Introduction

The atmosphere and seawater are facing an increase in carbon dioxide (CO2) concentrations. Global carbon analyses have reported that atmospheric CO2 levels increase from 280 parts per million (ppm) in preindustrial times to 400 ppm during the Anthropocene, reaching a peak of 425 ppm in February 20241. This CO2 excess is primarily introduced by human emissions, mainly fossil fuel combustion, industrial processes, and land-use changes2. This increase in CO2 is identified as the major cause of global warming, which has been buffered by oceans, which capture and store ~26% of anthropogenic emissions. However, ocean CO2 uptake is not uniform and has two consequences. High-latitude, cold-water oceans, where CO2 solubility is higher, captured more CO2 than temperate and low-latitude waters. While ocean CO2 uptake provides negative feedback to the climate change process, it also leads to a decrease in ocean pH, known as ocean acidification3.

CO2 is a major greenhouse gas which contributes to climate change by causing positive radiative forcing and warming the atmosphere. As a result, the storage of carbon in ecosystems has become increasingly important. It has been classified both as an ecosystem service and, in a more inclusive definition, as nature contribution to people (NCP)4. The biological pump plays a crucial role in the ocean carbon cycle, as it facilitates the influx of CO2 into the ocean through biological processes. This process begins with photosynthesis, after which the products are transported from the surface to deeper waters via a combination of sinking particles and animal activity. Different terms are associated with the different phases according to how long carbon will be stored. The carbon in phytoplankton and benthic algae biomass represents fixed carbon, which remains there for a short period (hours to months). In contrast, the carbon in benthic animal biomass is considered to be stored carbon that persists for a longer duration (months to years). Finally, the fraction of carbon that is buried in anoxic sediments (preventing aerobic respiration and thereby decreasing the remineralization rate) or reaches a depth of approximately 1000 m is considered sequestered carbon because it has been removed from the carbon cycle for at least 100 years5. This long-term storage of carbon has an important role in the global carbon cycle and climate change mitigation, as the ocean sequesters one-fourth to one-third of the carbon released by human activities. Therefore, estimating the carbon storage capacity of ecosystems is pivotal for ecosystem management, assessing the NCP, and achieving climate change mitigation goals.

The term ‘blue carbon’ was coined in 2009 to refer to organic carbon (OC) stored in the biomass and sediments of marine ecosystems. Although blue carbon assessments have principally focused on OC, more recently, inorganic carbon (IC) has started to be considered due to the increasing number of studies acknowledging the relevance of this fraction in carbon budget assessment6. Vegetated and coastal ecosystems, such as mangrove forests, seagrass meadows, and salt marshes, are particularly recognized as carbon sinks and are therefore considered as Blue Carbon Ecosystems (BCE). This recognition is due to their capacity and effectiveness in storing carbon produced within the ecosystem and from other sources7. This process primarily results from the canopies and complex root systems that alter the water flow above them, acting as a filter and stabilizing the sediment, preventing resuspension8. These sediments, which can be several meters deep, are often anaerobic, preventing remineralization and allowing OC to be sequestered for hundreds of years. However, coastal ecosystems with high carbon stocks face numerous threats due to population expansion, climate change, and industrial development, diminishing their effectiveness by reducing their surface area and releasing CO2 into the atmosphere9.

In the last decade, the blue carbon concept has been extended to include carbon fixation, storage, and sequestration by marine ecosystems, i.e. the biologically mediated carbon of marine ecosystems10. New systems such as kelp forests or deeper and colder ecosystems have also been proposed as BCEs, e.g. Antarctic benthic ecosystems11,12. Moreover, cold-water coral reefs and sponge grounds have been identified as potentially important ecosystems for carbon storage and sequestration13. Unlike the vegetated ecosystems previously mentioned, Antarctic benthic species might colonize new habitats due to the opening of new areas for new production, caused by ice losses, ice shelf collapses, and glacier retreat14. Thus, fixing, storing, and finally sequestering more carbon is crucial for providing negative feedback to climate change15. The protection of these mentioned high-latitude ecosystems is necessary because, although polar ecosystems have slower storage rates than tropical coastal ecosystems, the total area occupied by the former is larger, as coastal vegetated ecosystems cover only < 2% of the ocean surface16.

