High vulnerability of the endemic Southern Ocean snail Neobuccinum eatoni (Buccinidae) to critical projected oceanographic changes Authors: Rosvita González, Luis R. Pertierra, Pablo C. Guerrero, Angie Díaz Published: 2024 ABSTRACT Climate change is projected to substantially alter the Southern Ocean’s physical and chemical properties, thereby impacting its marine ecosystems and species, particularly those in Antarctic and sub-Antarctic regions. This study focuses on Neobuccinum eatoni, a polar marine ‘true whelk’ endemic to these regions, utilizing 166 spatially independent occurrence data records to model potential distribution shifts under future climate scenarios. Employing Species Distribution Models (SDMs) on spatially cross-validated occurrences, we achieved high predictive accuracy, identifying "sea water salinity range" at mean bottom depth as the most significant predictor of habitat preferences. Additionally, dissolved iron (minimum), ocean temperature (range), and pH (long-term maximum) emerged as critical factors influencing the species’ modeled distribution. By 2050, future projections under the SSP2-4.5 scenario predict an eastward expansion, particularly in the Antarctic Peninsula, the Scotia Arc and the Weddell Sea, with an expansion in the latter region also predicted under the SSP5-8.5 scenario. However, in both scenarios, a reduction in habitat suitability is expected in certain sectors around the Antarctic continent and the Kerguelen Archipelago. By 2100, under the moderate emissions scenario (SSP2-4.5), the species is projected to move to deeper areas and lower latitudes, with notable expansions in the Weddell Sea and in the Southern Ocean surrounding the Kerguelen Archipelago. However, under the SSP5-8.5 scenario, expansion is projected in the Weddell Sea and reductions in Antarctic and subantarctic regions. This study highlights the critical influence of changing salinity on N. eatoni’s distribution, predicting a significant habitat reduction under high CO2 emissions scenarios (SSP5-8.5). The findings underscore the urgent need for focused research on the vulnerability of endemic marine invertebrates to develop effective conservation strategies in the face of rapid climatic changes. Introduction The Southern Ocean is a critical water mass for sustaining marine life and climate on Earth1. The Southern Ocean is facing significant environmental shifts including sea-level rise, increasing sea surface temperatures, reducing sea ice extent, and ocean acidification, with pH projected to decrease by 0.1 to 0.6 units by 21002. These changes are occurring at a rapid pace, affecting the region more quickly than many others worldwide, as evidenced by the rapid regional warming temperatures, changing wind patterns, sea level rise, and sea ice fragmentation3,4. Such shifts have profound implications for marine ecosystems, particularly impacting benthic fauna known for their slow reproduction, longevity, and narrow thermal ranges5. Alterations in marine ecosystems will force species to readapt, migrate, or face extinction, thus affecting their latitudinal and bathymetric distributions6,7. However, the capacity for adaptation is generally limited by slow generational turnover and frequent stenothermy5,8, raising concerns about biodiversity loss. Despite their exceptional diversity in the Southern Ocean and its wide provision of nature’s contributions to people, the fate of marine invertebrates under global climate change remains largely understudied, presenting challenges for their conservation and management9,10. The Southern Ocean, encompassing the southernmost reaches of the Atlantic, Indian, and Pacific Oceans, covers approximately 22 million square kilometers or 6.1% of the global ocean surface. It is delineated oceanographically by the Antarctic Polar Front (APF), characterized by high current velocities and a pronounced thermal gradient, and dominated by the Antarctic Circumpolar Current (ACC), which plays a crucial role in global thermohaline circulation1,11–13. The ACC and APF significantly influence species distribution, contributing to the high endemism within the region. The isolation facilitated by the opening of the Drake Passage at around ~ 50 million years ago14and subsequent evolutionary pressures has led to a unique fauna, with endemism levels around 50% overall and even higher within specific groups like gastropod mollusks15,16. Global climate change pressures in the region pose significant challenges to these finely balanced ecosystems. The expectations for the impact of global change on Antarctic and sub-Antarctic marine ecosystems are complex and varied. In the sub-Antarctic regions warming could pose the risk of reducing habitats for marine fauna as sea level shifts and changes in oceanographic conditions alter the seafloor landscape, potentially isolating populations and preventing them from moving to new habitats that meet their ecological requirements. This is particularly critical for species with limited dispersal capabilities, specific depth preferences, or those that rely on