Navigating sustainability and health trade-offs in global seafood systems

Seafood is expected to play a key role in improving access to healthy diets while providing food products with relatively low rates of greenhouse gas emissions. However, both nutrients and carbon footprints vary among species and production methods, and seafood consumption is further influenced by price and consumer preference, such that it is unclear which species are best placed to provide low-emissions nutritious seafood. Here, we use seafood production data to assess the nutritional value, carbon emissions, sustainability, affordability, and availability of seafood available to UK consumers. Globally, most seafood products are more nutritious and emit lower greenhouse gases than terrestrial animal-source foods, particularly small pelagic fishes and bivalves that contributed to recommended intakes for 3–4 essential dietary nutrients at the lowest emissions. For seafood products relevant to UK markets and consumers, Atlantic mackerel had the highest availability (i.e. landings) of all wild-caught UK seafood and lowest carbon footprint of all finfish, with one fillet portion exceeding recommended intakes of three nutrients (selenium, vitamins B12 and D). We found that price and sustainability of UK seafood, both factors in consumer demand, had considerable trade-offs with nutrients, carbon footprint, and availability. Farmed salmon, for example, were produced in large volumes but were relatively more expensive than other seafood, whereas highly nutritious, low-emissions farmed mussels had limited production volumes. The UK’s seafood system is therefore not currently optimised to produce nutritious, low-emissions seafood in large amounts. Policies that promote local consumption of affordable species already produced in high volumes, such as mackerel, could improve intakes of nutrients that are deficient in the UK population at relatively low environmental cost.


Introduction
Food systems must be transformed if countries are to achieve net-zero greenhouse gas emissions targets by 2050 (Clark et al 2020, Rockström et al 2020, Halpern et al 2022, while also addressing growing malnutrition by improving access to healthy diets (Haddad et al 2016). Most animal-source foods, particularly livestock, have substantially higher greenhouse gas emissions than plant-source foods Clark 2014, Xu et al 2021), such that large-scale dietary shifts towards plants could substantially reduce food system emissions (Crippa et al 2021). However, animal-source foods provide concentrated, bioavailable sources of important dietary nutrients (calcium, selenium, fatty acids), some of which are not available in plant-source foods (e.g. vitamins B12, D) (Miller et al 2022), and deliver positive health outcomes for vulnerable populations, such as young children (Headey et al 2018). Transitioning towards low-emissions food systems while protecting access to healthy diets thus remains a significant global challenge (Rockström et al 2020), requiring analyses that assess both the nutritional value and environmental impact of diverse food products (Clark et al 2022).
Aquatic animals are increasingly recognized by the research community as nutritious animal-source foods that are critical to food and nutrition security (Belton andThilsted 2014, Hicks et al 2019), produced for (relatively) low greenhouse gas emissions (Hallström et al 2019, Koehn et al 2022, and have potential to sustainably contribute to growing global food demand (Béné et al 2015, Costello et al 2020. Seafood is a rich source of protein and essential micronutrients, produced locally (Thilsted et al 2016) and traded globally (Gephart and Pace 2015) via capture fisheries and aquaculture that are both expected to have key roles in transitioning towards sustainable global food systems (Costello et al 2020, Naylor et al 2021. The carbon footprint (Hilborn et al 2018, Parker et al 2018, Gephart et al 2021 and nutrient content (Hicks et al 2019, Golden et al 2021) of seafood, however, vary considerably among species and production methods. For example, capture fisheries for crustaceans can produce 40x more greenhouse gas emissions than those catching small pelagic finfish (Parker et al 2018), whereas seafood farmed using feeds and requiring land conversion, such as shrimp, tend to perform poorly compared to unfed products that have negligible production emissions (MacLeod et al 2020). Nutrient content of these products also vary among species (Hicks et al 2019, Bernhardt and O'Connor 2021), and comparative analyses have shown that small pelagic fishes are among the most nutritious and lowest emissions seafood globally (Hallström et al 2019, Bianchi et al 2022, Koehn et al 2022. However, the potential for lowemissions seafood products to contribute to nutritious and climate-friendly diets will depend on their relative affordability (Headey and Alderman 2019) and availability to consumers (i.e. production and trade) , which are rarely integrated into seafood carbon assessments . It therefore remains unclear which seafood species can contribute to nourishing, low-emissions diets, within local contexts, and how current seafood systems could be shaped to achieve these goals.
