Introduction

Improving the capacity of natural ecosystems to assimilate carbon dioxide (CO2) and store organic carbon (Corg) long-term is an important strategy for mitigating climate change1. As such, restoring the carbon sequestration capacity of terrestrial ecosystems, such as tropical rainforests and farmland, has been the focus of many conservation efforts2. Vegetated coastal ecosystems, particularly mangroves, salt marshes, and seagrasses also store large amounts of organic matter3,4,5. Seagrass communities account for up to half the global carbon accumulated in vegetated coastal habitats6. However, resolving the climatic benefit of marine ecosystems requires knowing the CO2 sequestration potential through long-term (>100 years) Corg burial7,8, as well as quantification of the methane (CH4) and nitrous oxide (N2O) flux during early diagenetic processes9,10,11.

Methanogens produce CH4 via hydrogenotrophic (CO2 reduced to CH4 using H2), acetoclastic (acetate disproportionation used to form CH4), and methylotrophic (methyl groups of methylated compounds used to form CH4) pathways12. N2O can be produced during a variety of processes but mostly nitrification and denitrification, e.g. 13. CH4 and N2O have a global warming potential (GWP) that is, depending on the 20-year or 100-year time-horizon, 79.7 to 27.0 for CH4 or 273 for N2O times respectively more powerful than CO214. As such, the production and release of CH4 and N2O from marine sediments has the potential to reduce (offset) the climatic benefit of Corg sequestered through long-term (>100 years) Corg burial (i.e., blue carbon)9,11. Similarly, an uptake of CH4 and N2O has the potential to enhance the climate benefit of vegetated coastal communities15,16,17. Although CH4 and N2O are released during the short-term decomposition of organic matter that is not buried, there long-term influence on the atmosphere supports a comparison to long-term Corg burial.

Three studies have used either CH4 fluxes18 or CH4 and N2O fluxes11,19 in seagrass communities to estimate offsets for individual sites that range from zero to 10% (GWP100), but none used measured long-term (>100 year) Corg burial rates. Two studies used seagrass CH4 fluxes to estimate global offsets that range from 0.5 to 4.5% (GWP100), but both these studies used global seagrass organic carbon accumulation estimates20,21, which do not represent long-term Corg burial rates. Considering the implications for policies and practice incorporating blue carbon habitat conservation and restoration as a natural strategy for CO2 removal, there is a clear need to re-assess the climate benefit of seagrass blue carbon based on a synthesis of seagrass CH4 and N2O fluxes and long-term Corg burial rates.

In this study we present a meta-analysis synthesizing seagrass long-term Corg burial rates (and Corg accumulation rates for comparison) and CH4 and N2O fluxes (Fig. 1 and Supplementary Tables 1–4), and use the combined CO2-equivalent CH4 and N2O fluxes to estimate the reduced climate benefit (offsets) to seagrass blue carbon (1) for individual seagrass species (2) for the Australian region, and (3) globally using the mean and median of all rates. We further present new water-air CH4 and N2O fluxes for one study site in Australia (Wallagoot Lake), which increased the limited number of seagrass N2O fluxes, added N2O fluxes for a new seagrass species, doubled the number of seagrass water-air N2O fluxes, and also added an additional seagrass CH4 flux. We found CH4 and N2O fluxes offset on average around 7% to 35% of the global climate benefits of seagrass blue carbon.

Fig. 1: Global distribution of CH4 and N2O fluxes, Corg accumulation and long-term Corg burial rates measured in seagrass communities.
figure 1

Seagrass areas (green areas) adapted from UNEP-WCMC, 2021 (https://data.unep-wcmc.org/datasets/7; data from Supplementary Tables 1–4). Long-term (>100 years).

Full size image

Results

Methane and nitrous oxide water-air fluxes in a Ruppia megacarpa community

Over the diel cycle (20 h), water overlying the R. megacarpa community in Wallagoot Lake had an average temperature and salinity of 25.8 ± 2.2 °C and 29.5 ± 0.2, respectively. The CH4 concentration varied between 50.4 and 86.2 nmol l−1 without following a diel pattern. In parallel, the N2O concentration slightly increased from 4.5 to 5.6 nmol l−1 in the first 4 h of the survey, thereafter gradually decreasing back to the initial concentration. The R. megacarpa community was a source of CH4 and a sink of N2O to the atmosphere. Water-air CH4 fluxes ranged from 16.8 to 57.3 µmol CH4 m−2 day−1 with a mean of 32.5 ± 0.4 µmol CH4 m−2 day−1 (±standard error (SE), n = 1202) (Supplementary Fig. 1). Water-air N2O fluxes ranged from −2.0 to −0.4 µmol N2O m−2 day−1 with a mean of −0.9 ± 0.01 µmol N2O m−2 day−1 (±SE, n = 1202) (Supplementary Figure 1).

