Introduction

In 2009 the term Blue Carbon was coined to highlight the importance of the ocean and its ecosystems as global carbon sinks (Nellemann et al. 2009). In particular, salt marshes, mangrove forests and seagrass meadows, were identified as key carbon sinks due to their capacity to accrete allochthonous and autochthonous carbon within their associated sediments (Duarte et al. 2013). Initially, macroalgae ecosystems such as kelp forests and fucoid beds were excluded from Blue Carbon frameworks as most do not accrete carbon locally, due to the rocky substrates they are often associated with, but rather export carbon as detritus (Hill et al. 2015). However, recent studies have argued for the inclusion of natural seaweed habitats because of their high rates of productivity and detrital export, considerable spatial extent, and their function as carbon donors in coastal ecosystems (Krause-Jensen and Duarte 2016; Pessarrodona et al. 2018; Duarte et al. 2022; Filbee-Dexter et al. 2022; Pessarrodona et al. 2022; Smale et al. 2022). As such, these ecosystems provide substantial nature-based solutions for curbing the rate of global warming (United Nations Environment Programme 2023). This means that, alongside reductions in greenhouse gas emissions, the restoration and protection of natural carbon sinks has been identified as an important component of mitigating anthropogenic climate change (IPCC 2022). Concurrently, the seaweed farming industry is expanding rapidly across the globe to meet the increasing demand for seaweed-derived products (FAO 2020). Carbon sequestration from seaweed farming is being increasingly proposed as a co-benefit that may contribute to climate change mitigation, although direct measurements of carbon uptake and release remain lacking which presents issues for decision-making around licenses, community engagement and developing carbon credit markets (Pessarrodona et al. 2024).

With regards to climate change mitigation, some representatives from industry, NGOs and academia have proposed large-scale cultivation of seaweeds (i.e. offshore afforestation) specifically for sinking biomass to the deep ocean to increase sequestration rates (as recently evaluated by Ould and Caldwell 2022; Ross et al. 2022)). However, the general consensus is that the evidence base, technological developments and ethical considerations do not support such an approach (Boyd et al. 2022; Ricart et al. 2022; Ross et al. 2022). Rather, where seaweed is grown for other purposes (e.g., human food, alginates, low-carbon products), some biomass is passively released into marine environment during the cultivation process. For cultivated kelp species, biomass may be released via chronic erosion, occurring mostly at the distal tips of the blades, and/or the dislodgement of whole (or partial) plants (Pedersen et al. 2020). Dislodgement and erosion rates are positively correlated with, often seasonal, increases in temperature and low nitrate levels, which lead to tissue degradation and therefore an increased likelihood of breakage (Rothäusler et al. 2009; Krumhansl and Scheibling 2012; Yorke et al. 2013). Heavy wave action associated with storm activity can also lead to high rates of dislodgement (Smale and Vance 2016). Meanwhile, epiphytic loading by bryozoans and other taxa can cause extensive defoliation, limit nutrient uptake by kelp and make tissue more brittle, again increasing susceptibility to breakage and erosion (Hurd et al. 1994, 2000; Krumhansl and Scheibling 2011). Grazing wounds, superficial excavations and perforations, also increase kelp erosion and dislodgement as weakened tissue becomes more vulnerable to breakage (Krumhansl and Sheibling 2011). The combination of epiphyte and grazer pressure can cause high rates of erosion and biomass release into the environment (Krumhansl and Scheibling 2011). As such, to optimise crop yield and quality, cultivated biomass is harvested before the onset of destructive biofouling and plant senescence (Wilding et al. 2021).

