Introduction

Climate change threatens many species globally through the expansion of unfavourable conditions and the loss of both terrestrial and marine habitats (Chin et al. 2010; Dirnböck et al. 2011; Manes et al. 2021). Therefore, there is an urgent need to identify and triage emerging climate-induced threats to wildlife so that viable conservation management solutions can be implemented. High on the list of climate threats is rising temperatures, which threaten entire ecosystems that cannot adapt quickly enough to changing environmental conditions, such as tropical coral reefs and temperate kelp forests (Cornwall et al. 2021; Ling and Keane 2024). However, in addition to temperature, other climate change threats exist. For example, coastal ecosystems are threatened by rapidly rising sea levels, which may lead to a loss of critical habitat for both terrestrial and aquatic species (Galbraith et al. 2002; Fish et al. 2005; Baker et al. 2006; Hunter et al. 2015; Kaplanis et al. 2020).

Sea turtles are one group where climate change poses threats across all life stages (Fuentes et al. 2023a). Rising temperatures are well known to threaten sea turtles by skewing hatchling sex ratios, since all sea turtle species exhibit temperature-dependent sex determination (e.g., Hays et al. 2025). Warming incubation temperatures also lead to increased embryo mortality in nests (Hays et al. 2017), with both these threats of increasing hatchling sex ratio skews and mortality unlikely to be mitigated by phenological shifts in nesting seasons (Fuentes et al. 2023b; Laloë and Hays 2023).

However, the impacts of sea level rise (SLR) on turtle nesting habitats and hatching success have not been studied to the same extent as the feminisation of hatchling sex ratios (but see for example Fish et al. 2005; Fuentes et al. 2010; Butt et al. 2016; Veelenturf et al. 2020; Rivas et al. 2023). SLR threatens sea turtles when it leads to a loss of suitable beach (Fuentes et al. 2023a) and then there is simply not enough nesting habitat for turtles to nest in successfully (Simantiris 2024). Furthermore, before this situation is reached, SLR might increase embryo mortality by increasing seawater over-wash of nests (Patino-Martinez et al. 2014; Limpus et al. 2021; Ware et al. 2021; Martins et al. 2022), erosion of nests during storm surges (Milton et al. 1994; Van Houtan and Bass 2007) and inundation of nests by seawater from below the nest during Spring tides and storm surges (Foley et al. 2006; Shaw 2013; Ware and Fuentes 2018; Booth et al. 2022). Given the significance of SLR as a threat to sea turtles, here we compile the available evidence for the magnitude of this threat around the world. We review the existing literature on the historical and projected impacts of SLR on sea turtle nesting beaches around the world. By collating this knowledge, we then identify the site types that are at the greatest risk of inundation from SLR. We outline the caveats to SLR projections on nesting beaches and review existing evidence for how turtles might mitigate SLR threats by relocating nesting sites. In reviewing the available information, we aim to identify major knowledge gaps in this field and highlight how addressing these knowledge gaps could lead to greater protection, management action and policy changes to protect globally important nesting habitats.

Methods

We conducted a literature review of peer-reviewed papers, theses and technical reports that examined the threat of SLR to nesting sea turtles. We searched for the terms “sea level rise”, “sea-level rise”, “climate change” and “sea turtle” on Web of Science™ (webofscience.com/wos/) and Google Scholar (scholar.google.com/) to identify relevant papers. We then ran forward, and backward citations searches from those papers to locate additional papers in this area and included measured statistics on SLR effects for nesting turtles. From the assembled papers we extracted information on; species, study site, country, methodology (i.e., historical or forecasted modelling), the modelled SLR scenarios and time period used (e.g., 0.3 m, 0.5 m and 0.9 m between 2010 and 2100), if beach recession was accounted for or estimated, the percentage of sea turtle habitat area lost and/or the percentage of nests flooded as a result of SLR and the overall trend (declining, stable or increasing) referred to in their summary of their findings. Nest flooding is typically defined by the complete and or partial inundation of a turtle nest by a rising watertable, tidal inundation and or storm surges (Limpus et al. 2021). Nesting (study) site was defined as the area identified by the authors of the research article where turtle nesting was occurring. For example, the nesting site identified may comprise of a single nesting beach or multiple islands in an atoll.

We grouped the scenarios used by different authors into four ranges for extent of SLR (change in water level); (C) conservative SLR ≤ 0.3 m; (M) moderate SLR = 0.31–0.59 m; (H) high = 0.6–0.89 m; (E) extreme ≥ 0.9 m. As modelled timescales vary between studies due to the model used and year of data collection, we then established the time period that each scenario was modelled for. For example, a study that recorded beach parameters in 2015 and forecasted a 0.5 m SLR by the year 2100 would result in an 85-year time period. This time period was then used to estimate the annual rate of change for the percentage of habitat area lost and or the percentage of nests flooded as the total projected habitat loss/number of years in the time period. In the instance where no end date was given, we referred to the IPCC report or SLR publication that was referenced in the text and extracted the date from the subsequent report. Where an end date range was used (e.g., 2046–2065), we used the mean year as a reference point to calculate the time period (e.g., 2056).

For each study site, we established the site type, where four main site types were identified: low-lying/coral island, barrier island, mainland, and elevated island which included islands formed from volcanic activity and continental islands formed via sedimentary processes. Where multiple species were documented for one site, the species were grouped together for analyses unless there were differences in modelled results. All extracted data were assessed for normality and variance homogeneity using the Shapiro-Wilk Test and the Levene’s test respectively. Due to the small sample sizes across site types, a comparative analysis between site types for projected change in habitat area loss and nest flooding was conducted using a Kruskal-Wallis test, followed by a Post-hoc Dunn’s test and a pairwise comparison using a linear mixed-effects model (LMM) in RStudio 2024.12.0.

