Shallow-water seagrasses capture and store globally significant quantities of organic carbon (OC), often referred to as ‘Blue Carbon’; however, data are lacking on the importance of deep-water (greater than 15 m) seagrasses as Blue Carbon sinks. We compared OC stocks from deep-, mid- and shallow-water seagrasses at Lizard Island within the Great Barrier Reef (GBR) lagoon. We found deep-water seagrass (Halophilaspecies) contained similar levels of OC to shallow-water species (e.g. Halodule uninervis) (0.64 ± 0.08% and 0.9 ± 0.1 mg C cm−3, 0.87 ± 0.19% and 1.3 ± 0.3 mg C cm−3, respectively), despite being much sparser and smaller in stature. Deep-water seagrass sediments contained significantly higher levels (approx. ninefold) of OC than surrounding bare areas. Inorganic carbon (CaCO3) levels were relatively high in deep-water seagrass sediments (8.2 ± 0.4%) and, if precipitated from epiphytes within the meadow, could offset the potential CO2-sink capacity of these meadows. The δ13C signatures of sediment samples varied among depths and habitats (−10.9 and −17.0), reflecting contributions from autochthonous and allochthonous sources. If the OC stocks reported in this study are similar to deep-water Halophila meadows elsewhere within the GBR lagoon (total area 31 000 km2), then OC bound within this system is roughly estimated at 27.4 million tonnes.
The tropical Indo-Pacific bioregion contains some of the largest and most diverse seagrass meadows on Earth . Located in this bioregion, the Great Barrier Reef (GBR) protects large areas of the northeast coast of Australia from wave exposure, providing habitat for seagrass between the reef and the mainland. Seagrasses in this region occur in shallow- (estuarine, coastal and reef) and deep-water (offshore) environments . It is only within the past decade that the large spatial extent of deep-water seagrasses in the GBR lagoon has been revealed at over 31 000 km2, an area roughly the size of Switzerland, making them among the most extensive on Earth [3,4].
Whether deep-water seagrasses provide the same important ecosystem services as their shallow-water counterparts has received much interest, especially given recent data demonstrating the potential for shallow-water seagrasses to act as globally significant ‘Blue Carbon’ sinks, thereby helping to mitigate climate change [5,6]. Blue Carbon broadly refers to organic carbon (OC) captured and stored within the oceans, by seagrass meadows, tidal marshes and mangrove forests . At present, information is lacking on the levels of OC storage for seagrasses growing in deep waters.
An argument against deep-water seagrasses in the GBR having a high OC sequestration capacity is they are smaller growth-form species with sometimes annual or ephemeral occurrence , and therefore have lower biomass and less capacity to trap particulate OC from the water column than larger, more persistent growth forms. However, Lavery et al.  found that Halophila ovalis in coastal waters—a small growth-form species—had the second highest OC stock among 10 Australian species tested, including large growth forms. Alternatively, Serrano et al.  found that OC stocks for Posidonia oceanicadeclined significantly with depth, with an approximately 15-fold decrease in sedimentary OC stocks from shallow (2 m) to deep (32 m) meadows. Besides these studies, data on the OC stocks of deep-water seagrasses are lacking.
Here, we compared carbon stocks from shallow- and deep-water seagrass meadows around Lizard Island, situated in the northern GBR (figure 1). We tested four hypotheses: (i) that OC stocks in seagrass habitat would be greater than adjacent bare sediment; (ii) that OC stocks for deep-water seagrasses would be lower than shallow-water seagrasses owing to the small stature and/or light-limited productivity; (iii) that the OC stocks in deep-water seagrasses would primarily derive from autochthonous sources; and, finally (iv) that owing to the region being a high-energy zone with large sediment grain size, the average OC stock values would be low compared with other reported values around Australia. Additionally, we also report on calcium carbonate (CaCO3) stocks in the region. Calcium carbonate is not currently considered in Blue Carbon accounting ; however, the process of CaCO3 precipitation is often very high for tropical seagrass meadows and it is a process that can exacerbate climate change .Figure 1.
