Main

With the global population expected to reach nearly 10 billion by 2050, there is a pressing need to transform food systems to make them more sustainable, efficient and equitable. Small-scale producers, who support 2–3 billion people (between 25% and 40% of the global population)1, are central to this transformation, contributing to nutrition, environmental sustainability, economic growth and the maintenance of cultural heritage. The 2030 Agenda of the United Nations (UN) recognizes their importance and advocates doubling their productivity and income by ensuring their access to land (Sustainable Development Goal 2 (SDG2): Target 2.3), marine resources and markets (SDG14: Target 14.b) and financial services (SDG8: Target 8.3). But despite their key contributions, policy and support for small-scale producers is constrained by the inability to tease apart their diversity, hindering the design of more specific actions tailored to their needs2,3,4 and constraining their full potential within the sustainable development agenda5,6,7,8.

A number of international commitments have been made to promote food systems transformation, support sustainable food and agriculture and guide food-related policy actions in line with the 2030 Agenda: for example, the UN Food Systems Summit9, the CFS Voluntary Guidelines on Food Systems and Nutrition10 and the Sharm el-Sheikh Joint Work on Agriculture11, which replaced the Koronivia Joint Work on Agriculture after COP27. Yet translating this into practice remains particularly challenging for small-scale producers, whose diversity and context-specific realities often resist simple categorization. To help bridge this gap, the UN Food and Agriculture Organization (FAO) Principles of Sustainable Food and Agriculture12,13 provide operational guidance to align food systems policies with SDG targets, including those focused on small-scale producers.

Small-scale producers vary widely within and across food sectors depending on geography, ecology, economic development and culture, among other factors4,5. The varying definitions of small-scale producers in farming2,14, fisheries15,16 and forestry17 give rise to confusion and lack of clarity in many legal documents and has been linked to policy paralysis and inaction18,19,20. The lack of global standardized definitions complicates the implementation of international or regional agreements and hinders data comparability across countries20,21. Although calls for a better understanding of what constitutes ‘small-scale’ have grown2,14,22, particularly in fisheries23,24, practical tools to systematically organize this diversity and make it actionable for policy remain lacking, creating a critical gap in efforts to achieve the SDGs25,26 (Extended Data Table 1).

Table 1 Potential policy actions for the five archetypes of marine small-scale fisheries identified following the FAO Five Principles of Sustainable Food and Agriculture12,13
Full size table

Despite common traits, such as small operational size and high reliance on family labour, the diversity of small-scale producers makes universal definitions challenging25,27. Compounding these challenges is the blurred use of terms like ‘smallholder’ and ‘small-scale producer’. In this Article, we use ‘small-scale producer’ to reflect a broader framing, particularly relevant in fisheries, that captures smallholder dimensions but also extends to governance, markets and production systems.

Small-scale producers often rely on diverse cultivation and exploitation strategies in geographically discrete areas, switch between subsistence and commercial uses and exhibit high seasonality—patterns that are hard to capture in surveys, particularly in low-income countries8—leading to simplified quantitative classifications based on criteria such as land-holding area or vessel size4,15,24. Such categorizations overlook socio-economic and governance attributes that are of fundamental importance in natural resource management and for the integrated perspective advocated by a food system approach28,29. There have been attempts to disaggregate differing scales of producer using multiple criteria21,22,30,31,32; however, data limitations with small-scale producers often prevent their implementation and can lead to inconsistencies when indicators conflict, such as a farm being small in land size but not in labour input20,25,33.

Researchers have engaged with this complexity, emphasizing that diversity among small-scale producers reflects adaptive strategies shaped by social, ecological and/or economic conditions rather than simple measures of scale34,35. In fisheries and aquaculture, recent work5 expanded on these insights by characterizing actors across eight dimensions of diversity through expert-based synthesis of 70 case studies. Although these conceptual frameworks underscore the richness of small-scale systems, translating this complexity into scalable groupings that can inform cross-country comparisons and policy action in a practical, low-cost manner remains unsolved. Continued reliance on overgeneralizable definitions constrains the design of effective management, policy and support programmes by making it unclear who within the sector is the target for intended benefits or costs. This lack of clarity often disadvantages the most marginalized actors (for example, women and indigenous groups)6,8.

Our paper addresses this lack of clarity by offering a systematic, quantitative method to identify the diversity of small-scale producers. By analysing more than 1,000 small-scale fishing producers, representing 66% of the global estimated marine small-scale catch8, we demonstrate how a characterization tool and a cluster-based model (multilevel latent class cluster analysis (LCCA)) can be applied together to generate nuanced, scalable groupings within the range of global small-scale fisheries. Rather than enforcing strict cut-offs between small- and large-scale operations, our approach prioritizes the most relevant attributes for each context. This characterization, which can complement other global sustainability frameworks, serves as a first step for policymakers to effectively connect diverse local realities with global food systems objectives.