Providing that kelp-derived carbon lost via erosion and dislodgement is not consumed, degraded or remineralised, it has the potential to be sequestered if transferred to and stored within carbon sink habitats (Krause-Jensen and Duarte 2016). To mitigate climate change and to be recognised as ‘additional’ carbon in crediting schemes, this carbon needs to be buried for geologically relevant timescales (> 100 years) (GESAMP 2019; WWF 2019). For natural seaweed ecosystems, Krause-Jensen & Duarte (2016) estimated that ~ 15% of particulate organic carbon (POC) exported from algal habitats is sequestered in either the deep sea and continental shelf sediments combined. Under farm conditions, the release of seaweed biomass (and therefore the amount of carbon potentially available for sequestration) is minimised to optimise product yield and quality (Fieler et al. 2021). However, different farming practices, from hatchery to harvest, affect the amount of biomass released into the environment and therefore potentially available for carbon sequestration (Forbord et al. 2020; Boderskov et al. 2021; Fieler et al. 2021; Matsson et al. 2021).

Estimates of the carbon sequestration potential of seaweed farming are highly variable and geographically biased, with the majority of studies originating from Asia (Duarte et al., preprint). In general, estimates of carbon sequestration potential have been derived from rates of primary production and carbon tissue content, although a recent study showed that the amount of carbon held within sediments directly beneath seaweed farms was greater than within reference sediments (Duarte et al., preprint). In 2021, Asia was responsible for ~ 99.6% of global seaweed aquaculture production, whilst the contribution from Europe was negligible, at ~ 0.1% (FAO 2023), indicative of the potential for growth of the seaweed industry across the region (Araújo et al. 2021). Seaweed farming in Europe is a rapidly growing industry driven by, in part, a prospective climate change mitigation strategy focussed on blue carbon. In light of geographical biases, the poor knowledge base underpinning blue carbon science and the emerging European industry, it is critical to address pressing knowledge gaps and to provide context-specific evidence. Here, we quantified carbon accumulation and loss from Saccharina latissima (sugar kelp), which is one of the most intensively farmed seaweed species in the north Atlantic and Pacific, due to its high growth rates and relative ease of culturing (Sæther et al. 2024) We examined carbon dynamics by monitoring growth, erosion and dislodgement rates through a growing season at a typical, small-scale low-trophic aquaculture site in the southwest of the UK. Such field-based measurements are needed to inform the carbon sequestration potential of kelp farming in Europe and thereby its contribution to climate change mitigation.

Material and methods

Study site

The study focused on cultivated Saccharina latissima (sugar kelp) and was conducted in Porthallow Bay, Cornwall, UK (50◦04’ N, 5◦04’ W) (Fig. 1A). The Bay is open and exposed to east and south-easterly fetch from the English Channel, although the Lizard Peninsula provides shelter from the prevailing south-westerly winds and swells from the Atlantic Ocean (see Fig. 1A). The farm site is situated ~ 500 m offshore, with multiple ~ 200 m header lines suspended ~ 1 m below the surface (Fig. 1B). Mytilus edulis (blue mussels) are predominantly farmed at the site (total farm footprint of ~ 16 ha), with ~ 2 ha of sugar kelp cultivated on dropper lines attached to pre-existing header lines (Fig. 1C). Beneath the farm, water depths vary from 5–15 m, with a ~ 5 m tidal range, and the seabed is comprised of a mix of rocky substrate, soft sediment and maerl gravel. In November 2022, twine seeded with sugar kelp sporophytes (material provided by Hortimare, Netherlands) was wound onto 200 braided dropper lines and then deployed across 2 ha of farm infrastructure. Cultivated kelp was then monitored for 5 months across the growing season (January-May 2023).

Fig. 1
figure 1

Study area indicating: A) the United Kingdom (UK); B) the location of the farm in Cornwall, UK; C) the position of the farm within Porthallow Bay with seaweed lines indicated in red; and D) a schematic of the infrastructure used for vertical sugar kelp cultivation (not to scale)

Full size image

Environmental data

To quantify key environmental variables during the experimental period, HOBO pendant loggers (64 k temperature/light sensors; UA-002–64) were deployed ~ 1 m below the water surface (i.e. at a comparable depth to the header lines) within the seaweed farm to measure temperature (°C) and light intensity (lux) every 15 min. Photoperiod was extracted for the same sampling period, at the appropriate latitude, using the R package meteor (Hijmans et al. 2023).

