Main

The realization that vegetated coastal habitats support globally relevant rates of organic carbon burial that rank amongst the highest in the biosphere1 led to the development of strategies to mitigate climate change through the conservation and restoration of seagrass, mangrove and saltmarsh habitats, termed blue carbon strategies2,3,4,5. Macroalgae, the most productive marine macrophytes on a global scale6,7, have been excluded from such blue carbon assessments as most macroalgae grow on rocks, where burial is precluded2,4,8. However, some macroalgae grow on sandy sediments9, with burial averaging 0.4% of net primary production7. More importantly, macroalgae export about 43% of their production7 both as particulate organic carbon (POC)10,11 and dissolved organic carbon (DOC)8,12,13,14. Some of this carbon may reach depositional areas and be sequestered in sediments, or reach the deep sea, where the carbon is locked away from exchange with the atmosphere. Macroalgae can thereby act as carbon donors to sink reservoirs located elsewhere6,8 and it has recently been argued that they should be included in blue carbon assessments8,15,16,17. However, the evidence required to estimate their contribution has been published under a range of research fields. For instance, macroalgal export has been studied because of its consequences for the dispersal of species and genes18,19, the relocation of rocks across the seafloor20, connectivity among habitats and the stimulation of secondary production in adjacent and distant habitats10,11, including the supply of food to deep-sea fauna21 and carbonate to the deep sea22.

Sequestration of macroalgal carbon in marine sediments

Macroalgae may have contributed to carbon sequestration for over 2.1 billion years, based on the oldest dating of a multicellular organism, Grypania spiralis, which is suggested to be a macroalga23. They have certainly done so through the past 500 million years, as macroalgae have been reported to be the source of a number of oil deposits24,25. One of the prerequisites for macroalgae to contribute to CO2 sequestration, that is, for their carbon be sequestered over centennial timescales, is thereby amply fulfilled.

Reports of the presence of macroalgal carbon in marine sediments are relatively few, but suggest that macroalgal carbon may be widespread, extending from shallow to deep-sea sediments and from polar to tropical regions, as well as across a broad range of depths into the sediment, from surface and subsurface layers down to deeper than a hundred metres into the sediment (Fig. 1, Supplementary Table 1). Macroalgal carbon is typically identified in depositional environments10,11,26, including anoxic basins, submarine canyons, sedimentation areas within complex rocky shores and the deep sea (Fig. 1, Supplementary Table 1). Macroalgal-specific markers such as stable carbon isotopes coupled with lipids, sterols and carotenoids have been used to trace the contribution of macroalgae to sediments9,27and food webs26. The preservation potential of macroalgal carbon in sediments depends on the lability of the organic carbon, which varies between species17.

Figure 1: Map of the locations where macroalgal carbon storage has been reported.
figure 1

The types of macroalgae are indicated for observations from sediment traps that are in the water column, on the sediment surface and buried in sediments. Inset, the frequency distribution of the water depths of macroalgae observations, with the majority representing the deep sea (<1,000 m). All references of observations are available in Supplementary Table 1.

Full size image

Export of macroalgal carbon to the deep sea

Multiple reports, including the presence of fresh Sargassum in the guts of abyssal isopods21, confirm the prevalent presence of macroalgal drift on the deep seafloor down to 6,475 m (Fig. 1, Supplementary Table 1). These reports are dominated by observations of brown algae, with abundant reports of Sargassum from subtropical latitudes, kelps in the temperate zone and kelps and Desmarestiales in the polar regions (Fig. 1, Supplementary Table 1). Macroalgal drift also appears to be particularly abundant on the discharge area of submarine canyons (Supplementary Table 1). Such canyons are widespread across all oceans (660 major canyons have so far been documented28) and are important conduits that focus the export of materials, including macroalgal carbon (Fig. 2), from the continental shelf to the deep sea29. For instance, more than 130,000 t of kelp is exported yearly through the canyon adjacent to the Monterey Peninsula30. In addition to observations of macroalgae on deep-sea sediments, sinking algal fragments have often been collected in open ocean pelagic sediment traps at depths of up to 1900 m (Fig. 1, Supplementary Table 1). This suggests two modes of transport: bed-load transport of drift material and sinking fluxes of negatively buoyant macroalgal detritus31 (Fig. 2).

