The empirical evidence for the social-ecological impacts of seaweed farming

Seaweed farming is widely expected to transform the way we approach sustainable developments, particularly in the context of the ‘Blue Economy’. However, many claims of the social and ecological benefits from seaweed farming have limited or contextually weak empirical grounding. Here we systematically review relevant publications across four languages to form a comprehensive picture of observed—rather than theorised—social and environmental impacts of seaweed farming globally. We show that, while some impacts such as improved water quality and coastal livelihoods are consistently reported, other


Introduction
Growing material, food, and energy demands [1,2] and the dwindling supply of available, unmodified land on which to expand terrestrial agriculture [3,4] have spurred inquiries into novel sources of future foods [5] and energy [6].Seaweeds are one such solution proposed as a means to tackle issues as diverse as food security [7,8], climate change [9], human health [10], eutrophication remediation [11], desertification [12], rural poverty [13], and gender inequities [14].In parallel, there is growing recognition of the diverse uses for seaweed biomass [15], and the corresponding commercial potential for seaweed farming and products [16].For example, some seaweeds, like those in the family Solieriaceae and Gracilariaceae, represent a fast-growing segment of global seaweed production, are primarily used for producing industrially-useful hyrocolloids such as carrageenans and agar, and are often farmed in shallow inter-tidal areas in the tropics [17].Others, such as those in the families Laminariaceae and Bangiaceae are more often used as food, and are produced extensively in East Asia, often on floating rafts and lines [17].
Historically, seaweed cultivation has been largely restricted to countries in Asia [17], but spurred on by recent insights, many countries around the world are showing an interest in, and beginning to launch, their own seaweed industries [18][19][20].Some of this industry growth will take place on land in growing facilities or in freshwater systems [21].However, it is likely that the majority will occur in coastal or ocean waters as part of emerging Blue Economy strategies; i.e. strategies aimed at expanding economic growth through sustainable development of marine sectors [22,23].Given the vast areas of ocean space available for development, the possible benefits are enormous-however, conversely, making an appreciable contribution to the material needs of society also requires this vast amounts of ocean space [7,8,[24][25][26].Likely, this will not only result in competition, or a need for integration, with other rapidly expanding sectors of the Blue economy [27], but, as we have seen on land with the rapidly growing palm oil sector [28][29][30], it will also likely change the social and ecological systems where farms are developed.Recognizing potential for both benefit and harm-as well as the power-dynamics and ethical issues that need to be considered in determining these normative views-there is a critical need to interrogate what we know about the social and ecological impacts of seaweed farming, given the sector's plausible global growth trajectory.
Seaweed farming is inherently embedded within diverse and complex social-ecological systems (e.g.[31,32]).The concept of social ecological systems emphasises integration of 'humans-in-nature' [33].It suggests that the two parts (i.e. the social system and ecological system) are equally important and that they function coupled and interdependently.Therefore, to understand the social-ecological impacts of seaweed farming, the natural and social science components and perspectives should be explored together.This paper brings the social and environmental impacts together, via an investigation of these impacts as documented in the literature.
Multiple studies have outlined the environmental impacts and ecosystem services provided by aquaculture generally [34,35], and seaweed farming specifically [9,18,[36][37][38][39][40][41][42].However, these reviews are limited in their relevance to the development of the seaweed sector.First, there is a lack of agreement in the suite of impacts assessed among studies, preventing a comprehensive, systemic understanding of how growth in seaweed farming will impact ecosystems and people; and in addition, knowledge of the environmental and social impacts of seaweed farming are rarely integrated or attributed equal weighting.Second, these studies heavily rely on evidence from natural seaweed ecosystems to extrapolate the services and impacts that are expected to be provided by seaweed cultivations, and largely ignore the many physical, biochemical, and social differences that may exist between wild and farmed systems.Third, these assessments do not delve into details about species diversity or farming techniques [43,44], which may obscure differences in impacts between different cultivation contexts [19].Finally, previous reviews have been largely limited to English-language publications.Since both historical seaweed cultivation and the majority of large-scale applications are based in East Asia [45], inclusion and investigation of the non-English literature from this region is essential for capturing all current knowledge [45,46].
Here we provide a comprehensive, systematic review of the empirical evidence for the social-ecological impacts (both positive and negative: normative views distinguished and determined by the perspectives of the papers themselves) of seaweed farms in marine systems.We systematically review and synthesize peer-reviewed literature from four languages (Chinese, Japanese, Korean and English) to ascertain the strength of evidence for, and identify gaps in, our knowledge of seaweed farming's social-ecological impacts; highlight possible threats to, and opportunities for, social-ecological systems from seaweed farming; and identify future directions towards a more comprehensive inquiry of emerging seaweed aquaculture systems.