Challenges in developing context-specific policies in small-scale fisheries

Small-scale fisheries account for around 40% of global catch and have the potential to satisfy the daily nutrient needs of 2.3 billion people for six nutrients commonly lacking in diets8. They act as a livelihood safety net, with almost the same number of fishers engaged in subsistence fishing as in commercial fishery employment36, and generate around 90% of employment for women in the entire fisheries sector8. In the global south, where small-scale fisheries are at greater risk from climate change than other food sectors37, the population heavily relies on them to avoid hunger and malnutrition38 and eradicate poverty36.

The need to improve definitions of small-scale fisheries was highlighted during the plenary discussions at the 2024 UN FAO Committee on Fisheries, the leading global fisheries body. Current classifications, often based solely on size, can undermine policy effectiveness by incorporating industrial-like fleets or excluding legitimate small-scale operations despite their larger size16 (Fig. 1). For example, the recent agreement on fisheries subsidies adopted by the World Trade Organization to reduce harmful subsidies. Automatically excluding larger vessels from the small-scale category could undermine policies for food security by overlooking fleets that play vital roles in local nutrition39,40. Moreover, accurately characterizing small-scale fisheries is also considered crucial for implementing regional policies such as those of the European Union41, the International Commission for the Conservation of Atlantic Tunas42 or the Indian Ocean Tuna Commission43. At the national level, corporate fleets of small vessels, driven primarily by profit, may provide less stable local benefits than family-run fleets, which contribute directly to local economies44. Treating them equally and ignoring such ownership differences can result in ineffective economic policies20.

Fig. 1: Alternative criteria for small-scale fisheries.
figure 1

The proportion of marine small-scale fisheries catch that would be included depends heavily on which attribute is used to define ‘small-scale’. For each country, we applied each criterion independently to the characterized fishery units and calculated the share of total characterized catch that meets that definition alone. This highlights how unidimensional definitions can underestimate or exclude fisheries that are operationally small-scale. For example, in Argentina and Kenya, some small vessels operate like industrial fleets, whereas in Bangladesh, larger cooperative-managed vessels contribute substantially to local nutrition. Total percentages of characterized catch varied by country (23–100%; see Supplementary Table 3).

Source data

Full size image

Previous efforts to broaden the characterization of small-scale fisheries to include environmental, social and/or political attributes have been confined to national analyses using quantitative and qualitative data, preventing cross-national comparability and often requiring costly survey data15,19,31,45,46. In contrast, our tool allows for objective comparisons by standardizing metadata collection and application of a scoring approach for each fishery, regardless of each country’s definition of ‘small-scale’ (see the section ‘Fishery Units’). This consistency allows fisheries to be compared on a shared basis, capturing diverse operational, socio-economic and technological attributes while reducing dependency on extensive empirical data, making it suitable for data-limited contexts. Our tool provides a characterization score that depicts the overall scale of the fishery: that is, the degree of ‘small-scaleness’31 and its place along the small- to large-scale continuum, enabling the identification of fundamental types of small-scale fisheries at global, regional and national levels to inform policy.

Results

How small is ‘small’?

Small-scale fisheries from 43 countries (n = 1,255 fishery units) exhibited a broad diversity of characterization scores globally (Fig. 2a) and regionally (Fig. 2b), along a small-scale large-scale continuum of operational scale, spanning from a score of 0 (corresponding to the ‘smallest’ operations, for example, gleaning fisheries in the Philippines and Fiji) to a score of 33 out of a possible 39 (the most mechanized and fully economically integrated operations, for example, the tuna fishery in Indonesia) (Fig. 2c).

Fig. 2: Distribution of catch by operational scale.
figure 2

The marine catch from small-scale fisheries shows great diversity along the small- to large-scale continuum of operational scale. Each dot represents the cumulative share of catch corresponding to a given total characterization score, grouping all relevant observations (whether at the level of regions, countries or individual fishery units within countries) that fall at that score. a–c, Distribution of marine small-scale fisheries catch at global (a; N = 43 countries), regional (b; Asia N = 11 countries, Americas N = 9 countries, Africa N = 19 countries) and national (c; based on fishery units grouped by country) levels across the total characterization score. Black circles indicate where 50% of the cumulative catch is reached along the total characterization score. By aligning panels vertically, readers can identify which countries contribute to the catch at each score: for instance, a score of 0 is present in the Philippines and Fiji. Total % of characterized catch varied by country (23–100%; Supplementary Table 3). Gabon, Congo, Congo DR and Liberia are excluded from c due to insufficient data. Greenland is grouped with the Americas following the FAO regional classification system. SSF, small-scale fisheries.

Source data

Full size image

Fishing operations at the lowest levels of operational scale—that is, the smallest scale of small-scale fisheries (score of 9 or lower), characterized by manual labour, minimal equipment and serving local needs—were present in 19 countries, contributing to 4% of the global marine small-scale catch (Fig. 2a). At the higher end of the small- to large-scale continuum, small-scale fisheries scoring 29 or higher are characterized by corporate ownership, multiple paid crew and broader market chains. Such fisheries were identified in only four countries (Indonesia, Maldives, Argentina and Peru), accounting for around 2% of the global small-scale catch (Fig. 2a).

