Diatom ooze—A large marine mercury sink

See allHide authors and affiliations

Science  24 Aug 2018:
Vol. 361, Issue 6404, pp. 797-800
DOI: 10.1126/science.aat2735

Mercury sinking

Mercury is a highly toxic, globally ubiquitous pollutant that can seriously damage human health. Most mercury pollution enters the atmosphere from burning coal and other fossil fuels and from industrial activity, but where does it all go? Zaferani et al. analyzed biogenic siliceous sediments (diatom ooze) from off the coast of Antarctica and found that they contained surprisingly large amounts of mercury. The results suggest that as much as 25% of mercury emissions over the past 150 years could be trapped in sediments like these, revealing the important role that the marine biological pump may play in the global mercury cycle.

Science, this issue p. 797


The role of algae for sequestration of atmospheric mercury in the ocean is largely unknown owing to a lack of marine sediment data. We used high-resolution cores from marine Antarctica to estimate Holocene global mercury accumulation in biogenic siliceous sediments (diatom ooze). Diatom ooze exhibits the highest mercury accumulation rates ever reported for the marine environment and provides a large sink of anthropogenic mercury, surpassing existing model estimates by as much as a factor of 7. Anthropogenic pollution of the Southern Ocean began ~150 years ago, and up to 20% of anthropogenic mercury emitted to the atmosphere may have been stored in diatom ooze. These findings reveal the crucial role of diatoms as a fast vector for mercury sequestration and diatom ooze as a large marine mercury sink.

Owing to a scarcity of archive data from the marine environment, the role of primary production and biogenic sediments for natural and anthropogenic mercury (Hg) sequestration in the oceans is poorly understood. In the open ocean, most Hg is received by atmospheric deposition and, to a minor extent, by discharge from rivers, and most atmospheric Hg is thought to be reemitted (1, 2). Available data on global Hg burial in marine sediments are based on model estimates (1, 3). These studies estimate that the amount of Hg buried in open ocean sediments is relatively low (190 to 200 Mg year−1) (2, 4). Several studies indicate that the marine Hg cycle is closely related to biological productivity and that Hg scavenging by organic particles is an important vector for Hg burial in sediments (57). However, the contribution of these processes to the accumulation of Hg in marine sediments and the overall marine Hg mass balance in the oceans is not known.

Diatoms are a major group of microalgae. Their remains make up pelagic sediments called diatom ooze, which cover ~11% of the ocean floor (8) and accumulate in largest amounts in the Southern Ocean (9, 10). Until now, no high-resolution (<50 years) Hg record—which allows calculation of global marine Hg sequestration—was available from ocean sediment, and data on Hg in the ocean were limited to water column measurements (7).

Here we show high-resolution Holocene Hg records from diatom ooze sediment cores taken at three basins around Antarctica: Adélie Basin (AB) [Integrated Ocean Drilling Program (IODP) Expedition 318 Site U1357], Prydz Bay (PrB) [Ocean Drilling Program (ODP) Expedition 119 Site 740], and Palmer Basin (PB) (ODP Expedition 178 Site 1098) (Fig. 1). From these cores and modeled marine biogenic silica (BSi) sequestration, we evaluated the role of diatom ooze as a Holocene Hg sink and estimated the global amount of Hg accumulated in diatom ooze. In addition, we reconstructed the chronology of Hg accumulation and anthropogenic atmospheric Hg pollution during the past 8600 years from annually laminated AB sediments at resolutions of 10, 20, and 200 years.

Fig. 1 Map of core sites.

Locations of the investigated cores in Antarctica: Prydz Bay core (ODP Expedition 119 Site 740), Palmer Basin core (ODP Expedition 178 Site 1098), and Adélie Basin core (IODP Expedition 318 Site U1357).


