While the climate sensitivity and importance of subsurface fluid and greenhouse gas reservoirs have received attention in the Arctic, the role of Antarctica in global methane emissions and the carbon cycle has largely been overlooked. Significant volumes of fluid and gases exist as subsurface reservoirs in the Antarctic, and geophysical surveys have identified that these intersect with hydrate, permafrost, and glacial ice systems, with connections to the coast1–5. These subglacial fluid reservoirs contain highly saline brines enriched in trace metals, nutrients, and gases, as a result of cryo-concentration through freeze-thaw cycles and extended rock-water interaction times2,6,7. Concurrently, an estimated 21,000 Gigatons of carbon (GtC) exists below Antarctic ice sheets, with up to 400 GtC immobilised as gas hydrates5 and a further 400 GtC of methane in subsurface marine reservoirs8. Also unique to the Antarctic is the potential for a shallow gas hydrate stability zone on the coastal margins9, which is driven by the uniquely persistent cold temperatures of the Southern Ocean. In contrast, gas hydrate reservoirs tend to be at much deeper (> 400 m) water depths in other areas. The existence of shallow gas hydrates in Antarctica could create vulnerability to shorter-term warming or pressure changes in shallow waters along the Antarctic coast, making them uniquely climate sensitive.
Here, we report discoveries of widespread, and recently emergent, seafloor fluid and gas seeps across the coastal Ross Sea region and discuss potential driving mechanisms linked to regional climate impacts. Methane is a powerful greenhouse gas with a warming potential 25 times that of CO2 over a 100-year time-scale10, and methane release from such seafloor seeps has the potential to initiate positive climate feedbacks with global consequences11,12. Equally important, the presence of these seeps indicates subglacial fluid routing that could be transporting key trace elements and nutrients to the ocean2, with potential to alter regional productivity.
Emergent methane seepage has recently been shown in coastal waters around Ross Island, potentially indicating a fundamental shift in regional methane release (Fig. 1). To date, dozens of individual seeps have been discovered, many at locations which have been repeatedly studied for years to decades with seeps not previously observed. The first seep (Cinder Cones, 10 m deep; Fig. S1), was initially observed in 2011 at a location that had been repeatedly visited by SCUBA diving surveys since 19678. Since 2011, and to the time of writing, this seep has been continuously releasing methane, with methane flux measurements in 2022 averaging 0.8 ± 0.3 mmol/m2/d. Another seep site at 7 m depth at Cinder Cones was discovered in 20168 and has similarly continuous methane release with measurements in 2022 revealing 5.0 ± 0.8 mmol/m2/d. In background (non-seep) sediments at both sites, methane concentrations were below detection limits. These seeps are visually recognisable through the presence of distinctive microbial mats (Fig. 1).
Since these initial discoveries, similar seep sites have emerged in the McMurdo Sound region at other routinely visited locations, including Cape Evans, Turtle Rock, Granite Harbour, Cape Barne and Dunlop Island (Fig. 1; S1), again recognisable by the presence of microbial mats. Benthic flux measurements from a microbial mat at Turtle Rock (Fig S1) in 2022 revealed a methane flux of 8 mmol/m2/d, again with background (non-seep) fluxes below detection limit. Microbial mats identified at Dunlop Island in 2021, were associated with seafloor pockmarks, potentially indicating fluid or gas expulsion13 (Fig. 1C). Intriguingly, video obtained using a Remotely Operated Vehicle (ROV) at a seep site identified at Cape Evans in October 2023 show gas bubbles escaping the seafloor and accumulating on the under-side of sea-ice above the seep site (Fig. 1B; S2; Supplemental Video 1) highlighting the potential for rapid transfer to the atmosphere.
Simultaneous with these increasing discoveries of emergent seepage at many routinely visited locations in the McMurdo Sound region, we have identified > 28 fluid and bubble plumes in 2021 and 2023 using shipboard acoustic techniques along the Northern Victoria Land coast (Fig. 1A; 1D; S5-S9). These features were discovered between 40 m (the shallowest extent of shipboard surveys) and 240 m water depth, with surveys in deeper waters not revealing any seep features. The bubble plumes were temporally variable ‘bubble bursts’ best observed at lower frequencies (< 38 kHz) indicative of bubbles > 1 mm in size and were first detected in 2023 (Fig. 1D; S8-S10). Camera surveys found the bubble plumes were associated with small white microbial mats, consistent with our observations from the shallower sites further south in McMurdo Sound (Fig. 1A-C) and also shallow marine gas seeps in other regions14,15.