Nearly three-quarters of the global small-scale catch comes from fisheries scoring 19 or lower, with representative fisheries from all countries characterized. Almost half of the global small-scale catch is associated with fisheries with a score of 13 or lower (Fig. 2a), a trend predominantly driven by China, which accounts for about 40% of the catch in our sample.

We also identified important regional differences, with nearly half of Asia’s small-scale from fisheries scoring 12 or less, indicating they are at the lower end of the small- to large-scale continuum. Conversely, half of the Americas’ catch comes from fisheries scoring 24 or less, indicating a much broader range of scale (Fig. 2b). Similar to the global trend, the lower scores in Asia are mainly influenced by China, with smaller contributions from the Philippines and Sri Lanka. We also found some fisheries in Asia that push the regional small-scale fisheries profile towards a higher score, such as tuna fisheries in Indonesia and the Maldives showing marked large-scale characteristics, with 92% of the Maldives’ catch coming from large pole-and-line vessels scoring over 19 (Fig. 2c). The high regional score for the Americas’ catch is predominantly shaped by the fishing fleets of Peru, Chile and Argentina, which represent the larger operational scale of small-scale fisheries. In Africa, nearly half of the catch falls around the midpoint (score of 16 or less), with countries such as Sierra Leone and Senegal showing larger operational scale compared with Madagascar and South Africa which exhibits traits of lower operational scale.

At the national level, fishing operations scoring 9 or less accounted for around 30% of the national small-scale catch in countries like the Philippines and Fiji and as low as 15% in others such as Argentina and Norway—an exception being Madagascar, where they constitute the entire small-scale catch (Fig. 2c). Catch volumes from these smaller operations are likely to be underestimated due to the lack of associated catch data, stemming from challenges in the implementation of monitoring systems8.

Archetypes

Although the total characterization score enables objective, multidimensional comparisons along the continuum of operational scale, two fisheries can achieve the same score through different combinations of attributes, making it challenging to capture their unique characteristics for tailored policy development. To address this issue, five global archetypes of marine small-scale fisheries were identified using multilevel LCCA (Fig. 3) based on 13 fishery attributes (Extended Data Table 2) in 1,255 fishery units from 43 countries. Archetypes were labelled by their main functions: (1) ‘nutrition and income safety nets’, (2) ‘in-between’ and (3) ‘economic engines’. Certain archetypes are more common in some regions than others (Supplementary Table 4).

Fig. 3: Global archetypes of marine small-scale fisheries.
figure 3

The five archetypes are identified based on their similarities on operational, economic, technological and post-harvest attributes. They can be grouped as follows: group 1 (small-scale fisheries acting as nutrition and income safety nets), group 2 (in-between seasonal small-scale fisheries) and group 3 (small-scale fisheries as economic engines). Archetypes were identified using multilevel LCCA (1,255 fishery units from 43 countries). The mean value and standard error of each attribute per archetypes is shown (Supplementary Table 5), where a value of 0 represents an absent/very low magnitude and 3 a high one (Extended Data Table 2). N is the number of fishery units per archetype. Arch., archetype.

Source data

Full size image

Small-scale fisheries as nutritional and income safety nets (Archetypes I and II)

These are low-technology, low-capital fishing operations in shallow coastal and estuarine waters (Fig. 3) from the global south (N = 14 countries) (Supplementary Table 7). Catch is usually for subsistence or traded or sold in local markets, with gear/boat owners usually working individually or in cooperatives without paid crew (Fig. 3). In Archetype I (N = 13 countries), fishing is an on-foot activity typically targeting inshore invertebrates (for example, clams, crabs, sea urchins) (Supplementary Fig. 2) using bare hands or simple hand-held tools (gleaning).

In Archetype II (N = 3 countries), demersal and pelagic fish are caught using labour-intensive (for example, cast nets, spears) or passive (for example, gillnets, traps) gear from non-motorized and unmechanized vessels (for example, paddled canoes) (Fig. 3). Both archetypes have limited integration into national economies and tend to have minimal government oversight (Fig. 3), often operating informally, resulting in many unreported catches. Oceania and Africa have a higher likelihood of such fisheries, with about half of Oceania’s small-scale fisheries in Archetype I and half of Africa’s in Archetypes I or II (Supplementary Table 4).

Small-scale fisheries as economic engines (Archetypes IV and V)

These fisheries are characterized by greater capital, investment and technology, which makes them more formalized and integrated into national economies through licence and landing fees (Fig. 3). Motorized vessels (typically with engines ≤100 horsepower) with better storage and refrigeration translate into a lower dependence on inshore waters and local markets, sometimes sustaining trading across distant locations. Archetype IV (N = 28 countries) features <12-m vessels operating up to 10 km from shore, conducting day trips with basic chilling facilities. Archetype V (N = 33 countries) typically operates beyond 10 km offshore, with vessels up to 12–24 m that might spend two to 4 days at sea and might be equipped with advanced refrigeration (Fig. 3). In Archetype IV, boat owners hire crew but also participate in fishing, whereas in Archetype V that is not usually the case (Fig. 3). Key targets are pelagic fishes, accounting for 43% of the total catch in Archetype IV and 60% in Archetype V, mainly comprising herrings, sardines and anchovies (Supplementary Fig. 2). Over 80% of the small-scale fisheries catches in the Americas and Europe fall into Archetype IV, whereas Archetype V is more common in Africa and, to a lesser extent, in Asia and Europe (Supplementary Table 4).