As CO2 enters seawater due to a partial pressure gradient at the atmospheric-ocean interface, the pH of the water decreases. Since preindustrial times, the ocean pH has decreased by 0.1 units, with a further decrease of 0.3–0.4 units predicted by the end of the century17,18. Ocean acidification also alters carbonate chemistry in seawater because an increase in CO2 causes a decrease in the CO3–2 concentration and, therefore, in the carbonate saturation state. The depth of the saturation horizons of calcite and aragonite, the most common carbonate forms in calcifying marine organisms and sediments, represents the boundary between precipitation (supersaturated water shallower than the saturation horizon) and dissolution (undersaturated water deeper than the saturation horizon) of CaCO3. Consequently, the saturation horizons are shoaling, which could decrease the rate of CaCO3 biogenic formation and promote dissolution. This is more evident in cold waters because of their higher CO2 solubility3.

Sedimentary carbonates are the major carbon reservoirs on Earth. During CaCO3 precipitation, two moles of HCO3 are taken up, and one mole of CO2 is released into the surrounding seawater. Conversely, as CaCO3 dissolves, the opposite reaction occurs. Consequently, precipitation decreases total alkalinity and dissolved inorganic carbon by a 2:1 ratio, leading to changes in seawater chemistry19. In other words, precipitation and dissolution could act as CO2 sources or sinks to the environment, respectively. Although the relationship between released CO2 and precipitated CaCO3 (ψ) depends on the saturation state, seawater conditions (e.g. temperature) and CO2 atmospheric concentration, 0.6 is the most commonly used value in seawater studies20. Despite the possible consequences of carbonate stock fate, there is no consensus on how carbonate stocks should be considered, and consequently, carbonate stocks have been largely overlooked in carbon budget assessments across different ecosystems21.

Marine protected areas (MPAs) are designed and managed with different aims that in general protect and conserve from charismatic or single species to entire ecosystems and their ecosystem services. Recently, there has been increasing interest in the role of MPAs as a tool to address the impacts of climate change, e.g. through maintaining the carbon storage and sequestration as a regulating NCPs22. Biodiversity loss and climate change are closely associated, and a broader approach is needed to address both threats. Identifying and subsequently protecting areas with high carbon content could prevent losses of carbon storage. For instance, protecting animals with a significant role in carbon fluxes, such as suspension feeders that connect pelagic and benthic environments, can contribute to these goals23.

Before the term “blue carbon” emerged, global carbon budgets focused on land ecosystems. Despite increased attention on stored carbon in oceans, only coastal and vegetated ecosystems have been considered, representing only a small fraction of the ocean surface. There have been few investigations in offshore ecosystems, which are deeper and/or colder than traditional BCEs. Moreover, there is missing information on IC stocks in both coastal and non-coastal ecosystems. Estimating the carbon content could underscore the value of these ecosystems to society and support their conservation. Designing MPAs based on such estimations could help prevent the release of carbon stored in organisms and sediments. For appropriate management of stored carbon, it is crucial to understand where it occurs and which processes promote it. This study aimed to assess carbon storage, as an ecosystem service, in both benthic assemblages and sediments of an open-sea MPA located in sub-Antarctic waters and in two nearby unprotected areas. We also included the inorganic carbon fraction in our evaluation and concluded that due to the release of CO2 during the formation of biogenic inorganic carbon and the high amount of CaCO3 found in the study areas, it is crucial to address this fraction in carbon assessments.

Methods

Sampling area

Carbon stocks were estimated for both benthic invertebrate biomass and sediments in the Namuncurá – Burdwood Bank I and II Marine Protected Areas in the Argentinean sub-Antarctic region (here referred to as NBB-I and NBB-II, respectively, or collectively referred to as BB) and neighbouring areas: the Beagle Channel (BC) and the Atlantic Shelf of Tierra del Fuego (AS) (Fig. 1). The main goal of MPA creation is benthic ecosystem protection (Argentine Laws 26,875 and 27,490). Benthic assemblages in these areas comprise a diverse range of long-lived organisms, including sponges, corals, and bryozoans, constituting animal forests and vulnerable marine ecosystems24,25. NBB-I and NBB-II were established in 2013 and 2018, respectively, but have been managed as a single MPA since 2019. The Burdwood Bank is a submerged plateau that is part of the northern Scotia Arc. NBB-I is bounded by the 200 m isobath, covering approximately 28,000 km2, and NBB-II is located on the southern slope of NBB-I and comprises approximately 32,300 km2 (a total area of ~ 60,300 km2 for both). Benthic assemblages on the plateau (NBB-I) are dominated by sponges and bryozoans, and those on the southern slope (NBB-II) are characterized by hard bottom dwellers such as cnidarian alcyonaceans26. The sampled area is mostly influenced by the Antarctic Circumpolar Current (ACC), which splits into two branches upon encountering the BB, wrapping around the edges of the bank. The BB is an active centre for the obduction of deep, fertile waters to the surface layers of the Drake Passage27.