Here, we compare the nutrient density and greenhouse gas emissions of 106 seafood products landed at fishing ports or produced at farm gates, and place these in context of availability (i.e. production and apparent consumption), affordability, and sustainability of seafood consumed in the UK. We use the UK as a case study because it has a productive and diverse seafood supply (Jennings et al 2016), high rates of animal-food consumption (Miller et al 2022), but long-term declines in seafood consumption (Franklin 1997, Watson 2022) and population-level deficiencies in nutrients that are concentrated in fish (Derbyshire 2018). The UK produces seafood through domestic fisheries landings (pelagic, demersal and shellfish species) and a large aquaculture sector dominated by Atlantic salmon (Garrett and Caveen 2018), while imports of salmon, cod, tuna and shellfish consistently exceed exports (Jennings et al 2016). These datasets are used to identify fish and invertebrate species low-emissions, affordable and nutritious, thus providing insights into how seafood could be harnessed to reduce food system carbon emissions while increasing supply of healthy animal-source foods.

Carbon footprint and nutrient density
We first assess associations between emissions and nutrients, using a global database of 106 seafood products and 98 fish and invertebrate species. All seafoods had a higher nutrient density than other animal-source meats, and their greenhouse gas emissions were similar to chicken, pork and dairy products, but only 25% the carbon emissions of beef and lamb (figure 1(A)). Small pelagic fish such as herrings, sardines and anchovies, and wild Pacific salmon species such as pink and chinook, were the most nutritious and lowest-carbon fishes (figure S1), reflecting the low greenhouse gas emissions per unit catch of pelagic fisheries Tyedmers 2015, Parker et al 2018). Invertebrate seafoods ranged from highly nutritious farmed mussels with negligible emissions output to crustaceans such as prawns and lobsters that are caught by high-emissions fisheries (average 11.6 kg CO 2 -eq per kg seafood). After accounting for expected processing and waste from seafood, small edible portions in products such as scallop (12%), lobster (25%) and mussel (26%) further raised emissions from invertebrates (figure S2).
Placing these values in context of recommended sustainable diet guidelines (EAT-Lancet, Willett et al 2019), a 100 g seafood portion would account for between 5% (small pelagic fish) and 85% (crustacean) of the daily greenhouse gas emissions per person (Kovacs et al 2021). As noted by several recent global seafood analyses (Hallström et al 2019, Bianchi et al 2022, Koehn et al 2022, nutrient and CO 2eq estimates averaged across wild and aquaculture obscured differences among species and production methods, with particularly large variation in greenhouse gas emissions among wild invertebrate fisheries and farmed fishes (figures 1(B) and S1). Such variability can be used to identify performance gaps (Gephart et al 2021), and here suggests that shifting production towards species with lower carbon emissions, within each taxonomic group, could still promote supply of nutritious seafood.
Next, we combined nutrient density and greenhouse gas emissions estimates to quantify the emissions per recommended nutrient target (NT, 15% of recommended intake) in a single seafood portion (100 g) (Bernhardt and O'Connor 2021), and The mean nutrient density and greenhouse gas emissions produced, for common seafood groups (red, live weight), terrestrial animal-source meat (black; beef, sirloin steak; chicken, average meat; lamb, mince; pork, mince) and dairy (black), (B) shows the mean and range of values of farmed and wild-caught fish and invertebrates. (C) is the greenhouse gas emissions per nutrient dietary target (averaged across species), coloured by the number of nutrient targets (>15% recommended daily intake) in a 100 g edible portion. Emissions were thus generated by live weights in A and edible weights in C. Nutrient density (A), (B) and targets (C) are recommended intakes of calcium, iron, selenium, zinc and omega-3 fatty acids for adults aged 18-65 (Drewnowski et al 2015). Animal-source foods (beef, chicken, lamb, pork) are included for comparison using CO2 values from (Clune et al 2017) and nutrient values from (Widdowson n.d.). See figure S1 for the nutrient density and greenhouse gas emissions of each seafood product, figure S2 for emissions corrected by edible portion, and figure S3 for greenhouse gas emissions per nutrient dietary target of each species. thus evaluate the potential for low-emissions seafoods to contribute to recommended intakes of specific nutrients. Across global seafood products with emissions data, wild-caught small pelagic fishes and farmed bivalves had the lowest emissions per NT, with a 100 g portion providing recommended intake for 3-4 nutrients at less than 0.3 kg CO 2 -eq per NT (figure 1(C)). All seafood products reached at least two NTs (selenium and omega-3 fatty acids), with the most nutritious seafood also reaching NTs for calcium (e.g. pelagic fishes), iron (e.g. bivalves) and zinc (pelagic fishes, crustaceans) (figure S3). Other animal-source foods only reached NTs for selenium (beef, chicken, pork) or zinc (beef, lamb, pork). In crustacean and livestock products, low nutrient content across multiple nutrients combined with high carbon footprints caused some crustaceans (e.g. Norway lobster, 3.1 kg CO 2 -eq per NT), beef (1.3) and lamb (2.5) to have the highest emissions per NT for animal-source foods in our analysis (figures 1(C) and S3). High content of selenium and zinc in livestock and poultry is similar to most seafoods but for a far higher carbon footprint.