A secondary, non-exhaustive literature search was conducted to collate information on global nesting sites and site type for all seven species of sea turtles. The objective of this secondary search was to identify key nesting sites and categorise them by their site type. To do this we searched for the terms “green turtle”, “hawksbill turtle”, “loggerhead turtle”, “flatback turtle”, “leatherback turtle”, “olive ridley”, “Kemp’s ridley”, “annual” and “nesting” on Web of Science™ and Google Scholar. Additional resources such as the State of the World’s Sea Turtles (SWOT) reports were used to identify location names to be used as additional search terms (for example SWOT 2020; 2021; 2022; 2024), however all of the data collated in the present study was extracted from the primary research article (see Supplementary Information Table S4). Sites were selected based on population size (annual clutches or nesting females) and the thresholds used were similar to those employed in the SWOT reports; green (Chelonia mydas) and loggerhead turtle (Caretta caretta) sites ≥ 1,000 clutches annually; hawksbill turtle (Eretmochelys imbricata) sites ≥ 100 clutches annually; olive ridley turtle (Lepidochelys olivacea) sites ≥ 5,000 clutches annually; leatherback turtle (Dermochelys coriacea) sites ≥ 500 clutches annually; flatback turtles (Natator depressus) sites ≥ 100 females annually. For each site, the site type was identified using one of the four categories previously outlined. GPS coordinates of each site were then digitized using QGIS Geographic Information System (2025, Open Source Geospatial Foundation Project http://qgis.org), and symbology for site type was used to create a map for each species. After conducting the meta-analysis of at-risk site types and combining it with these data on large nesting populations, a better assessment of each species’ relevant risk to inundation from SLR could be conducted.

Results

Impacts of sea level rise on sea turtles

SLR was reported as having a variety of documented negative impacts for sea turtles. For example, an increase wave wash-over events of nests, sub-surface inundation from the watertable and or erosion of nests during storm surges (Fig. 1a-c). These impacts have direct mortality effects, whilst barriers that are man-made (e.g., seawall or beach armouring) or natural (e.g., cliff) can have indirect effects on mortality by limiting natural beach recession, reducing available area in the nesting habitat which results in density-dependent effects and reduced hatchling production (Fig. 1d). Several studies emphasised that low-lying nesting islands have an increased risk to SLR (Baker et al. 2006; Fuentes et al. 2010; Rivas et al. 2023). Examples of low-lying coral islands include Raine Island in the northern Great Barrier Reef, Australia and Aves Island, Venezuela (Fig. 1e-f).

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

Documented examples of the impacts of sea level rise (SLR) on sea turtle nests, nesting females and the types of beaches most at risk. Embryonic mortality can increase due to (a) wave over wash of nests during spring tides and storm surges; (b) increased erosion of the nesting habitat as a result of SLR coupled with increased storm intensity. The formation of erosion banks and cliffs also create natural barriers and reduce access to turtles. (c) Coastal urbanisation and beach “armouring”, prevents natural shoreline retreat with SLR (i.e., “coastal squeeze”) and restricts beach access for nesting turtles. (d) Low-lying islands, such as Aves Island, Venezuela, are high-risk sites because they lack available elevation for shoreline retreat. Image credits are as follows: (a) Matthew Ware; (b) David Booth; (c) Wilma Katz, Coastal Wildlife Club; (d) Rodrigo Lazo

Trends in sea level rise impacts on global nesting habitats

The literature search was conducted between December 2024 – March 2025 and updated in August 2025. The search terms used in Web of Science™ (WOS) returned 131 unique articles, of which 23 were extracted and included in this study. The search terms used in Google Scholar (GS) returned 130,000 + studies, and the first 10 pages (90 articles) were processed. Of these 90, there were 22 articles included in the study, seven were not previously identified in the WOS search. Three additional articles were sourced only through secondary survey of the articles found in WOS or GS. Studies were included if they contained data on historical habitat loss and or projected habitat loss or nest flooding. Studies that included SLR as a parameter within a vulnerability index or risk assessment were excluded, as the effect of SLR alone could not be distinguished. Studies that had determined historical or projected area loss, as opposed to nesting habitat of sea turtles, were largely excluded, with the exception of two studies in the Red Sea and Arabian Gulf, as they were conducted on small sand islands. From the literature search, we identified 34 unique research articles that had investigated the historical (n = 9) and future (n = 24) impacts of SLR on sea turtles and included 66 assessments across six species and 40 global nesting sites. One study by Reece et al. (2013) conducted both historical and forecasted modelling of SLR impacts to loggerhead turtles nesting on Melbourne Beach, Florida. Of the 25 studies that had forecasted SLR impacts, 12 (48%) had used Digital Elevation Model (DEMs) to project future sea levels on a topographic model of the beach. The number of assessments for each species on the historical and projected loss of nesting habitat and inundation of nests were: green turtle (n = 22 assessments); loggerhead turtle (n = 20); hawksbill turtle (n = 11); leatherback turtle (n = 7); olive ridley turtle (n = 5); and the Kemp’s ridley turtle (Lepidochelys kempii) (n = 1). Of 66 species assessments, 15 were extracted from Rivas et al. (2023), as these authors had conducted analyses across six individual sites in the Americas and one Pacific nesting site, most of which had more than one species nesting. Most studies found in the literature search were focused on a singular beach or island in their analyses, with very few comparative studies demonstrating differences between sites. Many of the research articles also did not provide a clear explanation of how “suitable nesting habitat” was defined and typically assumed that the area between the mean high tide line and the dense vegetation zone or a barrier (e.g., infrastructure) as the suitable habitat for nesting. The majority of these studies were able to support their SLR analyses with data on nesting density and activity.