Coring sites around Lizard Island, where sediments were collected in bare sediment and seagrass (Halophila) habitats at: (a) shallow (0–3 m below MSL), (b) mid (8–12 m) and (c) deep (18–24 m) sites, and (d) Halodule uninervis and (e) Thalassia hemprichii habitat at shallow sites.
2. Material and methods
The study was conducted at Lizard Island, in the northern GBR (figure 1), which was dominated by three species of seagrass—Halophila species: H. ovalis, H. decipiens and H. spinulosa—with some Halodule uninervis and Thalassia hemprichii in shallow waters. Three sites in Halophila habitat and three in bare sediment habitat were sampled for sediment carbon characteristics at each of three different depth ranges (shallow: 0–3 m, mid: 8–12 m and deep: 18–24 m), with the exception of deep bare habitat where only two suitable sites were found with this habitat. One site each of H. uninervis and T. hemprichii habitat were also sampled in shallow water (see  for habitat characteristics). At each of the 19 sites, three sediment cores (30 cm depth) were collected on SCUBA using methods described in Macreadie et al. , and then prepared and analysed for elemental carbon, isotopic signature (δ13C) and loss on ignition. Comparisons of sediment carbon attributes among habitat types (seagrass versus bare sand) and water depth (shallow-, mid- and deep-water) as well as comparisons among seagrass species (Halophila, Halodule and Thalassia) were made with permutational multivariate analysis of variance. See the electronic supplementary material, for more details.
3. Results and discussion
This is one of the first studies to look at sedimentary OC stocks in tropical deep-water seagrasses. Down-core variation in carbon characteristics did not vary greatly with sediment depth (electronic supplementary material, figures S1–S3). Levels of carbon densities (Corg, mg cm−3) and the percentage of OC were significantly greater in Halophila habitats compared with bare sediment (pseudo-F1,33 = 9.15, p = 0.013, and pseudo-F1,33 = 14.23, p = 0.004, respectively; electronic supplementary material, table S1). Corg was also significantly higher in shallower habitats (pseudo-F2,33 = 5.18, p = 0.027; electronic supplementary material, table S1), owing mainly to decreases in bare sediment at greater depths (figure 2d). The greatest differences in both Corg and %OC occurred in deep water, where Halophila habitats (Corg = 0.88 mg cm−3, %OC = 0.64) had approximately nine times the amount of OC of bare sediments (Corg = 0.10 mg cm−3, %OC = 0.08) (figure 2a,d).Figure 2.
Differences in (a) sediment OC content; (b) sediment IC content; (c) δ13C signatures; and (d) sedimentary carbon stocks across different depth gradients (deep: 18–25, mid: 8–12 m, shallow: 0–3 m) and among seagrass and bare sediment habitats in GBR lagoon (Lizard Island). Values are mean ± 1 s.e.
On a per unit area basis, the OC levels were low compared with other seagrasses ; however, the vast areas covered by deep-water communities could equate to a substantial regional scale Blue Carbon reserve. In the GBR region of Queensland, 31 000 km2 of deep-water seagrasses have been described . While acknowledging the shortcomings of extrapolating from one location across the entire GBR, if we scale-up our mean value of 0.8839 mg Corg cm−3 for deep-water Halophila habitat, this would equate to an OC stock of 27.4 million tonnes in the surface 1 m of sediments of the reef. To gain a much more accurate estimate of these stocks, it is important to quantify carbon stocks over a much greater geographical range, incorporating different seagrass species and densities and sediment types. Blue Carbon stocks are also generally reported to 1 m sediment depth and we have continued that protocol for our GBR-wide estimate. Given that we have only sampled to 30 cm depth in our study, this creates further uncertainty in our estimate. Deeper cores at a range of locations and habitats would also reduce these uncertainties.