Fig. 1
figure 1

Areas and sampled locations: Beagle Channel (BC), Atlantic Shelf of Tierra del Fuego (AS) and Marine Protected Area Namuncurá – Burdwood Bank I and Burdwood Bank II (BB). The different colours indicate the origins of the samples used for the different analyses. IA isotopic analysis, LOI loss on ignition method.

Full size image

The BC is a coastal zone with estuarine features that connects the Pacific and Atlantic Oceans and is influenced by glacial melting, river runoff and the Cape Horn Current28. Benthic assemblages at depths similar to those in our study are represented mainly by asteroids, decapods, barnacles, bivalves, and bryozoans29. Finally, the AS is characterized by a gentle slope and bottoms of coarse sand, where sponges and crustaceans are the most abundant reported taxa; AS is affected by the diluted waters of the Cape Horn Current26.

Sampling was conducted during research cruises aboard the RV “ARA Puerto Deseado” in December 2016 and August–September 2018 and aboard the RV “BIPO Víctor Angelescu” in October–November 2019 and November 2019. Benthic assemblage samples were obtained at depths ranging from 90 to 710 m in the BB, 95 to 115 m in the AS, and 80 to 223 m in the BC. Sediment samples were obtained at depths between 90 and 1000 m, 75 and 115 m and 15 and 223 m, respectively. Benthic assemblage samples were collected using a bottom trawl equipped with a rectangular otter-board trawl measuring 6 m in total length, with a headrope and footrope length of 6 m each, a wing mesh of 25 mm, a cod-end mesh of 10 mm, a horizontal opening of 1.8 m, and a vertical opening of 0.60 m. Sediment samples were obtained from the top layer of the sea-bottom by using a benthic grab sampler with an approximate volume of 400 cm3. All samples (organisms and sediment) were stored at − 20 °C until analysis.

Sampling procedure and methods to estimate carbon

We employed two methods to analyse the carbon content of the samples:

  1. (1)

    Isotopic Analysis was used solely for determining the carbon stock in the sediments to estimate the organic and inorganic carbon contents. A total of 33 sites were sampled (7 in BC, 4 in AS and 22 in BB) by using a grab sampler. As this sampler does not retain sediment stratification, it was assumed that the samples represented the 10-cm upper layer of the seabed. Two subsamples were used to distinguish between OC and IC. For OC determination, carbonates were removed by HCl fumigation30. During this procedure, each sediment sample was oven-dried at 60 °C for 48 h, and ~ 40 mg of dried sediment was placed in Eppendorf tubes and moistened with distilled water (~ 75 µL). These samples were then exposed to HCl (12 N) vapour for 12 h in a closed glass desiccator cabinet. After acid fumigation, each sample was dried again at 60 °C for 4 h. The second half of the sample was used to estimate the total carbon content and was only oven-dried at 60 °C for 48 h. The acidified and nonacidified samples were then weighed in tin capsules. The weighted percentage carbon concentration for each sample was analysed with a Thermo Scientific DELTA V Advantage IRMS coupled with an elemental analyser at “Laboratorio de Isótopos Estables en Ciencias Ambientales” (LIECA-CONICET, Argentina) and at the Center for Stable Isotopes (University of New Mexico, USA). The IC was calculated as the difference between the total percentage carbon concentration determined in the nonacidified sample and the OC determined in the acidified sample.

  2. (2)

    Loss on ignition method was used to estimate the organic matter and inorganic carbon contents in the biomass samples and sediments. For biomass samples, a total of 30 sites were sampled to estimate macrozoobenthic carbon standing stocks (5 in BC, 2 in AS and 23 in BB) (Fig. 1). The bottom trawl was deployed and kept on the seabed for 5 to 10 min and was towed at approximately 2.5 knots. All organisms were distinguished and identified to the lowest possible taxonomic level based on macroscopic and/or microscopic morphological features, and the wet mass of each taxon was measured. A total of 1205 samples of macrozoobenthic organisms were processed. In the first burning, the sample with a known weight was heated at 450–500 °C for 5 h to estimate the organic matter content as the difference between the initial and final masses, which was 0.00001 g. In the second burning, the sample was reheated between 900 and 950 °C for 5 h to estimate the IC fraction by calculating the difference in weights before and after burning to eliminate CO2 from carbonates31. The organic matter content in the biomass samples was converted into OC by a factor of 0.532.