Most assessments of the impacts of SLR have been conducted in Central, South and North America (62.1%) (Fig. 2). The majority of assessments made showed declining trends (78.8%) in nesting habitat and nest survival from inundation was historically or projected (Fig. 2). Most assessments (84.8%) were of the projected changes in either habitat area or nest flooding under various SLR scenarios. In contrast, historical changes to the available nesting habitat due to SLR in recent decades were shown for 10 assessments (15.1%). For example, on Aldabra, a green turtle nesting site located on a coral island in the Seychelles, Western Indian Ocean, there was an average shoreline recession of 0.25–0.36 m yr−1 for 24% of the island’s nesting beaches recorded over a 51-year period (Constance et al. 2024). Rates of shoreline recession are much greater for high energy beaches, such as the loggerhead turtle nesting beaches on St. Catherines Island (Georgia, U.S.A.), which is a barrier island that has a reported average of 3 m of beach recession each year (Meyer et al. 2014). As a direct result, nests closest to the high-water mark are periodically inundated during spring tides and storms, with hatching success dropping as low as 23% for loggerhead turtle nests for one year on the island (Meyer et al. 2014). Significant erosional processes accelerated habitat loss in this study and highlights a need to decipher between coastal erosion and the effect of SLR on nesting beaches.

Fig. 2
Fig. 2The alternative text for this image may have been generated using AI.
Full size image

Global map of sea turtle nesting sites with trends of available nesting habitat based on global sea level rise (SLR). The status for each nesting site’s habitat was based on the author’s conclusion of the findings using either the proportion of nesting habitat lost or proportion of nests flooded due to SLR in either hindcasted or projected models. Coloured circles represent the species, with inset circles representing a single site with two or more species. Numbers indicate site number which is referenced in Supplementary Information Tables S1-S3. The * denotes a study that analysed historical beach habitat loss, all others used a forecast modelling approach to assess SLR risk

Habitats most at risk of SLR inundation

Forecasting SLR scenarios to model habitat loss and nest inundation was the primary method of investigation amongst the articles when we reviewed the literature. Typically, studies included more than one SLR scenario, and the most used SLR simulation was ~ 0.5 m, which was considered an intermediate increase equivalent to the global mean sea level (GMSL) projection under the Shared Socio-economic Pathway (SSP) SSP2-4.5 in the IPCC climate report (IPCC 2021).

Of the 48 estimates (all scenarios, species and sites) produced from 26 assessments (sites and species) of projected habitat loss, three estimates were excluded from these analyses due to outlier effects. These included the two conservative scenario estimates from Bioko Island for green and leatherback turtle nesting areas (1.73% and 1.28% per year respectively), and the moderate scenario estimate for the east coast of U.S.A. (2.21% per year) and included nesting area for three species; loggerhead, green and leatherback. These three assessments all had a mid-century end point, meaning the rates were calculated over a much shorter period (i.e., 40 and 36 years respectively). With the 45 remaining estimates, our analyses found that appreciable nesting habitat loss was predicted for all site types, regardless of the SLR scenario (Fig. 3a). The mean annual rate of change for habitat loss was greatest in the island sites compared to the mainland sites (Fig. 3a). Mean annual projected habitat loss was greatest for coral islands under extreme SLR (n = 2 estimates), and elevated islands under high SLR scenarios (n = 5 estimates) (0.73% and 0.76% per year respectively) (Fig. 3a). From the year 2025 to 2100, this would equate to 54.8% and 57% projected habitat loss for coral islands and elevated islands respectively. Whilst mainland sites, even under extreme SLR scenarios (n = 3), had a projected annual increase in habitat loss of 0.25%, which equated to 18.8% habitat loss on average by the year 2100 (Fig. 3a).

Fig. 3
Fig. 3The alternative text for this image may have been generated using AI.
Full size image

The number of studies with projected nesting habitat loss and percentage of nests that would be flooded under conservative (< 0.29 m, green), moderate (0.30–0.59 m, teal), high (0.6–0.89 m, purple) and extreme (> 0.9 m, pink) sea level rise (SLR) scenarios. Projections have been converted to rates of change across the reported time period. (a) Projected rate of nesting habitat loss (% per year) as a proportion of nesting area that will be inundated over time; and the (b) projected rate of nests likely to experience flooding (% per year) on the nesting beach under the four SLR categories. Each study was categorised into a site type, whereby elevated islands include volcanic, continental and sedimentary islands. A singular line represents the value from one study’s assessment, the X represents the mean value of a sample and the line inside of the box represents the median. The lower and upper hinge positions correspond to the first and third quartiles (i.e., 25th and 75th percentiles)

The use of proportional data and the presence of small sample size for some site types necessitated the use of a non-parametric test. An initial Kruskal-Wallace test indicated that rates of habitat loss on coral island and elevated island sites were significantly affected by SLR scenario (P = 0.0212, P = 0.0186 respectively), and when scenario was removed as a term from the test, site type had a weak effect (P = 0.0731). Dunn’s post hoc multiple comparisons determined that the coral islands were not significantly different from other site types once the p-values were adjusted using the Bonferroni method. The difference in projected rates of habitat loss from the mainland and elevated islands were close to significant (Z = 2.53, P = 0.068). Comparisons were also done between linear models to confirm if scenario had a significant effect on the rate of habitat loss. Three models were tested, (1) a simple linear model with only site type used as the predictor, (2) a linear mixed effects model with scenario added as a random effect, and (3) a final linear model with scenario added as a fixed effect. The AIC values were as follows, 18.47, 27.67, 9.51. Hence, the model with scenario as a fixed term was the most effective (F6,41 = 4.047, R2 = 0.28, P = 0.0028), and determined that mainland sites had significantly less projected habitat loss than the other island types (P = 0.0187) (Supplementary Information Fig. S1a). Both high and extreme SLR scenarios significantly increased the projected rates of habitat loss (P = 0.0006, P = 0.0096 respectively).

