slower than the predicted rate of climate change (Huntley et al., 1989; Etterson and Shaw, 2001). Because intergenerational selection or selection at shifting species distribution margins is required for acclimation and/or evolutionary change to take place, long-lived species, such as woody trees and shrubs versus herbaceous annuals and perennials, and poor dispersers, such as non-fleshy fruited species requiring either gravity or non-flying mammals for dispersal, are less likely to undergo such changes (Pearson and Dawson, 2003). All the species in this study may be considered “long-lived” being woody trees and shrubs but little work has been done to date evaluating their dispersal potential.

328

C.M. Costion et al. / Biological Conservation 191 (2015) 322–330

4.3. Biotic interactions Acclimation is not the only requisite for surviving drastic and/or rapid climate changes. A key issue when considering the viability of species in changing climates is whether species can still be competitive in novel species assemblages created by climate change, i.e., current lower elevation species migrating to higher elevations (Colwell et al., 2008; Smith et al., 2009). Large tracts of un-fragmented rain forest in well-established conservation areas surround and include each of the mountaintops in our study. These large areas of rain forest habitat ranging from 700 to 1000 m contain hundreds of species well adapted to the rain forest habitat at this elevation. Climate modelling predicts that these species will migrate higher up to the mountaintops and subsequently compete with the species modelled here and possibly even interfere with existing inter-species interaction. The ultimate result of these changing vegetation dynamics as new ranges overlap is unknown. Climate change is also predicted to affect ecological processes such as plant–pollinator interactions (Fleming and Kress, 2013). A recent study showed that even marginal temperature changes (1–4 °C) can significantly affect plant productivity and phenology in the tropics (Pau et al., 2013). Temperature increases may further affect obligate and/or generalist insect–plant pollination interactions critical for the survival of species. An increase of 3 °C for example, has been shown to potentially threaten the survival of an obligate fig wasp species in Singapore by decreasing the active adult life span of the wasps (Jevanandam et al., 2013). The phenology and pollination biology of the northeast Queensland mountaintop endemic plants have also not been studied to date and work is needed to assess the resilience of such mutualistic relationships in the montane tropics in the face of projected climate changes. The lack of consideration of biotic interactions in species distribution modelling research is problematic (Davis et al., 1998; Guisan and Thuiller, 2005). However, many bioclimatic models have successfully predicted current species distributions at different scales and demonstrate that many species distributions can indeed be predicted using only climatic factors (Pearson and Dawson, 2003). Data on biotic interactions for inclusion in models is rare, especially for tropical regions. Until such data become available, the inclusion of biotic interactions as a key variable in species distribution modelling will remain limited.

4.4. Dispersal potential With our modelling results predicting such drastic declines in suitable habitat within the Wet Tropics bioregion, the dispersal potential of species to other suitable habitat outside their existing ranges must be considered. Species with high dispersal potential can access all future potential climate space whereas species with lower dispersal potential will be restricted to suitable climates within their existing range (Peterson et al., 2001). Evidence of dispersal potential exists for three of the 19 taxa we examined. Cryptocarya bellendenkerana, Elaeocarpus linsmithii, and Uromyrtus metrosideros all have large current distributions across the bioregion with populations occurring both north and south of a recognized biogeographical barrier known as the Black Mountain Corridor (See Fig. 1). This area did not contain rainforest at the last glacial maximum (Hilbert et al., 2007) and has been shown to be a biogeographic barrier for many vertebrate species (Joseph et al., 1995; Schneider et al., 1999). Both C. bellendenkerana and U. metrosideros have current populations within the Black Mountain Corridor area suggesting that dispersal occurred here after the last glacial maximum. For the remaining 16 species, populations occur either north or south of the Black Mountain Corridor and most are restricted to one to two populations. These 16 species may have had previously larger distribution ranges and the contraction of rainforest at the last glacial maximum may have reduced them to their current range.