In-between small-scale fisheries (Archetype III)

Exclusively found in Asian countries in our sample (China, India, Iran and Turkey), these fisheries have the key differentiating attribute of being full-time but of a seasonal nature (Fig. 3). These fisheries target a diverse range of taxa, dominated by pelagic species but also including coastal and demersal fishes and invertebrates (Supplementary Fig. 2). Despite their taxonomic diversity, these fisheries share a common feature of seasonality, likely shaped by a combination of biological and environmental drivers, such as reproductive cycles, monsoonal dynamics and possibly management-related constraints (for example, rotational harvest, seasonal closures). Technologically, they resemble Archetype IV, with motorized vessels <12 m, fishing grounds <10 km from shore and participation of paid crew. However, socio-economically, they align more with Archetypes I and II, as they target fish for direct human consumption and present low levels of ownership and integration into national economies.

Archetypes at national levels

Beyond global and regional patterns, our approach also sheds light on the diverse roles of small-scale fisheries within countries, providing important insights for local tailored management policies. For example, our analytical approach identifies countries where small-scale fisheries primarily function as economic engines, such as Kenya, where Archetypes IV and V dominate small-scale fisheries catch (Fig. 4a), or where they act as nutrition and income safety nets, as seen in Madagascar, where Archetype II is most prevalent, followed by Archetype I (Fig. 4b). Understanding of the archetypes identified through our approach can be further enhanced by integrating additional country data when available, offering more comprehensive insight into how archetypes are distributed across ecosystems and catch (Fig. 4a), administrative or regulatory jurisdictions (Fig. 4b) and gender (Fig. 4c), alongside other relevant policy variables, thereby supporting regulations that are more effectively aligned with local needs and realities.

Fig. 4: National small-scale fisheries archetypes.
figure 4

Mapping national small-scale fisheries archetypes offers valuable policy insights, showing their relevance across different organizational structures, such as habitat and species types (Kenya), administrative and regulatory jurisdiction (Madagascar) or gender (Philippines). a–c, Marine small-scale fisheries archetypes identified in multilevel LCCA for Kenya (a; 29 fishery units characterized, accounting for 70% of the country small-scale fisheries catch), Madagascar (b; 286 fishery units characterized, accounting for 79% of the country small-scale fisheries catch) and the Philippines (c; 16 fishery units characterized, accounting for 100% of the country small-scale fisheries catch). Marine small-scale fisheries catch is displayed in tonnes at the centre of each circle. The gender data from the Philippines refer to participation in harvest, excluding pre- and post-harvest activities.

Source data

Full size image

Discussion

By identifying archetypes of small-scale fishing producers, we illustrate that the great heterogeneity and local specificity that define these producers across geographical levels (from global to national) can be systematically organized and aggregated into a continuum of small- to large-scale operational scale. Importantly, our multidimensional characterization approach and resulting archetypes enable their objective comparison across countries and regions. In doing so, our analysis challenges the utility of the conventional small- versus large-scale dichotomy by revealing a continuum of production strategies, social and governance arrangements within the small category itself. Our findings resonate with the global call for context-sensitive characterization of small-scale fisheries27 while offering an empirical approach to support implementation and policy alignment at national to global levels. The archetype analysis can guide regional and global policies for fisheries with shared characteristics, as illustrated in Table 1 and explained later, while also identifying specific national needs. Although our analysis is grounded in small-scale fisheries, the method offers insights applicable to small-scale producers in other food systems facing similar challenges of heterogeneity and data limitations. This in turn supports the design and operationalization of international policy recommendations and ensures their coherence with global sustainability goals.

The characterization matrix applied in this study, developed and tested by FAO47,48, addresses several interrelated challenges typical of data-limited contexts in the global south. Unlike previous multidimensional approaches15,21,30,31,32,46, it is less reliant on survey data, and it is cost-effective and easy to implement, with freely accessible online training through the FAO website (https://elearning.fao.org/course/view.php?id=1050). In this study, the matrix was applied in nine low-income countries, as classified by the World Bank, where limited institutional capacity often leads to scarce fisheries data, and its structure allowed for application based on expert knowledge even in the absence of systematic data collection. The matrix has also been piloted in several regions with a high prevalence of data-limited fisheries, including the Fishery Committee for the Eastern Central Atlantic, the Fisheries Committee for the West Central Gulf of Guinea and the Indian Ocean Tuna Commission. The matrix offers flexibility, bringing comparability to different ways of defining fishery units and effectively balancing local specificities with broader comparisons across fisheries at national, regional and global levels. It also accommodates varying levels of data availability, from entire fisheries to individual vessels or fishers (‘Fishery units’), enabling increasingly detailed assessments as more data become available. By focusing on specific attributes, such as catch use and market disposal, rather than relying on a total score or a single metric such as size or power, our analytical approach enables policymakers to identify small-scale fisheries that align with policy objectives, thereby increasing the effectiveness of the policy to be implemented.