The percentage of nests flooded was predicted to increase with SLR across at all site types (Fig. 3b). Barrier island sites were excluded from these analyses as there was only a single study (Rivas et al. 2023). Rivas et al. (2023) conducted their study on St George Island, Florida U.S.A., a loggerhead nesting site, where a 0.43% increase in nest flooding per year under a moderate SLR scenario was calculated, which equated to 38.4% of nests flooded by the year 2100. The greatest rate of increase in nest flooding was predicted for coral island sites, where annual nest flooding was calculated to increase by 0.77%, 1.02% and 1.04% per year under the moderate (n = 5), high (n = 2) and extreme (n = 2) SLR scenarios respectively (Fig. 3b). Between 2025 and 2100, this would equate to a total of 57.6%, 76.5% and 78% of nests flooded on average on coral islands under the moderate, high and extreme SLR scenarios respectively. Mainland sites had the lowest rates of projected nest flooding. Nest flooding was calculated to increase by 0.37% and 0.48% per year under moderate (n = 8) and extreme (n = 5) SLR scenarios respectively (Fig. 3b). This would equate to an average of 27.9% and 35.85% of nests flooded by 2100 under moderate and extreme SLR scenarios respectively.

Although the initial Kruskal-Wallace test indicated no significant effect of scenario on the rates of nest flooding between site types, further assessment with scenario removed as a vector from the test showed that site type was significant (P = 0.002), and the Dunn’s post hoc test indicated significant differences between coral island and mainland sites (Z = 3.5, P = 0.0005) and near significant differences between coral island and elevated islands (Z = 2.21, P = 0.08). A comparative analysis between three linear models was conducted; (1) a linear model with site type used as the predictor, (2) a linear mixed effects model with scenario as a random effect, and (3) a linear model with scenario added as a fixed effect. The comparison showed the most effective model to be where only site type was the explanatory variable in the model, followed closely by model 3, where scenario was a fixed effect (AIC = 12.18, 23.34, 12.61 respectively). The linear model for projected rate of nest flooding was significant (F2,32 = 11.2, R2 = 0.38, P = 0.0002), and a pairwise comparison showed significant differences between the coral island and elevated island sites (t ratio = 3.13, P = 0.011) and the coral island and mainland sites (t ratio = 4.66, P = 0.0002) (Supplementary Information Fig. S1b).

The greatest data coverage for projected nest flooding across the site types were mainland sites (47.1% of assessments), which also produced a great deal of variation across the scenario categories. Some sites such as Piro and Pejeperro, mainland beaches on the Osa Peninsula in Costa Rica, have both green and olive ridley turtles nesting but both showed variable responses to the same SLR projection due to differences in nest depth of the two species (Beber et al. 2024). At this site, it was calculated that there would be an increase in nest flooding by 0.49% and 0.53% per year for green turtles, and lower rate calculated for olive ridley turtles, with an annual increase of 0.13% and 0.15% per year in nest flooding under the high and extreme scenarios respectively (Beber et al. 2024).

Populations most at risk of SLR inundation

Due to the differing effects that SLR has on different site types (see analysis above), we investigated the relative proportion of large nesting populations that breed on each of these site types. Site names, nesting densities and references have been included in the Supplementary Information Table S4. This assessment included 11 nesting sites across seven sea turtle species, the number of sites per species were as follows: green = 41, leatherback = 23, hawksbill = 19, loggerhead = 12, olive ridley = 15, flatback = 7, and Kemp’s ridley = 1 (Fig. 4). Of the 117 major nesting sites identified in this study, 18 (15.4%) have had at least one research study conducted on either the historical or future impact of SLR on the nesting population. Green turtles had predominantly coral island nesting sites (51.2%), similar to hawksbill turtles (31.6%) (Fig. 4b-c). The largest being on Raine Island in the northern Great Barrier Reef, where between 5,000 and 100,000 green turtles are estimated to nest on this small coral cay per year (Dunstan et al. 2020). In contrast, most large nesting populations of leatherback, olive ridley and Kemp’s ridley turtles choose to nest on the mainland (87%, 93.3% and 100% respectively). Two of the largest being the Gahirmatha rookery in Orissa, India, where 50,000–386,000 olive ridley turtles nest annually during arribada events (Shanker et al. 2004, 2024). Loggerhead and flatback turtles preferably nest on elevated islands, typically close to the mainland and made up 50% and 57.1% of the nesting sites respectively. Important elevated island sites include Masirah Island in Oman, which has ~ 23,500 loggerhead turtle clutches laid annually (Willson et al. 2020). However, barrier island sites were also recognised as hugely important sites for nesting loggerhead turtles, particularly along the eastern coastline of continental U.S.A., where an estimated ~ 71,000 clutches are laid annually (Ceriani et al. 2019).

Fig. 4
Fig. 4The alternative text for this image may have been generated using AI.
Full size image

Geographic locations of nesting sites with large number of nests (as defined in the methods) for all seven sea turtle species; (a) leatherback turtle (Dermochelys coriacea); (b) green turtle (Chelonia mydas); (c) hawksbill turtle (Eretmochelys imbricata); (d) light blue = olive ridley turtle (Lepidochelys olivacea) and dark blue = Kemp’s ridley turtle (Lepidochelys kempii); (e) loggerhead turtle (Caretta caretta); and (f) flatback turtle (Natator depressus) in northern Australia. This figure serves to illustrate the range of nesting habitat types used by each species. Symbol shapes represent the type of nesting site. References for sites are included in Supplementary Information Table S4. Turtle images adapted from NOAA

The categorisation of globally important nesting sites by the site type demonstrates that green and hawksbill turtle populations may be at the highest risk of SLR impacting their reproductive success. Similarly, leatherback and olive ridley turtles, preferentially nest away from vegetation and on wide, flat beaches (Whitmore and Dutton 1985; Behera et al. 2013; Barik et al. 2014; Veelenturf et al. 2020; Shanker 2021), so despite nesting primarily on mainland beaches, they are likely also at higher risk from sea level rise.