The fluid plumes observed rise from the seafloor to approximately mid-water column height before becoming entrained presumably in prevailing currents (Fig. 1D; S5-S9). Features off Cape Hallet Peninsula gave a similar reflectivity to subglacial discharge measured from the nearby Tucker Glacier (Fig. S8) and to signals obtained from stratified turbulences in other global regions16–18. We consistently found metazoan dead zones at the centre of fluid seeps, with transition zones dominated by opportunistic mega-epifaunal taxa that were relatively rare in the surrounding biologically-rich area at these sites (Fig. 1D; S3; S4). Repeat transects in 2021 and 2023 at a fluid plume site at Cape Hallett indicate continued fluid seepage over this time-period (Fig. S3; S7)
Seeps can have significant local impacts on the overall function of marine ecosystems and shift larger-scale biogeochemical cycles, although some mechanisms remain unclear. In addition to the biological dead-zones at the fluid plume sites in Northern Victoria Land, the Ross Island seep sites have seen an outbreak of sea star wasting disease, potentially related to subdermal hypoxia or organic matter enrichment on sea star health19. Indeed, studies at global seep sites have shown benthic oxygen demand that is by two orders of magnitude greater than non-seep sites20. Seafloor seeps on the Arctic shelf are not only recognised sites of oxygen depletion20, but also ocean acidification21 and sources of dissolved iron22 that may enhance primary production23. While seafloor seeps are known to influence benthic systems and biogeochemistry at seep sites globally20,22–25, this is the first evidence of an impact on benthic fauna in the Antarctic.
The parallel discovery of seeps in multiple locations in the coastal Ross Sea in a relatively short period of time is suggestive of a shifting underlying driver of subsurface fluid and gas flux in the region. Known mechanisms determining seafloor seepage can span the land-sea continuum, connecting dynamic processes across cryosphere-lithosphere-ocean systems. We hypothesise that the emergent seafloor seeps reported here may be a response to cryospheric cap degradation (Fig. 2), or the the melting of ice, permafrost, and gas hydrates, which form ice-bound barriers that effectively trap climate-reactive gases and nutrient-rich fluids in subsurface reservoirs at high latitudes.
As rising atmospheric temperature enhances glacial ice melt in the terrestrial realm, this reduces the weight of ice sheets and glaciers. This can promote the degradation of subglacial cryospheric caps, thereby increasing the permeability of the subsurface environment and enhancing subglacial flux and subsequent fluid or gas discharge at the coast26–29 (Fig. 2). Similarly, in the marine realm, the reduction in ice mass weight can result in isostatic rebound outpacing eustatic sea-level rise and lowering coastal sea-level; this then decreases hydrostatic pressure and contributes to dissociation of hydrate deposits30,31. Impacts in the marine realm may also be further intensified by the local effect of ocean warming.
As demonstrated in the geologic record, the cryospheric cap system is highly sensitive to climate warming28,29. In the last 30,000 years, warm water incursion and isostatic rebound in the Arctic has led to episodic methane release30,32. In the mid-Pleistocene transition, methane was released from the seafloor during episodes of climate warming33, with methane release during the most recent glacial terminations correlated with loss of ice sheet mass33,34. Abrupt release of gas from permafrost and hydrate destabilisation in response to past ice-sheet retreat is also evident as seafloor depressions (pockmarks) in both the Antarctic and Arctic35,36, as was seen at Dunlop Island in the present day (Fig. 1C).
A focus of global efforts in polar research is to quantify the loss of ice sheets and shelves, measuring present-day cryospheric cap degradation in response to anthropogenic climate change. Widespread reduction in ice mass in the Antarctic includes the ice sheets and shelves in the Ross Sea regions studied here37,38, providing a mechanism for onset of subterranean fluid release. If the presence and emergence of these seafloor seeps, now observed throughout this region, is regulated by cryospheric cap degradation, as we hypothesise, there may be profound local and potentially global implications as the climate continues to rapidly warm. If shallow water methane release in the Ross Sea is indicative of the wider Antarctic continent, there is the potential for rapid transfer of methane to the atmosphere, as reported at other shallow global seep systems39–41. Regardless of the formative mechanism, these novel yet widespread discoveries of Antarctic seafloor seeps highlight the significant gaps of understanding in what appears to be an increasingly common phenomenon. We recommend coordinated, international research efforts to address this important topic.