With respect to dispersal to suitable habitat outside the bioregions, modelling predictions imply that dispersal within the Wet Tropics would not be sufficient for all but three out of the 19 mountaintop endemic species. The closest area south of the Wet Tropics with N1000 m elevation is Mt. Elliot just south of the southern Wet Tropics bioregion boundary. Mt. Elliot which has 367.8 ha of terrain above 1000 m is a potential refuge for the species modelled in our study due to its close proximity. Further work and climate data is needed to assess the likelihood that Mt. Elliot will maintain a suitable climate for these species. The next nearest areas with substantial upland terrain are the Border Ranges followed by the Nightcap National Park (Queensland– New South Wales border) approximately 1400 km to the south. North of the Wet Tropics bioregion there is no habitat N 800 m thus these species would have to disperse across the Torres Strait to the highlands of New Guinea roughly 900 km away. Current climate data for Australia outside the Wet Tropics bioregion is only available at much coarser scales that are not sufficient for accurate species distribution modelling. Thus dispersal likelihood outside of the Wet Tropics bioregion at this stage is largely open to speculation. Considering the existing biogeographic constraints, it appears that human-assisted translocation may be required. What about more localized dispersal at finer scales than current climate models can detect and make accurate predictions for? Recent studies have highlighted the importance of integrating habitat complexity into climate change assessments for particular species (Scheffers et al., 2013; Suggitt et al., 2011; Williams et al., 2008). Microhabitats have been shown to buffer the effects of climate change for certain animal species such as lizards and amphibians (Scheffers et al., 2014), which would otherwise be predicted to decline much faster based solely on macroclimate data. Studies to date on the resilience of microhabitats have been limited to vertebrates occupying a small niche within the canopy profile (e.g. forest litter). The species in our case, trees and shrubs, have long generation times requiring seedling recruitment in favourable climates or microclimates at sufficient rates to account for the changing climate in order to successfully migrate. These species also occupy a much larger proportion of the entire canopy profile and would require much larger microhabitats such as small gorges or gullies that retain moisture. 4.5. Conclusions and future actions This study presents a first approximation of how climate change will affect the distribution of plant species unique to Australia's tropical mountaintop environments. Although it is now well recognized that accurate predictions of how species will respond geographically to climate change are not possible, the bioclimatic envelope approach, as utilized here, is considered the best currently available guide for policy and conservation planning (Hannah et al., 2002). Our results are also consistent with earlier projections on the decline of endemic vertebrates of the region (Williams et al., 2003), which have been recently verified with observational data (Williams and Scheffers, 2013). We suggest that priority funding be directed to additional research on tropical mountaintop species and other species considered to be at most risk from climate change and that policy plans be drafted pending the outcome of such research. Specifically, we strongly advocate the following: 1.) A research programme that establishes whether or not the distribution of tropical mountaintop species is currently determined primarily by climate or by other constraints. 2.) A preliminary draft ex situ conservation plan for each of the 19 species assessed here utilizing existing public botanical garden facilities at lower, temperate latitudes that can be executed pending the outcome of target (1). 3.) Additional research that investigates the acclimation potential of the mountaintop species from our study and/or from other mountaintop habitats around the globe.