Discussion

In this review, we compiled the available evidence for the magnitude of SLR threats from 34 unique research articles and related this evidence to the historical and projected impacts of SLR on sea turtle nesting beaches around the world. This compilation of research extends upon previous reviews completed on the effect of climate change on sea turtles (Patrício et al. 2021). By collating this knowledge, we identified that sea turtle nesting beaches on islands, and more specifically low-lying coral islands are likely to be at the greatest risk to inundation from SLR. From a geophysical perspective, this is intuitive because as the sea level rises, low-lying coral cays and barrier islands are likely to be submerged if accretion and sedimentation rates cannot maintain the island’s elevation (Fig. 5a). Examples of low small lying islands that might be particularly susceptible to SLR include Raine Island (Australia), the world’s largest green turtle rookery (Robertson et al. 2024) and Aves Island (Venezuela) in the Caribbean, another large green turtle rookery (García-Cruz et al. 2015) (Fig. 1e-f). Whereas along mainland beaches with dunes and natural areas backing the beach, the nesting habitat can retreat with sea level rise, so there is unlikely to be any substantial loss of nesting habitat (Fig. 5b). The same could be said for other elevated nesting islands such as St Eustatius Island, in the Atlantic Ocean, where the island’s elevation is ~ 600 m above sea level at the volcanic peak. Hence, sea turtles are likely to continue nesting at this remote oceanic site even under the more extreme SLR scenarios if the nesting beaches simply move inland.

Fig. 5
Fig. 5The alternative text for this image may have been generated using AI.
Full size image

Schematic to illustrate the recession of nesting habitats over time based on projected sea level rise (SLR) and comparing the impacts on (a) remote coral sand islands; (b) beaches on continental or volcanic islands which are elevated; or (c) beaches with man-made barriers and or a natural barrier, such as a cliff. The key difference between these habitat types is the available area at altitude. (a) Coral islands are low-lying and reliant on sediment accretion to maintain above sea level, meaning turtle nests are more likely to be flooded and the island may be lost due to SLR. (b) Beaches on elevates islands and the mainland can recede with SLR. However, (c) coastal urbanisation and natural barriers can create a “coastal squeeze” effect, minimizing available nesting habitat and reducing hatchling production due to increased density of nesting activity. Without natural recession of the beaches, the turtles will be forced to seek out nesting habitat elsewhere. Graphic generated in Canva®

For other elevated island sites such as Barbados, beach retreats to higher elevations are not always possible because these islands can be heavily populated and most urban areas line the coast, so man-made barriers prevent natural recession with SLR (Fish et al. 2008) (Fig. 5c). This is a well-documented issue, known as ‘coastal squeeze’ and threatens many ecosystems (e.g., mangroves and seagrass) as well as human societies along coastal areas (Pontee 2013). The term ‘coastal squeeze’ refers to the process whereby rising sea levels drive coastal ecosystems landward, whilst man-made or some natural barriers can create a static border between land and sea and force these habitats to become narrower or disappear altogether (Pontee 2013). Nesting beaches with the under the greatest threat of coastal squeeze from man-made barriers are located in Florida, U.S.A., Jamaica, and the Cayman Islands, with loggerhead and hawksbill turtle nesting beaches likely to experience coastal squeeze effects more than other species (Biddiscombe et al. 2020). Coastal squeeze can also be exacerbated by the introduction of beach armouring. However, Nelson Sella et al. (2023) recently analysed the reproductive success of loggerhead turtles nesting on ‘urban pocket’ beaches, which are formed by a similar process to natural pocket beaches. Overall, loggerheads preferentially nested in urban pocket beaches and had higher hatchling success than both unarmored and armored beaches (Nelson Sella et al. 2023), highlighting that there are mitigation strategies against SLR already available.

Similar to man-made barriers, natural barriers such as cliffs or waterways can also prevent the natural recession of a nesting habitat (Fig. 5c). Cliff faces commonly back the beaches of islands that have been formed through volcanic activity and beaches with high wave energy. Katselidis et al. (2014) identified that natural cliffs would prevent the future land-ward shift of important loggerhead turtle nesting habitat on Zakynthos Island, in Greece. Under a 0.6 m SLR simulation between 2009 and 2100, these authors projected a 32.2% loss of nesting habitat and 44.7% of nests would be flooded as a result (Katselidis et al. 2014). This is substantially higher than another Mediterranean nesting site, Samandag Beach, Turkey which has been projected to lose 13.1% of green turtle nesting habitat and have ~ 5% of nests inundated under the 0.63 m SLR scenario between 2014 and 2100 (Sönmez et al. 2021). Samandag Beach is a mainland site with no natural barriers, however there are sections with substantial urban development backing the beach, and so some areas of the beach are likely to be more impacted by SLR than others. Similarly, some sea turtle species may be more impacted by SLR than others due to their innate preference for certain habitat types, beach slope angle or the natural nest depth. It is clear from our review of the globally important nesting sites that green and hawksbill turtles may be more susceptible to SLR as they primarily nest on low-lying coral islands (51.2% and 31.6% respectively), which are also most likely to experience greater rates of habitat loss and nest flooding.