Seagrass organic carbon accumulation and burial rates

Global seagrass community Corg accumulation rates ranged from 1.3 to 175.3 g C m−2 year−1 with a mean (±SE) of 41.4 ± 7.4 g C m−2 year−1 and a median (IQR) of 20.7 (53.8) g C m−2 year−1 (Fig. 2 and Supplementary Table 1). Global seagrass community long-term Corg burial rates (>100 years) ranged from 1.7 to 112.9 g C m−2 year−1 with a mean (±SE) of 25.1 ± 5.3 g C m−2 year−1 and a median (IQR) of 11.3 (20.2) g C m−2 year−1 (Fig. 2 and Supplementary Table 2). Although the mean and median Corg accumulation rates were higher than mean and median Corg burial rates they were not significantly different (p = 0.78). 14C measured long-term burial rates (mean ± SE 24.5 ± 6.7 g C m−2 year−1; median (IQR) 9.0 (16.8) g C m−2 year−1) were also not significantly different (p = 0.94) than 210Pb measured rates (mean ± SE 27.1 ± 6.6 g C m−2 year−1; median (IQR) 24.4 (25.3) g C m−2 year−1).

Fig. 2: Boxplots CH4 and N2O fluxes, Corg accumulation and long-term Corg burial rates.
figure 2

Boxplots showing median, mean (red circle), 25th and 75 percentiles (boxes), 10th and 90th percentiles (bars) and data points (circles) of CH4 (n = 35; green) and N2O fluxes (n = 9; pink) and Corg accumulation (acc.) (n = 36; blue) and long-term (>100 years) Corg burial rates (n = 33; blue) in seagrass communities (Supplementary Tables 1–4).

Full size image

Global seagrass methane and nitrous oxide fluxes

Global seagrass community CH4 fluxes ranged from −27.8 to 401.3 µmol CH4 m−2 day−1 with a mean (±SE) of 79.0 ± 16.8 µmol CH4 m−2 day−1 and a median (IQR) of 48.2 (96.3) µmol CH4 m−2 day−1 (Fig. 2 and Supplementary Table 3). Mean (±SE) (103.6 ± 39.0 µmol CH4 m−2 day−1) and median (IQR) (67.4 (135.2) µmol CH4 m−2 day−1) water-air CH4 fluxes were higher than mean (70.5 ± 19.5 µmol CH4 m−2 day−1) and median (IQR) (46.8 (97.6) µmol CH4 m−2 day−1) sediment-water CH4 fluxes, but they were not significantly different (p = 0.34). Seagrass community N2O fluxes ranged from −7.2 to 3.7 µmol N2O m−2 day−1 with a mean (±SE) of −1.4 ± 1.0 µmol N2O m−2 day−1 and a median (IQR) of −0.9 (1.0) µmol N2O m−2 day−1 (Fig. 2 and Supplementary Table 4). Mean (±SE) water-air N2O fluxes (−1.2 ± 0.3 µmol N2O m−2 day−1) were similar to mean sediment-water N2O fluxes (−1.4 ± 1.3 µmol N2O m−2 day−1). In contrast, median (IQR) N2O water-air uptakes (−1.2 (0.3) µmol N2O m−2 day−1) were higher than median N2O sediment-water uptakes (−0.6 (2.7) µmol N2O m−2 day−1).

Reduced climate benefit of seagrass communities

Measurements of CH4 and N2O fluxes and long-term Corg burial rates were available for only three seagrass species (Supplementary Tables 2–4). All three species were a source of CH4 and a sink for N2O. Zostera sp. and Halophila sp. were a CH4 + N2O-CO2-equivalent (CO2e) source when using 20-year global warming potential (GWP20) and a CH4 + N2O-CO2e sink when using the 100-year global warming potential (GWP100), and Posidonia sp. was a CH4 + N2O-CO2-equivalent (CO2e) source (GWP20,100). All three species were net CO2e sinks when long-term Corg burial was included. The mean seagrass community CH4 + N2O-CO2e sources reduces the climate benefit of Corg long-term burial 15.8 to 58.3% (GWP20) (average of means 34.7%) across the three seagrass species (median 25.8 to 89.9% (average of medians 54.0%) (Table 1). CO2e offsets were much lower using GWP100 reducing the climate benefit of Corg long-term burial from −16.9 to 18.5% (average of means 1.0%) across the four seagrass species (median −6.5 to 28.6% (average of medians 7.1%)).