C.M. Costion et al. / Biological Conservation 191 (2015) 322–330

4.) Research that investigates the importance of biotic interactions for species survival as climate change prompts species to migrate. 5.) Research that investigates the dispersal capacity and likelihood for species predicted to be at most risk from climate change. As the world continues to experience rapid changes in climate, public botanical gardens and seed banks may play a crucial role in the conservation of plants threatened by climate change (Primack and Miller-Rushing, 2009; Seaton et al., 2010). Public outreach and developing community interest in the effects of climate change on species is a key issue, especially for rare taxa that only occur on isolated mountaintops difficult for the public to access. Studies and educational programmes that emphasize ecological services of environments particularly threatened by climate change to everyday people may help bring a sense of value to species in remote environments, as would research on the ethnobiology of such taxa. The local economy of the region in this case study is largely based on tourism. Exploring potential eco-tourism prospects for tropical mountaintops may provide a platform for community engagement in this issue. Much insight could be gained through a comprehensive modelling analysis incorporating acclimation, interspecific interactions, dispersal limitation and evolutionary adaptation data (Chevin et al., 2010; Rowland et al., 2011). Unfortunately in our case and for the vast majority of the world's tropical species (Corlett, 2011), these types of data are simply unavailable. Environmental niche modelling methods using the current climate-envelope approach, as done in this study, can provide a quick and broad-scale assessment of multiple species at the same time. This approach can enable the identification of species most likely to be at risk in the future and identify important knowledge gaps most relevant to biodiversity conservation. Acknowledgements We would like to thank the National Environment Research Program of Australia for funding this research and supporting climate change and threatened species research in general. We also wish to thank Andrew Ford and Jeremy VanDerWal for providing helpful comments and recommendations for improving the manuscript. References Alley, R.B., 2000. Ice-core evidence of abrupt climate changes. Proc. Natl. Acad. Sci. U. S. A. 97, 1331–1334. Brooks, T.M., Mittermeier, R.A., da Fonseca, G.A.B., Gerlach, J., Hoffmann, M., Lamoreux, J.F., Mittermeier, C.G., Pilgrim, J.D., Rodrigues, A.S.L., 2006. Global biodiversity conservation priorities. Science 313, 58–61. Chen, I.C., Shiu, H.J., Benedick, S., Holloway, J.D., Cheye, V.K., Barlow, H.S., Hill, J.K., Thomas, C.D., 2009. Elevation increases in moth assemblages over 42 years on a tropical mountain. Proc. Natl. Acad. Sci. U. S. A. 106, 1479–1483. Chevin, L.M., Lande, R., Mace, G.M., 2010. Adaptation, plasticity, and extinction in a changing environment: towards a predictive theory. PLoS Biol. 8, 8. Colwell, R.K., Rangel, T.F., 2010. A stochastic, evolutionary model for range shifts and richness on tropical elevational gradients under Quaternary glacial cycles. Philos. Trans. R. Soc., B 365, 3695–3707. Colwell, R.K., Brehm, G., Cardelus, C.L., Gilman, A.C., Longino, J.T., 2008. Global warming, elevational range shifts, and lowland biotic attrition in the wet tropics. Science 322, 258–261. Corlett, R.T., 2011. Impacts of warming on tropical lowland rainforests. Trends Ecol. Evol. 26, 606–613. Costion, C.M., Edwards, W., Ford, A.J., Metcalfe, D.J., Cross, H.B., Harrington, M.G., Richardson, J.E., Hilbert, D.W., Lowe, A.J., Crayn, D.M., 2015. Using phylogenetic diversity to identify ancient rain forest refugia and diversification zones in a biodiversity hotspot. Diversity and Distributions 21, 279–289. Crayn, D.M., Costion, C., Harrington, M.G., 2015. The Sahul–Sunda floristic exchange: dated molecular phylogenies document Cenozoic intercontinental dispersal dynamics. J. Biogeogr. 42, 11–24. Cunningham, S., Read, J., 2003a. Comparison of temperate and tropical rainforest tree species: growth responses to temperature. J. Biogeogr. 30, 143–153. Cunningham, S.C., Read, J., 2003b. Do temperate rainforest trees have a greater ability to acclimate to changing temperatures than tropical rainforest trees? New Phytol. 157, 55–64. Davis, A.J., Jenkinson, L.S., Lawton, J.H., Shorrocks, B., Wood, S., 1998. Making mistakes when predicting shifts in species range in response to global warming. Nature 391, 783–786.