Figure 2: Conceptual diagram of the pathways for export and sequestration of macroalgal carbon.
figure 2

KING ABDULLAH UNIVERSITY OF SCIENCE AND TECHNOLOGY

Air bladders are common among brown algal taxa and facilitate their long-range transport (i). Langmuir circulation forms windrows of macroalgae (ii) and can force the algae to depths where water pressure makes the air bladders burst and the algae then sink. Macroalgal carbon can be sequestered either via burial in the habitat or by transport to the deep sea where it is sequestered whether buried or not (iii).

Full size image

Through shedding of old fronds, kelps support a continuous flux of export material. The substantial drag of large kelps leads to detachment from the substrate during high-energy events as well as removal by moderate swells10,32. The gas vesicles characteristic of many brown algae (pneumatocysts, Fig. 2) favour the formation and long-distance drift of floating aggregates of macroalgae18,19. This ability to drift, in combination with their relative unpalatability due to phenols and refractory carbon compounds such as fucoidan17, explain their prevalent role as carbon source in deep sediments (Figs 1 and 2, Supplementary Table 1). Detached macroalgal tissue is transported offshore by currents, which support long-range export (Fig. 2). For instance, a 588 km2 patch of detached brown algae, Colpomenia sp., organized into windrows by wind-driven Langmuir circulation, was reported off the Great Bahamas Bank33. Drifting rafts of giant kelp may occur at very high densities, with 39,000 to 348,000 rafts identified in the Southern California Bight alone, exporting the kelp more than 300 km offshore34. Drifting surface mats of Sargassum are also abundant22,35,36.

A number of mechanisms have been identified for the delivery of drifting macroalgae to marine sediments. Wind-induced Langmuir circulation can entrain floating macroalgal fragments at depth, where pressure can collapse their gas vesicles, rendering the macroalgae negatively buoyant and removing them from the neuston (Fig. 2). For instance, the gas vesicles of Sargassum have been found to collapse in 5 h under a pressure of only 30 dbar, although those of the Sargasso Sea Sargassum seem to be more resistant37. Another delivery mechanism is the ballasting of floating macroalgae by the stones dislodged by excessive drag forces (Fig. 2) — a phenomenon of global geological relevance that results in deep-sea soft sediment plains being paved with stones20. The growth of calcifiers on macroalgal surfaces can also add to their density and contribute to their subsequent sinking22.

The offshore export of macroalgal fragments from the coastal zone fuels a potentially large flux of macroalgal carbon to the deep sea: there are reports of 16.5 gC m−2 d−1 of giant kelp being exported through the Carmel Canyon, California38, and of 0.4 gC m−2 yr−1 of Sargassum reaching 3,600 m depth in the Northwest Atlantic35. These fluxes can also be highly episodic, such as the estimated input in excess of 7 × 1010 gC potentially reaching the seafloor at 1800 m depth off the Bahaman shelf33 during a storm.

Global carbon sequestration by macroalgae

Macroalgae are the dominant primary producers in the coastal zone1,6 with a global net primary production (NPP) of 1,521 TgC yr−1 (range: 1,020–1,960 TgC yr−1; Fig. 3) over an estimated area of 3.5 million km2 (range: 2.8–4.3 million km2; see Tables 1 and 2). There are few studies that document the fate of NPP and export to the deep sea, but they do help to provide a first-order estimate of the contribution of macroalgae to carbon sequestration from burial in coastal sediments and export to the deep sea (defined as >1,000 m depth), where the carbon is precluded from exchanging with the atmosphere over extended timescales even after being remineralized (Fig. 3). We combine existing information on the fate of macroalgal carbon (Table 1) and propagate uncertainties through the calculations using a Monte Carlo approach (see Methods, Tables 1 and 2) to derive this crude estimate of the contribution of macroalgae to carbon sequestration.

Figure 3: Pathways for the sequestration of macroalgal carbon in the ocean.
figure 3

Each step of the carbon flow from global macroalgal net primary production (NPP) to carbon sequestration (in blue) is supported by the literature or inferred by a difference between a total and subcomponents supported by literature (Table 1). The means (with 25 to 75% quartile ranges in parentheses) shown are derived from an uncertainty propagation analysis (Methods), except for those fluxes not conducive to carbon sequestration (all values are in TgC yr−1). As the estimates have been derived independently, their total does not necessarily match to the mean global NPP estimate. Grazing (33.6% of the NPP) and remineralization (37.3% of the NPP) in the algal bed are adopted from a previous budget7.