Table 1 Seagrass long-term burial rates, CH4 and N2O fluxes, net climate benefit and offsets.
Full size table

All seagrass communities in Australia were a source of CH4, except for one site (Supplementary Table 3), and a sink for N2O, except for one site (Supplementary Table 4), and an overall CH4 + N2O-CO2e source. Seagrasses in the Australian region were a net CO2e sink when long-term Corg burial was included. Using the GWP20, the mean Australian seagrass community CO2e source reduces the climate benefit of Corg long-term burial by 41.3% (median 33.9%) (Table 1). Seagrass CH4 + N2O-CO2e offsets in Australia were much lower using the GWP100 with a mean reduction of the climate benefit of Corg long-term burial of 1.1% (median 2.1%).

Globally, seagrass communities were a source of CH4, except for one site (Supplementary Table 3), and seven of the nine seagrass communities were a sink for N2O (Supplementary Table 4), and an overall CH4 + N2O-CO2e source. Seagrasses globally were a net CO2e sink when long-term Corg burial was included. Using the GWP20, the mean global seagrass CH4 + N2O-CO2e source reduces the climate benefit of Corg long-term burial by 44.6% (median 33.9%). Seagrass CH4 + N2O-CO2e offsets were lower using the GWP100 with a mean reduction of the climate benefit of Corg long-term burial of 8.8% (median 7.0%).

Discussion

Burial, methane and nitrous oxide processes in seagrass communities

The range of seagrass long-term Corg burial rates (1.7 to 112.9 g Corg m−2 year−1) in this synthesis (Supplementary Table 2) is similar to previous global ranges of long-term burial rates (e.g., 9 to 122 g Corg m−2 year−1 5); However, our mean ± SE global long-term Corg burial rate (25.1 ± 5.3 g Corg m−2 year−1; n = 33) is much lower than a previous estimate of the mean long-term burial rate based on only P. oceanica from the Mediterranean (58 g Corg m−2 year−1 5); Similarly, our mean seagrass accumulation rate (41.4 ± 7.4 g Corg m−2 year−1; n = 36) is lower than previous estimates that have also included Corg mass balances (53 g Corg m−2 year−1 5); Using the most recent global seagrass area of 160,387 km2 22 gives mean and median global seagrass long-term Corg burial of 4.0 and 1.8 Tg Corg year−1 respectively, which represents a minor coastal sink compared to the 250 Tg C year−1 uptake of CO2 on the continental shelf23. Most importantly, our mean ± SE global seagrass Corg burial rate of 25.1 ± 5.3 g Corg m−2 year−1 is also much lower than mean global accumulation estimates that have been used in previous global offset estimates (e.g., 119 g Corg m−2 year−1 20; 138 ± 38 g Corg m−2 year−1 21. Global seagrass Corg accumulation estimates of refs. 20,21 do not reflect long-term Corg burial because many of the estimates are based on carbon mass balances, not long-term (<100 year) 14C and 210Pb Corg burial rates, and are biased toward the “matte”-forming P. oceanica7,8. Therefore, we argue that the long-term Corg burial rates we used here are a better estimate of blue carbon7 to compare to Green House Gas-CO2e fluxes and to estimate offsets.

Table 2 Published reduction in the net climate benefit of seagrass communities due to methane and nitrous oxide fluxes (% offsets).
Full size table

Our revised global mean ± SE seagrass community CH4 flux of 79.0 ± 16.8 µmol CH4 m−2 day−1 (median(IQR) 48.2 (96.3) µmol CH4 m−2 day−1; n = 35; Supplementary Table 3) in this study is lower than the most recent seagrass compilation (mean 112.5 ± 62.5 µmol CH4 m−2 day−1; median 81.2 (87.5) µmol CH4 m−2 day−1; n = 1824; and based upon almost twice as many measurements. Although seagrass CH4 emissions are lower than from mangroves and salt marshes24, seagrasses cover a larger surface area, globally. Using the most recent global seagrass area of 160,387 km2 22 gives mean and median global seagrass CH4 fluxes of 0.074 and 0.045 Tg CH4 year−1, respectively. Our revised global CH4 emission estimates are about 40% of the most recent estimate24, due to a lower mean and median CH4 flux rate and a smaller surface area of seagrass, and as such, represents a small contribution to global estuarine CH4 emissions (mean 0.9 Tg CH4 year−1 24.