Small-scale fisheries archetypes provide useful insights across different organizational structures, such as habitat types (Kenya), administrative and regulatory jurisdictions (Madagascar) or gender (Philippines) (Fig. 4). These organizational differences reflect variations in data availability, cultural views and local priorities. Identification of small-scale fisheries archetypes also facilitates the development of gender-sensitive policies49 by clearly identifying the characteristics of those fisheries where most women are involved. The characterization matrix and archetypes can support two complementary applications. Countries can use the matrix to develop national typologies by incorporating locally relevant variables and simpler clustering approaches. Alternatively, they can rely on the assignment probabilities to identify which global archetypes are represented nationally (Supplementary Table 7). In Argentina, researchers applied the characterization matrix and used factor analysis of mixed data along with k-means clustering to develop a national typology50. Their analysis identified three archetypes, each also associated with varying degrees of informality and management intensity. This matches the number of archetypes we found for Argentina using our global model (Supplementary Table 7). Although the first approach enables deeper, more context-specific insights, the second situates national diversity within a broader global and regional frame, facilitating cross-country comparisons.

Transformation pathways in food systems are complex and shaped by numerous disruptors, such as market and technological changes, which affect different types of small-scale producers in varied ways51. Yet policies often adopt a ‘one size fits all’ approach that usually fails to consider the unique needs of diverse producers5 at varying operational scales. Our approach tackles this by developing archetypes that combine information from multiple variables, responding to calls for integrating socio-economic criteria in small-scale producers’ policies20,21,31,45. These archetypes enable the proposal of targeted policy actions (Table 1), offering an illustrative action roadmap that serves as a starting point for further refinement, helping avoid maladaptive practices while promoting resilient outcomes. To ensure alignment with global sustainable development agendas, these actions are linked to relevant SDG targets and organized through the FAO Five Principles of Sustainable Food and Agriculture12,13, which serve as a bridge between context-specific priorities and global goals (Extended Data Table 1).

When a region, country or local jurisdiction seeks to advance a specific development priority for small-scale producers, the archetype-informed actions in Table 1 can help tailor responses to the particular context. For instance, in efforts to promote gender equality in resource access (SDG5: Target 5.a), fishers in Archetype I benefit from recognition of informal and unpaid work and integrating customary laws into legal systems5,6,7,8, whereas in Archetype IV, efforts focus on strengthening women’s roles in cooperatives and expanding access to digital tools that enhance participation and decision-making. These differences reflect broader patterns: for example, most on-foot fishers in our analysis fall into Archetype I, often engaging in informal subsistence activities (for example, gleaning in the Pacific), whereas others, such as full-time paid fishers in higher-income countries targeting highly profitable species like Chilean abalone or Spanish barnacles, align more closely with Archetype IV. Likewise, to address post-harvest losses (SDG12: Target 12.3), actions for Archetype I involve improving storage and processing infrastructure supported by low-interest loans52, whereas Archetype IV benefits from alternative marketing programs, such as eco-labelling, that highlight the origin and quality of legally harvested products to gain consumer trust and access new markets53,54. These examples illustrate how archetype-based insights can guide differentiated policy responses while contributing to shared global targets.

The five archetypes identified in this study effectively capture the complexity and diversity of over 1,000 small-scale fisheries around the world. Although our attribute set partially overlaps with those used by ref. 5, our analysis goes further by statistically deriving globally validated archetypes, distinguishing subtypes within nutrition- and income-oriented fisheries and supporting spatial comparisons across countries and regions. Because our approach prioritized broader comparisons by respecting national definitions, we acknowledge that this may have excluded some fisheries considered small-scale under alternative criteria and equally included others that might be classified as large-scale elsewhere. Clustering quality and model validation metrics (Extended Data Table 4) offer confidence in the identification of these fundamental archetypes, illustrating the nuanced continuum from smallest to largest scale of operations19,30. Nonetheless, incorporating additional data, especially from underrepresented regions such as Oceania and the Americas (Extended Data Table 3), could uncover new divisions or subdivisions of archetypes not fully captured in our analysis. Although we focused on marine fisheries, inland fisheries were explored using the same archetypes clustering approach. It was found that the results were shaped by too few countries and reflected their national characteristics rather than revealing distinct inland fishery archetypes. Future application to inland fisheries could benefit from additional data to account for important features such water body type and national development context to better capture their diversity.