certain seabed features for feeding, reproduction, or shelter. Over time, these changes can fragment populations and reduce genetic diversity, increasing the risk of local extinctions17,18. Conversely, in Antarctica, emerging suitable conditions from retreating ice sheets may reveal new habitats for colonization. Overall, distributional changes over latitudes are expected to go poleward19–22whilst net bathymetric changes are harder to anticipate due to the complex interplay of factors such as geological and oceanographic processes, including plate tectonics, sedimentation, and erosion, which may be altered by climate-induced variations in ocean currents and sea levels. Additionally, the precision of models that predict these changes is constrained by the limited availability and low resolution of bathymetric data, which complicates efforts to accurately represent the temporal and spatial variability of these processes23. The pivotal role of the ‘Hutchinsonian niche-biotope duality,’ conceptualized as the correspondence between multidimensional niche and geographic spaces, is fundamental for forecasting species’ potential distribution shifts under environmental changes24. Species distribution models (SDMs) leverage this duality to associate occurrence data with environmental variables, enabling precise projections of species distributions across various geographical areas and climate scenarios25–27. Ongoing alterations in oceanic conditions due to climate change may affect habitat availability, leading to the displacement, expansion, reduction, or complete loss of suitable areas, all of which directly impact species survival28. These shifts significantly alter community structure and biogeographical patterns29. In this context, SDMs have emerged as essential tools for simulating the biogeographical responses of species to climate change30–32. By integrating statistical analyses with Geographic Information Systems, SDMs provide an effective way to forecast potential shifts in species distributions under varying climate change scenarios. Despite considerable advancements in our comprehension of climate change impacts on diverse taxa of the Southern Ocean7,33,34, there remains a conspicuous gap in our understanding of how specific biological groups, notably marine gastropods (75% endemism15), navigate the evolving environmental dynamics within the ACC and the APF regions. This gap underscores a demand for focused research endeavors aimed at inferring the ecological consequences of climate change on the biogeography of marine gastropods and other less-studied functional groups. Among the diverse benthic fauna, the ‘true whelk’ gastropod Neobuccinum eatoni (E.A. Smith, 1875) exemplifies the endemic marine life of the Southern Ocean. It has a broad distribution including sub-Antarctic regions and an ample bathymetric range, making this species an excellent candidate to study potential shifts in the distribution of marine benthic fauna that share its biological and ecological characteristics in response to environmental changes. Additionally, as a scavenger, it is particularly adept at adjusting to fluctuations in food availability, enhancing its utility as an indicator of ecological changes within marine ecosystems35,36. Our study aims to assess the impacts of climate change on the distribution of N. eatoni, focusing on its vulnerability to local extinction in sub-Antarctic regions and distributional shifts in Antarctic waters. Through SDMs, we explore future shifts in the potential distribution patterns of N. eatoni under two IPCC climate change scenarios, SSP2-4.5 and SSP5-8.5, for the periods 2040–2050 and 2090–2100. Our predictions pose that N. eatoni,a gastropod endemic to the Southern Ocean, will migrate poleward to higher latitudes in response to projected climate change. This distributional shift is expected under the climate scenarios SSP2-4.5 and SSP5-8.5 for the forecast periods of 2040–2050 and 2090–2100. Specifically, in regions such as the Kerguelen Plateau, global warming is likely to affect the growth rates, population structures, prey availability, and reproductive success of various marine species37. In this regard, the irregular sea urchin, Abatus cordatus (Verrill, 1876), endemic of the coastal regions near the Kerguelen Plateau has been identified as especially susceptible to future environmental changes7. The study of the distributional shifts of the Southern Ocean benthic fauna can shed light in comprehending broader biological responses to ocean warming, including new species interactions, disrupted ones, or both38. Moreover, changes in some cases can impose the very risk of localized or even global extinctions7,20,39. For these reasons broader ecological niche forecasts are deemed essential for understudied functional groups, particularly those involved in nutrient recycling and ecosystem health maintenance. Methods Spatial occurrence data The collection of N. eatoni occurrence records underwent a detailed multi-step filtration process from the Global Biodiversity Information Facility (GBIF; http://gbif.org) database, accessed in April 2023, which initially contained 574 records of the species. Additionally, museum specimens were screened and excluded when photographic evidence suggested mismatches with N. eatoni identification or when their geographic information was found to be inaccurate. This filtering yielded 349 georeferenced occurrence points. Once the data were downloaded, they were subsequently filtered, eliminating duplicate points at the resolution of the oceanographic variables to avoid model inflation from spatial pseudoreplication40. Records older than 60 years were also discarded, considering a temporal resolution of presence of the species between 1961 and 2019, reducing to 126 the occurrence points obtained from GBIF. This dataset comes from various repositories, such as: Southern Ocean Mollusc Database (SOMBASE), International Barcode of Life project (iBOL), National Institute of Water and Atmospheric Research (NIWA), and from data centralization initiatives such as The Antarctic Biodiversity Information Facility (ANTABIF), together with complementary data from sampling events published by SCAR-AntOBIS. We complemented the records acquired from GBIF with additional occurrence records sourced from the National Museum of Natural History in Paris, France, and the Scientific Collection of the University of Vigo in Vigo, Spain. Details are provided in Table 1. Furthermore, we incorporated 15 additional localities sampled during our fieldwork on Antarctica, Heard, and Kerguelen Island, resulting in a comprehensive dataset of 166 occurrence records (Fig. 1a; Supplementary Information 1).Table 1Oceanographic variables used in species distribution models for Neobuccinum eatoni and their contribution to the models.VariablesAverage contributionMean bottom depth1. Dissolved molecular oxygen (minimum)0.052. Dissolved molecular oxygen (range)0.073. Salinity (range)0.324. Ocean temperature (minimum)0.08Sea surface5. Dissolved iron (minimum)0.146. pH (long term maximum)0.137. Salinity (maximum)0.028. Sea ice cover (minimum)0.039. Ocean temperature (range)0.13Fig. 1Geographic occupancy across the Southern Ocean of the marine snail, Neobuccinum eatoni: (a) occurrence points used in Species Distribution Models with different colors depicting various data sources—Global Biodiversity Information Facility (GBIF) in brown, National Museum of Natural History in Paris in pink, Scientific Collection of the University of Vigo, Spain in yellow, and new samples obtained through our fieldwork in orange. (b) Current potential distribution. Map was generated using QGIS 3.30 (https://www.qgis.org/). Table 1Oceanographic variables used in species distribution models for Neobuccinum eatoni and their contribution to the models.VariablesAverage contributionMean bottom depth1. Dissolved molecular oxygen (minimum)0.052. Dissolved molecular oxygen (range)0.073. Salinity (range)0.324. Ocean temperature (minimum)0.08Sea surface5. Dissolved iron (minimum)0.146. pH (long term maximum)0.137. Salinity (maximum)0.028. Sea ice cover (minimum)0.039. Ocean temperature (range)0.13 Oceanographic variables used in species distribution models for Neobuccinum eatoni and their contribution to the models. Geographic occupancy across the Southern Ocean of the marine snail, Neobuccinum eatoni: (a) occurrence points used in Species Distribution Models with different colors depicting various data sources—Global Biodiversity Information Facility (GBIF) in brown, National Museum of Natural History in Paris in pink, Scientific Collection of the University of Vigo, Spain in yellow, and new samples obtained through our fieldwork in orange. (b) Current potential distribution. Map was generated using QGIS 3.30 (https://www.qgis.org/). We applied spatial cross-validation methods to address spatial autocorrelation and sampling bias in the occurrence data, splitting the data into distinct geographic regions41. This procedure mitigates the effects of spatial clustering, which can lead to over-optimistic model evaluations. Following these corrections, SDMs were conducted using four clusters of occurrence data derived from these analyses (Supplementary Information 2), ensuring more robust and reliable model predictions. Oceanographic variables The marine environmental data selected were those gridded datasets from BioORACLE v3.0 database42that were both available for the present and for the Shared Socioeconomic Pathway (SSP) scenarios SSP2-4.5 and SSP5-8.5 of the latest Intergovernmental Panel on Climate Change Assessment Report43 for two future periods (2040–2050 and 2090–2100). A preliminary list of sixty global-scale raster layers, each with a grid-cell resolution of 0.05 degrees (approximately 5.5 km at the equator), were analyzed and trimmed to the Southern Ocean region (defined here as −180°/ + 180° longitude; −45°/−78° latitude; Supplementary Information 3). This step was taken to construct the background environmental space specifically available for the Southern Ocean, thereby avoiding the inclusion of artificial projected presences in disjunct regions with