Substantial blue carbon in overlooked Australian kelp forests

Recognition of the potential for vegetated coastal ecosystems to store and sequester carbon has led to their increasing inclusion into global carbon budgets and carbon offset schemes. However, kelp forests have been overlooked in evaluations of this ‘blue carbon’, which have been limited to tidal marshes, mangrove forests, and seagrass beds. We determined the continental-scale contribution to blue carbon from kelp forests in Australia using areal extent, biomass, and productivity measures from across the entire Great Southern Reef. We reveal that these kelp forests represent 10.3–22.7 Tg C and contribute 1.3–2.8 Tg C year−1 in sequestered production, amounting to more than 30% of total blue carbon stored and sequestered around the Australian continent, and ~ 3% of the total global blue carbon. We conclude that the omission of kelp forests from blue carbon assessments significantly underestimates the carbon storage and sequestration potential from vegetated coastal ecosystems globally.


Results and discussion
We calculated that Australian kelp forests store an aboveground biomass of 10.3-22.7 Tg C and contribute 1.3-2.8 Tg C year −1 in sequestered production (see Supplementary Data). This represents 11-13% of the total standing stock of blue carbon and 27-34% of the annual blue carbon sequestration reported for the Australian continent (Fig. 1). The total surface area of kelp forests in Australia is 3.2 to 7.1 Mha 20 . This is comparable to seagrass beds and 4 and 7 times higher than the extent of tidal marshes and mangrove forests, respectively (Table 1). Importantly, the distribution of kelp forests is largely disjunct from the other vegetated coastal ecosystems in Australia, with ~ 36% of kelp forests occurring in areas with no tidal marshes, mangrove forests or seagrass beds (Fig. 1a). This extensive ecosystem holds between 10 and 23 Tg C in aboveground biomass, which is similar to that of seagrass beds in Australia (Fig. 1b). We calculated that annual production per unit area of the dominant kelp species (Ecklonia radiata) on Australian reefs averages 3.9 Mg C ha −1 year −1 (± 0.9 SD) ( Table 1). Based on the current best-estimate of proportion of net primary production (NPP) to become sequestered through burial in deep ocean sediments or transport below the mixed layer in the deep sea 19 , this represents an average sequestration rate per unit area of kelp forest of 0.39 Mg C ha −1 year −1 (± 0.09 SD). Although a coarse estimate, this rate is within the range of carbon sequestration per unit area of tidal marshes and seagrass beds and lower than mangrove forests, but when extrapolated over the total habitat area in Australia it forms a significant proportion (31%) of the total blue carbon sequestration rate (Fig. 1c). Indeed, our calculation may even underestimate the  www.nature.com/scientificreports/ blue carbon contribution from Australian seaweed habitats substantially, as it does not include the extensive beds of Sargassum on tropical reefs in the north e.g., 22 or even more dominant fucoids 23 and deep beds of red algae 24 along the southern part of the continent, which, when combined with kelp forests, have been estimated to represent a total of 110 Tg C in aboveground biomass 11 .
A key challenge of including kelp forests in blue carbon assessments is that kelp carbon may end up in, and be accounted for indirectly in, estimates from other blue carbon ecosystems, because significant amounts of seaweed detritus (i.e., epiphytic and drifting seaweed) can be buried in tidal marshes, mangrove forests and seagrass beds 5,[25][26][27] . According to the estimate from Krause-Jensen and Duarte 19 , 11% [range = 4-18%] of all seaweed NPP is sequestered, and this percentage is almost entirely composed of NPP that reaches the deep ocean (> 1,000 m). Only 0.9% of NPP is buried on the entire continental shelf, such that an even much smaller proportion of this 0.9% would deposit in shelf habitats such as tidal marshes, mangrove forests and seagrass beds and be at risk of double counting. To ensure that kelp sequestration was not already accounted for as allochthonous seaweed derived carbon in estimates of carbon burial in other blue carbon systems (e.g., 21 ), our calculations conservatively excluded all burial on the continental shelf (0.9% of NPP) by using a sequestration rate of 10.1% NPP 19 . A more important challenge, however, is that the best estimates of the proportion of seaweed NPP sequestered in deep marine habitats are rudimentary. This represents a significant knowledge gap that must be closed to increase the confidence in estimates of kelp-derived blue carbon.
Conservation and restoration of blue carbon ecosystems are now being included in strategies to mitigate CO 2 emissions 3,6 . There is current debate surrounding the application of these blue carbon strategies to coastal ecosystems other than tidal marshes, mangrove forests and seagrass beds 6,11,15 . Rooted vegetated marine ecosystems share commonalities with terrestrial ecosystems because they sequester carbon through local burial in accreting sediments, which is similar to carbon burial on land, such as in soil 28 . In contrast, accounting for carbon that is mainly sequestered as allochthonous detritus in the deep ocean 12,19 is challenging for blue carbon policy because it is difficult to trace and to attribute a source to the site of storage, because of the risk of double-counting of material that ends up in other blue carbon ecosystems, and because sink habitats in the open ocean do not fall within national jurisdictions 6,8 . These are challenges for all blue carbon ecosystems, not only kelp forests. Export of detritus from tidal marshes, mangrove forests and seagrass beds is currently not considered to contribute to carbon sequestration, although detrital production from these habitats is likely substantial 29 . At the same time, the inability to trace allochthonous sources of buried carbon within tidal marshes, mangrove forests and seagrass beds currently prevents both accurate blue carbon accounting and allocations of carbon offset credits under many frameworks 6,30 . Regardless of the pervasive practical challenges around accounting for allochthonous carbon, kelp forests constitute important standing stocks of organic carbon and key components of organic carbon cycling in the coastal zone. Policy barriers and existing frameworks should not preclude their inclusion in our attempts to understand, quantify and manage carbon sources and sinks in the ocean.
Like most other blue carbon ecosystems, kelp forests follow a global trend of deterioration and decline, which is projected to worsen in the coming decades 7 . Australian kelp forests have been some of the worst impacted by human activities globally, and most regions of the Great Southern Reef have experienced kelp declines over the past decades 31 . Australia-wide ~ 1,000 km of coastline has been affected by kelp loss, totaling at least 140,187 ha ( Table 2). Drivers of loss include an extreme marine heatwave 32 , coastal pollution 33,34 , warming and drought 35 , sea urchin overgrazing from climate-driven changes in the Eastern Australia Current 36,37 , and the influx of tropical herbivores with warmer waters 38 . In total these declines represent 0.45 Tg C of lost standing stock and 0.06 Tg C of lost annual sequestration. Importantly, these recorded losses come from reefs in intensively researched areas, and it is possible that similar declines have occurred throughout less studied regions of this remote ecosystem.
When kelp forests are lost, most of their carbon (89%) 19 is incorporated into marine food webs and eventually remineralised as CO 2 , which can enter the atmosphere. As a result, potential changes in kelp forest area have important ramifications for carbon accounting strategies and predictions of carbon stocks in coming decades. By Table 2. Consequences of past (a) and future (b) losses of kelp forests in Australia on carbon standing stock and sequestration rates. 1 Wernberg et al. 32 , 2 Connell et al. 33 , 3 Carnell and Keough 35 , 4 Ling and Keane 37 , 5 Vergés et al. 38 , 6 Martínez et al. 39 . Calculations are provided in the Supplementary Data. www.nature.com/scientificreports/ 2100 Australia's Ecklonia radiata kelp forests are predicted to lose 49 to 71% of their current distribution under the RCP 2.6 and RCP 6.0 CO 2 emission scenarios, respectively 39 . Even under the most optimistic scenario (RCP 2.6), this implies a loss of 6% of total blue carbon stock and a 15% loss of blue carbon sequestration rates for Australian vegetated coastal ecosystems (Table 2). Kelp forest management and restoration programs show potential to revert this alarming trajectory [40][41][42] . Restoration and proactive management actions could help minimize increased CO 2 emissions from loss of standing stock and maintain valuable carbon sequestration rates from kelp forests, including along Australia's Great Southrn Reef. In order to scale up these national estimates to a global level, higher quality data on the areal extent and standing stock, as well as production, export and burial rates of kelps, such as those that exist for Australian kelp forests, are needed. Comprehensive and accurate estimates of blue carbon at large scales are critical for the success of blue carbon mitigation strategies and must include kelp forests if they are to fully capture the intense carbon storage and sequestration potential of the coastal zone.