The observed social and ecological dimensions of seaweed farming
We identified 186 studies describing a diverse range of observed social-ecological impacts (n = 553) from seaweed farming (S1 Table ).Of these, 132 investigated environmental impacts covering 354 unique impact observations which we classified into 11 categories.Fifty-six studies investigated social (including economic) impacts with 199 unique impact observations across nine categories (Fig 1).Just two studies (1%) investigated both social and environmental impacts (See S1 Fig) .These interdisciplinary studies highlight how the social and environmental impacts of seaweed farming are inherently intertwined within complex socio-ecological systems, as for example in Fig 2 .Thus, environmental impacts from seaweed farming have been the subject of nearly double the research attention relative to social impacts to date, and very few studies have sought to evaluate the interplay between environmental and social dimensions.Further, the degree of evidence for each impact category varied greatly, from well-established benefits or variable outcomes to negligible influence and deleterious impacts.

Well-established impacts
Numerous promulgated social and ecological benefits of seaweed farming have robust empirical grounding.For instance, 99/143 (70%) water quality observations reported improvements in association with seaweed farming, particularly with respect to nutrient reduction (nitrogen and phosphorous), the sequestration of organic particulates from co-cultivars in integrated multi-trophic aquacultures (IMTAs), and increasing dissolved oxygen [48][49][50][51][52][53].There is also substantial evidence that seaweed farms can drive carbon cycling on a local scale and that expanded seaweed farming will allow the capture of large quantities of carbon (e.g.[49,[53][54][55][56][57][58][59]) or lead to an increase in the carbon flux from the air to the sea [60,61].The latter may lead to long-term sequestration as standing biomass erodes and winds up in the recalcitrant dissolved organic carbon (DOC) pool [62].In terms of biosecurity, while there was evidence for some gene flow from cultivated to wild seaweed populations [63,64], in general, seaweed cultivation had led to little or no spread of non-natives or negative impacts on wild seaweed populations [65][66][67][68][69][70].This is despite the well-documented persistence of invasive seaweed populations [71][72][73] and may reflect either that most seaweeds have low invasive potential, improved biosecurity practices in recent years, or simply a lack of available data documenting invasions.Similarly, whilst there is limited field evidence to suggest that seaweed farms inhibit bloomforming microalgae [74,75], there is a preponderance of evidence that shows that, if poorly managed, seaweed farms can enable blooms of pest macro-algae species [76][77][78][79][80][81][82][83][84][85].Twenty social and environmental impact categories of seaweed farming as identified by our systematic review.Impact polarity based on available data is noted by the colour of each circle.For impact categories with greater than one observation, we either report the polarity with a majority of observations supporting it (>50%), or we report a 'Variable' impact.For impact categories with only one observation, we report 'Data Deficient'.Confidence scores are based on a cumulative probability binomial test that the identified impact is incorrect (Medium: 0.05 > p > 0.005; High p < 0.005).The numbers in parentheses refer to the number of observations from each impact category.https://doi.org/10.1371/journal.pstr.0000042.g001 For local people and communities that practice seaweed farming, the derived income provisioning and livelihood opportunities are, on the whole (87% of included social studies), beneficial to coastal communities (e.g.[86][87][88][89][90][91]). Less work has investigated the gendered distribution of benefits, but of those that do, 88% show that women are the primary practitioners and beneficiaries of seaweed farming (e.g.[86,[92][93][94]).Farming allows many women to secure independent access to funds and affords them empowerment within patriarchal cultures [89,94].Many of these studies, notably those based in the Global South, describe the flow-on benefits of increased income on living standards, including greater household food security and affluence, and increased access to education and material goods (e.g.[47,92,95]).In many communities, social cohesion has improved as a result of the community groups and collectives that have been established as a way of sharing resources, including equipment, sporophytes (reproductive material), and labour for seaweed farming operations (e.g.[87,88,[95][96][97]). Nonetheless, the importance of context is critical for realising these benefits.For example, water quality benefits may be negligible where strong seasonal pulses, such as those from seasonal monsoons and local currents are the key determinants of nutrient concentrations [98], or where seaweed production is small-scale or ephemeral e.g.Xiangshan Bay, China [99].Further, in places where seaweed farming adoption has been rapid or heavily industrialized, family farming traditions and community management has decreased in lieu of privatised, fixed location farming that has weakened social cohesion and led to population displacement [47,[100][101][102].Thus, while the benefits of seaweed farming are supported in this review, so too are the negative impacts increasingly evidenced.