From a broad perspective, the scale of operations in small-scale marine fisheries across the regions mirrors the patterns observed in agriculture, with Asia generally skewed towards smaller and the Americas leaning towards larger operational scales, and with Africa more evenly distributed across the continuum of scale33. Similarly, in high-income countries (for example, Norway, UK, Greenland) the profile of small-scale fisheries tends towards larger operational scale driven by intensification, regulation, technological and economic factors, with smaller-scale operations playing a minor role20,33. These parallels suggest that lessons learned from small-scale fisheries can offer valuable insights into broader efforts to support small-scale food producers within the context of global transformation of food systems. Targeted policy actions are needed as part of this transformation that are capable of feeding a growing population in a way that is both healthy and sustainable55.

Small-scale producers, key to this global shift and with interconnected livelihoods across food sectors37, can learn from each other’s successes and challenges to accelerate adaptation and resilience. The use of the characterization matrix and analysis of archetypes represents a step forward by organizing the diversity of small-scale producers, enabling more structured approaches to understanding their needs and tailoring policy actions. If archetypes are developed across food sectors, potential links between them can be explored, fostering national policy coherence as well as broader cross-sectoral learning and collaboration. Although efforts to characterize agricultural small-scale producers do exist, such as FAO’s work on SDG 2.3.156) and typology development22, these tend to rely on a narrower set of indicators than those captured in our matrix, offering opportunities to adapt or align broader ways of capturing producer diversity across food sectors. Small-scale producers, whether in terrestrial or aquatic systems, face shared challenges, such as insecure land and access rights (particularly for women4), heightened vulnerability to climate change compared to large-scale counterparts57 and limited voice in policy decisions3,8,23, all of which highlight the need for better coordination. By linking governance reforms, income diversification and cross-sector policy coherence, the current momentum—reflected in the 2030 Agenda’s recognition of small-scale producers—can support the development of a more resilient, equitable and sustainable global food system.

Over four decades ago, ref. 18 noted that while broad definitions of small-scale food producers can foster unity during the initial stages of policy development, they often create challenges during implementation. This issue becomes evident in instances where small farms are equated with family farms, community forestry with smallholder forestry or small-scale fishing with subsistence fishing. Although in many cases these concepts overlap, they are not synonymous, and conflating them can lead to substantial misunderstandings16,17,20,24,33. To avoid this, policymakers should prioritize establishing clear, precise definitions of ‘small’ that reflect the unique characteristics, context and goals of each policy. Our results, focused on small-scale fisheries, provide an empirical foundation for this by identifying common types of producers based on recurring patterns in production, social and governance attributes.

Our findings demonstrate that small-scale producers can be grouped into meaningful categories. This key insight challenges the narrative that addressing the needs of small-scale producers is intractable and instead highlights their potential to contribute meaningfully to food systems transformation through well-targeted policy action at regional, national and local levels. Our approach is low-cost, simple to apply and well-suited for decision-making processes in data-limited contexts, particularly in the global south. The case of small-scale fishing serves as an inspiration for future work aimed at better understanding and categorizing small-scale producers across other food sectors, paving the way for more precise policy-making. A clear characterization of small-scale producers is a crucial aspect for implementing international agreements (for example, the World Trade Organization’s agreement on fisheries subsidies), improving data comparability across countries and regions and effectively monitoring progress toward the SDGs (for example, SDG 2, SDG 8 and SDG 14 explicitly mention ‘small-scale food producers’). All of these are critical processes affecting the lives of tens of millions of people around the world.

Methods

Small-scale fisheries data were empirically compiled through country-level case studies and characterized using a characterization matrix tool as part of the Illuminating Hidden Harvests (IHH) initiative. More than 1,000 marine fishery units (n = 1,255) from 43 countries were analysed using a multilevel LCCA technique to identify stable groupings or archetypes as described below. Case study countries were selected through a ranking process based on their absolute and per capita contributions to production and employment and the importance of fish as a source of protein in peoples’ diets as per ref. 8. Small-scale fisheries included for analysis represented approximately 88% of the marine small-scale catch for those countries and 66% of the global estimated marine small-scale catch for the period 2013–20178. This global estimate (25.1 million tonnes, representing 31% of the global marine catch) was derived from observed data across 52 case studies and model-based extrapolations for an additional 100 countries (see section B1.1 of the supplementary material in ref. 8 for a detailed description).

Characterization matrix tool

Marine small-scale fisheries were characterized in a systematic and objective manner using the characterization matrix tool adapted by Funge-Smith et al.48 from Funge-Smith47. This matrix captures 13 fishery attributes related to the operational (for example, fishing trip duration), technological (for example, refrigeration capacity), socio-economic (for example, disposal of catch) and post-harvest use of the catch to create a standardized framework for intercomparability across countries and regions. The definitions of the fishery attributes are provided in the Supplementary Table 1. Each of the 13 attributes of the matrix can be individually scored (0, 1, 2, 3) (Extended Data Table 2). The aggregation of all the scores from the 13 attributes provides a total characterization score for the fishery that ranges from a minimum of 0 (for example, the most artisanal-like fishing operations) to a maximum of 39 (the most industrial-like fishing operations). The total score of each fishery or fishery unit (described below) depicts its degree of ‘small-scaleness’31. The 13 attributes include common features found in the peer-reviewed and grey literature31,47.