environmental characteristics similar to those of the Southern Ocean. The oceanographic descriptors include various set of layers around sea floor conditions: dissolved molecular oxygen, salinity, and ocean temperature; as well as layers depicting conditions estimated at the sea surface, such as: dissolved Iron, dissolved molecular oxygen, pH, salinity, sea ice cover, ocean temperature and total phytoplankton. Each variable set comprise six different constructed parameters: long-term average, minimum and maximum records, long-term average of minimum and maximum records per year, and range, thus obtaining 60 preliminary climate layers for the present and for the SSP2-4.5 and SSP5-8.5 scenarios, modeled for the years 2050 and 2100, all downloaded from the Bio-ORACLE v3.0 database. Before executing the SDMs, we conducted multicollinearity analysis to refine our selection of oceanographic variables. This step was critical for eliminating variables with high correlation, adhering to the principle that most statistical methods require predictive variables to exhibit low correlation for optimal performance44. Utilizing the ‘vifstep’ function within the ‘usdm’ package in R45, we identified variables exhibiting a Variance Inflation Factor (VIF) less than 2. This process resulted in a refined set of oceanographic variables, characterized by minimal correlation, thereby enhancing the reliability and interpretability of our SDMs. The final selection was composed of 4 variables of mean bottom depth: dissolved molecular oxygen minimum, dissolved molecular oxygen range, salinity range and ocean temperature minimum; and another set of 5 variables of sea surface: dissolved iron minimum, pH long term maximum, salinity maximum, sea ice cover minimum and ocean temperature range. This meticulous variable selection process was instrumental in mitigating potential multicollinearity, thereby safeguarding the robustness of our modeling efforts. To model the projected future potential distribution of N. eatoni in the Southern Ocean, Shared Socioeconomic Pathway (SSPs) scenarios outlined by the IPCC in its Sixth Assessment Report from 2022 were utilized. These scenarios depict theoretical trajectories associated with varying concentrations of greenhouse gasses in the atmosphere, each characterized by specific levels of Radiative Forcing (RF). RF quantifies the disruption in the equilibrium of Earth’s atmospheric system resulting from alterations in atmospheric composition due to human activity46. It is quantified as the rate of change in net radiative energy flux per unit area of the planet’s surface, measured at the top of the atmosphere, in watts per square meter (W/m2). There are five projection scenarios, ranging from a low future RF scenario of 1.9 W/m2 (SSP1-1.9) to a very high RF scenario of 8.5 W/m2(SSP5-8.5) by the year 210043. For our study, we selected SSP2-4.5, where RF stabilizes at approximately 4.5 W/m2 after 2100, and SSP5-8.5, where RF exceeds 8.5 W/m2 in 2100 and continues to rise. These scenarios anticipate CO2 levels reaching 538 and 936 parts per million (ppm) respectively19, leading to global mean surface temperature increases of 1.8 °C and 3.7 °C for the period 2081–2100 (relative to 1986–2005) and global mean sea level rises of 0,44—0,76 m and 0,63—1,01 m43. Recent years have seen a steady increase in atmospheric CO2 concentrations, with annual averages reaching 410 ppm in 2019, 418 ppm in 2022, and a monthly average of 420 ppm CO2in April 202347. These levels have already surpassed the lower CO2 concentration scenario (SSP1-2.6 with 421 ppm) projected by the IPCC AR6 for the year 2100, rendering it an unrealistic future scenario for the purposes of this research. Consequently, this scenario has been excluded from our modeling efforts. Instead, our work has focused on the SSP2-4.5 and SSP5-8.5 scenarios, which predict moderate (538 ppm) and very high (1100 ppm) CO2 concentrations by 2100, respectively. Species distribution models From 166 occurrence records of N. eatoni and the 9 largely uncorrelated oceanographic variables, the potential distribution of the species was modeled for the present and then projected for future scenarios, using the Biomod2 package in R Studio48. Pseudo-absences were generated by randomly sampling background data from areas where the species is not known to occur. Specifically, the background data were sampled using the Kernel Density Estimation (KDE) layer, which estimates the probability of finding presence data in each pixel. For each replicate, 500 background points were randomly selected from the KDE layer, ensuring that these pseudo-absences were distributed across the environmental space. Five modeling algorithms were tested (Random Forest (RF), Gradient Boosting Machine (GBM), Generalized Linear Models (GLM), Multivariate Adaptive Regression Splines (MARS), Maximum Entropy Network (Maxnet)) and 5 runs of each algorithm. The combined modeling results were used to construct an ensemble modeling consensus approach for predicting N. eatoni’s distribution. This decision was influenced by