For sediment carbon stock estimation, a total of 24 sites were sampled (9 in BC, 3 in AS, and 12 in BB) with a grab sampler, as for IA (Fig. 1). Some of these sites were the same as those in which we estimated the biomass carbon stock and the same in which we performed IA in sediments. Approximately 30 g dry mass of each sediment sample was used and dried at 60 °C until a constant weight was reached. A conversion coefficient was calculated for each area (BC, AS and BB) based on the relationship between the percent organic matter and OC from samples processed by both the LOI and IA methods. This coefficient was used to infer the OC content for those samples solely analysed by the LOI (0.25 for BC, 0.49 for AS and 0.27 for BB). Dry bulk density (g cm–3) was determined at 14 sites (3 in BC, 3 in AS and 8 in BB) to calculate the carbon stock (kg m–2) by dividing the dry mass of the sample by its volume after drying at 60 °C. Carbon contents were converted from percentage to stock by multiplying the carbon content (percent dry weight) by the dry bulk density. The average dry bulk density in each area was used to estimate the remaining carbon content. Since sediments were mostly composed by rest and debris of benthic organisms (corals, sponges, bryozoans, molluscs, etc.), we assumed that carbonate sediment stocks had a biogenic origin, with a negligible lithogenic fraction.

To estimate and compare the total Carbon storage, the mean value of the carbon stock in BB, both in sediments and biomass, was scaled up by the total area of the MPA Namuncurá-Burdwood Bank.

We estimated the carbon content in different ways to compare and discuss how carbonates should be considered part of the blue carbon. First, we used formulas according to different previous studies15,33,34,35,36. Second, we propose a new equation that considers the carbonates and the CO2 released into the ocean as a consequence of carbonate formation. Since the ratio between released CO2 and precipitated CaCO3 (ψ) is variable throughout the ocean, we used a coefficient ψ = 0.8 for the three sampled areas20. We used the appendixes corresponding to the South Atlantic, South Pacific and Antarctic Oceans, and we chose a ψ value by approximation to the average water temperature at our sampling sites.

Statistical analysis

In the sediment samples, we performed one-way analysis of variance (ANOVA) to compare the carbon stocks and carbon contents among the areas (both organic and inorganic carbon). Since the sample size differed between areas and to avoid differences in unbalanced samples, we performed repeated comparisons by randomly selecting a pool of samples to obtain a balanced comparison, which was then replaced and repeated until all the possible combinations were used. We considered that comparisons were different when > 50% of the trials yielded significant differences. We also compared the OC and IC stocks inside the areas. For the BC and BB data, we applied a fourth root transformation to achieve homogeneity of variance, whereas Welch’s t test (T statistic) was used for the AS data. The homogeneity of variances was tested using Levene’s test at a confidence level of 0.05.

To test for differences in organic and inorganic carbon stocks in biomass among BB and BC (AS was excluded since n = 2) and between IC and OC content inside the areas, we also used ANOVA following the same criteria recently mentioned. We applied a transformation of the fourth root of the BC data to achieve homogeneity of variance. The SciPy Python library was used for all mentioned analyses. Data manipulation and visualization were carried out using the NumPy, Pandas, Matplotlib and Seaborn libraries.

Results

Carbon stock in sediments

The percentage of inorganic carbon was similar between the two different methods of estimation (LOI and IA; ANOVA; F = 0.38; p = 0.54; Fig. 2A); therefore, the data were pooled. The percentages of organic matter (OM), OC, and IC were similar among the three areas (p > 0.05, FBC = 2.52, FAS = 1.17 and FBB = 2.04, respectively), as was the percentage of carbonates between samples analysed by different methods (IA and LOI) (p = 0.54, F = 0.38) (Fig. 2). Moreover, the total carbon stock (organic and inorganic) was similar among the three areas (p > 0.05, F = 1.78) or between the OC and IC stocks in the AS (p > 0.05, t = -0.26) and BC (p > 0.05, H = 0.28) (Fig. 3). However, the IC stock was greater than the OC stock in BB (p < 0.05, F = 39.19) (Fig. 3). By expanding our carbon stock estimations to the total area of BB, these MPAs are storing 933,258,336 ± 394,428,631 Mg C (± s.d.) (188,089,629 ± 82,397,598 Mg of OC and 745,168,707 ± 312,031,033 Mg of IC) and 325,200,671 ± 139,811,308 Mg C (188,089,629 ± 82,397,598 Mg of OC and 137,111,042 ± 57,413,710 Mg of IC) considering CO2 production through CaCO3 precipitation.

