Conflict with industrial fleets undermines the contributions of small-scale fisheries to sustainable development in vulnerable regions

Abstract

Equitable and sustainable governance of small-scale fisheries (SSFs) is essential for achieving food security and conservation in coastal areas. Yet in increasingly industrialized oceans, competition and conflict with large-scale fisheries (LSFs) threaten the contributions of SSFs to the United Nations Sustainable Development Goals. Here we have developed a small-scale and industrial fisheries interaction index and applied it to 126 countries using 2 recent global datasets. Vulnerability to conflict is highest in countries where SSFs are vital for nutrition, but governance is weak, especially in Africa, Oceania and Asia. We also identify low-exposure outlier countries with unexpectedly low risk of SSF–LSF interactions, providing insights for proactive policy. Preferential access areas granting spatial rights to SSFs are common in both low- and high-vulnerability countries, suggesting such policies are often reactive and may be insufficient alone without strong enforcement. As ocean industries expand, it will be necessary to strengthen SSF access rights and limit LSF overcapacity to reduce conflict and advance sustainable development.

Small-scale fisheries are vulnerable to conflict with industrial fishing. The small-scale and industrial fisheries interaction index reveals regions with weak governance and emphasizes the need for enhanced access rights and management strategies.

Main

Conflict between large- and small-scale fisheries threatens the contributions small-scale fisheries (SSFs) can make to sustainable development¹. Globally, nearly half a billion people rely on SSFs for food or livelihoods, indicating their critical role towards meeting multiple United Nations (UN) Sustainable Development Goals (SDGs), especially in low-income coastal countries where dependence on SSFs is highest^2–4. Despite their importance, SSF contributions can be hampered by exposure to other marine-based activities (for example, large-scale fishing, oil and gas, tourism, shipping and so on), which, if negatively interacting, can completely erase SSF contributions to human health, economies and cultural well-being^5–7. Interactions with industrial or large-scale fisheries (LSFs) are unique from other ocean activities because they can lead to direct competition for targeted species, space and fishing rights^8–11. These interactions can occur via multiple pathways, including direct physical contact and conflict, socioeconomic mechanisms such as labour shifts or market competition^12–15, and/or via ecological pathways like increased pressure on target species and trophic cascades¹⁶. They can also lead to conflict between countries, vessel collisions, destruction of fishing gear and life-threatening confrontations at sea, escalating safety risks in an already hazardous occupation^8,17,18.

Over the coming century, fishery-related conflict is projected to increase in frequency and severity as target species’ distributions shift due to climate-driven fish migrations and rising demand for seafood¹⁹. While previous case studies have provided valuable local insights^18,20–22, a systematic approach is needed to better understand the underlying drivers of vulnerability to these conflicts and their impact on the SDGs, especially as blue economy initiatives risk accelerating industrialization and marginalizing coastal communities^23–25. In parallel, there has been growing interest in governance responses aimed at mitigating these conflicts, including spatial measures that secure small-scale fishing access, such as small-scale preferential access areas where industrial fishing is limited or prohibited²⁶. However, evidence remains limited on the contexts in which such preferential access areas emerge, and whether they are associated with reduced exposure to SSF–LSF conflict. This need was underscored recently during the 2024 meeting of the UN Sub-Committee on Fisheries Management, which urged countries to prioritize addressing competition and potential conflicts between these SSFs and LSFs²⁷ (Box 1). Identifying areas and drivers of high vulnerability to SSF–LSF interactions is a first step to inform how countries fulfil this recommendation.

The nature and magnitude of these effects will depend on many factors, including the intensity and type of LSF presence; the degree of spatial and resource overlap between both subsectors; the economic and nutritional value of a country’s SSFs; and a country’s geopolitical influence and ability to regulate access^15,28. Understanding these dynamics requires a framework that captures both the extent of SSF–LSF interactions and how they shape SSF vulnerability. One way to contextualize these drivers is the exposure–sensitivity–adaptive capacity framework, where vulnerability to these interactions is driven by exposure, or the degree to which SSFs interact with LSFs, and sensitivity, which reflects the extent to which communities rely on SSFs for livelihoods and food security²⁹. In contrast, adaptive capacity, shaped by governance strength and institutional support, can mitigate these risks by enabling SSFs to respond to and recover from external pressures³⁰.

The goal of this study is to develop and test an index of countries’ potential vulnerability to interactions between SSFs and LSFs. The small-scale and industrial fishing interaction (SIFI) index of vulnerability is based on the exposure–sensitivity–adaptive capacity framework³¹ because it has been empirically applied to quantify globally relevant patterns of fisheries vulnerability to anthropogenic pressures^32–35 (Tables 1 and 2). Here we apply this framework to estimate proximate drivers of SSF–LSF interaction vulnerability and identify low-exposure outlier countries that may serve as illustrative bright spots to inform proactive policy. This framework provides a foundation for understanding the multifaceted risks of SSF–LSF interactions in an increasingly industrialized ocean.