Our compilation demonstrates that CH4 production in seagrass communities is widespread, with rates up to 401.3 µmol CH4 m−2 day−1 (Supplementary Table 3), despite seagrasses being found in marine waters where sediments are considered to have negligible CH4 emissions25. The release of methylated compounds from plants may fuel CH4 production; a recent study showed that seagrass CH4 production was sustained exclusively via methylated compounds produced and released by the plant18. The more common pathways of hydrogenotrophic and acetoclastic CH4 production in anoxic sediment were undetected in seagrasses and were most likely outcompeted by sulfate-reducing bacteria18.

This is the first compilation of measured N2O fluxes for seagrass communities (Supplementary Table 4). Although26 included seagrass N2O fluxes in the estimation of coastal and inland water N2O emissions, a separate seagrass N2O flux was not presented. A recent synthesis of N2O rates for seagrasses27 were based on denitrification rates in seagrasses and N2O:N2 ratios and not on direct N2O flux measurements. Using the most recent global seagrass area22 gives mean and median global seagrass N2O uptakes of −0.004 and −0.002 Tg N2O year−1 respectively, which represents a minor sink in marine N2O emissions26,27. Previous global estimates have suggested that seagrasses are a source of N2O27, which highlights the problem with using N2O:N2 ratios and denitrification rates to estimate seagrass N2O fluxes.

Seven of the nine seagrass sites showed an uptake of N2O, with uptakes as high as −7.2 µmol N2O m−2 day−1 in the Noosa Estuary, Australia (Supplementary Table 4). This is similar to mangrove and estuarine systems that show an uptake of N2O when NO3− concentrations are below 5 µmol l−1 28,29. However, some of the N2O uptakes in the seagrasses were greater than N2O uptakes in mangroves15,29. This may reflect even lower water column NO3- concentrations at some seagrass sites. For example, NO3− concentrations in Wallis Lake and the Noosa and Maroochy estuaries were all below 1.0 µmol l−1 16,17. This combination of low NO3− concentrations and high rates of denitrification in seagrasses30,31,32 may result in N2O consumption via denitrification. Conversely, nutrient enrichment would increase N2O fluxes and reduce the climate benefit of seagrass communities.

Implications for the climate benefit of seagrass blue carbon

We present three approaches for estimating the mean and median reduction of the climate benefit of seagrass Corg long-term burial (offsets). Mean and median offsets were similar for the individual seagrass species and globally (Table 1). Although this might be expected since the individual seagrass species were calculated from a subset of the data used in the global offsets. The reduced climate benefit of the Australian region was also calculated from a subset of the global data, but the mean and median GWP20 offsets were much higher mostly due to lower long-term burial rates, but the GWP100 offsets were lower due to the N2O-CO2e sink reducing the CH4-CO2e source. As such, the resolved geographic variability in long-term burial rates and greenhouse gas fluxes has important implications for evaluating the climate benefit of seagrass across different regions. Offsets calculated using the GWP100 were much lower than those calculated using GWP20 because the CH4 GWP100 multiplier (27.0) is much lower than the CH4 GWP20 multiplier (79.7), while the N2O multiplier is the same for GWP20 and GWP100. By reducing the GWP100 CH4-CO2e source it allows the GWP100 N2O-CO2e sink to greatly reduce the combined GWP100 CO2e CH4 and N2O reduction of the climate benefit of long-term seagrass Corg burial.

Despite accounting for N2O uptakes in seagrass (i.e., increasing the climate benefit seagrass Corg long-term burial) the mean global seagrass CH4 and N2O offset (33.4% GWP20) is still higher than the global CH4 offset estimated for mangroves (20.5% GWP20 9). Although CH4 fluxes are higher in mangroves (0 to 2127.2 µmol CH4 m−2 day−1), the lower mangrove offset reflects the much higher Corg burial rates in mangroves (56.6 to 651.0 g C m−2 year−1 9; compared to seagrass communities (Supplementary Table 2). The mangrove CH4 offset would also be further reduced, particularly for GWP100, if N2O uptake, which is common in mangrove waterways15, were included.