the complexity of environmental impacts on its habitat preferences and limited to those models of high performance (values > 0.65) across TSS and ROC metrics (Supplementary Information 4). Such an approach ensures robustness and reliability in projections by mitigating the limitations associated with single-model predictions and enhancing the overall confidence in the model outputs. Projections of the potential distribution were made for the present and for future climate scenarios (2050 and 2100 for SSP2-4.5 and SSP5-8.5). In addition, the median bathymetric depth at each scenario was calculated by averaging the scores of all predicted ‘presence’ cells of the region within the binarization threshold of maximum sensitivity and specificity. Finally, N. eatoni occurrence probability projection maps for the present and future were visualized using the free and open-source Geographic Information System, QGIS v3.30 (https://www.qgis.org/). Results All compiled N. eatoni occurrences are dispersed across circum-sub-Antarctic and Antarctic regions (Fig. 1a), being recorded between 47.12°S and 78.26°S latitudes and spanning a bathymetric range of 0 to 900 m. The species is naturally found on the sub-Antarctic islands of South Sandwich, Kerguelen, and Heard and McDonald, as well as on the Antarctic Peninsula, the Antarctic South Orkney Islands, and throughout the seafloor around the continent. All models showed high values (> 0.65) when analyzed with both evaluation metrics (TSS and ROC), indicating a good modeling performance41. The average contribution of oceanographic variables to the models was highest for the variable “salinity range of mean bottom depth” contributing in 0.32 points (Table 1), indicating its paramount importance in determining the species’ habitat preference. Temperature variables, particularly the range of the sea surface, and pH also show contributions that stand out in the ensemble, with the contributing observed of 0.13 points. The minimum iron (Fe) dissolved in the sea surface is another significant variable since it contributes 0.14 points to the models. This highlights the crucial role of nutrient availability as a determining factor in the productivity and structure of marine ecosystems. The spatial and temporal variability of nutrients impacts primary production, the composition of phytoplankton communities49and consequently, directly and indirectly influences the distribution and larval survival of organisms, as well as the availability of food among benthic species50. In high-nutrient, low-chlorophyll (HNLC) regions such as the Southern Ocean, Fe fertilization in particular, induces large phytoplankton blooms confirming its role as a limiting micronutrient for biological production51,52. The current potential distribution within the binarization threshold encompasses a total area of approximately 5,112,597 km2 ± 761,249 km2 (Fig. 2), surrounding the Antarctic continent, also including South Orkney Island, the sub-Antarctic islands of the Kerguelen Archipelago and Heard, McDonald and South Sandwich Islands. Additionally, the SDM revealed other potential areas for current distribution, contingent on suitable abiotic conditions, in the South Georgia Islands, southern South America and New Zealand (Fig. 1b).Fig. 2Potential distribution areas projected to be suitable for occupation by Neobuccinum eatoni currently and under varying global socioeconomic developments (SSP2-4.5 and SSP5-8.5) in the short term (2050) and long term (2100). Error bars represent the standard deviation of projected areas (± 761,249 km2), and the horizontal line indicates the average across scenarios, providing a measure of uncertainty and a reference for comparison. Potential distribution areas projected to be suitable for occupation by Neobuccinum eatoni currently and under varying global socioeconomic developments (SSP2-4.5 and SSP5-8.5) in the short term (2050) and long term (2100). Error bars represent the standard deviation of projected areas (± 761,249 km2), and the horizontal line indicates the average across scenarios, providing a measure of uncertainty and a reference for comparison. The projection of the distribution of N. eatoni´s under the intermediate scenario (SSP2-4.5) for the short term (year 2050) suggests an expansion of its distribution at higher latitudes (~ 30%), particularly an increase in the probability of occurrence in the eastern sector of the Antarctic Peninsula, Weddell Sea and in the Scotia arc (Fig. 3a; Supplementary Information 5). The decrease in distribution of the species would reach ~ 11% and is expected to occur around the Antarctic continent (Antarctic Peninsula, Weddell Sea and in the eastern sector of the Antarctic continent) and furthermore in the sub-Antarctic region of the Kerguelen Archipelago, projecting to find the species in deep and shallow areas surrounding Kerguelen, Heard and McDonald Islands (Fig. 4a). Under SSP5-8.5 scenario, short-term projections reveal a shift in potential distribution of the species towards the pole and to shallower areas, expanding mainly in the