329

Elith, J., Graham, C.H., Anderson, R.P., Dudik, M., Ferrier, S., Guisan, A., Hijmans, R.J., Huettmann, F., Leathwick, J.R., Lehmann, A., Li, J., Lohmann, L.G., Loiselle, B.A., Manion, G., Moritz, C., Nakamura, M., Nakazawa, Y., Overton, J.M., Peterson, A.T., Phillips, S.J., Richardson, K., Scachetti-Pereira, R., Schapire, R.E., Soberon, J., Williams, S., Wisz, M.S., Zimmermann, N.E., 2006. Novel methods improve prediction of species' distributions from occurrence data. Ecography 29, 129–151. Etterson, J.R., Shaw, R.G., 2001. Constraint to adaptive evolution in response to global warming. Science 294, 151–154. Fleming, T., Kress, W.J., 2013. Ornaments of Life: Coevolution and Conservation in the Tropics. University of Chicago Press (588 pp.). Foster, P., 2001. The potential negative impacts of global climate change on tropical montane cloud forests. Earth Sci. Rev. 55, 73–106. Ghannoum, O., Phillips, N.G., Sears, M.A., Logan, B.A., Lewis, J.D., Conroy, J.P., Tissue, D.T., 2010. Photosynthetic responses of two eucalypts to industrial-age changes in atmospheric CO2 and temperature. Plant Cell Environ. 33, 1671–1681. Gordon, H., Farrell, S.O., Collier, M., Dix, M., Rotstayn, L., Kowalczyk, E., Hirst, T., Watterson, I., 2010. The CSIRO Mk3 . 5 Climate Model [electronic resource]. CAWCR Technical Report No. 021. Centre for Australian Weather and Climate Research, Melbourne, Australia. Guisan, A., Thuiller, W., 2005. Predicting species distribution: offering more than simple habitat models. Ecol. Lett. 8, 993–1009. Gunderson, C.A., O'Hara, K.H., Campion, C.M., Walker, A.V., Edwards, N.T., 2010. Thermal plasticity of photosynthesis: the role of acclimation in forest responses to a warming climate. Glob. Chang. Biol. 16, 2272–2286. Hannah, L., Midgley, G.F., Millar, D., 2002. Climate change-integrated conservation strategies. Glob. Ecol. Biogeogr. 11, 485–495. Hilbert, D.W., Graham, A., Hopkins, M.S., 2007. Glacial and interglacial refugia within a long-term rainforest refugium: the wet tropics bioregion of NE Queensland, Australia. Palaeogeogr. Palaeoclimatol. Palaeoecol. 251, 104–118. Huntley, B., Bartlein, P.J., Prentice, I.C., 1989. Climatic control of the distribution and abundance of beech (Fagus) in Europe and North America. J. Biogeogr. 16, 551–560. Hutchinson, M.F., Houlder, D.J., Nix, H.A., Mcmahon, J.P., 2000. ANUCLIM user's guide Version 5.1. Centre for Resource and Environmental Studies, Australian National University, Canberra, Australian Capital Territory, Australia. Jevanandam, N., Goh, A.G.R., Corlett, R.T., 2013. Climate warming and the potential extinction of fig wasps, the obligate pollinators of figs. Biol. Lett. 9, 4. Joseph, L., Moritz, C., Hugall, A., 1995. Molecular support for vicariance as a source of diversity in rain-forest. Proc. R. Soc. B Biol. Sci. 260, 177–182. Jouzel, J., Masson-Delmotte, V., Cattani, O., Dreyfus, G., Falourd, S., Hoffmann, G., Minster, B., Nouet, J., Barnola, J.M., Chappellaz, J., Fischer, H., Gallet, J.C., Johnsen, S., Leuenberger, M., Loulergue, L., Luethi, D., Oerter, H., Parrenin, F., Raisbeck, G., Raynaud, D., Schilt, A., Schwander, J., Selmo, E., Souchez, R., Spahni, R., Stauffer, B., Steffensen, J.P., Stenni, B., Stocker, T.F., Tison, J.L., Werner, M., Wolff, E.W., 2007. Orbital and millennial Antarctic climate variability over the past 800,000 years. Science 317, 793–796. Krishnaswamy, J., John, R., Joseph, S., 2014. Consistent response of vegetation dynamics to recent climate change in tropical mountain regions. Glob. Chang. Biol. 20, 203–215. La Sorte, F.A., Jetz, W., 2010. Projected range contractions of montane biodiversity under global warming. Proc. R. Soc. B Biol. Sci. 277, 3401–3410. Laurance, W.F., Useche, D.C., Shoo, L.P., Herzog, S.K., Kessler, M., Escobar, F., Brehm, G., Axmacher, J.C., Chen, I.C., Gamez, L.A., Hietz, P., Fiedler, K., Pyrcz, T., Wolf, J., Merkord, C.L., Cardelus, C., Marshall, A.R., Ah-Peng, C., Aplet, G.H., Arizmendi, M.D., Baker, W.J., Barone, J., Bruhl, C.A., Bussmann, R.W., Cicuzza, D., Eilu, G., Favila, M.E., Hemp, A., Hemp, C., Homeier, J., Hurtado, J., Jankowski, J., Kattan, G., Kluge, J., Kromer, T., Lees, D.C., Lehnert, M., Longino, J.T., Lovett, J., Martin, P.H., Patterson, B.D., Pearson, R.G., Peh, K.S.H., Richardson, B., Richardson, M., Samways, M.J., Senbeta, F., Smith, T.B., Utteridge, T.M.A., Watkins, J.E., Wilson, R., Williams, S.E., Thomas, C.D., 2011. Global warming, elevational ranges and the vulnerability of tropical biota. Biol. Conserv. 144, 548–557. McJannet, D., Wallace, J., Fitch, P., Disher, M., Reddell, P., 2007. Water balance of tropical rainforest canopies in north Queensland, Australia. Hydrol. Process. 21, 3473–3484. Metcalfe, D.J., Ford, A.J., 2008. Floristics and plant biodiversity of the rainforests of the Wet Tropics. In: Stork, N.E., Turton, S.M. (Eds.), Living in a dynamic tropical forest landscape. Blackwell Publishing. Parmesan, C., Yohe, G., 2003. A globally coherent fingerprint of climate change impacts across natural systems. Nature 421, 37–42. Pau, S., Wolkovich, E.M., Cook, B.I., Nytch, C.J., Regetz, J., Zimmerman, J.K., Wright, S.J., 2013. Clouds and temperature drive dynamic changes in tropical flower production. Nat. Clim. Chang. 3, 838–842. Pearson, R.G., Dawson, T.P., 2003. Predicting the impacts of climate change on the distribution of species: are bioclimate envelope models useful? Glob. Ecol. Biogeogr. 12, 361–371. Peterson, A.T., Sanchez-Cordero, V., Soberon, J., Bartley, J., Buddemeier, R.W., NavarroSiguenza, A.G., 2001. Effects of global climate change on geographic distributions of Mexican Cracidae. Ecol. Model. 144, 21–30. Phillips, S.J., Anderson, R.P., Schapire, R.E., 2006. Maximum entropy modeling of species geographic distributions. Ecol. Model. 190, 231–259. Pounds, J.A., Fogden, M.P.L., Campbell, J.H., 1999. Biological response to climate change on a tropical mountain. Nature 398, 611–615. Pounds, J.A., Bustamante, M.R., Coloma, L.A., Consuegra, J.A., Fogden, M.P.L., Foster, P.N., La Marca, E., Masters, K.L., Merino-Viteri, A., Puschendorf, R., Ron, S.R., Sanchez-Azofeifa, G.A., Still, C.J., Young, B.E., 2006. Widespread amphibian extinctions from epidemic disease driven by global warming. Nature 439, 161–167. Primack, R.B., Miller-Rushing, A.J., 2009. The role of botanical gardens in climate change research. New Phytol. 182, 303–313.

