Apex predators include some of the world’s most imperiled species and climate change is predicted to have devastating impacts on some members of this important trophic group. Perhaps the best publicized example of climate impacts on an apex predator is the polar bear (Ursus maritimus), which undergoes population decline in years with low levels of arctic ice coverage due to reduced reproduction and adult female survivorship (Hunter et al., 2010; Laidre et al., 2020). Projections of continued reductions in arctic ice area indicate that the polar bear will experience severe population declines by 2100 (Hunter et al., 2010; Molnár et al., 2011). Climate change is also predicted to impact tropical apex predators. For example, in the mangrove Sundarbans of southern Bangladesh, sea-level rise is predicted to eliminate all suitable habitat for the Bengal tiger (Panthera tigris tigris) by 2070 (Mukul et al., 2019). Climate change impacts are also likely to extend to reptile apex predators, including the world’s largest lizard, the Komodo dragon (Varanus komodoensis). The Komodo dragon occurs on five islands in Indonesia and climate change under moderate emissions scenarios is predicted to reduce suitable habitat by ∼90% by 2050 (Jones et al., 2020). From these examples, it is apparent that climate change will impact apex predators via several pathways and across multiple biomes.

Crocodilians, with 25 extant species currently recognized, function as a crucial apex predator in semiaquatic ecosystems, and are also anticipated to be negatively impacted by climate change. For example, 37-year observations in Florida showed that Crocodylus acutus hatching shifted to earlier dates by 1.5 days, every 2 years with increased sea surface temperature (Cherkiss et al., 2020). Similar results show that increased temperature led to shorter incubation periods for the same species in Mexico (Charruau et al., 2017). Moreover, the sex ratio is determined, in crocodilian species, by incubation temperature, and climate warming is expected to interfere with this (Maciejewski, 2006; González et al., 2019; Bock et al., 2020). For Alligator mississippiensis in Florida, it is estimated that a temperature rise by 1.1–1.4°C in 2040–2050 may skew the sex ratio initially to 95.6% males and then to 97.8% females with a temperature rise by 1.6–3.2°C in 2090–2100 (Bock et al., 2020).

Global warming also can be a threat to crocodilians by destroying their habitat, in particular freshwater swamps or floodplains, through saltwater inundation (SWI) as a result of the sea level rise (SLR). Most species require freshwater habitat for breeding and nesting and such areas typically lie at a low elevation along coasts or rivers and are, therefore, vulnerable to imminent SWI (Pezeshki et al., 1990; Mulrennan and Woodroffe, 1998; Pettit et al., 2018). Despite the adaptation to the saline environment, as implied by their common name, saltwater crocodiles, C. porosus is one such species and requires constant or regular access to freshwater for breeding.

Here, we examine the impacts of climate change on C. porosus, the largest extant crocodilian species, and quantify the potential loss of nesting habitat through SWI and SLR in the Northern Territory (NT), Australia. This region is of global significance, as it supports the most extensive freshwater wetlands and floodplains and contains the largest population of this apex predator in the world (Webb et al., 2010; Fukuda et al., 2021).

Our model of current saltwater crocodile nesting habitat performed well (test AUC of 0.958 ± SD 0.003) and showed suitable areas around the coasts and floodplains of the Top End of the NT (Figure 1). The most important variable was distance to perennial lakes (71.0% contribution, 87.5% permutation importance), with a negative logistic relationship with nesting habitat suitability (Figure 2A). Early wet season rainfall was the next most important variable (12.9% contribution; 0.9% permutation importance), with a positive logistic relationship between Early wet season rainfall and nesting habitat suitability (Figure 2B). The freshwater swamp variable (8.3% contribution; 1.2% permutation importance) showed much higher suitability than other landform types (Figure 2C). Seasonality in rainfall (BIO15) showed 6.6% contribution; 3.0% permutation importance, with a negative logistic relationship to nesting habitat suitability (Figure 2D). All other variables contributed < 5% to the model.

For the whole of the Top End, our thresholded Maxent model predicted a total of 32,306.91 km² of suitable nesting habitat across the study area for the current time. For the period 2081–2100, using the predicted future temperature and rainfall patterns derived from the BCC-CSM2-MR climate model but retaining the other variables at their present values, our Maxent model predicted an area of suitable habitat of 32,199.32 km² under the low emissions scenario (SSP 126) and 42,740.78 km² under the high emissions scenario (SSP 585) across the study area.

Within the Kakadu Region, a total area of 4232.08 km² was assigned as suitable nesting habitat by our current model (Figure 3A). However, the hydrodynamic models by Bayliss et al. (2016) show that as at 2013 sea level, approximately 16.40% of the suitable habitat was already affected by < 0.25 m SWI, and 3538.10 km² remains suitable for nesting (Figure 3B). After 1.1 m SLR (that is, sea levels forecast for 2100), only 1775.70 km² remains suitable (Figure 3C). This represents a 58.04% reduction in predicted suitable nesting habitat over 87 years, between 2013 and 2100).

Our model of suitable nesting habitat for the saltwater crocodile revealed a dominant importance of abundant, perennial lakes in Australia’s Top End. The estimate of nesting suitability rapidly dropping beyond a 50 m range from perennial lakes (Figure 2A) is consistent with field observations that a vast majority of nests are made around the edges of waterbodies during the breeding season (Webb et al., 1977, 1983; Fukuda and Cuff, 2013). The perennial lakes appear to be important as females guarding their nests during the incubation period, typically 2–3 months, require access to freshwater (Webb et al., 1977; Magnusson, 1980). The model also identified the importance of early rains in the wet season (Figure 2B), supporting previous findings that rains in early wet season (Oct–Dec) trigger the highly seasonal breeding of C. porosus, and higher precipitation during this period results in higher abundance of hatchlings in the following dry season (Fukuda and Saalfeld, 2014).