Table 1.

Glossary of terms and associated definitions used in this study

Term	Definition used in this study	Relevant references
Small-scale fishing	Small-scale fishing, sometimes referred to as artisanal fishing, is variably defined but is generally characterized as fishing that requires relatively low capital investment and capacity, smaller or no fishing vessels, and manual gear operation. Frequently, small-scale fishing involves households rather than commercial companies that make short fishing trips and operate relatively close to shore. Countries may use different definitions of small-scale fishing; this study follows the definitions used by the Illuminating Hidden Harvests project2. Also refers to SSFs.	Refs. 44,63–65
Large-scale fishing	Large-scale fishing, sometimes referred to as industrial fishing, is variably defined but generally characterized as fishing that requires relatively high capital investment, large vessels, and advanced fishing technology for catching and storage for large-scale processing. Also refers to LSFs.	Refs. 66,67
Adaptive capacity	The degree to which a nation is able to effectively and resiliently mediate the impacts of LSF–SSF interactions.	Refs. 68–70
Exposure	The degree to which LSF activities within a nation’s waters could interact with SSF activities in a country; in this study, exposure is defined as a function of LSF density within 25 km of the coastline and SSF catch as determined by the Illuminating Hidden Harvests2 findings.	Ref. 71
Sensitivity	The degree to which a nation depends on SSFs, that is, the importance of fisheries to national economies, employment and food security.	Ref. 72
SIFI index of vulnerability	A composite measure assessing vulnerability to SSF–LSF interactions; defined as the degree to which a nation’s SSFs are at risk from interactions with LSFs, influenced by exposure of SSFs to LSFs, the dependence of SSFs on national economies, employment and food security, and the nation’s ability to mitigate and adapt to these interactions effectively via strong governance.	Refs. 73–79

Table 2.

Summary of vulnerability components, drivers and variables for inclusion in construction of the SIFI index of vulnerability

Component	Driver	Variable	Source
Exposure	Density of nearshore large-scale fishing activity	AIS-broadcasting large-scale fishing density: annual average of fishing vessels detected by automated identification systems in nearshore area (2017–2021), scaled to size of country’s available nearshore area (km2) Non-broadcasting large-scale fishing density: average number of non-broadcasting fishing vessels detected by SAR in nearshore area (2017–2021), scaled to size of country’s available nearshore area (km2)	Global Fishing Watch; refs. 50,80
	SSF catch	Marine SSF production/country nearshore area	Refs. 1,2
Sensitivity	SSF value: employment and economic dependence on the SSFs sector	Contribution to SSF global marine fisheries landed value (% of global landed value) Contribution to SSF global marine fisheries employment (% of global marine SSF employment) Contribution of marine SSF employment in fisheries to total employment (% of country’s labour force) Contribution to marine subsistence employment (work for own consumption) (% of global subsistence workers)	Refs. 1,2,81
	SSF nutrition: coastal nutritional supply and nutritional need available from SSFs	Quantity: contribution to marine SSF production within a country (kg per capita per year for coastal population within 100 km to shore) Quality: nutrient supply of catch (calcium, iron, selenium, zinc, vitamin A, omega-3 fatty acids) for coastal residents (compiled domestic proportions that would meet 25% reference daily intake for coastal population) Hunger that could be addressed by SSFs: Prevalence of Inadequate Micronutrient Intake Index for six micronutrients	Refs. 1,2,41,82,83
Adaptive capacity	HDI	HDI (which includes: health, measured by life expectancy at birth; education, measured by a combination of mean years of schooling for adults aged 25 and older and expected years of education for children entering school; and standard of living, measured by gross national income per capita, adjusted for purchasing power parity)	Ref. 84
	Governance	Corruption estimate Political stability and absence of violence estimate Government effectiveness estimate Regulatory quality estimate Rule of law Voice and accountability	Ref. 85

These variables were combined to produce a composite index of vulnerability as a function of countries’ exposure to SSF–LSF interactions, sensitivity and adaptive capacity.

Box 1 Competing priorities in fisheries governance.

During a discussion over SSF–LSF conflicts at the UN FAO Council Committee on Fisheries Meeting in 2024, country delegates expressed different priorities related to the importance of fisheries subsectors, reflecting global tensions in priorities related to SSFs and LSFs:

“For whom is a fishery? For the fishers whose lives are intertwined with oceans and seas since times immemorial? The traditional fisheries that have been fishing in the coastal areas… But, however, we have not secured their food security.” Delegate from India.