Experimental set-up

At the start of each month of observation, 5 ‘mature’ plants (total length > 15 cm) within a 1 m marked section of a labelled dropper line were tagged using a combination of cable-tie and surgical tubing around the stipe. We focused on ‘mature’ plants as the vast majority of individuals were from the initial cohort of seeded plants, being similar in length and development time, although some ‘sub-canopy’ small plants were also present. Using a modified version of the hole-punch method, holes were excised at 10 and 15 cm from the base of each blade (Parke 1948)—the second hole was used to identify growth outside of the meristematic tissue area – and the length of the blade measured and recorded. At the end of the month (see Table S1 for exact sampling intervals), the position of the holes relative to the base and distal tip of the blade was recorded to estimate rates of elongation and loss of material (see Fig. S1). Dislodgement was determined from the number of lost tagged plants each month. Moreover, the density of mature plants was quantified by enumerating individuals within a 20 cm section of the dropper line. This approach to measuring growth, erosion and dislodgement was repeated on four dropper lines distributed across the seaweed cultivation area (n = 20 tagged plants per month).

At the end of each month, tagged kelp were retrieved and returned to the laboratory for further biometric analysis. Subsequently, new plants were selected (5 plants on each of 4 droppers), tagged and hole-punched to capture growth, erosion and dislodgement rates for the following month. On return to the laboratory, three sections of blade tissue (basal, mid and distal, 5 cm in length) as well as the stipe-holdfast complex were excised from each plant and weighed to obtain fresh weight (FW). Samples were then dried (at 60 °C for 48 h) and reweighed to obtain dry weight (DW) values, the average ratio of FW to DW was then calculated for each month (Table S2).

Conversion to carbon dynamics, scaling-up and statistical analysis

In cultivated kelp, biomass accumulation primarily occurs through both elongation of the blade and allocation of biomass to the stipe-holdfast complex. In natural kelp populations, biomass accumulation also occurs via a ‘thickening’ of blade tissues as sugars and other storage compounds are laid down throughout summer in advance of the following winter/spring growth season (Bartsch et al. 2008). In the cultivation context, plants are generally harvested whilst still in the fast-growth phase and so this process of biomass accumulation via blade thickening was considered to be negligible and not directly measured here. Biomass accumulation related to blade elongation was calculated by converting the daily elongation rate to a biomass value obtained from the basal segment of the blade tissue. Biomass accumulation related to the stipe-holdfast complex was estimated by calculating the difference in average biomass of this structure between consecutive months. Biomass loss via erosion was estimated by converting the daily loss rate of tissue to a biomass value obtained from the distal segment of the blade tissue. Rates of biomass accumulation (via blade elongation and stipe-holdfast growth) and biomass loss (via erosion) were multiplied by the density of plants per dropper to yield a process rate per month. Dislodgement rates were converted to biomass release values by multiplying average plant biomass each month by the number of plants lost (i.e. dislodgement rate x density). The FW:DW described above and the carbon content of sugar kelp derived from the literature (29.75%; Brady-Campbell et al. 1984; Gevaert et al. 2001) was then applied to elongation, growth, erosion, dislodgment and biomass values to estimate carbon accumulation and loss rates. Conservative first-order estimates of the amount of carbon captured and released by the farm per hectare (t C ha−1 y−1) were calculated using the monthly averages of dropper carbon accumulation and loss rates multiplied by the number of droppers across the farm (200) which corresponded to its size (2 ha). Processes are presented as annual rates although in actuality the growth season of sugar kelp is ~ 6 months and for the remaining half-year seaweed biomass is currently not cultivated.