Nutrient content and carbon footprint of UK seafood production
We compiled seafood production data for the UK (figure S4), where demand for wild and farmed seafood is declining (Seafish 2019b) and populationlevel intakes of nutrients concentrated in seafood are suboptimal (Gibson and Sidnell 2014, Derbyshire 2018). Seafood production was defined as the combined seafood available annually from total landings at UK ports, aquaculture in UK fish farms, and imported products. We also extracted data on five additional nutrients (iodine, vitamins A, D, B12, and folate) from UK and Norwegian food tables (Norwegian Food Safety Authority 2021, Widdowson n.d.).
These nutrients are concentrated in seafood but were unavailable for all species in our global database (figure S1). Almost all seafood products provided 4-6 nutrient targets for less than 0.5 kg CO 2 -eq per target, with pelagic fishes (skipjack tuna, herring) and bivalves (mussels) containing the most nutrient targets at lowest carbon emissions (figure S5). These seafood species could therefore contribute to alleviating population-level inadequate nutrient intakes at lower carbon cost than other animal-source foods.
In the UK, one in two women are deficient at least one essential micronutrient (Stevens et al 2022), with high deficiency rates for selenium (50%), vitamin D (22%), iron (21%), and folate (19%) (Derbyshire 2018, Stevens et al 2022), all of which are concentrated in low-emissions seafood already available to consumers. For example, Atlantic mackerel had the lowest carbon emissions and highest nutrient density, providing over 100% the recommended intakes of selenium, vitamins B12 and D, 69% of omega-3 fatty acids, and 19% of iodine, for 0.25 kg CO 2 -eq (figures 2(A) and (C)). Furthermore, 91% of UK children between 18-35 months are estimated to have inadequate dietary vitamin D intakes (Gibson and Sidnell 2014), yet a child's portion (40 g) of herring or mackerel contains 49%-57% of the reference vitamin D intake for children between 1 and 3 years old. These low-emissions wild-caught fish thus provide greater nutritional benefits than other animal-source foods (beef, chicken, lamb and pork < 5% of reference vitamin D intake) at far lower greenhouse gas emissions. Oily fish such as mackerel, salmon and herring also contain toxic dioxin-like compounds that can produce negative health effects (Nøstbakken et al 2021), though risks from high oily fish consumption may be outweighed by their health benefits (Tuomisto et al 2020). Policies recommending future seafood consumption will nevertheless require guidance from both fisheries scientists and health professionals.
The potential for low-emissions seafood to contribute to healthy diets, however, also depends on its relative availability for domestic consumption, and consumer preference for those products (Jennings et al 2016, Parodi et al 2018, Zander andFeucht 2018). In the UK, four wild fish species (cod, haddock, mackerel, skipjack tuna) and farmed Atlantic salmon accounted for half of total available seafood in 2019 (figure S4). These top five species had similar average nutrient densities (284%-410%, average = 350%) and carbon footprints (0.25-3.95 kg CO 2 -eq, average = 2.6) (figure 2). Mackerel had the lowest carbon footprints of any wild-caught species and high contributions to recommended intakes for iodine, selenium, omega-3 fatty acids, and vitamins B12 and D (figure 2). Promoting access and consumption of mackerel in the UK could improve diets with relatively low environmental impact, although currently the UK exported ∼43% of its available mackerel (i.e. imported and landed) in 2019 (figure S6), limiting its potential as a locally produced, low-emissions nutritious food. Nutritious and low-emissions seafoods often had low apparent consumption relative to their total production (e.g. herring, mackerel) (figure S6), perhaps reflecting limited consumer appeal of some species and products. Outside these five highproduction species, UK seafood produced in lower volumes had more diverse CO 2 emissions and nutrient density, including the highest (Norway lobster and wild-caught shrimp species) and lowest emissions for capture fisheries (mussels and Atlantic herring) (figures 2 and S1).