Mean offsets for the individual seagrass species from our synthesis are much higher than previous offset estimates (Table 2). The higher offsets in this study were mostly due to the use of long-term Corg burial rates, compared to short-term Corg accumulation rates being used in previous studies. Our mean global offsets (GWP100) in seagrasses are only a little higher than two previous global offset estimates20,21; Table 2). However, this similarity is a coincidence, and not an agreement, due to the high Corg accumulation rates used in the previous offset estimates and the N2O fluxes used in our estimates, which cancel each other out when using GWP100. The difference between previously published offsets and our offsets due to different combinations of Corg long-term burial rates, and CH4 and N2O fluxes, highlights importance of measuring all three parameters when assessing the net climate benefit (Blue Carbon) of seagrass communities (Table 1), and when deciding which seagrass communities to include in restoration projects. We argue that our global seagrass Green House Gas offsets are more realistic because we include a larger number of CH4 fluxes, we include N2O fluxes and we only include long-term Corg burial rates. In addition, we also provide the first estimates of seagrass global offsets using GWP20.

Uncertainties in offsets and future research

Our three offset estimates show that CH4 can reduce the climate benefit of seagrass communities for GWP20 and with a reduced effect for GWP100 and N2O fluxes can enhance the climate benefit of seagrass communities (GWP20,100). However, there are a number of uncertainties in our synthesis that should be considered, as they may result in either over- or under-estimations of seagrass offsets (Table 3). These uncertainties also highlight areas for further research.

Table 3 Uncertainties in the reduction of the net climate benefit of seagrass communities due to methane and nitrous oxide fluxes.
Full size table

The seagrass species approach for estimating offsets shows a large range in the possible seagrass blue carbon offset by CH4 and N2O fluxes. Posidonia had the highest offsets (GWP20 and GWP100) due to high CH4 fluxes. However, this high offset estimate may be due to the limited data availability, with Posidonia estimate based on only two measurements of CH4 fluxes and one N2O flux (Supplementary Tables 3 and 4). Halophila sp. showed a net CH4 + N2O-CO2e uptake (GWP100) due to low CH4 fluxes and high N2O fluxes, but this was based on only two measurements of N2O fluxes and one long-term burial rate. Globally, there was limited N2O flux data (9 sites globally). In particular, there were few, or no, measurements from Africa, South America and Southeast Asia (Fig. 1). There were also no simultaneous measurements of CH4 and N2O fluxes and Corg long-term burial from the one location which may result in a bias in our offsets33. Clearly, more measurements of CH4 and N2O fluxes and Corg long-term burial in the same seagrass communities, and across a range of seagrass species and locations, is required for accurate blue carbon assessments.

CH4 offsets were mostly based on seagrass sediment-water CH4 fluxes (26 sites) with less water-air CH4 fluxes (9 sites; Supplementary Table 3). Although it is unknown if sediment-water CH4 fluxes from seagrass communities reach the atmosphere, water column CH4 oxidation rates at salinities >6 are typically low34. This is consistent with negligible water column oxidation at the Wallis Lake site16. Mean and median seagrass water-air CH4 fluxes were also higher, but not significantly different (p = 0.34), than seagrass sediment-water CH4 fluxes suggesting sediment-water fluxes do not systematically over-estimate seagrass CH4 fluxes to the atmosphere. Similarly, some of the CH4 and N2O flux studies were only undertaken in one season, and some were undertaken multiple seasons. However, the was no significant difference between single season and multiple season CH4 fluxes (p = 0.64). As such, differences in the temporal scale of collection, does not appear to have added any bias into our analysis.

This study has only considered two (long-term burial, CH4 and N2O fluxes) of the many processes that determine the net climate benefit of seagrass communities. For example, we have not considered CaCO3 burial which greatly reduces, and in some cases completely removes, the climate benefit of seagrasses (e.g. refs. 16,35,36). Seagrasses also emit other climate warming greenhouse gases like isoprene37, which we have not considered. In addition, we have not compared the net climate benefit of seagrass communities to the net climate benefit of adjacent unvegetated sediments. Without comparing CH4 and N2O fluxes to adjacent bare sediments the CH4 and N2O fluxes in this synthesis are a maximum attributable to the seagrass community (see refs. 11,38). Similarly, without comparing long-term burial rates to adjacent bare sediments the long-term burial rates in this synthesis are a maximum attributable to the seagrass community.