Experts were instructed to characterize their country’s small-scale fisheries using the characterization matrix (Extended Data Table 2). In situations where fishery units (see definition of fishery units below) included diverse characteristics, experts were guided to choose the score that most accurately reflected the group as a whole.

Fishery units

In recognition of the vast diversity and complexity of small-scale fisheries, the IHH initiative did not prescribe a rigid definition of the subsector. Instead, it allowed country experts to use the legal or most commonly used definition of small-scale fisheries within their countries, which ensured that the characterization of each fishery was contextually accurate. This approach respected national nuances while maintaining comparability across different fisheries by consistently gathering the same metadata for each unit, regardless of how ‘small-scale’ was defined in that particular context8. Although this approach ensured that the characterization of small-scale fisheries reflected national contexts, it also carries some limitations. Relying on country-specific definitions may have led to the exclusion of fisheries that would be classified as small-scale under other criteria and equally to the inclusion of fisheries that might be considered large-scale elsewhere. However, we deliberately prioritized the use of national definitions to avoid imposing external standards (for example, vessel size, fishing gear restrictions) that could misrepresent the local realities of small-scale fisheries. This was both an ethical and a practical choice, as collecting primary data across fisheries in all countries was not feasible given resource and logistic constraints.

A ‘fishery unit’ is a unit of assessment at the data collection phase for a relatively homogenous group of fishing activities. Each country’s case study expert selected these units based on the type of management and available data, ensuring that they reflected characteristics relevant to the country’s small-scale fisheries sector. Fishery units could encompass one or multiple fisheries with variations in location, gear types, fishing methods, vessel types, target species or how catch data was originally aggregated (for example, by fishing fleet). Experts organized species and fleet data into fishery units, balancing data limitations with the aim of maintaining the greatest possible disaggregation for catch, an essential part of the case study protocol. Typically, the experts used bio-ecological (for example, taxa, habitat type) and/or technological (for example, gear or vessel type) attributes to define them (Supplementary Table 2). Under the highest-resolution scenario, a fishery unit is composed of a single or closely related species targeted by a particular gear in a specific location. Under a lower-resolution scenario, a fishery unit might group multiple species targeted by different types of gear at a larger scale.

As part of the data collection, catch for the years 2013–2017 disaggregated by functional groups of species was also collected at the level of fishery units using official and unofficial data sources8. The number of fishery units and associated catch per country are detailed in Supplementary Table 3.

Kenya, Madagascar and the Philippines were used as examples to illustrate how additional data could be incorporated into our archetype analysis. In Kenya and Madagascar, we used the countries’ organization of fishery units (by taxa and habitat and by administrative regions, respectively; Supplementary Table 2), whereas in the Philippines, we included gender-disaggregated data on harvesting participation provided by country experts at the fishery unit level.

Assessment of the degree of ‘small-scaleness’

The degree of ‘small-scaleness’ in a fishery refers to the extent to which a fishing operation is characterized by features typical of small-scale fisheries such as low levels of investment and technology, short supply chains and family-owned or community-managed operations, among others31. Using the total score of the characterization matrix, we explored how the degree of ‘small-scaleness’ varies at the global, regional and country levels. To understand how catch is distributed across different degrees of ‘small-scaleness’, we aggregated the annual average catch (2013–2017) of all fishery units with the same total characterization score at the different geographical scales. We then scaled the data from 0 to 1 to adjust for differences in catch magnitude across countries and regions, enabling effective comparisons.

Archetype identification via multilevel LCCA

We used multilevel LCCA to identify small-scale fisheries archetypes using the 13 attributes of the characterization matrix tool. LCCA is a model-based clustering method for multivariate categorical data that fits a probabilistic model to the data. The goal of LCCA is the same as in traditional clustering methods like hierarchical and k-means clustering: to identify homogeneous groups within a larger population. The main difference between LCCA and traditional clustering approaches is that instead of finding clusters based on an arbitrary distance metric, a model is used to represent the data as a weighted combination of cluster-specific distributions. The weights indicate cluster abundance, and the distributions describe the probability of attribute combinations within each cluster: that is, a mixture model. Based on the estimated model, one can calculate the probability of each observation belonging to each cluster (or ‘latent class’). The use of a statistical model addresses two of the most controversial topics in clustering. The first is the lack of characterization of uncertainty in clustering estimates and derived inferences. LCCA addresses this issue by providing an assessment of the clustering results using the model fit via likelihood-based statistics. The second is how to determine the right number of clusters. The use of LCCA enables the calculation of the number of clusters using an extended analytical strategy based on model fit and cluster separation (see below)58. LCCA is therefore considered a more statistically robust method for clustering than traditional cluster analysis59,60.