Sustainability and affordability of low-emissions nutritious seafood
Consumer demand for seafood in the UK is primarily influenced by price, with consumers favouring more affordable products (Seafish 2019a). Across Western Europe, preference for sustainable products is also a key influence on consumer behaviour (Menozzi et al 2020), as reflected by the rapid growth in seafood eco-labels (Roheim et al 2018). Indeed, low-emissions nutritious seafoods can contribute to healthy diets where they are affordable (Springmann et al 2021), and sustainable ecolabels can both promote consumption of certain seafood products (Honkanen and Young 2015, Jacobs et al 2018) and incentivize rebuilding of certified stocks (Gutiérrez et al 2012). We assessed these factors by compiling data on average price (£ per kg) (Watson 2021) and (perceived) sustainability of seafood consumed in the UK, as defined by a ratings scheme designed for UK consumers (The Good Fish Guide) (see section 3).
Wild-caught seafood was, on average, cheaper than farmed seafood, owing to the dominance of farmed Atlantic salmon in domestic seafood production, which is associated with (relatively) higher prices (figures 3 and S8). Average sustainability ratings were similar between wild-caught and farmed seafood, but varied considerably between species (figure 3) and production methods (figure S8). Sustainability of wild-caught seafood was particularly variable, owing to spatial variability in stock status of species such as cod and herring ( figure  S8(B)). No species maximised all five desirable variables, underlining existing trade-offs between production, carbon footprints, price, nutritional value, Figure 3. Blue food profile of 11 major UK seafood products. Radar plots show the average carbon footprint (live weight kg CO2-eq, inverse), nutrient density (10 nutrients), sustainability rating (Good Fish Guide), production volume (annual tonnes), and price (£ per kg, inverse) for wild-caught species and farmed species on average, and by product. All variables are scaled between 0% and 100%, and CO2-eq and price are scaled to their inverse (i.e. 100% is the least CO2-eq use and lowest price), such that species with the largest radar areas are low-emissions, nutritious, sustainable, high-production, and affordable. For average wild/farmed seafood, kg CO2-eq, nutrient density, sustainability and price are mean values weighted by species' annual production. and sustainability. These trade-offs reveal limitations of certain production systems (e.g. high emissions of Norway lobster) but also highlight potential for improving the environmental performance of highvolume foods (e.g. farmed salmon), exploitation of overfished stocks (e.g. cod), and the production of nutritious future foods (e.g. mussels) (Parodi et al 2018). Alaska pollock was the most affordable and sustainable seafood product for UK consumers but had relatively low availability (i.e. imports), suggesting that increasing Alaskan pollock imports could improve supply of affordable low-emissions seafood in the UK.
Nutritious, cheaper, and low-emissions wildcaught fishes such as mackerel and haddock had high sustainability ratings (figure S8(B)), due to use of low-impact fishing gears (pelagic trawls) and low number of overfished stocks. Indeed, fisheries assessments show that low-emissions UK fisheries have steadily improved stock status since 1990, with high stock biomass and all mackerel stocks and 30% of herring stocks recently fished within sustainable levels (figure S9). These trends underline the effectiveness of fisheries management in rebuilding depleted fish populations when harvest control rules are implemented (Melnychuk et al 2021), and thus the benefits to food supply when stocks are sustainably fished (Costello et al 2016, Jennings et al 2016. Bringing the remaining overfished UK-sourced stocks within sustainable limits would therefore improve domestic supply of nutritious low-emissions food to UK consumers, and also reduce greenhouse gas emissions from fishing vessels by improving fuel use per unit catch (Hornborg and Smith 2020). Further gains in nutritious seafood production could be achieved by incorporating nutrient-based reference points (e.g. Maximum Nutrient Yield) into fisheries assessments that assess strategies for enhancing nutrient-rich catches. In North Sea fisheries, for example, nutrient yields could be increased by prioritising long-term catch of resilient and nutritious species such as herring and sprat .