Variable outcomes
The role of seaweed farming on most other categories (9/20) were inconclusive despite a wealth of research attention.For instance, impacts on marine biodiversity have been heavily studied but are highly variable between and within taxonomic groups (Fig 3).Mobile herbivorous finfish stand the most to gain from the proliferation of seaweed farms, likely due to the increase in grazing opportunities and habitat [103][104][105][106].But the benefits for planktonic communities, including improved community diversity and stability, were heavily dependent on the seaweed species being cultivated [99,[107][108][109][110][111][112]; and outcomes were more frequently negative for benthic invertebrates and seagrass whom particularly suffered from habitat disruption due to 'off-the-bottom' farming techniques [113][114][115][116].
Biogeochemical and hydrodynamic outcomes were also among the most variable.Numerous studies have found that seaweed farms produce increases in pH compared to reference Where seaweed farming was associated with a benefit to any of these biodiversity indicators, the impact was classified as beneficial.Where seaweed farming was negatively associated with the biodiversity indicator, the impact was defined as deleterious.Other cases were defined as either neutral (neither positive or negative impacts) or mixed (both positive and negative effects).'No impact' was defined for observations where the seaweed farms had no impact on the biodiversity indicator.https://doi.org/10.1371/journal.pstr.0000042.g003sites [117][118][119][120][121][122][123], while others report no statistically clear effect [98,111,124] and one records a decrease in pH [125].Sediments underneath and near seaweed farms undergo various shifts including an increase in the acid volatile sulfide content of the benthos [126], a decrease in bacterial production and nitrogen content in sediments [127], a decrease in organic matter content [128] and a redistribution of fine and coarse grained sediments [114,129,130], with many of these changes being associated with clear changes in faunal communities [114,126,127,130].Sedimentary community shifts may also be linked to hydrodynamics due to seaweed farms altering water flow [131][132][133][134].In some cases this can trap sediments and thereby protect corals [135], increase the settlement of commercially valuable juvenile bivalves [136], or decrease turbidity [134].Whilst this decreased turbidity can increase light penetration [134], in other instances the physical presence of seaweed farms can also physically block the amount of photosynthetically active radiation reaching the benthic layer [42,137].Whether or not these changes benefit or harm existing biota will depend on the species assemblages and the nature of limiting pressures affecting them.
For communities where seaweed farming as become an important livelihood, the outcomes for human health, traditional practices, vulnerability to shocks, and overall resilience are inconsistent.Evidence suggests seaweed farming can lead to better access to healthcare (e.g.[89,94,138]).However, the physical demands of seaweed farming can be detrimental to the well-being of practitioners; a range of ailments have been reported including musculoskeletal pains from hard physical labour, exposure to marine parasites during tending, and respiratory problems from the off-gassing of vapours during the drying process [90,95,100,[139][140][141].Similarly, for some, seaweed farming has allowed a reliable income source; creating more consistent quality seaweeds relative to wild collection and safer production in the face of harmful algal blooms compared to fisheries sectors [142].However in contrast, in some tropical areas, seaweed farming is vulnerable to environmentally-mediated diseases, such as 'ice-ice' and other pests [138] rendering communities with an increased reliance on seaweed farming vulnerable to climate-driven disease burden in the long-term [143,144].This may be compounded by economic contexts; farmers attached to underdeveloped local markets and inequitable supply chains may struggle to negotiate for higher wages and could become increasingly vulnerable as global commodity prices fluctuate [47,92,139,145,146].It can also be difficult for some communities to resume previous livelihood activities in the face of these shocks, either due to lost infrastructure [47] or disrupted transfer of cultural knowledge [47].Therefore, understanding how and where seaweed farming can strengthen local traditional cultures [147] and diversify livelihoods from current success stories (e.g.[148]) is critical for improving community resilience from this industry.
Anticipating sustainability outcomes among impact categories will require careful consideration of the prevailing environmental and social conditions in which seaweed farming is or will be embedded.For example, several studies from a range of countries demonstrate that seaweed farms can have higher levels of biodiversity than surrounding areas only if the surrounding areas are already low in terms of structural and ecological complexity; where seaweed farms are located in relatively biodiverse or structurally complex areas, they end up with similar or lower biodiversity compared to their nearby comparison sites [104,105,149].Similarly, the global economic fluctuations that have impeded farming in some locations have not influenced farmers in other areas, such as in India, where there is high local demand for seaweed products and relatively low levels of export [86,150].Moreover, case studies from Korea illustrate that a shift from artisanal to industrialised farming can weaken community relationships and lead to a loss of traditional management practices [101,102,151], illustrating that scale of production will also be an important mediating factor for ensuring sustainable seaweed aquaculture.