Fig. 2
figure 2

Carbon contents in sediments. (A) Percentage of inorganic carbon of dry mass: only samples analysed by both LOI and IA are shown. (B) percentage of organic matter of dry mass by the LOI method; (C) percentage of organic carbon of dry mass by the IA method; diamonds are outliers.

Full size image
Fig. 3
figure 3

Carbon stock in sediments. (A) organic carbon and (B) inorganic carbon stock (n = 10 in BC, n = 4 in AS, n = 26 in BB). Diamonds are outliers.

Full size image

Carbon stock in biomass

The stock of carbon, both OC and IC, was greater at BB than at BC and AS. Both OC (p < 0.05, F = 14.59) and IC (p < 0.05, F = 12.02) storage in biomass was greater in BB than in BC (Fig. 4A and B). Moreover, we observed significant differences between the OC and IC stocks in BB (p < 0.05, F = 13.29) but not in BC (p > 0.05, F = 1.46). Considering the total area of BB, these MPAs stored 52,085.78 ± 43,539.70 Mg C (± s.d.) (34,964.16 ± 30,448.54 Mg of OC and 17,121.62 ± 13,091.16 Mg of IC) in benthic assemblages, which decreased to 38,114.53 ± 32,857.31 Mg C (34,964.16 ± 30,448.54 Mg of OC and 3150.37 ± 2408.77 Mg of IC) by subtracting CO2 production through CaCO3 precipitation.

Fig. 4
figure 4

Carbon stock in biomass. Comparisons between BB and BC. AS was included as a point because of the small sample size, n = 2). Diamonds are outliers. Stars represent the raw data.

Full size image

The different phyla contributed differently to the OC and IC stocks in terms of biomass across the three sampled areas due to variations in their abundances (phylum distributions) and proportions of stored carbon (Table 1). Among the BB assemblages, Porifera was the phylum with the highest OC stock (42%), predominantly on the plateau. The category ‘broken pieces’ (mainly live fragments of Bryozoa and Porifera) stored the most IC stock (43%), followed by Echinodermata, ~ 16%, of which 68% were Ophiuroidea, 16% were Echinoidea and 15% were Asteroidea; and Cnidaria, ~ 15%, of which 62% were Anthozoa, and 32% were Hydrozoa (Fig. 5A and B). However, two zones with different OC stock contributions of the phyla can be distinguished in the BB: the plateau (in NBB-I) was dominated by Porifera, while the upper slope (in NBB-II) was dominated by Echinodermata (Fig. 5A and C). On the other hand, in the AS, the majority of the OC and IC were stored by Echinodermata (38% and 43%, respectively). Chordata (class Ascidiacea) accounted for 36% of the OC, whereas Echinodermata accounted for 43% of the IC. In the BC, the phyla that stored the most OC and IC were Arthropoda, represented only by Crustacea and mainly by the squat lobster Grimothea gregaria (~ 42% OC), and Bryozoa (~ 55% IC) (Fig. 5A–D).

Table 1 Percentages of OC or IC for each phylum (± s.d.).
Full size table
Fig. 5
figure 5

Organic and inorganic carbon per phylum. Panels (A) and (B) show the relative percentages of organic and inorganic carbon per phylum across the four locations. Panels (C) and (D) depict the stock of carbon at each sampling station arranged by location. BC Beagle Channel, AS Atlantic coast, S: BB slope, P: BB plateau. “Broken pieces” refers to parts of largely live animals fragmented by trawl, mainly sponges and bryozoans.

Full size image

Carbon stock in biomass and sediments

When considering different approaches to calculate the IC fraction, the variation in carbon stock was an order of magnitude greater for benthic assemblages in the BC, while minor variations were observed for the BB assemblages (Table 2). Furthermore, a variation of one order of magnitude was also observed in the sediments of both BC and BB, with sediments even serving as a source of carbon (Table 2, equation (III)).

Table 2 Comparisons of total carbon stored between areas according to different approaches, considering (or not considering) inorganic carbon.
Full size table