Weddell and Ross seas (Fig. 3b, Table 2).Fig. 3Range shift maps for the projected potential distribution of Neobuccinum eatoni under varying global socioeconomic developments (SSP2-4.5 and SSP5-8.5) in the short term (2050) and long term (2100). Colors depict as follows: pink areas indicate extinction; yellow areas denote stability; blue areas represent expansion. Map was generated using QGIS 3.30 (https://www.qgis.org/).Fig. 4Continuous distribution maps of Neobuccinum eatoni projected for future climate change scenarios SSP2-4.5 and SSP5-8.5 in the short and long term, in the sub-Antarctic region of the Kerguelen Archipelago. Map was generated using QGIS 3.30 (https://www.qgis.org/). Range shift maps for the projected potential distribution of Neobuccinum eatoni under varying global socioeconomic developments (SSP2-4.5 and SSP5-8.5) in the short term (2050) and long term (2100). Colors depict as follows: pink areas indicate extinction; yellow areas denote stability; blue areas represent expansion. Map was generated using QGIS 3.30 (https://www.qgis.org/). Continuous distribution maps of Neobuccinum eatoni projected for future climate change scenarios SSP2-4.5 and SSP5-8.5 in the short and long term, in the sub-Antarctic region of the Kerguelen Archipelago. Map was generated using QGIS 3.30 (https://www.qgis.org/). Long-term projections for the year 2100 under the intermediate emissions scenario (SSP2-4.5) indicate a loss of circum-Antarctic habitat, including the South Sandwich Islands and South Orkney Islands, potentially decreasing the total area of presence by ~ 13% (Figs. 3c). However, specific deeper areas may temporarily become favorable within the Kerguelen Archipelago, driving the species to lower average latitudes and greater depths (Fig. 4c, Table 2). Projections also show an overall shift towards shallower depths and poleward movement under both moderate (SSP2-4.5) and high (SSP5-8.5) emissions scenarios, with notable expansions in the Ross Sea and Weddell Sea by 2100. However, certain areas like the Kerguelen Sea may see deeper habitats temporarily becoming favorable (Fig. 4c, 4d). A decrease in the probability of occurrence is expected in the eastern sector of the Antarctic continent and the Antarctic Peninsula, and in the Weddell Sea, continuing the trend in the South Orkney Islands and the sub-Antarctic islands (Figs. 3b, 4b), reaching a decrease of ~ 14%. Furthermore, under SSP5-8.5 by 2100, the SDMs predict that the species will move to new, shallower areas located mainly in the Ross and Weddell Seas (Fig. 3d, Table 2). In the long term, the reduction of suitable habitat for N. eatoni would reach ~ 21%, with losses concentrated in the eastern part of the Antarctic continent, in the Bellingshausen and Ross Seas, as well as the sub-Antarctic islands (Fig. 3d, 4d). Additionally, significant decreases are expected in the South Orkney Islands and the Scotia Arc. One area predicted to remain mostly stable in terms of species presence is the Antarctic Peninsula.Table 2Bathymetric centroid of the Neobuccinum eatoni occupation range in each niche potential scenario using the binary modelization output from the ensemble forecasting.TimeframeEmissions ScenarioMean bathymetryMean latitudeMin / Max latitudeSpatial trendMain expanse areasPresentPresent−708−65.1−44.9 / −78.72050SSP2-4.5−1091−65.2−44.9 / −78.7deeper polewardWeddell seaScotia arc2050SSP5-8.5−590−65.4−44.9 / −78.7shallower polewardRoss seaWeddell sea2100SSP2-4.5−1065−64−44.9 / −78.7deeper tropical wardWeddell seaKerguelen2100SSP5-8.5−448−65.8−44.9 / −78.7shallowerpolewardWeddell sea Table 2Bathymetric centroid of the Neobuccinum eatoni occupation range in each niche potential scenario using the binary modelization output from the ensemble forecasting.TimeframeEmissions ScenarioMean bathymetryMean latitudeMin / Max latitudeSpatial trendMain expanse areasPresentPresent−708−65.1−44.9 / −78.72050SSP2-4.5−1091−65.2−44.9 / −78.7deeper polewardWeddell seaScotia arc2050SSP5-8.5−590−65.4−44.9 / −78.7shallower polewardRoss seaWeddell sea2100SSP2-4.5−1065−64−44.9 / −78.7deeper tropical wardWeddell seaKerguelen2100SSP5-8.5−448−65.8−44.9 / −78.7shallowerpolewardWeddell sea Bathymetric centroid of the Neobuccinum eatoni occupation range in each niche potential scenario using the binary modelization output from the ensemble forecasting. Weddell sea Scotia arc Ross sea Weddell sea Weddell sea Kerguelen shallower poleward Discussion The joint modeling approach demonstrated high predictive accuracy for the potential distribution of N. eatoni, highlighting the interplay between species’ ranges and its oceanographic environment. Although individual models are accurate, ensemble models were constructed to address emerging uncertainties that may affect potential distribution inferences53. The SDM predictions accurately reflect the well-known circumpolar distribution of N. eatoni, including its presence in sub-Antarctic areas36,54,55. Overall, our results underscore the dynamic changes in bathymetry and latitudinal distribution of N. eatoni. These changes indicate a dynamic