Gaps in the distribution of empirical evidence
The degree of evidence supporting our understanding of the environmental or social impact categories we identify is highly uneven (Fig 4 ).For example, observations focused on water quality represented 40% of all observations across ecological and social dimensions, whereas two impact categories contained only one study each.We found one observation on the potential for large-scale seaweed farms to generate substantial quantities of iodine-based particles, which could have implications for local weather patterns and/or global climate [152], and another that examines the contribution of seaweed farm infrastructure (e.g.plastic ropes and floats) to the global pool of marine microplastics [153].Similarly, whilst we found evidence from 16 genera and 14 families, no individual family has been comprehensively studied across all impact types, with the families Soliariaceae and Gracilariaceae providing the broadest distribution across categories (Fig 4).
We found no region in the world where there has been a comprehensive investigation of all 20 impact categories.The regions with the greatest research effort (Eastern Asia, East Africa, Southern Asia & South-eastern Asia), are also those where a greater range of impact types and The intensity of the fill in each box corresponds to the number of observations within the specified impact for that family.For example, in Gracilariaceae there are numerous observations of beneficial effects on biodiversity, water quality and carbon, very few beneficial effects of acidification and algal blooms, and no observations in other impact categories.For the same family there are some observations or mixed/neutral impacts, some instances of no impact being observed, and one instance of light being negatively impacted.The coloured bar on the x-axis shows which broader group each family belongs to (Brown, Red, and Green algae).Boxes with points represent unique observations only available in non-English literature.n = 553.https://doi.org/10.1371/journal.pstr.0000042.g004impact polarities have been observed, and where a lower proportion of impacts are beneficial (Fig 5).This is consistent with the current global distribution of production [18], but highlights the need for more research in the new geographies of this industry's growth, and the importance of considering findings from the non-English literature to highlight unseen opportunities and threats otherwise lacking from the emerging English-language discourse on the sustainability of seaweed farming.