Discussion

Our results show the important role of three sub-Antarctic areas, the MPAN-BB, the BC and the AS, in carbon storage, with higher carbon content in sediments than in the benthic biomass and, therefore, highlighting their value as a regulating NCP. Despite the differences between traditional BCEs (mangroves, salt marshes and seagrasses), which have been noted for their potential to retain large amounts of carbon, and benthic ecosystems in the BC and BB (e.g. depth, primary production, latitudinal location, etc.), both are important for carbon storage and could be potential hotspots for carbon sequestration. The total carbon storage per unit area in the sediments of the studied areas was similar to that reported in some BCEs as mangroves and higher than those reported in other BCEs as salt marshes or seagrasses, and it was also higher than those registered in some Antarctic areas as Weddell and Amundsen seas (see Table 3). However, major differences were detected between the OC and IC fractions. Our studied sub-Antarctic systems presented much greater amounts of inorganic carbon and consequently lower amounts of organic carbon than those reported in the BCE, including southern Atlantic salt marshes7,37,38,39. High carbonates contents in sediments may correspond to the high carbonate content in benthic taxa. This pattern is similar to that observed in Antarctic ecosystems, where appear to be a relation between macrozoobenthic stocks and carbonate contents in sediments, given the reduced biomass of phytoplanktonic calcifiers40,41,42. This can also point to a higher remineralization rate than in compared ecosystems.

Table 3 Total carbon stock of blue carbon ecosystems according to their area and comparison with the total carbon stock in BB.
Full size table

Compared to the Antarctic locations, the BB and the BC showed similar organic matter (OM) contents in sediments to those reported in the Bellingshausen Sea and higher organic matter contents than those registered in the Bransfield Strait and three fjords in the West Antarctic Peninsula11,50. This comparison is based on OM instead of carbon because carbon values were not available in those studies. Since there are no direct measurements of vertical sedimentation rates for the studied areas, we can hypothesize that the BC might present higher sedimentation rates due to its coastal, glacial, and riverine inputs than the other location, particularly the oceanic BB. According to a modelling projection, sedimentation rates in BB could be comparable to those in off shore Antarctic areas around the Antarctic Peninsula51. In contrast, sedimentation rates in the BC could be higher than in Antarctic fjord areas. Based on these assumptions, comparisons of accumulated carbon stocks in these locations would represent similar periods of carbon52,53,54. Unlike benthic invertebrate biomasses, which generally consist of approximately 50% of carbon in the OM, the proportion of carbon in sediments is more variable. Our studied areas presented different carbon contents in the OM of sediments (BB ~ 8%, AS ~ 38% and BC ~ 15%). These values are comparable to the C:OM fraction reported for mangroves and sea grasses, where the carbon in the OM was reportedly 14% and 4.2%, respectively7,55. In Antarctic fjords, an estimated 20% of the carbon in the OM has been reported11,55,56.

In our study area, the biomass carbon storage was lower than that reported for BCE carbon stocks7,57,58,59,60 (see Table 3). Nevertheless, the benthic biomass was not evenly distributed among sites and areas, presenting marked differences among samples at our three studied sites (Fig. 5). For instance, in the coastal areas of the BC, the forests of the giant kelp Macrocystis pyrifera yield high organic carbon content and high abundances of associated fauna, whereas some other areas of the BC presented very low biomasses61. A similar pattern of marked differences in carbon content was recorded in the BB, with stations differing by more than an order of magnitude in carbon storage (Fig. 5C). However, the benthic assemblages of these sub-Antarctic areas still have lower carbon contents than those of the BCE. However, there are also other factors to consider that could represent advantages for long-term carbon storage in these high-latitude areas, especially the oceanic MPAN-BB. On the one hand, carbon contained in deeper benthic organisms is likely to be isolated from the ocean–atmosphere carbon cycle for longer time periods. On the other hand, BCEs are exposed to more risks, especially anthropogenic activities that threaten entire ecosystems and therefore carbon stocks, which are not present or are at least reduced, at the moment, in our sampled areas, especially BB, which are MPAs with a high protection status62.

Although the carbon content in the benthic biomass in the three sampled areas was lower than that in BCEs and some high latitude areas as the Barents Sea (from 3 to 20 g/m2)36, it was similar to the levels reported in other high-latitude regions (see Table 3). In the BB, benthic assemblages had a higher carbon content (0.8 g/m2, Fig. 4) than some near Antarctic locations, such as the Bellingshausen Sea (0.7 g/m2), the West Ross Sea shelf (0.68 g/m2) or even continental shelves at South Georgia (0.00017 kg/m2), but lower than others, such as in the South Orkneys Islands (2 g/m2) and the Weddell Sea shelf (1 g/m2)33,34. The amount of carbon per unit area is lower compared to typical BCE; however, the total carbon stored in these Antarctic and sub-Antarctic areas becomes substantial when considering their vast extent. Therefore, several of these areas have been designated or proposed as Marine Protected Areas aimed at conserving biodiversity and ecosystem services. The South Orkney Islands southern shelf MPA, established in 2009, covers 94,000 km2. Benthic assemblages in this area were identified as a carbon immobilization hotspot, storing 119,380 Mg C34. The South Georgia and the South Sandwich Islands MPA spans 307,517 km2 that store 52,277 Mg C63. Additionally, the Ross Sea MPA, with a coverage of 1.2 million km2 at its General Protection zone, store 816,000 Mg C33. Although MPAs are shielded from direct human impacts such as fishing and mineral extraction, they remain highly susceptible to indirect anthropogenic threats, including climate change and ocean acidification. This vulnerability is exacerbated by the generally slow rate of biological processes and lower temperatures in these regions, which reduce carbonate saturation states. Consequently, these systems are more prone to the effects of ocean acidification and climate change compared to their counterparts at lower latitudes.