response to different climate scenarios, showing both shallowing and deepening trends along with shifts towards higher and slightly lower latitudes. Notably, the model predicts potential habitat availability in the South Georgia Islands, where the species has not been documented. This suggests that predictive modeling identified new potential distribution areas for N. eatoni, guiding future empirical studies to explore these habitats. Factors such as competitive exclusion or dispersal constraints may influence these potential expansions. Barriers like the Antarctic Circumpolar Current, the thermal gradient of the Antarctic Polar Front, and the deep basins in the Drake Passage, which exceed 4,000 m, hinder the colonization and dispersal of benthic fauna, including eurybathic species like N. eatoni56–60. These physical barriers maintain clear biogeographic separation between the biota of the Southern Ocean and lower latitudes. Under the SSP2-4.5 for the short-term (2050), the potential distribution of N. eatoni is projected to expand by approximately 30% at higher latitudes, particularly in the sector of the Scotia Arc and Weddell Sea. However, a decrease of around 11% is anticipated in the northern Antarctic Peninsula and some parts of the edge of the Antarctic continent, as well as in some parts near to the sub-Antarctic region of the Kerguelen Archipelago, where the species is expected to find new habitats in deep and shallow areas surrounding Kerguelen, Heard, and McDonald Islands. Similarly, under the SSP5-8.5 for the same period, projections indicate a shift in potential distribution towards the pole and shallower areas, with expansions primarily in the Weddell and, to a lesser extent, Ross Seas. Despite this, a continued decrease in the probability of occurrence is expected in the eastern Antarctic continent, Antarctic Peninsula, and eastern edge of the Weddell Sea, mirroring the trend observed under SSP2-4.5. Deep water temperature is expected to increase and accumulate on the continental shelf influenced by the current and projected future strengthening of the westerly winds61. This increase in sea temperature would lead to the reduction of ice shelves, exposing new benthic habitats and altering the distribution areas of species inhabiting the seafloor2. Moreover, as temperatures increase, shell-breaking predators like crabs are returning to Antarctica, reversing historical trends of reduced predation pressure on suspension-feeding organisms, which are expected to reduce the endemic character of Antarctic subtidal communities62. Furthermore, various marine organisms, including Antarctic krill, notothenioid fishes, and Adélie penguins, as well as benthic invertebrates such as sponges and corals, may expand their potential distributions to higher latitudes in search of more favorable conditions. This shift is typically driven by adaptations to the extreme cold of Antarctic waters, reliance on ice for life cycle stages, and changes in sediment and nutrient dynamics due to melting ice13,63. All these changes will alter trophic webs, with potential synergistic effects combined with oceanographic changes making N. eatoni highly vulnerable to future biotic and abiotic changes60,64,65. Under the SSP2-4.5 for the long-term (2100), a further loss of approximately 13% of N. eatoni’s circum-Antarctic habitat is projected (Fig. 3c). However, the Weddell Sea and specific deeper areas in the Kerguelen Archipelago are identified as having the potential to become temporarily favorable, indicating some potential for habitat stability at greater depths. In contrast, under the SSP5-8.5 for the same period, the species is expected to face a significant reduction in suitable habitat, reaching approximately 21% loss, with significant decreases expected in the South Orkney Islands and the Scotia Arc. The species is projected to move to shallower areas mainly in the Weddell Seas. Other studies corroborate the projected habitat loss in this region due to the limited resilience of marine fauna to climate change (e.g., the irregular sea urchin Abatus cordatus7,39). The Kerguelen Plateau hosts a vast diversity of species in both pelagic and benthic ecosystems, including one of the largest populations of the commercially important Patagonian toothfish (Dissostichus eleginoides66), making it a critical area for monitoring the impacts of climate change. Also, the sub-Antarctic islands of Heard and McDonald, part of this plateau, are home to vulnerable, slow-growing benthic organisms, endemic fish species, and rich populations of seabirds and mammals67. Subramaniam et al.37 identified key drivers affecting ecosystem dynamics in this area, including increases in sea surface temperature (SST), strengthening of zonal winds, and shifts in the Southern Annular Mode (SAM) towards negative phases. These factors impact nutrient availability, primary productivity, and the structure and composition of benthic marine communities, contributing to a potential decay in N. eatoni sub-Antarctic populations in the long term. In addition, dissolved iron