UK aquaculture was less diverse than its capture fisheries, with domestic and imported Atlantic salmon together representing 62% of available farmed seafood. Sourcing low carbon-emissions inputs to aquaculture feeds, such as avoiding inputs associated with land-use conversion (Ziegler et al 2013) and improving feed conversion ratios (MacLeod et al 2020) would have significant benefits for improving UK aquaculture emissions. In contrast, farmed mussels were the highest-ranking seafood in 4 of 5 categories, but had the lowest production volume of all top 11 products (figure 3). Enhancing bivalve production and consumption globally has been proposed as a means of increasing global food supply with minimal environmental impacts (Willer and Aldridge 2019), and could contribute to production of more nutritious farmed seafood in the UK (Willer et al 2022). However, increasing prevalence of disease, toxic algal blooms and extreme weather has caused declines in Europ mussel production (Avdelas et al 2021) and, in the UK, several additional factors have hindered marine aquaculture expansion, including competition for coastal space and poor water quality (Huntington and Cappell 2020).
Reductions in livestock consumption, particularly beef, through demand-side policies have been proposed as a means of improving dietary health while reducing food-system carbon emissions (Bajzelj et al 2014, Springmann et al 2018. However, in the UK, seafood products are the most expensive protein food, above red meat and chicken (Watson 2021), while seafood retail prices increased by 31% from 2010 to 2020, exceeding general inflation (21%, Consumer Price Index) and terrestrial meat (11%) (Department for Environment, Food and Rural Affairs 2022). This likely contributes to long-term declines in seafood consumption, particularly for poorer households and younger age groups (Watson 2021(Watson , 2022. The UK's capacity to transition towards low-carbon animal-source foods is thus limited by low affordability of desirable high-volume seafood, such as salmon (£17.01/kg) and cod (£8.61/kg), and lower appeal of more affordable products (∼£5.60/kg: Atlantic herring, farmed mussels). Positioning seafood as 'climate smart' will depend on the availability of nutritious, low-emissions products that offer consumers value for money compared to other proteins. This could be incentivised directly through increased production of low cost species, but also indirectly through food labelling, education campaigns, and taxation (Springmann et al 2021).
Collectively our findings suggest wild caught pelagic fishes and farmed bivalves have the greatest potential to be sustainable, nutritious, and low-emissions animal source foods, corroborating previous research (Hallström et al 2019, Bianchi et al 2022, Koehn et al 2022. By placing nutrient and carbon footprints in the context of seafood production volumes, we also reveal opportunities for transitioning seafood systems towards low-emissions, healthy foods. Information on long-term patterns in supply, affordability, sustainability, and consumption will develop deeper understanding of the drivers of seafood systems, and thus inform efforts to promote low-emissions seafood consumption. We expect our UK case study to be representative of seafood products in other high-income countries in the Global North where seafood sectors supply both wild-caught (e.g. whitefish, pelagic species) and farmed seafood (e.g. Atlantic salmon). In these countries, policies that support less welldeveloped sectors (e.g. farmed mussels) could reduce food sector emissions, while policies that help inform consumer choice of existing products (e.g. expanding certification schemes to include carbon emissions (Madin and Macreadie (2015)) could nudge consumers towards low-emissions, nutritious seafood (Bucher et al 2016).

Carbon footprints and nutrient data
We extracted estimates of greenhouse gas emissions relative to live weight wild-caught or farmed seafood from data modelled in the Seafood Carbon Emissions Tool (Monterey Bay Aquarium Seafood Watch and Dalhousie University) (Seafood Watch and Dalhousie Universit 2021). This dataset was initially compiled to focus on seafood products relevant to the United States, but overlaps substantially with key species for other regions. Modelling underpinning emissions estimates was based on reported fuel use intensity (L/t) values for marine fisheries (Parker and Tyedmers 2015, Parker et al 2018), including emissions associated with bait use (e.g. tuna longlines, lobster traps). Emissions from aquaculture production were estimated with Monte Carlo analyses based on data extracted from published life cycle assessments and other sources. Input parameters accounted for consistently recognized drivers of greenhouse gas emissions in culture systems for which data were available across species and systems: feed conversion ratios, general feed composition, feed ingredient impact factors, rates of on-farm energy use, relative use of electricity or fuels, and impact factors for fuels and country-specific electricity grids.