An input of allochthonous Corg is major process not considered by this study. Although we only used 210Pb and 14C measured rates of seagrass Corg long-term burial across a broad range of seagrass species and locations, these estimates may still be too high because of an input allochthonous Corg, both recalcitrant and labile8,39. For example, up to 50% of the Corg in seagrass communities may be derived from non-seagrass sources, including macroalgae (e.g. refs. 40,41,42). In some seagrass communities that are net heterotrophic there can even be a net flux of CO2 due to an input of allochthonous organic carbon43. If we assume 50% of the seagrass Corg long-term burial is from non-seagrass sources, this would increase our global mean offsets from 33.4% GWP20 and 7.0% GWP100 to 66.8% GWP20 and 14.0% GWP100. As such, our seagrass offsets should be considered a lower estimate as an input of any allochthonous Corg would result in higher offsets. However, this assumes that the allochthonous Corg is not included as seagrass “Blue Carbon” but the CH4 and N2O produced during its decomposition are included. It could be argued that the CH4 and N2O fluxes associated with the allochthonous Corg should also not be included, and therefore the allochthonous component would not affect the offsets, unless it produces CH4 and N2O differently to the autochthonous Corg (Table 3). Similarly, Corg produced by the seagrass community, but transported offsite and buried is unaccounted for by the long-term burial rates measured in the seagrass community. However, when this material is buried offsite it will still result in the production or consumption of CH4 and N2O, which would also contribute to an offset. The local conditions of long-term burial and associated production or consumption of CH4 and N2O would determine if this results in an over- or under-estimate of the offset (Table 3).

The long-term burial rates may also be too low. Although there was no significant difference between the 14C and 210Pb measured long-term burial rates in this synthesis, 14C long-term burial rates are up to 1000’s of years old, and may be lower than seagrass long-term burial rates over 100’s of years (e.g. ref. 44), resulting in an over-estimate of offsets (Table 3). There are also timescale differences between the long-term burial rates and CH4 and N2O fluxes. Long-term burial is Corg that has been sequestered for >100 years compared to CH4 and N2O fluxes which are a recent measurement, typically only undertaken once or over one annual cycle (Supplementary Tables 3 and 4). It is unknown if the recently measured CH4 and N2O fluxes are representative of the seagrass community for the longer period over which Corg has been buried. In particular, recent disturbance of the seagrass community may change the CH4 and N2O fluxes. For example, warming and shading could result in higher seagrass CH4 fluxes20,45 and nitrogen enrichment may reduce the uptake of N2O in seagrasses17, both of which would result in an over-estimate of offsets (Table 1).

Our choice of global warming potential (i.e., IPCC 2021) also influences the net climate benefit offset estimates. For example had we chosen to use the sustained global warming potential (SGWP46) our global mean offset would have increased from 33.4% GWP20 and 7.0% GWP100 to 41.8% GWP20 and 14.8% GWP100. Despite a higher SGWP100 value of 349 for N2O uptakes, which would reduce the offset, the offset is still higher due to the higher SGWP100 value for CH4 (454).

In summary, this study shows that CH4 fluxes can reduce, and N2O fluxes can enhance, the climate benefit of seagrass blue carbon. We also highlight the importance of using long-term Corg burial rates when assessing the climate benefit of seagrass blue carbon. These findings will contribute to policies and practice incorporating blue carbon habitat conservation and restoration as a natural strategy for CO2 removal.

Materials and methods

Wallagoot Lake study area, experimental procedure, and calculations

Wallagoot Lake is a small slightly degraded temperate Intermittently Closed and Open Lake and Lagoon (ICOLL) on the southern New South Wales coast. R. megacarpa occurs in the shallow water (<2 m) around the edge of the whole lake. A 20-h time-series measuring continuous water CH4 and N2O concentrations over a R. megacarpa seagrass community (S36°47’39”; E 149°56’19”) was undertaken in summer 2016. Water was pumped from 30 cm water depth and passed through a shower-head equilibrator connected to a cavity-ring-down spectroscopy (Picarro, G2308) analyzer. CH4 and N2O concentrations were measured at ~1 Hz (precision at 1 min <7 ppb ± 0.05% of reading). Wind speed was measured hourly with a digital anemometer (Q1411, Dick Smith Electronics), as well as water temperature and salinity at 15 min intervals (Hydrolab HL4 Sonde, Aqualab).