We formally examined statistical differences between droppers and months using univariate permutational analysis of variance (PERMANOVA), which was carried out in PRIMER v7 using the PERMANOVA add-on (Clarke and Gorley 2015). For response variables measured at the plant level (i.e. elongation and erosion rates, plant biomass) we used a full model with ‘month’ as a fixed factor and ‘dropper’ as a random nested factor, to examine variability across the farm. Where response variables were calculated at the dropper level (i.e., stipe-holdfast growth, dislodgement, density, total carbon accumulation and loss) we used a simpler model with ‘month’ as a single fixed factor. In all cases, permutations (9999 under an unrestricted model) were based on Euclidean distances between untransformed data. Monte Carlo simulations were used when permutations were low to determine significance to mitigate the effect of small sample sizes. Where a significant effect of the fixed factor was detected (p < 0.05), pairwise comparisons were conducted to determine which months differed from one another. Where it was not possible to collect monthly data due to adverse weather conditions, an estimate was derived from the means of the month prior and following. These estimates were excluded from statistical analyses but included in figures and descriptions of trends. All summary statistics were carried out in R v4.3.1 using the package dplyr (R Core Team 2023; Wickham et al. 2023).

Results

Environmental data

Environmental data collected from 1 m below the water surface at the seaweed farm in Porthallow Bay, showed that average daily (07:00–19:00) irradiance increased from 9.65 µmol photons m−2 s−1 in January to its peak in May at 312.73 µmol photons m−2 s−1 (Fig. 2A). Photoperiod increased from ~ 8 h in January to ~ 16 h in May (Fig. 2A). Average daily sea temperatures remained stable at around 10 °C from January to April, and optimal temperatures for plant growth (< 15 °C) were not exceeded until June, by which point sampling had concluded (Fig. 2B).

Fig. 2
figure 2

The A) daily photoperiod (left y-axis) and average daytime (07:00–19:00) light availability (lux) ± S.D. (right y-axis) and B) daily average temperature ± S.D., recorded at 1 m below the water surface within the seaweed farm from January to July 2023

Full size image

Rates of biomass growth and loss

At the start of the monitoring period, sugar kelp individuals were ~ 41 cm in length and grew, in terms of blade elongation, at an average rate of 1.26 ± 0.35 cm day−1 (Fig. 3A). Compared to January–February, growth rates increased ~ 1.5 and ~ 1.8 fold in February–March and March–April, respectively. By April–May, growth rates began to decline. Growth rates differed significantly between months, with no effect of dropper line (Table S3). The average growth, recorded as fresh weight biomass accumulation, of the holdfast-stipe complex almost quadrupled from January (0.028 ± 0.002 g FW day−1) to April (0.076 ± 0.037 g FW day−1) (Fig. 3B). By April–May, growth declined to 0.009 ± 0.022 g FW day−1, slower than the initial rates recorded in January–February. While there was an overall significant difference between months, variability increased throughout the monitoring period, and growth of the stipe-holdfast complex did not differ between March and April (Fig. 3B; Table S4).

Fig. 3
figure 3

The monthly A) blade elongation, B) holdfast-stipe (Hf-St) growth, C) erosion rates of cultivated S. latissima, D) percentage dislodgement, E) biomass of plants and F) density per 20 cm of cultivation line. The box plots show medians, 1st and 3rd quartiles, with whiskers showing maximum and minimum values, and dots representing outliers. Diamonds indicate mean values, with dashed-line diamonds representing the modelled mean for missing months (see method for calculation), whilst ‘na’ indicates unavailable data. The level of significance is indicated with asterisks, where ‘*’ is p < 0.05, ‘**’ is p < 0.01 and ‘***’ is p < 0.001 and significance between months is denoted by different lettering

Full size image

Whilst average growth (elongation and holdfast-stipe complex growth) slowed in April–May, mean erosion peaked here at 0.80 ± 0.91 cm day−1 (Fig. 3C). Throughout the growing season erosion were consistently low, ranging from 0.48 ± 0.37 to 0.55 ± 0.36 cm day−1, with no significant difference identified between months (Table S3). Similarly, there was no significant difference in percentage dislodgement between months, with values highly variable throughout (Table S4). Nevertheless, a trend of decreasing dislodgement can be observed throughout the growing season, with average loss of ~ 20% in January–February and ~ 5% in April–May (Fig. 3D).