Methods
Kelp forest area was determined from suitable reef area and bounded by a lower depth limit of 30 m 20 . This represents a conservative estimate because kelps are often found to 45 m depth or more in several places along the Great Southern Reef 43 45 and from density data collected from 558 plots spread across New South Wales, Tasmania, South Australia and Western Australia (3 locations × 3 sites × 5-6 quadrats in each state) (Wernberg unpublished data). We calculated net primary productivity using 1,577 individual plant growth rates 31 and 558 plots of kelp densities from across Australia. Carbon content in kelp tissue was assumed to be 30% of dry weight 46 . Carbon production rates were calculated using average net primary production measured from 7 separate tagging field studies across 7° longitude of coast 31 . We compared the contribution of kelp forests to other blue carbon habitats in Australia using data from Serrano et al. 21 .
We calculated historic blue carbon loss using estimates of the areal extent of kelp loss along the Great Southern Reef 32,33,35,37,38 . For time series data we averaged kelp abundance from the first 3 records and last 3 records (where available, see Supplementary Data). We also estimated future losses using the areal extent of range shifts modeled for the Great Southern Reef under different CO 2 emission scenarios 39 . For studies not reporting the actual area of reef lost, we estimated reef area from coastline length using the average coastline to reef ratio reported from other regions of the Great Southern Reef 20,32 . We calculated the impacts of these events on standing stock and sequestration rates using average per area estimates for the entire Great Southern Reef (Table 1). Source data and calculations are provided as a Supplementary Data file.

Data availability
A Supplementary Data file containing the raw data and calculations presented in the figures and tables is provided. Additional information can be obtained from the authors.