330

C.M. Costion et al. / Biological Conservation 191 (2015) 322–330

Raxworthy, C.J., Pearson, R.G., Rabibisoa, N., Rakotondrazafy, A.M., Ramanamanjato, J.B., Raselimanana, A.P., Wu, S., Nussbaum, R.A., Stone, D.A., 2008. Extinction vulnerability of tropical montane endemism from warming and upslope displacement: a preliminary appraisal for the highest massif in Madagascar. Glob. Chang. Biol. 14, 1703–1720. Rowland, E.L., Davison, J.E., Graumlich, L.J., 2011. Approaches to evaluating climate change impacts on species: a guide to initiating the adaptation planning process. Environ. Manag. 47, 322–337. Scheffers, B.R., Brunner, R.M., Ramirez, S.D., Shoo, L.P., Diesmos, A., Williams, S.E., 2013. Thermal buffering of microhabitats is a critical factor mediating warming vulnerability of frogs in the Philippine biodiversity hotspot. Biotropica 45, 628–635. Scheffers, B.R., Edwards, D.P., Diesmos, A., Williams, S.E., Evans, T.A., 2014. Microhabitats reduce animal's exposure to climate extremes. Glob. Chang. Biol. 20, 495–503. Schneider, C.J., Cunningham, M., Moritz, C., 1998. Comparative phylogeography and the history of endemic vertebrates in the Wet Tropics rainforests of Australia. Mol. Ecol. 7, 487–498. Schneider, C.J., Smith, T.B., Larison, B., Moritz, C., 1999. A test of alternative models of diversification in tropical rainforests: ecological gradients vs. rainforest refugia. Proc. Natl. Acad. Sci. U. S. A. 96, 13869–13873. Seaton, P.T., Hu, H., Perner, H., Pritchard, H.W., 2010. Ex situ conservation of orchids in a warming world. Bot. Rev. 76, 193–203. Sheldon, K.S., Yang, S., Tewksbury, J.J., 2011. Climate change and community disassembly: impacts of warming on tropical and temperate montane community structure. Ecol. Lett. 14, 1191–1200. Smith, M.D., Knapp, A.K., Collins, S.L., 2009. A framework for assessing ecosystem dynamics in response to chronic resource alterations induced by global change. Ecology 90, 3279–3289. Still, C.J., Foster, P.N., Schneider, S.H., 1999. Simulating the effects of climate change on tropical montane cloud forests. Nature 398, 608–610. Suggitt, A.J., Gillingham, P.K., Hill, J.K., Huntley, B., Kunin, W.E., Roy, D.B., Thomas, C.D., 2011. Habitat microclimates drive fine-scale variation in extreme temperatures. Oikos 120, 1–8. Sunday, J.M., Bates, A.E., Dulvy, N.K., 2012. Thermal tolerance and the global redistribution of animals. Nat. Clim. Chang. 2, 686–690.