Composition of the sediment is similar at each of the three locations (Fig. 2), with high SiO2 concentrations of 53% (PB), 66% (PrB), and 70% (AB), respectively, identifying the sediments as diatom ooze (1113). Principal components analyses reveal that Hg generally shows low or no positive covariance with lithogenic elements (e.g., Al, Ti, Zr) (fig. S1). This indicates that binding of atmospheric-derived Hg to microalgae particles is the dominant process of Hg accumulation in these sediments and that terrestrial influences are negligible. The upper core sections contain median Hg concentrations between 55.6 and 70.9 ng g−1 and thereby exceed the median background (lower core section) concentrations (which are between 32.1 and 36.0 ng g−1) by a factor of 2.2 (Fig. 3). Maximum concentrations reached 73.0 (PB), 84.5 (AB), and 86.6 ng g−1 (PrB). This strong increase in Hg concentrations in the upper core sections likely results from the two- to fivefold increase in global atmospheric Hg loads due to industrial Hg emissions in the past two centuries (14). These concentrations are similar to those found in other marine sediments (15). In the lower core sections, Hg concentrations vary little (Fig. 3) but exhibit dilution effects at higher silica concentrations (see Si-normalized Hg concentrations in fig. S2). Similar Hg concentrations in the three cores taken at locations separated by thousands of kilometers indicate the absence of strong local effects and suggest that Hg concentrations in diatom ooze are related to concentrations of dissolved Hg in seawater, which are controlled by atmospheric Hg deposition and the Henry’s law constant, as well as scavenging and mineralization processes of sinking particles.

Fig. 2 Silicon concentration in diatom ooze sediments.

Downcore records of silicon (Si) concentrations in Adélie Basin (AB), Palmer Basin (PB), and Prydz Bay (PrB) sediments (note the differences in core depth).

Fig. 3 Mercury concentration and accumulation rate in diatom ooze sediments.

(A to C) Downcore records of mercury (Hg) concentrations in Adélie Basin (AB), Palmer Basin (PB), and Prydz Bay (PrB), and mercury accumulation rates (HgAR) in AB sediments. The shaded area in (A) denotes the industrial period. An average sedimentation rate of 2 cm year−1 was used for AB chronology reconstruction (13). Median background Hg concentrations equal 32.1, 34.5, and 36.0 ng g−1 with median absolute deviations of 2.74, 2.21, and 1.20 for AB, PB, and PrB, respectively (note the differences in core depth).

Owing to the consistency in Hg concentrations, the different sedimentation rates of the three cores control the large differences in Hg accumulation rates (HgAR). The extremely high sedimentation rates (~2 cm year−1) of AB sediments lead to HgAR that are the highest ever reported from remote marine areas [median rates of 576 and 1200 μg m−2 year−1 for the preindustrial and industrial (after 1850) periods, respectively] (Fig. 3). These values even exceed those reported from estuaries and shelf areas (60 to 700 μg m−2 year−1) (16), which have hitherto been ranked as the largest marine Hg sinks (17). The lower sedimentation rates of the PrB and PB cores produce lower HgAR (PrB median: 16.8 and 33.3 μg m−2 year−1; PB median: 124 and 192 μg m−2 year−1 for the lower section and upper section of the cores, respectively) than AB cores. However, the PB and PrB rates are still high relative to previously reported values, indicating areas with high accumulation of diatom ooze as large Hg sinks.

To quantify the role of diatom ooze sediments for Hg sequestration in the oceans, we used an estimated total area of pelagic diatom ooze sediments of 23.3 million and 31.1 million km2 (10, 18). Owing to the high sedimentation rates of our Antarctic cores, which could lead to large overestimation of global HgAR in diatom ooze (DOHgAR), we applied two independent approaches for DOHgAR calculation. For calculation one, we used the diatom ooze sedimentation rate suggested for the Southern Ocean (0.75 mm year−1) (19) (dry mass accumulation rate of 0.68 kg m−2 year−1) in combination with median Hg concentrations (preindustrial and industrial period) derived from our Antarctic cores. For calculation two, we used published estimates of global annual BSi accumulation rate and different diatom preservation rates (30, 60, and 80%) combined with median Si/Hg ratios from our cores for the preindustrial and industrial periods, respectively (Table 1).