Biomass and density

Individual plant biomass increased significantly over the growing season, from an average 35.9 ± 12.4 g to 187.0 ± 150.0 g (Fig. 3E; Table S3). Similarly, a significant increase in plant density per 20 cm of dropper line was observed, from 8.67 ± 1.15 to 54.8 ± 10.9 mature plants (Fig. 3F; Table S4).

Carbon accumulation and loss

Using monthly average biomass and density data, standing stock (i.e. biomass × density) in terms of carbon was calculated for a dropper line. We recorded an increase from ~ 0.48 kg C dropper−1 in February–March to ~ 2.9 kg C dropper−1 in April–May (Fig. 4A). Carbon accumulation average, relating to blade elongation, increased steadily from 0.92 ± 0.25 g C day−1 dropper−1 in January–February and peaked at 14.90 ± 5.47 g C day−1 dropper−1 in April–May (Fig. 4B). Carbon accumulation via elongation was significantly different between months, although post-hoc comparisons showed that March–April and April–May were not significantly different from each other (Table S5). Carbon accumulation average, in terms of the holdfast-stipe complex, were initially low at 0.06 ± 0.01 g C day−1 dropper−1 (Fig. 4C). Accumulation average continued to increase to a peak of 0.66 ± 0.32 g C day−1 dropper−1 in March–April, after which they declined to 0.13 ± 0.34 g C day−1 dropper−1 in April–May. Here, rates differed significantly among months, except between March–April and April–May (Table S5).

Fig. 4
figure 4

The A) standing stock of carbon per dropper of cultivated S. latissima from January – May 2023, shown with carbon accumulation rates for B) blade elongation and C) holdfast-stipe (Hf-St) growth, carbon loss rates via D) erosion and E) dislodgement. The box plots show medians, 1st and 3rd quartiles, with whiskers showing maximum and minimum values, and dots representing outliers. Diamonds indicate mean values, with dashed-line diamonds representing the modelled mean for missing months (see method for calculation), whilst ‘na’ indicates unavailable data. The level of significance is indicated with asterisks, where ‘*’ is p < 0.05, ‘**’ is p < 0.01 and ‘***’ is p < 0.001 and significance between months is denoted by different lettering

Full size image

The average loss of carbon via chronic distal erosion was 0.26 ± 0.18 g C day−1 dropper−1 at the beginning of the monitoring period (Fig. 4D). Like carbon accumulation via elongation, this increased gradually to its peak, 7.40 ± 8.70 g C day−1 dropper−1, in April–May (Fig. 4D). Here, monthly rates were significantly different (Table S5). Carbon lost via dislodgement was highly variable across the growing season, with no significant difference between months (Table S5). The average rate of loss in February–March was 2.87 ± 3.67 g C day−1 dropper−1, almost doubling by April–May (Fig. 4E).

Annual accumulation and loss of carbon

Using conservative first-order estimates, farm-wide carbon accumulation and loss were calculated (Table 1). This estimate suggests that ~ 0.14 t C ha−1 y−1 are accumulated within the biomass of the sugar kelp farmed in Porthallow Bay, of which ~ 70% is lost, almost equally accounted for by erosion and dislodgement.