significantly contributes to the models by influencing N. eatoni’s potential distribution. Variation in dissolved iron impacts habitat quality through its effect on marine food web productivity52. In the Southern Ocean, where iron is a limiting factor for phytoplankton growth, reduced iron availability can decrease primary production, resulting in less organic matter reaching the benthic zone where N. eatoni inhabit. This reduction in food supply, coupled with shifts in trophic interactions due to changes in phytoplankton composition, can reduce habitat suitability for N. eatoni, ultimately affecting its distribution. Among the oceanographic variables selected for use in SDMs, the range of seawater salinity at the mean bottom depth is the major contributor in determining the habitat suitability of N. eatoni.Antarctic Bottom Water (AABW) properties are significantly affected by and contribute to climate change, with important salinity changes occurring between the eastern and western Southern Indian Ocean68,69. Seawater salinity is a critical abiotic factor for marine organisms, significantly influencing species physiology and geographic distribution. When salinity experiences drastic alterations from its optimal level, organisms face osmotic stress that negatively impacts their functioning70. The projected variation in salinity for polar regions, due to the interaction among ocean warming, ocean circulation and accelerated melting of sea ice and icebergs71,72, directly influences the distribution of species, especially if they depend on a specific salinity range, forcing them to move to depths where salinity levels are more stable and close to their preferred range, in the case of N. eatoni, models indicate that while rising temperatures may push the species to seek cooler waters, changes in salinity could restrict its ability to inhabit deeper areas. This would lead the species to retreat to shallower zones that align with its physiological limits, explaining the potential contraction of its distribution under extreme future climate scenarios.Experimental freshening conditions have shown that marine gastropods experience high mortality due to anoxia and acidosis70,73. Additionally, reduced salinity due to sea ice melting can negatively affect zooplankton abundance and phytoplankton biomass, altering trophic interactions and community composition69. Future salinity fluctuations would also increase ocean acidification, negatively impacting calcifying organisms by challenging their growth and the integrity of their calcium carbonate structures74,75. The surface ocean pH has decreased by 0.1 units since the industrial era, with further acidification projected by the end of the century, especially in higher latitudes where cold waters allow for greater CaCO3 solubility76–78. This acidification significantly affects mollusks, reducing calcification, growth, and development, particularly in larvae and juveniles79. Research shows that gastropod mollusk embryos, including species like N. eatoni, which lack a larval dispersal stage and have intracapsular metamorphosis, are negatively affected by reduced pH levels, experiencing lower viability, prolonged development times, and reduced locomotion80,81. These effects are compounded by the lack of data on species’ long-term responses to unsaturated environments, where reduced pH decreases carbonate ion availability, which is essential for calcium carbonate structures82. As ocean pH drops, more areas become unsaturated, making it more difficult for organisms to build and maintain their shells and skeletons, leading to potential population declines. In the Southern Ocean, saturation horizons will drastically shoal under explored scenarios. Therefore, the overall impact of pH is likely underestimated. Studies project pH declines of up to 0.36 units by 2100 for the top 200 m, causing severe ocean acidification throughout the water column and ubiquitous end-of-century aragonite undersaturation under the highest emission scenarios83. In conclusion, our study underscores the severe potential distribution shifts expected for N. eatoni under future climate scenarios in the Southern Ocean. We identified several key factors influencing the species’ range: dissolved iron, pH, and temperatures at the surface, and salinity at mid-depth. The interaction of these factors is critical and becomes unfavorable for the species distribution under the very high RF scenario. However, in an intermediate scenario, the probability of suitable habitat may increase in some areas. These findings underscore the complex challenges in conserving marine invertebrates, suggesting that the magnitude of environmental changes will be crucial for their physiological and ecological adaptations. This emphasizes the need of comprehensive conservation initiatives that incorporate less charismatic species, whose ecological roles are essential not only for maintaining ecosystem efficiency but also for supporting ecosystem complexity and resilience. Supplementary Information Supplementary Information. Supplementary Information.