Thomas, C.D., 2010. Climate, climate change and range boundaries. Divers. Distrib. 16, 488–495. Thompson, L.G., Mosley-Thompson, E., Brecher, H., Davis, M., Leon, B., Les, D., Lin, P.N., Mashiotta, T., Mountain, K., 2006. Abrupt tropical climate change: past and present. Proc. Natl. Acad. Sci. U. S. A. 103, 10536–10543. VanDerWal, J., Shoo, L.P., Johnson, C.N., Williams, S.E., 2009. Abundance and the environmental niche: Environmental suitability estimated from niche models predicts the upper limit of local abundance. American Naturalist 174, 282–291. VanDerWal, J., Beaumont, L., Zimmermann, N.E., Lorch, P., 2011. Climates v.0.1-1.4: Methods for working with weather and climate. Software package [online resource] www.rforge.net/climates/. Williams, S.E., Pearson, R.G., 1997. Historical rainforest contractions, localized extinctions and patterns of vertebrate endemism in the rainforests of Australia's wet tropics. Proc. R. Soc. B Biol. Sci. 264, 709–716. Williams, S.E., Bolitho, E.E., Fox, S., 2003. Climate change in Australian tropical rainforests: an impending environmental catastrophe. Proc. R. Soc. B Biol. Sci. 270, 1887–1892. Williams, J.W., Jackson, S.T., Kutzbacht, J.E., 2007. Projected distributions of novel and disappearing climates by 2100 AD. Proc. Natl. Acad. Sci. U. S. A. 104, 5738–5742. Williams, S.E., Shoo, L.P., Isaac, J.L., Hoffmann, A.A., Langham, G., 2008. Towards an integrated framework for assessing the vulnerability of species to climate change. PLoS Biol. 6, 2621–2626. Williams, K.J., Dunlop, M., Bustamante, R.H., Murphy, H.T., Ferrier, S., Wise, R.M., Liedlo, A., Skewes, T.D., Harwood, T.D., Kroon, F., Williams, R.J., Joehnk, K., Crimp, S., Smith, M.S., James, C., Booth, T., 2012. Queensland’s biodiversity under climate change: impacts and adaptation synthesis report. A Report Prepared for the Queensland Government, CSIRO Climate Adaptation Flagship, Canberra, Brisbane. Williams, S.E., Scheffers, B.R., Scheffers, 2013. As climate changes, animals move fast to escape the heat. The Conversation https://theconversation.com/as-climate-changes-animals-move-fast-to-escape-the-heat-18511. Winter, J.W., 1997. Responses of non-volant mammals to late Quaternary climatic changes in the wet tropics region of north-eastern Australia. Wildl. Res. 24, 493–511. Woodward, F.I., 1990. The impact of low-temperatures in controlling the geographicaldistribution of plants. Philos. Trans. R. Soc. Lond. B Biol. Sci. 326, 585–593.