Table 1 Global annual mercury accumulation and total accumulated mercury in diatom ooze sediments (DOHgAR).

Values for the industrial (after 1850) and preindustrial (before 1850) periods were determined by two approaches. For calculation one (Calc. 1) we used a diatom ooze area of 23.3 million and 31.1 million km2 (10, 18) and a sedimentation rate of 0.75 mm year−1, as suggested for Antarctic diatom ooze (19), and average Hg concentrations for the preindustrial and the industrial period derived from AB, PrB, and PB cores. Calculation two (Calc. 2) was based on the annual accumulation of biogenic silica (BSi) in the world’s oceans (23), assuming BSi preservation of 30, 60, and 80% and average Si/Hg ratios for the preindustrial and industrial periods derived from AB, PrB, and PB cores.

View this table:

Average DOHgAR for the industrial period ranges between 850 and 1166 Mg year−1 (Table 1), which are, respectively, a factor of ~4.5 and 6.1 higher than estimates of existing model approaches (190 Mg year−1) (1, 3) for burial of Hg in deep ocean sediments, indicating that Hg sequestration in biogenic marine sediments appears to have been largely underestimated to date. Preindustrial DOHgAR values range between 497 and 689 Mg year−1, which are a factor of 1.7 lower than those calculated for the industrial period. Although diatom ooze sedimentation and preservation in the Southern Ocean is known to be generally high (>86%) (20), we additionally calculated two conservative approaches using the lowest reported diatom ooze sedimentation rate (~0.01 mm year−1) (21, 22) and the lowest published BSi preservation rate (3%) (23). These calculations reveal DOHgAR of 18 and 38 Mg year−1, which still account for 9 and 20% of the model estimate (190 Mg year−1), respectively (1, 3). These findings emphasize the importance of Hg sequestration in diatom ooze in particular and the role of marine primary production in general. However, the amount of data used for these calculations is still small and may exhibit large uncertainties.

Streets et al. (24) estimated that ~1130 Gg of Hg were released from anthropogenic sources to the environment between 1850 and 2010, and 336 Gg of this amount were emitted directly to the atmosphere. On the basis of the minimum and maximum DOHgAR obtained from calculations one and two (397 to 1322 Mg year−1), the total amount of Hg accumulated in diatom ooze in the past 150 years is 22 to 84 Gg, or about 6.5 to 25% of all Hg emitted to the atmosphere and 2.0 to 7.4% of all Hg released to the environment through anthropogenic emissions in the past 150 years. The remote location and large distance of most diatom ooze sediments from coastlines indicate atmospheric fluxes as the predominant Hg source. Under this assumption, ~20% of the anthropogenic Hg emitted to the atmosphere since 1850 may have been buried in diatom ooze alone. This amount might have been even higher considering that DOHgAR in areas such as AB largely exceeds the rates used in our calculation. The unequal distribution of atmospheric Hg concentrations during the industrial period as well as that of diatom ooze sediments between the hemispheres suggest that Hg sequestration by diatom ooze might be more important in the Southern Hemisphere than in the Northern Hemisphere, which might reduce the amount of anthropogenic Hg accumulated in diatom ooze. It is further unknown to what extent the 1 to 5.5 Gg year−1 of global river discharge of Hg, mainly released to the Northern Ocean (25), contributes to Hg burial in diatom ooze. However, recent studies indicate that most of this Hg is buried in estuaries and on the continental shelf (17).