Assessing the historic loss of sea turtle nesting areas

Analyses of the impact of recent sea level changes are needed to assess whether SLR is already negatively impacting sea turtle nesting and the speed with which nesting area can be lost. However, such studies are relatively rare. Although not for the purpose of mapping changes to sea turtle nesting habitat, Albert et al. (2016) analysed historical aerial imagery (1947–1962) combined with satellite imagery (2002, 2011 and 2014) and showed that there has been substantial loss of beaches on 33 island sites in the Solomon Islands, South Pacific, including the complete loss of five vegetated islands. However, major loss of nesting habitat can also occur very rapidly, sometimes within a couple of months primarily due to storm surges from cyclones and hurricanes (Marra et al. 2022). Storm frequency and intensity is predicted to increase with climate change, and so SLR will also likely exacerbate storm inundation levels (Knutson et al. 2020). Marra et al. (2022) recently reported on satellite imagery taken of East Island in the French Frigate Shoals (Lalo), Hawai’i and showed the complete demolition of the entire island after hurricane Walaka in 2018. Similarly, Baker et al. (2020) reported the progressive loss of terrestrial habitats in Lalo across several decades, including the inundation of Whaleskate and Trig Island, which are also important basking and breeding habitats for Hawaiian monk seals (Neomonachus schauinslandi) and various seabird species. This loss has pushed the breeding and basking activities of turtles, seals and seabirds to Tern Island, accounting for ~ 75% of the remaining terrestrial habitat in Lalo (Baker et al. 2006). These examples highlight the rate of rapid loss of beach habitats in the Pacific just in recent decades, but the rate at which this is occurring for other nesting sites around the world is less well known. Satellite imagery is available globally going back many decades, with higher resolution imagery (15 m) introduced with Landsat 7 (1999), this approach could be used to measure the historic loss of nesting area for sea turtles around the world. In this way, sites that have lost nesting area particularly quickly in recent decades could be triaged for conservation intervention.

Ability of nesting turtles to locate alternative nesting habitat

An important factor that will drive whether sea turtles can mitigate threats of SLR is how they will respond to loss of nesting habitat. One possibility is that turtles might change their preferred nesting location within the beach profile. Often turtles already crawl far up the beach until either vegetation impedes their progress (Stokes et al. 2024) or until they are still on the open beach but far enough from the ocean to minimize seawater overwash of nests. The ability of females to adapt their nest positioning in the face of SLR warrants further study. Another key issue here may be the extent to which turtles can colonise new nesting areas if their original beaches are lost. We can piece together possible responses to loss of nesting beaches by considering the extent of inter-nesting movements of turtles, their navigational inaccuracies leading to new sites being located and case-studies of responses when a nesting habitat has been lost and a new nesting habitat has become available. Christiaanse et al. (2024) recently conducted an in-depth global analysis of nesting habitat use for five species of sea turtles and found that the fundamental niche was 23% of the coastal regions between − 39˚ and 48˚ latitude, but the realized niche was only 7% of this same region. Given the current threat to nesting sites, the findings from Christiaanse et al. (2024) highlight the need to protect existing suitable sites, as they may be needed as more current beaches are lost.

From satellite tracking and flipper tagging studies it has been shown that turtles may nest on several beaches throughout the nesting season (e.g., Hays et al. 1991; Perez et al. 2022). In extreme cases, individual turtles have been recorded nesting on beaches > 1000 km apart (Le Gall and Hughes 1987). There is some evidence that nesting turtles have the capacity (given the energy requirements) to find a new site, should a turtle arrive at her natal site which is no longer suitable. Sloan et al. (2022) observed the inter-nesting movements of 14 green turtles in southwest Florida, U.S.A., and documented the distance travelled for individuals within the nesting season ranged from 277 to 1,489 km (average travel distance ~ 21 km day−1). Leatherback turtles have been documented to have even larger inter-nesting movements (Shillinger et al. 2010; Robinson et al. 2017; Asada et al. 2021). A leatherback female that was tagged in the Caribbean Sea travelled ~ 56 km day−1 between nesting events (Byrne et al. 2009), and so some species may have an even greater propensity for sporadic nesting site changes in response to SLR.

Turtles migrating from a foraging area to a nesting site may also have navigational inaccuracies that lead to new potential nesting areas been discovered before their first clutch of a season is laid. For example, migrating turtles travelling to isolated islands sometimes miss their target, even overshooting by ~ 1,000 km before eventually finding the intended destination (Hays et al. 2020). Likewise, turtles migrating to mainland sites often travel long distances, following along the coastline and building up their knowledge of alternative nesting sites which could be used in future (e.g., Schofield et al. 2013). In some cases, migrating turtles arrive at non-suitable breeding sites due to navigational inaccuracies. For example, every year during the nesting season at Ascension Island, a very remote volcanic island in the South Atlantic, green turtles are also seen mating at St Helena, the nearest alternative island which is ~ 800 km away but surrounded by cliffs with no suitable nesting beaches (Hays et al. 2025). These turtles arriving at St Helena are thought to be lost turtles that have missed Ascension Island (Hays et al. 2025). These navigational imperfections of turtles are likely linked to their geomagnetic senses only allowing crude, rather than pinpoint, homing (Lohmann et al. 2001). Previous experiments have shown that post-nesting turtles are able to navigate back to their foraging grounds even when magnetic disturbances are applied (Papi et al. 2000). Further, between nesting seasons or in the long interval between a hatchling leaving a nesting beach and then attempting to breed for the first time, the Earth’s magnetic field moves, a process known as secular variation (Lohmann et al. 2008). So, turtles returning to the magnetic geocoordinates (magnetic inclination and intensity) that characterised their nesting beach years previously, may end up several hundred km distant to that beach (Lohmann et al. 2008). Turtles are described colonising new nesting beaches, for example in western Mediterranean, which can only occur due to breakdowns in natal homing (Carreras et al. 2018). As another example, few loggerhead clutches were laid (< 10 per year) further south of the Wongarra coast and the Capricorn Bunker group of central Queensland, Australia prior to the 1990 s (Tarvey 1993; Limpus 2008). However, none of these clutches produced hatchlings because nest temperatures were too low to support successful development (Tarvey 1993). In more recent times (since ~ 2010), clutches laid over 300 km south of these traditional nesting sites have begun to produce viable hatchings because of increased sand temperatures (West 2019). So, while navigational imperfections may help sea turtles to locate new nesting areas, more information is needed on the rate of colonisation of new nesting sites and the extent to which this will mitigate loss of existing nesting beaches. Furthermore, there have been several documented examples of situations where nesting populations of turtles have shifted their nesting site, typically by a few kms of tens of km, due to changes in beach dynamics (Baker et al. 2006; Pritchard 2008; Shanker 2021), but this is still not well understood.