Inorganic carbon content showed a significant interaction between water depth and habitat (pseudo-F2,33 = 4.14, p = 0.046; electronic supplementary material, table S1). There was a general trend of increasing total inorganic carbon (TIC) with water depth, and at mid- and deep-water depths bare sediments had a higher TIC than Halophilahabitats; however, this was reversed in shallow water (figure 2b). There was a significant difference in %TIC among seagrass habitats (pseudo-F4,22 = 4.97, p = 0.040; electronic supplementary material, table S2), with deep-water Halophila containing higher levels (8.17%) of inorganic carbon (IC) than shallower seagrass habitats (2.27 to 5.18%) (figure 2b). Higher inorganic carbon (calcium carbonate—CaCO3) in deep-water seagrass sediments could potentially be precipitated from epiphyte assemblages within the meadows, which would reduce their ability to act as a CO2-sink, given calcium carbonate precipitation emits CO2 [10,14].
There was a significant interaction among water depth and habitat in relation to δ13C signatures (pseudo-F2,33= 5.27, p = 0.025; electronic supplementary material, table S1). Signatures varied little between Halophila and bare sediment in mid- to deep-water habitats; however, there was considerable separation in shallow waters, where bare sediment signatures were significantly lower than Halophila (figure 2c). There was no significant difference in δ13C among seagrass habitats (pseudo-F4,22= 3.86, p = 0.069; electronic supplementary material, table S2), although a spread of signatures were seen, particularly in shallow water, where mean values of H. uninervis and T. hemprichiihabitats ranged from −17.0 to −10.9, respectively (figure 2c). The δ13C signatures in seagrass sediments all fall within the ranges of the plant material cited for the seagrass species dominating each habitat (see electronic supplementary material, table S3), indicating there is potentially a high degree of autochthonous contribution from seagrasses in the sedimentary carbon pool. As isotopes for the different seagrass species and other potential primary sources of carbon at Lizard Island were not analysed, we cannot define the contribution of autochthonous versus allochthonous carbon in each habitat, although it is likely to have originated from a range of sources, including seagrass, macroalgae and terrestrial sources.
The results from our study are somewhat surprising, with our initial expectations that the sparse deep-water Halophila communities would be unlikely to contain more sedimentary OC than bare areas, and would contain substantially lower OC than nearby larger-growing species in shallow water. Halophila decipiens is a rapid-growing species with a very fast turnover time and can completely turn over its biomass within a few days . Perhaps this rapid turnover delivers substantial plant material to the sediments.
Our results contrast with those of , where a gradient of reducing sedimentary carbon in seagrass meadows occurred with depth. At our study location, this did not occur. The contrasting seagrass communities (temperate Posidonia versus tropical Halophila) and sediments (terrigenous versus carbonate) between the two studies may explain these differing results. Regardless, our results emphasize the dangers of adopting a universal paradigm on the value of deep-water seagrasses in carbon capture.
The results point to the requirement for a greater effort to assess the roles of deeper seagrasses in sedimentary carbon capture across a more extensive geographical and ecological range. Furthermore, the role of carbonate material in Blue Carbon cycling warrants further investigation to determine whether tropical seagrass meadows are facilitating carbonate sequestration. Currently, global tropical deep-water seagrasses are poorly quantified, with most discoveries relatively recent and the likelihood that many more areas exist in unsurveyed waters. Our study suggests these areas potentially harbour significant undescribed global seagrass carbon sinks.
P.H.Y. and P.I.M. conceived the idea; all authors contributed to fieldwork and analyses and to writing and improving the manuscript. All authors approved the final version of the manuscript and agree to be held accountable for its content.
The authors have no competing interests to declare.
P.H.Y. and P.I.M. were supported by an Isobel Bennet Marine Biology Fellowship with funds supplied by the Hermon Slade Raiatea Foundation to the Lizard Island Reef Research Foundation. M.A.R. and P.I.M. are supported by an Australian Research Council Linkage Grant (LP160100492). P.I.M. was supported by an Australian Research Council DECRA Fellowship (DE130101084).
We thank the staff at the Lizard Island Research Station and C. Bryant for assistance with fieldwork.
A contribution to the special feature ‘Blue Carbon’ organized by Catherine Lovelock.
Electronic supplementary material is available online at http://dx.doi.org/10.6084/m9.figshare.c.4302653.