We found that projected changes in temperature and rainfall under climate change could impact the area of suitable nesting habitat negatively or positively (0.33% decrease under a low emission scenario SSP 126 and 32.30% increase under a high emission scenario SSP 585). It is worth mentioning that, apart from the adverse impact of SLR, the projected nesting habitat would otherwise be increased to some extent under the higher future emission scenario, because of some positive effects on the climate variables such as increased early rains and decreased seasonality. Nevertheless, these effects will be canceled by the much higher, negative impact by SLR.

Most importantly, our analysis for the Kakadu Region showed that almost 50% of the suitable nesting habitat in 2013 would be lost to the 1.1 m SLR by 2100 (Figure 3). While fine-scale SLR data and forecasts are not currently available outside of the Kakadu region, if a similar proportion of the habitat is affected across the larger study area, 16,522 km² would be lost to SLR across the Top End. This is of significant concern as the majority of the Australian population of C. porosus resides in this area (Fukuda et al., 2007, 2021). This suggests that SLR by climate change may represent the major threat to the species. Other crocodilians that build mound nests in freshwater wetlands or floodplains, such as A. mississippiensis and C. moreletti, could also be at similar risk, as their nests are prone to flooding (Kushlan and Jacobsen, 1990; Platt et al., 2021).

One important caveat to the predicted loss of saltwater crocodile nesting habitat is that the creation of new freshwater habitats in the Kakadu region has not been accounted for. Bayliss et al. (2016) were unable to predict new areas of freshwater habitat created through SLR because the Lidar-derived fine-scale elevation data did not extend beyond the current floodplain. Undoubtedly, SLR will result in the formation of some areas of new habitat suitable for saltwater crocodile nesting which may somewhat offset the areas lost. However, the sandstone escarpments that occur further inland (Northern Territory and Geological Survey, 2006) form a physical barrier to the formation of extensive freshwater floodplains. In addition, while small billabongs or lakes may form on the escarpment, the steep and rocky terrain curtails the dispersal of saltwater crocodiles, and to date, they have never been observed in these areas (Letnic and Connors, 2006). For areas outside of the sandstone escarpments, the potential creation of new floodplain habitat is an important avenue for further research. Additional data such as LiDAR (Dong and Chen, 2017) would facilitate more accurate forecasts of potential future suitable habitat.

Another source of uncertainty is the degree of future SLR. Bayliss et al. (2016) adopted a value of 1.1 m in 2001 based on a range of published projections (Solomon et al., 2007; Jevrejeva et al., 2010). The IPCC recently released their 6th Assessment Report which includes predictions of 0.63–1.01 m by 2100 under the moderate emission scenario and 0.66–1.33 m by 2150 under the very high emissions scenario (Masson-Delmotte et al., 2021). Therefore, the 1.1 m SLR used in our study represents likely SLR between 2100 and 2150. However, several studies document the acceleration of SLR in recent decades (Nerem et al., 2018; Dangendorf et al., 2019), and in addition, the IPCC stated that due to the substantial uncertainty in global ice sheet dynamics, SLR approaching 2 m by 2100 and 5 m by 2150 cannot be ruled out (Masson-Delmotte et al., 2021). The substantial uncertainty in the potential creation of new habitat and the degree of SLR means that estimates of the loss of saltwater crocodile nesting habitat should be recalculated as new data become available.

It should be noted that the change from a freshwater to more saline habitat is not anticipated to be monotonic, and gradual replacement of freshwater plant species with those that are more saline tolerant is the likely scenario. While the floodplain vegetation is determined by fine-scale variation in topography (Finlayson et al., 2006) and the global mean SLR is 4.8 mm annually (Dangendorf et al., 2019), even infrequent SWI can lead to sudden changes in salinity (Finlayson et al., 2006; Bayliss et al., 2016). Freshwater plant taxa have a unique tolerance to salinity and water depth (Cowie, 2003), and most freshwater plant species are unable to tolerate salinity much in excess of 1 ppt (Pettit et al., 2018). Previous work has shown an association between crocodile nesting and Oryza dominated tall tussock and Melaleuca open grassland with an understory comprised of species also present in the Oryza dominated grasslands (Fukuda and Cuff, 2013), and Oryza species are unable to tolerate salinity above 1 ppt (Pettit et al., 2018). In contrast, mangrove species can tolerate salinity to 50 ppt (Ball, 1998), and are not obligate halophiles, that is, they are also able to live in freshwater conditions (Woodroffe, 1988).

Although fossil records and molecular analysis indicate that extant crocodilians were capable of tracking changes in their distribution in response to drastic climate changes in the last 100 million years (Roos et al., 2007; Brochu et al., 2009; Ryberg and Lawing, 2018), whether they can continue to track changes in the spatial distribution and quality of suitable nesting habitat is unknown. Crocodylus porosus is one of the oldest extant species in the world, little changed since dinosaurs roamed the planet, yet their survival during the Anthropocene may now depend on how quickly humanity can transition away from the burning of fossil fuels and thus stabilize global warming.

The original contributions presented in the study are publicly available. This data can be found here: Fukuda and McDonald (2022).

Ethical review and approval was not required for the animal study because this study did not involve live animals and only used the location data of crocodile nests that were reported to the Northern Territory Government, Australia.

YF, PM, and BC contributed to conception and design of the study, and wrote the manuscript. YF organized the data access. YF and PM performed the spatial analysis. All authors contributed to manuscript revision, read, and approved the submitted version.

BC was affiliated with Complex Systems Consulting. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.