While these various lines of evidence suggest that turtles may be able to move to alternative nesting beaches if their original sites disappear, the full extent of this movement remains unknown. This response of turtles to beach loss remains a key unresolved question. Satellite tracking turtles could provide a better understanding of how nesting beaches are selected by individuals. Almost all tracking studies with adults focus on equipping turtles on nesting beaches and tracking their interesting and post-nesting movements (Hays and Hawkes 2018). More challenging is to equip turtles on their foraging grounds and track their migrations to breed. Such studies have rarely been performed (see for example James et al. 2005) and may shed light on the tendency of turtles to visit potential alternate nesting sites on their breeding migrations.

Complexities with estimating SLR effects on nesting beaches

There are some studies that have pointed to the dynamic nature of beaches and how new beaches may emerge as others disappear. There are particularly interesting complexities associated with studying the geo-morphological dynamics of coral islands, as some studies indicate that while one exposed side of the island actively erodes away, the lagoon-side of the island actively accretes sediment (Stoddart and Steers 1977). This dynamic nature of beaches points to a limitation of current estimates of the impacts of SLR on nesting turtles. Of the 25 studies that had conducted forecast projections of SLR and the impact to nesting turtles, 48% had used Digital Elevation Models (DEMs) to construct a topographic model of the nesting beach. Typically, studies take an existing beach profile and then raise the sea level to assess how much of the nesting area would be lost or what percentage of nests would be inundated under various SLR scenarios (for example Varela et al. 2019; Veelenturf et al. 2020; Beber et al. 2024). This ‘bathtub’ methodology has some limitations, as it cannot account for the coastal processes that occur with SLR and, and may overestimate the number of nests inundated (Hawkes et al. 2009), as the assumption is that turtles would not nest higher up the beach in response to SLR. However, without a barrier, it is widely accepted that SLR will cause beaches to retreat and move in an inland direction, allowing turtles to continue to nest until a barrier is met and the turtles are forced to nest elsewhere.

Some studies, such as Lyons et al. (2020) accounted for the landward retreat of the nesting area by using mapping tools, whilst others opted to use the Bruun Rule (Bruun 1988) to estimate the rate of beach recession (Fish et al. 2008; Reece et al. 2013). The Bruun Rule has been widely used but also regarded as a simple two-dimensional model (Cooper and Pilkey 2004), as it projects coastal recession by calculating the upward and landward displacement of the seabed profile in response to SLR (Bruun 1988). Some authors have justified excluding the Bruun model from their analysis due to its simplicity (for example Veelenturf et al. 2020). A better understanding of how beaches will respond to SLR might be possible through more recent modelling approaches. The probabilistic coastline recession (PCR) model developed by Ranasinghe et al. (2012), is a process-based model that estimates dune recession under SLR scenarios based on physical erosion and accretion processes acting on the dunes (e.g., storms, dune recovery). Makris et al. (2023) developed a high-resolution model called ‘coastFLOOD’ to predict coastal inundation from storm surges, which has already been used to produce more advanced risk assessments of nesting beaches for loggerhead turtles in the Cabo Verde archipelago, Atlantic Ocean (Sousa-Guedes et al. 2025). Similarly, ‘XBeach’ is a nearshore two-dimensional model that can be used to simulate coastal hydrodynamics and transport of sediments and has also been validated with field test cases (Roelvink et al. 2009; Patricio et al. 2021). Such methods would provide a more holistic account of dune retreat under future SLR scenarios and storm surges and allow for more robust assessments of the impacts this would have on the nesting population. In addition to work associated with sea turtles, there is also huge interest in SLR-induced coastal change in the context of coastal management, including threats to human infrastructure and society (Martyr-Koller et al. 2021; O’Donnell et al. 2024). Collaborations with those working in this wider area, may provide fruitful insights for sea turtle conservation. Hence, we can expect that existing nesting beaches where there is natural elevated land backing the dunes would be less vulnerable to SLR than low lying islands, assuming that there are also no natural or human-made barriers to prevent beach retreat.