This database contained greenhouse gas estimates for 98 fish and invertebrate species, representing 151 seafood products at the point of production (i.e. fishing port or farm gate), standardised as CO 2 equivalents per kg of seafood (kg CO 2 -eq). A seafood product was one species produced by a specific production method (e.g. capture: longline, trap, trawl; farmed: pond, cage, net pen), and each species-method combination had median values and lower and upper limits of carbon emissions (25th and 75th quantiles). In cases where production was heavily skewed towards certain production systems, those systems were selected for inclusion in further analysis, excluding uncommon production methods (e.g. recirculating systems producing Atlantic salmon). These data were used to generate the range of expected greenhouse gas emissions produced by wild and farmed seafood products (table S1). Most species had multiple emissions estimates collated across studies of different seafood production methods and locations, and we did not consider emissions generated in distribution, transport, and processing of seafood products. Our carbon footprint analysis thus represents the potential emissions generated by seafood production at port (capture fisheries) or farm gate (aquaculture), per kg of unprocessed fish or shellfish. By addressing emissions up to the point of landing or harvest, these estimates thus omit potentially important sources of emissions (e.g. distribution of products), and are insufficient for broad-scale carbon footprint modelling (e.g. biogenic emissions and land-use change emissions from converting mangroves for pond culture). However, this database provided a methodologically consistent approach among diverse fish and invertebrate species, and sufficient resolution of data to differentiate between related species. We estimated the minimum and maximum kg CO 2 -eq for each species, and the midpoint of those values, separately for wild and farmed (n = 106 seafood products), and for related species groups (e.g. bivalves, whitefish, small pelagics) (n = 10 seafood groups) (table S1). These values capture the range in species-level, live weight emissions between wild-caught and farmed seafood, across diverse production methods.
Nutrient data were extracted from Fishbase (Froese and Pauly 2022), providing estimates of calcium (mg), iron (mg), selenium (µg), zinc (mg), and omega-3 fatty acids (g) per 100 g of muscle tissue. Invertebrate nutrient content were the genera-or family-level mean nutrient concentrations from the FAO/INFOODS database of 195 samples of 45 species (FAO 2016, Rittenschober et al 2016. We estimated the nutrient density of each seafood product, defined as the combined contribution of a 100 g portion to recommended daily intakes of all five nutrients (Drewnowski et al 2015, Hicks et al 2021, based on nutrient reference values for adults aged 18-65 (FAO/WHO 2004).
We visualised nutrient density and greenhouse gas emissions (kg CO 2 -eq) in a biplot alongside values for terrestrial animal-source foods, including dairy (cheddar cheese, whole eggs, semi-skimmed milk) and livestock (beef, sirloin steak; chicken, average; lamb, mince; pork, mince), based on a meta-analysis of carbon emissions data in (Clune et al 2017) and nutrient values in UK food composition tables (Widdowson n.d.). For carbon emissions, we used median values for each product, corrected to represent emissions from farm to farm gate (using table 2 in Clune et al (2017)). Note that terrestrial meats were per kg of bone free meat whereas seafood values were per kg of unprocessed whole fish. We then combined these metrics to measure the greenhouse gas emissions (kg CO 2 -eq) per nutrient target (15% of recommended daily intake) of each terrestrial animal-source food and seafood product, following Bernhardt and O'Connor (2021). These emissions estimates were corrected to reflect the edible fraction of each species (Seafood Watch and Dalhousie University 2021). Edible fractions were initially derived the UN Food and Agriculture Organization (FAO 1989) as well as from multiple government, industry, and nongovernmental organisation datasets (P. Tyedmers pers. comm. 2017). Adjusting for edible fraction allows for emissions to be communicated relative to the edible unit against which nutritional values are also communicated, and accounts for variation in yield of edible product among species of fish and shellfish. This metric thus expresses the greenhouse gas emissions required to meet one dietary target, based on recommended adult intakes (18-65 years old) contained in a 100 g edible portion.