Full size image
Table 1 Details of the data used in the uncertainty propagation analysis.
Full size table
Table 2 Distribution of values originating from the uncertainty analysis.
Full size table

On average, about 0.4% of macroalgal NPP is buried directly in the habitat7 for macroalgae that grow on soft sediments (a mean of 6.2 TgC yr−1, Fig. 3). An estimated 43% of macroalgal NPP is exported, supporting a global flux of about 679 TgC yr−1 (Fig. 3). We estimated the fraction exported as DOC by combining the mean area-specific estimate of macroalgal DOC release (101 gC m−2 yr−1)12 with the estimated global macroalgal area (Tables 1 and 2), yielding 355 TgC yr−1 or 52% of the total export (Fig. 3). The remaining export (48%, 323 TgC yr−1) is in particulate form (Fig. 3, Tables 1 and 2).

We estimate that 33% of the DOC flux (117 TgC yr−1) is exported below the mixed layer, representing an upper boundary for the amount of macroalgal DOC reaching the deep sea (Fig. 3). This value is supported both by the finding that the net oceanic primary production (approximately 50 PgC yr−1), of which about 13% (that is, 6.5 PgC yr−1) is released as DOC, results in a downward DOC export of 2 PgC yr−1 (approximately 30%) below the mixed layer (details in Table 1) and the large inputs of DOC from ocean margins to the ocean interior39. We assume that the same fraction of macroalgal DOC is exported below the mixed layer and potentially reaches the deep sea (Table 1).

Regarding the fate of macroalgal POC export, three independent studies suggest that about 11% (35 TgC yr−1) reaches the deep sea. One study reports that around 10% of drift Sargassum reaches the deep seafloor as particulate material35, a second reports that approximately 3% of NPP31 (that is, equivalent to around 10% of the POC export) reaches the deep seafloor as phytodetritus and a third work finds that around 13% of drift kelp is exported through canyons (an average of two surveys before hurricane40) (Fig. 3, Table 1). The remaining 89% of the export POC flux is assumed to stay in the coastal ocean. Of this fraction, an estimated 4.6% (14 TgC yr−1) is buried in shelf sediments41 (Fig. 3, Tables 1 and 2) and we assume that the rest (95.4%) is mineralized.

Together these findings yield a first-order estimate of the contribution of macroalgae to carbon sequestration of about 173 TgC yr−1 (range: 61–268 TgC yr−1), of which about 88% is sequestered in the deep sea (Fig. 3). This estimate exceeds that for carbon buried in angiosperm-based coastal habitats (111–131 TgC yr−1)1 and provides evidence of the importance of macroalgae in biological CO2 sequestration. However, the range around this estimate varies by an order of magnitude, highlighting the need for targeted efforts to address the main sources of this uncertainty, which include the area covered by macroalgae, the amount of macroalgal-derived carbon that is sequestered in sediments and the fate of macroalgal-derived DOC exported from the mixed layer.

An assessment of the potential changes in the global rates of CO2 sequestration of macroalgae requires a global evaluation of the trends and drivers of this sequestration, which, unlike those of seagrasses, mangroves and salt-marshes4, has not yet been attempted. Climate change leads to the loss of kelp forests near their southern distribution limit42,43, but may favour their poleward expansion into the Arctic44,45 and may change macroalgal NPP and detrital export in the future46. Other global drivers of change47, including eutrophication48 and the growing macroalgal harvest and aquaculture industry49,50, may also influence the contribution of macroalgae to carbon sequestration by affecting the future area of macroalgal growth and production. Such changes in the sequestration of macroalgal carbon should be monitored and macroalgae should be considered both in carbon accounting reports and within blue carbon conservation and restoration strategies to mitigate climate change.

Methods

Estimates of the uncertainty in the global net primary production of macroalgae and its fate required the mean and range for the macroalgal carbon sequestered annually either in sediments or the deep sea to be calculated. An uncertainty propagation analysis was then undertaken using Monte Carlo simulations. For each component of the global net primary production of macroalgae and its fate, 1,000 randomly generated values were obtained by sampling randomly from a normal distribution with the corresponding mean and standard deviation (Table 1). Individual estimates were calculated by combining each of the 1,000 simulated values, thereby yielding 1,000 estimates for each step in the calculations that combined all of the uncertainties from the terms entering these calculations (Table 1). We then retrieved the mean from the 1,000 estimates of macroalgal production, which is generated by combining the uncertainties in the area covered globally and the NPP per unit area (Table 1) and carbon flux, and characterized the uncertainty by the central 50% interquartile range of the values (that is, the 25% to 75% quartiles of the values generated; Table 2).

Data availability.

The data that support the findings of this study are available within the text and Supplementary Information.