Table 1 Average dropper carbon accumulation and loss rates (to 2 decimal places) and standard deviation, scaled up across the study site, over the 2023 growing season (January-May)
Full size table

Discussion

Our field-based measurements of carbon accumulation and release at a small-scale commercial cultivation site informed representative estimates of the carbon sequestration potential of kelp farming. Biomass accumulation by cultivated sugar kelp at Porthallow Bay captured ~ 0.14 t C ha−1 y−1, with ~ 0.09 t C ha−1 y−1 released into the marine environment as detritus (i.e. POC). The proportion of carbon accumulated by farmed kelp subsequently released into the marine environment at the cultivation site (i.e. ~ 70%) was lower compared to both natural kelp beds (~ 82%) (Krumhansl and Scheibling 2012) and a recent comparable estimate for a sugar kelp farm in Northern Ireland (~ 84% NPP) (Dolliver and O’Connor 2022). With regards to natural kelp beds, the greater relative rates of carbon release may be related to the highly dynamic, shallow, wave and tide-exposed conditions that often characterise these habitats (Smale et al. 2022). The reasons for the higher rates of release at the farm site in Northern Ireland remain unclear but may include differences in tidal dynamics, biofouling, nutrient availability or seeding and monitoring techniques (Dolliver and O’Connor 2022). Based on these conservative first-order estimates, carbon capture and release at the farm in Porthallow Bay falls within the range reported across other European studies, which are generally lower than those reported in Asia (Table 2). By applying a previous estimate of ~ 15% of seaweed-derived POC that is released into the environment being sequestered (Krause-Jensen and Duarte 2016), we estimated a sequestration rate of ~ 0.014 t C ha−1 y−1 (71.98 g C m dropper−1 y−1) for the farm in Cornwall, southwest UK. Despite this relatively low sequestration rate, this study provides a starting point for predicting the carbon sequestration potential of a scaled-up seaweed farming industry. With 2,900,000 ha suitable for farming sugar kelp and Laminaria digitata (oarweed), in English waters alone (MMO 2019), the maximum hypothetical carbon sequestration potential is estimated at 40,600 t C y−1, or 149,002 t CO2e. Clearly, this potential sequestration rate does not take in to account the carbon emissions needed to establish and maintain farm infrastructure, nor to seed and harvest the seaweed lines, which would be needed for a full ‘forensic’ account of the net carbon gains. Rather, this is an estimate of the additional potential carbon co-benefit derived from pre-existing, operational farm sites.

Table 2 Estimated rates of carbon capture and loss by kelp farms in Europe and Asia
Full size table

Our study highlighted pronounced seasonality in core processes. Throughout the growing season, the light and temperature environment varied markedly, although conditions recorded remained favourable for growth of sugar kelp (Fortes and Lüning 1980; Peteiro et al. 2016). Growth rates related to both blade elongation and holdfast-stipe growth increased markedly from winter (i.e. January) to spring (i.e. March), reflecting seasonal patterns of growth observed for natural populations in the northeast Atlantic (Parke 1948; Sjøtun 1993). After March growth rates slowed significantly, which may be related to reduced nutrient availability, responses to photoperiod or endogenous circannual growth rhythms (Bartsch et al. 2008; Peteiro et al. 2016). In terms of release of biomass, erosion rates were consistently low throughout the growing season. This highlights the chronic nature of distal erosion, however, this is expected to increase as temperatures and biofouling increase, and nutrient availability decreases, which normally coincides with the end of the growing season (Krumhansl and Scheibling 2011, 2012). As such, chronic erosion of the distal tips of blades may represent a relatively unimportant mechanism of detritus release through the growing season in this case, but would be expected to increase in relative importance as temperatures and biofouling and grazer pressure increase through summer. Extending the monitoring period past harvest (April–May) would likely capture this pattern.