In Antarctica, the high biomass of benthic organisms is related to high phytoplankton blooms64. In the MPAN-BB surface, phytoplanktonic production is low; however, a high zooplankton abundance in spring and summer, together with high benthic secondary production, suggests the presence of allochthonous energy sources40. However, carbon sources from the tychoplanctonic diatom Rhizosolenia crassa should not be disregarded, as they may constitute 98% of the phytoplanktonic subsuperficial biomass of 120 µg Chlorophyl L–141. On the other hand, in Antarctica, the benthic assemblages at > 1000 m showed high variability in the IC stock: in the eastern Weddell Sea and at the tip of the western Antarctic Peninsula (WAP), the IC in the benthic assemblages was one order of magnitude greater than our estimates for the BB and the BC. Notwithstanding, our figures of IC of benthic biomass were similar to those from the WAP south of 64 °S42. These estimations in Antarctica have not corrected for the CO2 resulting from the production of CaCO3, and thus, they may be overestimated.

For many years, blue carbon studies across different ecosystems have neglected or did not consider the inorganic carbon fraction. Only recently have carbonates entered the scene of blue carbon assessments. This could be due, at least in part, to the inherent difficulties in assessing straightforward ways to properly include carbonates in carbon estimations. An appropriate IC estimation could involve considering one atom of carbon in a CaCO3 molecule and then proceeding with the calculation, where 12% of the CaCO3 mass is C. Finally, this can be estimated as the net uptake of atmospheric CO2. This relation has been calculated and reported in many studies15,36,42. However, biogenic CaCO3 production and IC chemistry in seawater are much more complex. There is still a debate about the net balance of carbon storage and how to consider carbonates in carbon stock biogenic CaCO3 production, including carbon capture and the reduction in CO2 solubility in sea water that causes the liberation of CO2 to the atmosphere (Table 2). As a general estimation, one mole of precipitated CaCO3 will release 0.6 mol of CO2 to the atmosphere. However, this proportion is dependent on many factors, such as temperature, which has an important impact on the carbonate saturation state65. Particularly in the three sampled areas, the factor used should be 0.8. This implies that for each precipitated molecule of CaCO3, 0.8% of carbon is released into the atmosphere20. On the other hand, CaCO3 dissolution has the opposite potential, with the sea acting as a CO2 sink in a reverse relation.

Despite its potential to offset the role of the BCE as a carbon reservoir, inorganic carbon has been overlooked66. Many studies on carbon stocks in the BCE have focused only on organic carbon, but few studies that have considered carbonate content have reported high carbonate percentages, similar to our data. In seagrasses, values of carbonate in sediments could suggest even a net production of CO2 (from ~ 30 to ~ 50% of dry mass)35,44,67,68. Moreover, the mean carbonate content in mangrove forests and tidal marshes was also high (from ~ 13 to ~ 38%), and some of them even grew over carbonate banks21,44,69. Since carbonate production releases CO2, it should be subtracted from the total carbon stock, and studies that included IC into carbon stocks used different calculation methods on the basis of different approaches (Table 2). In terms of total stored carbon, the most negative assessment is directly subtracting the IC from the total carbon estimation35. Carbonates are considered exclusively a source of atmospheric CO2. In contrast, some studies have highlighted the potential advantages that carbonates can provide to carbon storage and sequestration, and these advantages should also be considered. For instance, carbonates in living organisms provide protection to organic carbon, a function that is consistent across calcifying taxa, facilitating organic carbon storage for a longer period. Therefore, these studies considered the IC (12% of CaCO3 mass) together with the OC in the total carbon assessment15,36. In sediments, carbonates can also protect particulate OC by adsorbing OC to mineral surfaces. Carbonates contain and preserve organic matter through both intracrystalline and nonintracrystalline structures. This serves as an efficient pathway for the preservation of organic matter, making remineralization more difficult and delayed. Consequently, this process promotes carbon sequestration70.