Discussion
This comprehensive review found consistent evidence for beneficial socio-ecological impacts of seaweed farming, including positive effects on water quality, carbon management, livelihood generation, gender equality, and living standards, a deleterious facilitation of algal blooms, and a benign impact in terms of invasiveness (Fig 1).There was a lack of consistent data to determine whether seaweed farming had an overall negative, positive, or neutral effect in several of the categories that emerged from this review, including biodiversity, acidification, hydrodynamics, human health, and community culture and resilience.Local context and conditions play a large role in mediating this variability (Fig 1).These results support the conclusions of Gentry et al (2019) that seaweed farming can provide numerous ecosystem services [35], but also underscores the risks highlighted by Campbell et al (2019, [36].In any case, this review clearly shows that seaweed farming will transform social and ecological systems where it is introduced and that the specific outcomes of this transformation will depend largely on the way farms are developed and managed. The evidence for the benefits of seaweed farming suggest that this transformation may, as many proponents of the Blue Economy suggest, be an important pathway towards economic development and could help many coastal nations make progress towards the United Nations' Sustainable Development Goals (SDGs) [155].However, the numerous impacts for which we found a high degree of uncertainty around polarity and magnitude, suggests that these benefits may come with trade-offs for other SDGs.These identified impacts highlight the potential for negative outcomes in the context of the Blue Economy; e.g. in contexts where farm management and development are not well-designed which could jeopardize the stability of socio-ecological systems.As we have shown here, certain socio-ecological contexts may be better positioned to reap the rewards of seaweed farming than others, although these benefits may be accompanied by costs that may or may not undermine the overall value of benefits.Comprehensive social and ecological impact and risk assessments, based on local contexts and with well-understood levels of uncertainty and risk acceptance, will enable decision-makers and practitioners to clearly communicate and manage the benefits and trade-offs of seaweed farming.
The already established seaweed industries of East Asia, provide a window into how these trade-offs may play out.For example, Steenbergen et al. (2017) highlight how the rise of seaweed farming over the previous several decades, spurred on by global demand for hydrocolloids, has led to a considerable increase in material wealth for the Indonesian community of Tanimbar Kei, but also how this initial success has led to overcrowding in the shallow bay where seaweed is farmed and has led to a general decline in environmental health and farm productivity [47] (Box 1).In Korea, rapid expansion of industrialized seaweed farming is accelerating the loss of traditional community rules that have been maintained to secure sustainable use of local marine resources (e.g., a periodic rotation of seaweed farming sites and community cleaning [151].Avoiding these poor outcomes may require re-orientation of the conditions for industry growth, from a top-down approach driven by external factors towards a bottom-up approach driven by the needs of communities for solutions that seaweed can provide [156].This may prove challenging within the context of the Blue Economy, which is already having social and ecological consequences.Safeguarding the sustainability of the seaweed farming industry will require understanding and incorporating social and ecological components together, through transdisciplinary efforts that draw on diverse disciplinary expertise as well as traditional and experiential knowledge [27].
However, given the relative nascency of seaweed farming in many parts of the world and the highly context-dependent nature of its impacts, more targeted research is needed to better understand how seaweed farms will transform social-ecological systems.Whilst the 16 genera of seaweeds that have been captured by our review account for more than 99% of the globally cultivated biomass [18], they represent only 19% of the 84 genera of seaweeds that are harvested or cultivated globally [44].As seaweed cultivation expands and we learn more about their utility, lesser-known species may be identified and prioritized for cultivation at scale [157].Further, as seaweed farming becomes more established throughout the world, changes in physiology, biochemistry and seeding/harvesting regimes due to domestication and optimized production will complicate much of what we already know [158].Similarly, some impacts have been discussed in the literature for which there is a paucity of empirical evidence that supports or refutes them.For example, it has been conjectured that seaweed farms could contribute to coastal protection [9,[159][160][161][162], and while there is evidence of such an effect from natural systems [163][164][165] and modelling that supports the idea [132,133,166,167], we did not find any empirical evidence that could confirm the impact.
Similarly, we found no research into 0the impacts of seaweed farming on non-finfish megafauna; it is unclear whether this reflects a lack of such impacts or a lack of reporting or research, but in either case, the production of empirical research demonstrating the innocuity for megafauna will be important for the development of new farms, especially near critical habitats for threatened and endangered species.Finally, many pathways for how seaweeds can potentially sequester carbon have been proposed in the literature [168][169][170], however, with the exception of funnelling carbon into the recalcitrant DOC pool [62], we found very little evidence of farms being able to directly contribute to long-term carbon sequestration in the context of an in situ farm.Given that this particular impact has been one of the most attractive selling points in the eye of the broader public [171,172], this gap in our understanding is especially important to address.
The ongoing failures of the palm oil industry to address environmental and social concerns provide a valuable warning for the nascent seaweed industry.Once hailed as a 'vehicle to eradicate rural poverty', provide valuable food and non-food products, and reduce greenhouse gas emissions [173], palm oil production is now plagued by issues relating to its role in accelerating deforestation and biodiversity loss, and for its negative impacts on the social fabric of local communities [28][29][30] that even recent sustainability certification schemes have failed to address [174].This review highlights that we should temper our expectations for the benefits of seaweed farming with a pragmatic view of the possible costs.Similarly to the palm oil, or other novel industries, as seaweed farming expands alongside other Blue Industries there will be numerous opportunities to transform the sustainability of socio-ecological systems.These gains will only be achieved through proper accounting and consideration of the range of local environmental and social costs.