The decrease in relative dislodgement rates throughout the growing season could be attributed to larger plants, indicated by the increase in biomass throughout the same period, being more resistant to breakage or removal by wave action than the smaller, more fragile sporophytes present in January. This is likely related to greater holdfast size and attachment strength as well as reduced storminess from winter through to spring/summer. Like biomass, plant density increased from January to May. This, in combination with decreasing dislodgement, resulted in more densely populated lines of larger sugar kelp plants by May. Like erosion, dislodgement is expected to increase with the onset of warmer temperatures, reduced nutrient availability and increased biofouling causing kelp to become more brittle and prone to breakage (Rothäusler et al. 2009; Krumhansl and Scheibling 2012; Yorke et al. 2013). This pattern could again be captured by extending the monitoring period and delaying harvest. Postponing harvest would likely increase the amount of dislodgment and erosion-derived POC released into the marine environment. However, extending the growing season would reduce the quality and quantity of harvestable, marketable biomass for which the seaweed is primarily grown, not the additional, passive co-benefit of carbon sequestration (Fieler et al. 2021).

Whilst these results show that kelp farming captures carbon, the associated sequestration potential is relatively low and farming techniques and harvesting schedules are not optimised with carbon capture in mind. Therefore, carbon capture should be viewed as a passive co-benefit of kelp farming rather than a major driver. Instead, the focus should be directed towards other benefits associated with low trophic marine aquaculture that support climate change mitigation to a greater degree, like future food security, bioenergy production and the replacement of high-carbon footprint products (Krause et al. 2022; O’Dell et al. 2023; Troell et al. 2023). Importantly, the loss of kelp-derived carbon as detrital POC, as presented here, is just one pathway by which carbon can be exported and potentially sequestered. To achieve a more holistic understanding of the carbon sequestration potential of kelp farming, other processes and pathways need to be considered, such as the release and fate of dissolved organic carbon (DOC), the transport and persistence of kelp-derived carbon in marine environments, rates of long-term burial and storage, and the flux of carbon dioxide between the atmosphere and sea (Hurd et al. 2022, 2023). Indeed, a major limitation of the current study was the application of an assumed sequestration rate (~ 15% of exported carbon, as proposed by Krause-Jense and Duarte 2016), rather than applying direct measurements of carbon burial and long-term sequestration within the study region. Further work should focus on addressing the current uncertainty around these processes. We also recommend conducting similar surveys across a range of farm sites in the UK and Europe that encompass different environmental contexts, both monoculture and co-culture of seaweeds, and diverse cultivation practices to build a more robust, context-specific evidence base. With seaweed farming in its infancy in Europe, there is an opportunity to collate carbon capture and storage estimates using standardised methodologies, as has been done elsewhere (Duarte et al., preprint 2023). This would support the credible and measurable growth of the sector while upholding transparency regarding potential benefits. Additionally, further research should focus on conducting life cycle assessments (LCAs) on functional kelp farms to highlight where emissions can be avoided through process optimization and by replacing carbon-intensive products with low-carbon, kelp-derived products (Duarte et al. 2021). This is important for ensuring the sustainable development of the seaweed industry in Europe, as well as identifying the climate change mitigation potential of the sector and products outside of the growing season.

In conclusion, we recorded high relative rates of kelp-derived POC release, with ~ 70% of total biomass accumulation being exported to the marine environment. However, when assessed as the absolute amount of carbon likely to be locally stored, based on an estimated sequestration rate and the scale and stocking density of the cultivation site, the carbon sequestration potential of UK seaweed farming is relatively low. Even so, small-scale low trophic aquaculture may have multiple ‘stacked benefits’, including habitat provisioning and biodiversity maintenance (Corrigan et al. 2022, 2023) as well as additional carbon accumulation and release, which need to be considered holistically to reliably assess the risks and benefits associated with continued expansion of the seaweed industry. In the UK, natural kelp habitats extend across much of the shallow rocky coastline (Smale et al. 2013), where they exhibit high rates of primary productivity (Pessarrodona et al. 2018; Smale et al. 2020) and support elevated biodiversity (Teagle et al. 2018; Bué et al. 2020; King et al. 2021). The amount of carbon released into the marine environment via natural kelp populations will likely be several orders of magnitude greater than farmed kelp, even if scaled-up, and understanding and protecting carbon stocks and flows through these critical ecosystems should therefore remain a priority (Smale et al. 2018; Queirós et al. 2023).