We accounted for the multilevel structure of our data by including region as a fixed-effect covariate, and the model was fit using the glca package in R61. This adjustment addresses the expectation that fishery units are likely to be more similar within regions than across them. Using the fishery-specific cluster probabilities estimated by LCCA, we assigned each fishery unit to its most likely archetype (cluster) and summarized these patterns across the different geographical levels. To further characterize each archetype, we calculated means and standard deviations of the attributes, better understanding the defining features of each cluster (Supplementary Table 6). Additionally, the cluster weights (representing the proportion of fishery units assigned to each archetype) were computed to illustrate the relative abundance of each cluster (Extended Data Table 5). To provide further insight into the distribution of fishery units across archetypes, we calculated the proportions of fishery units assigned to each cluster within each country and region, along with their corresponding standard errors and 95% confidence intervals. These measures help quantify the uncertainty in archetype assignments and highlight the variability across regions (Supplementary Table 6).

The archetype designations were typically clear, with more than 92% of fishery units having an estimated probability greater than 0.95 for their assigned archetype and more than 86% exceeding a probability of 0.99. In contrast, fewer than 0.1% of fishery units had a probability below 50%. For comparison, if archetype designations were completely uncertain, each of the five categories would have an equal probability of 20%.

To further explore the relationships among the identified archetypes, we also performed hierarchical clustering based on the mean attribute values obtained from LCCA (Supplementary Table 5a). Using Euclidean distance and Ward’s method, via the dist and hclust functions in base R, we calculated the proximity between archetypes, providing an additional perspective on their relationships. The grouping patterns identified by hierarchical clustering were used to create the dendrogram in Fig. 3.

Determining the optimal number of archetypes

We chose the final model based on model fit (Akaike information criterion (AIC) and Bayesian information criterion (BIC))58 and cluster separation (average silhouette width (ASW) and entropy)62,63. The glca fit reports region-specific probability matrices, so to assess the cluster structure, we consider how distinctly each region’s points are clustered. ASW was calculated for each region separately so that SW is computed for individual data points relative to other points within the region. Then, the ASW is computed as the average of the SW values over the regions. The reasoning behind this is that for each region, some clusters are 0 or have low probabilities, so we do not want to compute the distance to points from other regions that fall into these clusters. Besides ASW, to assess cluster separation we also considered the number of observations with a negative value of ASW: that is, the number of observations that are likely to be placed in the wrong cluster. The information-based measures to assess model fit for each number of clusters are detailed in Extended Data Table 4. In addition to the use of the mentioned metrics to select the number of clusters, the number of countries and regions resulting in each cluster was also considered to prioritize solutions in which clusters included fisheries from more than one region when possible.

Data coverage and methodological considerations

The IHH marine small-scale fisheries dataset used in this study is the most globally comprehensive available to date8, and although LCCA is a superior approach for cluster (archetype) identification, there are limitations to note. The incorporation of additional data not available at the time of the study may uncover patterns that are not fully captured in the current analysis. This is particularly relevant for regions like Oceania and the Americas, which are less represented in our sample (covering only 2% and 35% of their regional marine small-scale fisheries catch, respectively) compared to better-covered regions like Africa (63%) and Asia (79%) (Extended Data Table 3). Despite regional differences in data coverage, the 43 countries included in our study were identified following a selection process based on the contributions of their fisheries sector to both global and national fisheries indicators (for example, catch and employment). This approach ensured that the sample captured both countries providing major contributions at the global level and countries where fisheries play a vital role locally8. Of the 52 countries originally identified in the country ranking process of the IHH initiative, Ghana, the USA, the Pacific Islands and Brazil were not included in our analysis. Ghana, the USA and the Pacific Islands were excluded due to the unavailability of characterization data and Brazil due to methodological considerations. Specifically, inclusion of Brazil’s data in the model resulted in two additional clusters composed exclusively of Brazilian units, reflecting national-level patterns rather than globally relevant archetypes, coupled with diminished model performance metrics. This is attributable to the large number of reported fishery units (652) and their limited variability: many shared identical characterization scores. Importantly, the five global archetypes, along with their defining attributes and country composition, remained consistent whether or not Brazil was included. Also, less than 1% of Brazil’s national marine small-scale fisheries catch could be linked to the characterized fishery units, making it difficult to meaningfully include the country in catch-based analyses.

Small-scale fisheries archetypes were empirically derived by grouping fishery units using the LCCA model. Importantly, the number of fishery units is not a proxy of catch, as some countries have a wide diversity of fishery units but have limited catch data (for example, Gabon), whereas others have a higher coverage of the national catch but concentrated in fewer fishery units (for example, Egypt, Bangladesh) (Supplementary Table 3). We defined archetypes by averaging attributes based on the number of fishery units, treating each unit equally regardless of how much fish they caught (Fig. 2). To ensure that our results were not skewed, we also recalculated the attributes for each archetype by weighting them according to each unit’s catch size. The comparison showed that although there were some differences, the overall patterns remained consistent (Supplementary Fig. 1), confirming that our archetype definitions accurately capture the diversity of marine small-scale fisheries worldwide.

The potential policy actions presented in Table 1 were developed by the author team drawing on expert knowledge gained throughout the IHH project. They reflect common challenges, opportunities and needs identified through close engagement with country case study authors and national experts. The list was refined through iterative internal discussions to ensure that the actions were grounded in observed realities while remaining broadly applicable across archetypes.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.