Methods
We conducted a systematic review of the peer-reviewed literature in four languages (Chinese, Japanese, Korean, and English), screened each article according to a specific set of criteria, and then used the resulting papers to classify and analyse the environmental and social impacts of seaweed farming.Here we are defining 'impact' as a direct consequence, either beneficial or deleterious, to a social or environmental system that arises from the cultivation of seaweeds.

Search and screening
We conducted a quantitative systematic review [175] of scientific literature, published at any time in a peer-reviewed journal which reported and analysed empirical evidence from existing commercial or experimental seaweed cultivations.We excluded studies that focused on cultivations primarily purposed for restoration (e.g.[176]), because this study focuses on the impacts that the growing commercial seaweed industry will have.As English-language-only reviews can exclude important insights from other languages [45], we assembled a multi-lingual team of eight people which included at least two reviewers fluent in each of the four target languages: Chinese, Japanese, Korean and English.These languages were chosen because they represent countries that have historically had well-developed seaweed farming industries [177] and well-developed scientific communities.
We designed search terms to identify all articles related to ocean and coastal seaweed farming in English, and each language team of reviewers translated these terms into their respective languages.These terms were input into at least two relevant scientific databases for each language in July 2019 and again in November 2021 (see S2 Table for databases used and search terms).It is possible that aquaculture and mariculture studies, which did not specifically identify seaweed as the organism being cultivated, may have been omitted from these searches.The results from these initial literature search yielded 5,634 unique English titles, 1,878 unique Chinese titles, 4,158 unique Japanese titles, and 4,680 unique Korean titles (S1 Fig) .Each reviewer then read through each title and screened papers based on whether they (i) focused on seaweed farming in a marine or coastal context (excluding studies into freshwater, tank-based, or other closed systems), and (ii) were likely to contain an assessment of an environmental or social impact of seaweed farming.
Two reviewers for each language independently conducted the screen and any titles included by at least one reviewer were included.The abstracts of the remaining papers were read and only papers that met all the following criteria were included: (i) not a review, (ii) a study about cultivating seaweed in an open marine context, (iii) contained an assessment of an environmental or social impact, (iv) the seaweed cultivation studied was not for the purpose of seaweed restoration.Again, two reviewers performed independent screens and papers accepted by at least one reviewer were included in the next step.During this stage, papers were divided into two categories, those investigating environmental or socio-economic impacts.Finally, reviewers read through the full text of each paper and again assessed the suitability of the paper for inclusion based on the above criteria.

Data extraction
Environmental and social impacts from the selected publications were summarized, along with information on (i) general study details, (ii) cultivation context, (iii) impact type, and (iv) impact polarity (i.e.whether the impact was generally beneficial, deleterious, mixed, or neutral) (See S1 Text for worksheets).We distinguish between these polarity categories based on the descriptions of the literature authors in describing the impact in the results and discussion of their paper.So for example, impacts that are described in favourable terms or in need of promotion were designated 'beneficial', impacts that were described negatively or in need of being avoided were designated 'deleterious', impacts that were reported as possibly being one or the other, were designated 'mixed', and those impacts which were described without subjectivity were designated as 'neutral'.Owing to the different structures of papers documenting social and ecological impacts, we adopted a slightly different approach for extracting data from each.For environmental impacts, reviewers recorded the impact, as described in the paper for all papers with no guidance, and then these impact categories were consolidated according to type post hoc.In some studies that investigated multiple impact categories or species concurrently, each documented impact and species was treated as a separate observation, meaning that multiple observations may have been drawn from individual studies.In the case where multiple papers examined a single event from one system, for example the harmful bloom of Ulva spp.from Chinese seaweed farms (e.g.[178]), one observation was recorded.While several papers examined growth rates and community characteristics of epiphytes growing on cultivated seaweeds, unless impacts were documented on ecosystems outside of farms, these studies were not included.
For social impacts, 20 papers were read by two reviewers independently who then identified and agreed upon impact categories.These were then used to create a worksheet that guided data extraction for the remaining papers, where each study was recorded as one unique observation.During the reviewing stage, it was found that the majority of studies focused on communities where seaweed biomass is being exported, often into global hydrocolloid markets; this theme was added to the extraction worksheet.