Low-emissions potential of UK seafood
Next, we placed carbon footprint and nutrient density scores in the context of seafood production , using the UK as a case study. We compiled annual landings, imports, exports, and aquaculture data for all UK seafood products from government databases (www.gov.uk/government/collections/uksea-fisheries-annual-statistics), Seafish (www.seafish. org/insight-and-research/market-supply-data-andinsight/), and the European Commission (https:// stecf.jrc.ec.europa.eu/reports/economic/-/asset_pub lisher/d7Ie/document/id/287 169). For each species group, we combined landings, import and export data for 2019 with the average annual aquaculture production across 2015-2018 (2019 data were unavailable), and matched these products to their average estimated carbon footprint and nutrient density. Where appropriate, species were combined into groups that aligned with commonly used product names (e.g. scallops, trout, shrimp). We estimated the annual seafood production available to the UK (sum of landings to UK ports, aquaculture produced in UK farms, and imported seafood), and apparent consumption of seafood by UK consumers (total production-exports, corrected for edible portion). These metrics quantify the composition and volume of seafood available to the UK per year, based on live weight production in 2019. Carbon emissions estimates were unavailable for farmed scallop, though this product contributed <1% of total UK scallop production (9.25 t).
We estimated the kg CO 2 -eq and kg CO 2 -eq per NT of all products that represented the top 90% of seafood availability in the UK. We used carbon emissions data that represented the dominant production method for each species (table S2), and thus capturing key impact drivers of UK seafood emissions . To assess potential for UK seafood to contribute to improving suboptimal nutrient intakes in adults and children (Gibson and Sidnell 2014, Derbyshire 2018), we extracted nutrient content for iodine and four vitamins (A, B12, D, and folate; µg 100 g −1 of raw flesh) from food composition tables for the top 90% seafood products available in the UK (Norwegian Food Safety Authority 2021, Widdowson n.d.). Nutrient density and nutrient targets estimates for UK seafood were recalculated including these five nutrients (i.e. across ten nutrients in total), and thus exceeded values of the global seafood analysis.
In addition to nutrients and health benefits, preference for affordable, quality seafood is a key driver of consumer behaviour in the UK (Seafish 2019a). Although less important than price, seafood ecolabels can also positively influence consumer preference across Western Europe (Zander andFeucht 2018, Menozzi et al 2020), and promote behaviour shifts towards more sustainable products (Jacobs et al 2018). To assess these factors in the context of carbon footprints and nutritional potential, we next examined the affordability and (consumer-labelled) sustainability of the 11 most-produced seafoods in the UK. Average price (£ per kg) was extracted from market surveys conducted by Seafish (Watson 2021) and sustainability scores were extracted from the Marine Conservation Society's Good Fish Guide (Marine Conservation Society 2022). We note that seafood sustainability is 'imperfectly measurable' (Roheim et al 2018), and ecolabels may target different aspects of sustainability, from sustainable fishing levels and habitat damage to pollution, bycatch and endangered species. Here, we use The Good Fish Guide sustainability metric as a standardised rating scheme with particular relevance for UK consumers, that qualitatively compares environmental impacts of processes specific to both wild (e.g. overfishing) and farmed (e.g. disease) products. Capture fisheries sustainability was assessed by ranking stock status (catch limits, biomass level, International Union for Conservation of Nature Red List status), management (existence of regulatory frameworks), and capture method (habitat impacts) (Marine Conservation Society 2018) for 94 stocks relevant to UK seafood supply. Aquaculture sustainability was assessed by scoring feed resource use (traceability, sourcing), environmental impacts (habitat, water quality, disease), fish welfare, and regulations and management (enforcement of standards and thirdparty certification) (Marine Conservation Society 2020) for 13 farm systems (Atlantic salmon = 9, Rainbow trout = 2, mussels = 2) relevant to UK seafood. To facilitate comparisons between these two methodologies, we rescaled all sustainability ratings between 0 (low) and 1 (high sustainability). For capture fisheries, we also extracted indicators of fishing pressure and biological status for stocks of UK interest. These metrics were extracted for 231 stock-year combinations of cod, herring, mackerel, haddock and Norway lobster over 1990-2019, and used to assess longterm trends in fishing pressure relative to maximum sustainable yield (F relative to F MSY ) and reproductive capacity (spawning stock biomass relative to B Lim ) (Lynam et al 2021).

Data availability statement
The data that support the findings of this study are openly available at the following URL/DOI: https:// github.com/jpwrobinson/UKseafood.