The timescales and locations of CaCO3 precipitation, whether from local production within the studied ecosystem or of allochthonous origin, have also been included in the carbon balance. This inclusion helps address whether a particular ecosystem can act as a sink or a source of carbon. Due to the highly complex carbonate cycle, which involves numerous uncertainties, such as the actual flux of oceanic CO2 to the atmosphere, it becomes challenging to establish a straightforward and standardized estimation method. As a result, there is still a debate on whether certain ecosystems, such as coral reefs or the BCE, can act as CO2 sources or CO2 sinks44,71. In our analysis, we incorporated the carbon trapped in CaCO3 along with the organic carbon for total carbon estimation. Additionally, we estimated and subtracted the CO2 produced in carbonate precipitation from the ratio (ψ) CO2 flux:CaCO3 precipitation according to depth and latitude, which in this case was 0.8 instead of the more extensively used 0.620. Using the BB as an example of a different method for estimating IC, the results vary by an order of magnitude for both biomass and sediment carbon storage; therefore, depending on the calculation method, the system could be a carbon source or sink (Table 2). We therefore emphasize the importance of taking into account the carbonate content in total carbon estimation.

Another key aspect of carbonates is their potential dissolution, mainly under current ocean warming and acidification conditions. The concentration of carbonate ions determines whether calcification or dissolution occurs. Since this depends on both the water pH and the shoaling depth of the carbonate saturation horizons due to ocean acidification, dissolution could be a possible scenario, mostly in cold waters where CO2 is more soluble72. This possible dissolution would have two main consequences. First, animals with hard skeletons (corals, molluscs, echinoderms) would have slower growth rates, reductions in size, and/or lower larval survival73. Second, carbonates could represent deposits of fossilized alkalinity favouring oceanic CO2 uptake71. Both of these consequences could be important in the BB. On the one hand, there are abundant phyla (Bryozoa, Echinodermata, Arthropoda) with hard skeletons and high carbonate content that can dissolve. On the other hand, we found high abundances of carbonates at our deepest sampling site (~ 9% of IC). Carbonates represent the major fraction of the carbon reservoir in sediments. Carbonates stored at greater depths and in colder waters will dissolve earlier than those stored in coastal and tropical or temperate ecosystems, such as mangroves, salt marshes and sea grasses. Currently, in the sampled areas, the aragonite saturation horizon (ASH) is approximately 1000 m deep, and projections made with the current CO2 emission rate indicate that ASH could be < 100 m deep for 210074. Therefore, the greatest abundance of scleractinian corals within the 500–1000 m depth, including our deepest site (710 m), may be constrained by the depth of the ASH75. Moreover, the skeletons of echinoderms, which constitute a high proportion of carbonate biomass content in the three sampled areas (~ 9% in BC,  ~ 16% in BB, and ~ 43% in the two sites of AS), have high Mg-calcite content. According to previous studies, this high Mg-calcite content could be the first to dissolve76. Indeed, the highest carbonate content in the BB sediments was found at the deepest site (1000 m).

According to our results, in the three sampled areas, the benthic biomass stores more organic carbon than inorganic carbon. Conversely, the opposite relationship was observed in the sediments. This suggests, on the one hand, a high remineralization rate at least in the upper 10 cm of the sea bottom and, on the other hand, that the accumulated carbonates are very old, suggesting a very low sedimentation rate. Moreover, in addition to the amount of carbon stored per unit area; it is very important to consider the total area of the ecosystems to understand their storage capacity. Considering the total area of these MPAs, our study revealed that the BB sediments store 25–30% of the carbon stored in mangrove, seagrass or salt marsh sediments (Table 3).

Understanding how climate change and ocean acidification will affect marine life and carbon stocks is critical for conservation management. Although the number of MPAs has grown exponentially worldwide, MPA effectiveness has been shown to be correlated with levels of protection77. The climate benefits strongly depend on the effectiveness of the MPA in protecting, for example, stored carbon in biomass from bottom trawls23. Our findings have important implications for the MPAN-BB management and for the assessment of carbon storage in future environmental changes. This study has estimated, for the first time, the carbon stock in macrobenthic biomass and in sediments of a sub-Antarctic channel and in an open sea marine protected area of Argentina from a natural resource management perspective by taking into account benthic assemblages and their relation to carbon storage. Moreover, this study contributes to the assessment of global carbon inventories under current climate change scenarios. These cold ecosystems, for which the carbon content had not been estimated prior to this study, could be included in the biological pump inventories due to their capacity to store carbon for long periods and their large area.