The CH4 and N2O water-air flux (F) was estimated as F = k K0 (Cwater–Cair), where k is the gas transfer velocity (m day−1), K0 is the solubility coefficient (mol kg−1 atm−1) at the measured water temperature and salinity47,48, and Cwater and Cair are the partial pressures (µatm) of CH4 or N2O in the water and air, respectively49. Due to variability in k obtained by different empirical models50, three models were selected as adequate for the investigated area to determine the gas transfer velocity51,52,53, and the resulting fluxes were averaged.

Blue carbon offset calculations

Scopus and google scholar were searched for “seagrass” AND methane”, “seagrass AND nitrous oxide” and “seagrass AND burial”. Additional papers were identified from reading the papers found in the searches. Sediment-water (cores, benthic chambers) and water-air (open water, floating chambers) seagrass CH4 and N2O fluxes were included in the synthesis. Because we were interested in the seagrass CO2e offset to long-term (>100 years) seagrass Corg burial, only seagrass accumulation rates obtained using 210Pb and 14C dating were included7. Studies that used sediment accumulation rates from the literature were excluded (i.e. refs. 54,55). Multiple CH4 and N2O fluxes and long-term Corg burial rates for a given seagrass species at a given site were averaged (site mean). Temporal CH4 and N2O flux measurements (e.g., diel, seasonal) were also averaged.

Each of the seagrass Corg accumulation studies identified were further assessed to determine if they could be considered long-term (>100 years) seagrass Corg burial following the best practice outlined in ref. 8. To be considered long-term (>100 years) seagrass Corg burial for cores dated using 210Pb the study had to show 210Pb concentrations profiles so we could determine that there was no sediment mixing and Corg concentration profiles to determine if Corg was buried below the zone of rapid carbon remineralisation (i.e., where concentrations no longer decrease and become constant8); If an average of the Corg concentration profile was used, but the Corg concentration profile increased or was constant with depth this was considered acceptable. To be considered long-term (<100 years) seagrass Corg burial for cores dated using 14C Corg concentration profiles were required. All studies that did not meet the long-term burial criteria were assigned to Corg accumulation.

CH4 and N2O fluxes were converted to CO2 equivalents using the 20-year and 100-year GWP of each gas. The GWP20 and GWP100 for CH4 are 79.7 and 27 respectively, and the GWP20 and GWP100 for N2O is 27314. The reduced climate benefit of seagrass blue carbon was calculated three ways. Firstly, seagrass species offsets were calculated using the mean and median of long-term Corg burial rates and CH4 and N2O fluxes rates for all individual seagrass species for which all three parameters were available (Supplementary Tables 2–4 and Fig. 1). Secondly, the Australian seagrass offset was calculated using the mean and median of all seagrass long-term Corg burial rates and CH4 and N2O fluxes for the Australian region (Supplementary Tables 2–4 and Fig. 1). This was done because Australia is the only region were there was sufficient data for this calculation. Thirdly, the global offset was calculated using the mean and median of all seagrass long-term Corg burial rates and CH4 and N2O fluxes (Supplementary Tables 2–4 and Fig. 1). Although we use the term “global” this doesn’t reflect the mean or median long-term Corg burial rates, CH4 and N2O fluxes and %offsets at any given location, as this will depend on site specific factors. Both the mean and median as the central statistic is presented as per24 to allow comparisons with past and future studies. The following six equations were used to calculate offsets from the mean and median of long-term Corg burial rates and CH4 and N2O fluxes:

$${{{{{\rm{Offset}}}}}}( \% )=({{{{{\rm{Total}}}}}}\,{{{{{{\rm{CO}}}}}}}_{2}{{{{{\rm{e}}}}}}\,{{{{{\rm{source}}}}}}/{{{{{{\rm{CO}}}}}}}_{2}{{{{{\rm{e}}}}}}\,{{{{{\rm{long}}}}}}\,{{{{{\rm{term}}}}}}\,{{{{{{\rm{C}}}}}}}_{{{{{{\rm{org}}}}}}}\,{{{{{\rm{burial}}}}}})\times 100 \%$$
(1)
$${{{{{\rm{Net}}}}}}\,{{{{{\rm{climate}}}}}}\,{{{{{\rm{benefit}}}}}}= {{{{{{\rm{C}}}}}}}_{{{{{{\rm{org}}}}}}}{{{{{{\rm{CO}}}}}}}_{2}{{{{{\rm{e}}}}}}\,{{{{{\rm{long}}}}}}\,{{{{{\rm{term}}}}}}\,{{{{{\rm{burial}}}}}}-{{{{{{\rm{CH}}}}}}}_{4}{{{{{{\rm{CO}}}}}}}_{2}{{{{{\rm{e}}}}}}\,{{{{{\rm{flux}}}}}}-{{{{{{\rm{N}}}}}}}_{2}{{{{{\rm{O}}}}}}{{{{{{\rm{CO}}}}}}}_{2}{{{{{\rm{e}}}}}}\,{{{{{\rm{flux}}}}}}$$
(2)
$${{{{{\rm{Total}}}}}}\,{{{{{{\rm{CO}}}}}}}_{2}{{{{{\rm{e}}}}}}\,{{{{{\rm{source}}}}}}={{{{{{\rm{CH}}}}}}}_{4}{{{{{{\rm{CO}}}}}}}_{2}{{{{{\rm{e}}}}}}\,{{{{{\rm{flux}}}}}}+{{{{{{\rm{N}}}}}}}_{2}{{{{{\rm{O}}}}}}{{{{{{\rm{CO}}}}}}}_{2}{{{{{\rm{e}}}}}}\,{{{{{\rm{flux}}}}}}$$
(3)
$${{{{{{\rm{CH}}}}}}}_{4} {{{{{{\rm{CO}}}}}}}_{2}{{{{{\rm{e}}}}}}\,{{{{{\rm{flux}}}}}}({{{{{\rm{g}}}}}}\,{{{{{{\rm{CO}}}}}}}_{2}\,{{{{{{\rm{m}}}}}}}^{-2}\,{{{{{{\rm{year}}}}}}}^{-1})= {{{{{{\rm{CH}}}}}}}_{4}\,{{{{{\rm{flux}}}}}}(\upmu {{{{{\rm{mol}}}}}}\,{{{{{{\rm{CH}}}}}}}_{4}{{{{{{\rm{m}}}}}}}^{-2}\,{{{{{{\rm{day}}}}}}}^{-1})\ \times {10}^{6}\times 365\times 16\times 79.7({{{{{{\rm{GWP}}}}}}}_{20})\;{{{{{\rm{or}}}}}}\times 27.0({{{{{{\rm{GWP}}}}}}}_{100})$$
(4)
$${{{{{{\rm{N}}}}}}}_{2}{{{{{\rm{O}}}}}} {{{{{{\rm{CO}}}}}}}_{2}{{{{{\rm{e}}}}}}\,{{{{{\rm{flux}}}}}}({{{{{\rm{g}}}}}}\,{{{{{{\rm{CO}}}}}}}_{2}\,{{{{{{\rm{m}}}}}}}^{-2}\,{{{{{{\rm{year}}}}}}}^{-1})= {{{{{{\rm{N}}}}}}}_{2}{{{{{\rm{O}}}}}}\,{{{{{\rm{flux}}}}}}(\upmu {{{{{\rm{mol}}}}}}\,{{{{{{\rm{N}}}}}}}_{2}{{{{{\rm{O}}}}}}\,{{{{{{\rm{m}}}}}}}^{-2}\,{{{{{{\rm{day}}}}}}}^{-1}) \times {10}^{6} \times 365\times 44\times 273({{{{{{\rm{GWP}}}}}}}_{20,100})$$
(5)
$${{{{{{\rm{C}}}}}}}_{{{{{{\rm{org}}}}}}} {{{{{{\rm{CO}}}}}}}_{2}{{{{{\rm{e}}}}}}\,{{{{{\rm{long}}}}}}\,{{{{{\rm{term}}}}}}\,{{{{{\rm{burial}}}}}}({{{{{\rm{g}}}}}}\,{{{{{{\rm{CO}}}}}}}_{2}\,{{{{{{\rm{m}}}}}}}^{-2}{{{{{{\rm{year}}}}}}}^{-1}) = {{{{{\rm{g}}}}}}\,{{{{{{\rm{C}}}}}}}_{{{{{{\rm{org}}}}}}}\,{{{{{{\rm{m}}}}}}}^{-2}\,{{{{{{\rm{year}}}}}}}^{-1}\times 12/44$$
(6)

Statistical analysis

The Welch two sample t-test was used to test whether the means of two sample populations are different, assuming independent samples for unequal variances and sample sizes under normality. The t-tests were performed in R (R version 4.0.3 (2020-10-10)56.