Semi-qualitative method for assessing impact polarity
For each of the 20 social and environmental impact categories (Gender, Carbon Assimilation, etc.), we assessed the polarity of the impact of seaweed farming, i.e. whether the reported impacts of seaweed farming were consistently positive, negative, variable, or negligible.We describe the process for designating impact polarity in detail in Fig 6 .For each impact category, the polarity of the impact is reported as either Beneficial, Deleterious, or Negligible if a statistically clear majority of observations (>50%) support that conclusion (see Fig 6).The categories Aerosols and Marine Plastics were excluded from this analysis because there was only one observation in each; all other categories had at least eight observations.Where there is no statistically clear majority, we report a Variable impact.
For impact categories characterized as Beneficial, Deleterious, or Negligible, a confidence score is derived from hypothesis testing with the binomial distribution.The null hypothesis is that the observations of Beneficial and Deleterious impacts in that category are equally likely.The probability of the number of Beneficial or Deleterious observations occurring is then calculated.Where the null hypothesis is rejected, the likelihood of false positive (i.e.rejection of the null hypothesis where it is true) is used to assign qualitative confidence scores (Medium: 0.05 > p > 0.005; High p < 0.005).Where the null hypothesis cannot be rejected, we re-assign the category as Variable.

Fig 1 .
Fig 1.Twenty social and environmental impact categories of seaweed farming as identified by our systematic review.Impact polarity based on available data is noted by the colour of each circle.For impact categories with greater than one observation, we either report the polarity with a majority of observations supporting it (>50%), or we report a 'Variable' impact.For impact categories with only one observation, we report 'Data Deficient'.Confidence scores are based on a cumulative probability binomial test that the identified impact is incorrect (Medium: 0.05 > p > 0.005; High p < 0.005).The numbers in parentheses refer to the number of observations from each impact category.

Fig 2 .
Fig 2. The interconnected socio-ecological impacts of seaweed farming in Tanimbar Kai, Indonesia.In this community, the rise in seaweed farming has led to increased income, which has raised living standards, and led to greater access to education, medical services and livelihoods for women.Conversely, this has also led to family members spending more time out of the house and having less time to prepare healthy meals or pass on cultural knowledge to youth.Increased reliance on seaweed farming and the raised expectations of income has also made former livelihood activities, such as copra production and shark fishing, less attractive, making it harder to 'fall-back' on these activities as seaweed production diminished due to falling global market prices and failed crops due to overexploitation of environmental resources.Finally, seaweed farming has reduced fishing effort, making stricter fishing regulations easier to adopt.Blue arrows indicate positive feedback, red arrows indicate negative feedback, and bidirectional red arrows indicate a balancing loop between nodes.Case study based on Steenbergen et al. 2017[47].Icons are from OpenClipArt or generated by DALLE2.https://doi.org/10.1371/journal.pstr.0000042.g002

Fig 3 .
Fig 3. Number of observations of seaweed farming impact on biodiversity: marine mammals, plankton, seagrass, invertebrates, and fish.Where seaweed farming was associated with a benefit to any of these biodiversity indicators, the impact was classified as beneficial.Where seaweed farming was negatively associated with the biodiversity indicator, the impact was defined as deleterious.Other cases were defined as either neutral (neither positive or negative impacts) or mixed (both positive and negative effects).'No impact' was defined for observations where the seaweed farms had no impact on the biodiversity indicator.

Fig 4 .
Fig 4. The frequency of observations of a given environmental impact for each family of seaweed studied.A single study may include multiple observations.The intensity of the fill in each box corresponds to the number of observations within the specified impact for that family.For example, in Gracilariaceae there are numerous observations of beneficial effects on biodiversity, water quality and carbon, very few beneficial effects of acidification and algal blooms, and no observations in other impact categories.For the same family there are some observations or mixed/neutral impacts, some instances of no impact being observed, and one instance of light being negatively impacted.The coloured bar on the x-axis shows which broader group each family belongs to (Brown, Red, and Green algae).Boxes with points represent unique observations only available in non-English literature.n = 553.

Fig 5 .
Fig 5. Regional distribution of empirical evidence for a) each impact type (environmental and social) where 'Not Studied' indicates no evidence for an individual impact type from that region, and b) the polarity of impacts.In b), the size of the pie represents n, the number of observations from that region.Countries were assigned to regions according to the 'subregion' designation from Natural Earth [154].https://doi.org/10.1371/journal.pstr.0000042.g005