Management and mitigations against sea level rise

Both short-term and long-term mitigations strategies have been developed to help alleviate the effects of SLR on coastal erosion and improve nest success (Fuentes et al. 2012). Short-term mitigation involves the manipulation of the nesting habitat or the relocation of clutches (Tuttle and Rostal 2010; Robertson et al. 2024), whilst long-term mitigation may involve policy changes to developments along the coastline that prevent future coastal squeeze effects (Fish et al. 2008), and more broadly there would be system changes to reduce carbon emissions and alleviate the impacts of climate change. Short-term measures have been hugely effective for nesting sites across the world. For example, the relocation of “doomed” clutches to higher dune areas is a widely accepted protocol in turtle conservation and often requires the construction of a hatchery above the spring tide level to protect nests (Pike 2008; Tuttle and Rostal 2010; Martins et al. 2021). However, this protocol is most practical for small beaches as it is very labour intensive and requires consistent night-time monitoring so that the eggs are relocated within the first few hours of oviposition to avoid movement-induced embryonic mortality (Limpus et al. 1979). Much more challenging are attempts to move clutches of eggs to alternate nesting beaches to establish new nesting sites. As the survival of eggs to adulthood is likely very rare (Frazer 1986; Parmenter and Limpus 1995; Santidrián Tomillo 2019), huge numbers of eggs need to be moved to establish a new rookery. While this approach was trialled in 1977 as part of the bi-national Kemp’s Ridley Restoration and Enhancement Program (KRREP), translocating 10 to 20 clutches from Kemp’s ridley turtles per year from Playa de Rancho Nuevo, Tamaulipas, Mexico to re-establish a nesting colony on Padre Island, Texas U.S.A. (Wibbels and Bevan 2015), the efficacy of this approach is still not conclusive (Shaver et al. 2016).

The mechanical manipulation of the beach environment, often termed “beach nourishment” is a relatively new concept for the specific purpose of elevating the turtle nesting beaches and is regarded as a “soft” solution to coastal erosion and SLR (Rupert 2011; Robertson et al. 2024). Historically, beach nourishment has indirectly caused negative effects to reproductive success of turtles (Rumbold et al. 2001; Leonard Ozan 2011), as the primary purpose was for coastal development and beach restoration, so therefore likely caused modifications to the beach sand composition (e.g., sand colour and size, moisture, salinity, compaction, mineral content, and porosity) (Greene 2002). In a large ongoing effort, the Raine Island Recovery Project was initiated in 2013 and involved the re-profiling of the nesting habitat for the world’s largest green turtle population located in Australia’s northern Great Barrier Reef, whereby sand from the lower part of the beach was moved with bulldozers to the upper part of the beach (Robertson et al. 2024). Over a five-year period, 40,000 m3 of sand was moved, effectively doubling the area of nesting habitat that was outside of the inundation zone (~ 35,000 m2 to ~ 70,000 m2). In the years that followed, the mean hatching success of nests in the re-profiled areas was greater than that of the control area nests (Robertson et al. 2024). Re-profiling and any other environmental manipulations require very stringent measures to ensure that the sand composition is viable for turtle nests, as there have been several cases in Florida where beach nourishment has initially resulted in the reduced nesting success of loggerhead and green turtles (e.g., 44.1% and 52.2% respectively) (Rumbold et al. 2001; Brock et al. 2009).

Long-term management for SLR and the protection of sea turtle nesting beaches can involve the establishment of building policies to restrict developments within a defined area or distance of the spring tide line (Fish et al. 2008). This measure can be put into place to allow for the natural landward recession of the dune area with SLR and provide a buffer against coastal squeeze effects. These buffer zones around nesting beaches are referred to as “setbacks” and are either used in formal regulations or informal guidelines for developers and or incorporated into coastal management plans in some countries (Fish et al. 2005, 2008). For example, Fish et al. (2008) projected that conservative setback regulations (10 and 30 m) were not adequate enough to prevent complete habitat loss for hawksbill turtles nesting on Barbados, in the Caribbean. However, setback regulations that were greater than 50 m resulted in over 50% greater protection of nesting habitat with projected SLR, and no habitat loss was projected with 90 m setback regulations in place (Fish et al. 2008). However, only five out of the 33 Caribbean islands surveyed had setback regulations greater than 50 m, and had an average setback of just 30 m, indicating that the majority of these islands would not be resilient to SLR and experience coastal squeeze effects or the complete loss of nesting habitat for critically endangered hawksbill turtles (Fish et al. 2008). In Baha, Brazil, Lopez et al. (2015) developed similar recommendations for safeguarding hawksbill turtle nesting beaches against SLR based on their threat level. These recommendations included setback regulations of 60, 120 and 180 m to allow for natural recession, as well as restrictions for beachfront lighting and beach access for the public to prevent disturbance to nesting turtles and hatchlings (Lopez et al. 2015).

Future investigations and conclusions

We have shown that most studies of potential impacts of SLR on sea turtles have involved projecting SLR scenarios to show how much of existing beaches will be lost. Key information gaps remain concerning:

  1. (i)

    Empirical data on the extent of nesting beach loss that has already occurred in recent decades. This question could be addressed by analysis of existing remote sensed imagery.

  2. (ii)

    Information on how beaches will re-configure with SLR. For example, where all the nearby land is not low-lying, beaches may migrate inland. Here collaborations with coastal geo-morphologists may provide an avenue for more accurately predicting how beaches will change in various SLR scenarios.

  3. (iii)

    The ability of turtles to colonise new nesting areas. Here key issues are the distance to suitable alternative nesting sites and the ability of turtles to move to those sites.

It might be expected that there will be differences in ability to locate new sites between remigrant turtles that have nested previously on a beach and so know that site better versus neophyte turtles that are breeding for the first time and so are less familiar with the nesting area and may be more likely to move elsewhere. Tracking both established breeders (remigrants) as well as first-time nesters (neophytes) on their migrations to breed, while challenging, might be an avenue to learn more about how they select nesting areas and their resilience to SLR.

In conclusion, this review showcases our current knowledge on the threat of SLR to turtles and also compiled available evidence from around the world. By collating this knowledge, we identified that sea turtle nesting beaches on low-lying islands are likely to be at the greatest risk to inundation from future SLR. Highlighting that the impacts on the nesting population from rising sea levels is likely to be most acute for low-lying islands with no nearby alternative nesting areas. We examined how natural processes (e.g., moving to alternative sites) and conservation interventions might mitigate this threat. However, we require more empirical data on the extent of historical nesting habitat loss and more informed models on the rates of projected habitat loss and nest flooding.