The high Hg accumulation rates in diatom ooze, which largely surpass the reported Hg deposition rates to the oceans (1), might be partly explained by increased Hg fluxes from the atmosphere into the seawater. Similar to CO2, such increased atmospheric Hg fluxes could be caused by a permanent shift of the dissolution equilibrium (based on Henry’s law constant) toward the dissolved phase if the dissolved phase is permanently removed through Hg scavenging by a large amount of algae during blooms. We assume that scavenging of water-phase Hg by sinking diatom organic matter, rather than active uptake [see (26) for Hg uptake by algae], is the dominant process of Hg binding in diatom ooze. Furthermore, dissolved Hg in the water column will provide a large Hg pool for scavenging by sinking particles, which could explain the high observed Hg sedimentation rates (see supplementary materials for further discussion). Previous mass-balance models estimated that 96% (~1300 Mg year−1) of the atmospheric Hg flux to the ocean (~1350 Mg year−1) is re-emitted (1). Even the conservative Hg accumulation rates in diatom ooze suggest that these high re-emission fluxes are likely overestimated if used on a global scale. A recent model approach on Hg mass-balance changes during eutrophication in the Baltic Sea estimated a 30 to 40% decrease in Hg evasion during diatom blooms (27), which was attributed to Hg sedimentation by algae organic matter. This is in line with findings of Hg concentrations in settling organic particulate matter in the Atlantic Ocean, which suggest that Hg partitioning to particles is up to 1000 times greater than suggested by models resulting in decreasing Hg availability for evasion (28), emphasizing the important role of the marine biological pump for the global Hg cycle.

The undisturbed 170-m-long annually laminated sediments of AB offer an ultrahigh resolution (~2 cm year−1) record of Hg accumulation in the oceans, spanning large parts of the Holocene (Fig. 3) (13). Hg concentrations in AB sediments show a median of 32.1 ng g−1 between ~8600 years ago and ~1850 AD, and Hg accumulation rates largely follow the Hg concentration record, with a median value of 556 μg m−2 year−1. Since 1850, Hg concentrations and accumulation rates have increased by a factor of ~2.2 to median values of 70.9 ng g−1 and 1296 μg m−2 year−1, with local peaks around 1890, 1950, and 1990, and highest values of 84.5 ng g−1 and 1830 μg m−2 year−1 in ~1990. The increase in Hg accumulation was coincident with the onset of the industrial period and intense coal burning, which has been the most important anthropogenic source of Hg emissions to the atmosphere and global Hg dispersion. The local peaks in Hg accumulation during 1900–1920 and 1950–1970 (Fig. 3) correspond to emission peaks calculated on the basis of global Hg production and consumption rates (24, 29). An indication of a global influence of Hg emissions from gold and silver mining in the Americas during 1850 and 1910, as estimated in recent models (29), which nearly reach or even surpass levels of those found at times of maximum Hg emissions during the 20th century, was not found in our cores. Moreover, no anthropogenic influence before ~1850 that could be linked to Hg emissions from colonial gold mining in South America, as suggested by model estimates, could be detected (16). The amount of Hg emitted to the atmosphere by colonial gold mining was probably too low to be detectable in remote Antarctica. Similarly, the increase in global atmospheric Hg loads, as revealed by our Antarctic marine sediment cores, attributed to all-time anthropogenic activities, is only about a factor of 2.2, which is slightly lower than in most Hg records from terrestrial archives (14, 3032) but distinctly lower than current model predictions. This might be due to the fact that most anthropogenic Hg has been emitted in the Northern Hemisphere so that a larger Hg portion is sequestered there. However, the Southern Ocean is one of the most productive areas on Earth and, because of the link between primary production and Hg sequestration, is likely a large sink for global atmospheric Hg.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S4

Tables S1 to S14

References (3354)

References and Notes

Acknowledgments: We thank P. Rumford (Golf Coast Repository), W. Hale, and H. Kuhlmann (Bremen Core Repository) for providing samples and A. Calean and P. Schmidt for technical assistance. Funding: This work was funded by Technische Universität Braunschweig. Author contributions: S.Z. carried out the laboratory analyses, and M.P.-R. performed the statistical analyses. H.B. conceived of the research. All authors contributed equally to data interpretation and manuscript writing. Competing interests: The authors declare no competing interests. Data and materials availability: All data are available in the supplementary materials.

Stay Connected to Science

Navigate This Article