“My country happens to base its economic well-being on largely large-scale fishing. We would prefer to have our own decisions on doing that rather than committing to some reducing of that capacity and going back to previous practices of small-scale and artisanal fishing.” Delegate from Iceland.

Results

The SIFI index was used to identify global patterns in vulnerability for 126 countries (Figs. 1 and 2); highlight low-exposure outlier countries to illustrate how the index can inform proactive policy (Fig. 3); identify key drivers of vulnerability (Fig. 4); and test hypotheses about how changes in exposure, sensitivity, and adaptive capacity influence each other and overall vulnerability (Fig. 5).

Fig. 1. Conceptual representation of the SIFI index.

a, Vulnerability to SSF–LSF conflict is estimated by combining six national drivers of exposure, sensitivity and adaptive capacity to SSF–LSF interactions. b, Exposure was defined as a composite value of the relative density of nearshore LSF density and SSF production, where countries with simultaneously dense LSF and high SSF production score highest.

Fig. 2. Global and regional variation to SSF–LSF conflict vulnerability using the SIFI index.

a–d, Geographic variation is highlighted using a diverging colour scale centred at 0.5 for the SIFI index of vulnerability (a) and its components: exposure (b), sensitivity (c) and adaptive capacity (d). e, The SIFI index indicates that vulnerability was highest in Africa (n = 36), followed by Oceania (n = 8), Asia (n = 29) and Central America (n = 27), and lowest in Europe (n = 24) and North America (n = 2). Examining the components of vulnerability together and individually can provide guidance for policy action; for instance, identifying countries with currently low exposure but high sensitivity and low adaptive capacity that may experience increased vulnerability in the future as exposure increases, provides policymakers the opportunity to put in place preventive measures. Each country contributed 1 index value (n = 126 countries). All scores are scaled from 0 to 1, with higher scores indicating greater vulnerability. Adaptive capacity indicators were reverse-coded such that higher scores indicate lower adaptive capacity. Countries shown in grey were excluded because they do not have marine fisheries or due to insufficient data. Basemap data in a–d from Natural Earth (www.naturalearthdata.com).

Fig. 3. Identifying low-exposure outliers and the association between preferential access policies and vulnerability.

a, Low-exposure outlier countries compared with others included in the SIFI index of vulnerability. Countries were selected if their SSF–LSF exposure was lower than the regional median within the same FAO region, while their sensitivity and adaptive capacity scores exceeded the regional median values. The y axis represents the percent difference in exposure from the median value of each region, and shaded boxes represent countries with preferential access areas for SSFs. b, Distribution of SIFI index and component scores for countries with (n = 34) and without (n = 92) SSF preferential access areas (PAA) for SSFs (shaded bars). Points represent individual countries and are horizontally jittered for visibility. Boxes show the 25th–75th percentiles and central lines indicate medians. Whiskers extend to 1.5× the interquartile range. The lowest and highest points represent the observed minimum and maximum values for each component of vulnerability, respectively. Two-sided t-tests compare countries with and without preferential access areas.

Fig. 4. Key drivers of vulnerability to SSF–LSF interactions as defined by the SIFI index.

The path diagram depicts relationships modelled by a confirmatory factor analysis for 126 countries, where each country contributed one set of country-level indicator values. Standardized relationships (factor loadings) between vulnerability (also called the latent factor, λ, represented as an ellipse) and its drivers (also called observed variables, represented as squares) are indicated as vectors, with blue indicating positive loading and red indicating negative relationship. The size of the vector lines for drivers is scaled to reflect their strength, with thicker lines representing stronger relationships. Solid vector lines represent statistically significant relationships at α = 0.05. The unstandardized loading of the first indicator was fixed to 1 to set the scale of the latent factor and was therefore not independently tested. Model fit was assessed using CFI, TLI, root mean square error of approximation (RMSEA) and standardized root mean square residual (SRMR) (Extended Data Table 1).

Fig. 5. Global estimations of the relationships among the three components of vulnerability and their drivers based on empirical data from 126 coastal countries.

a–c, Direct bivariate relationships between the components estimated using linear models suggest that sensitivity is positively associated with exposure (a) and negatively associated with adaptive capacity (b), and that exposure and adaptive capacity are negatively associated (c). Shaded areas represent 95% confidence intervals and colour denotes whether the relationship is positive (blue) or negative (red). d, Indirect relationships estimated using a structural equation model indicate that sensitivity mediates the relationship between exposure and adaptive capacity. Ellipses represent the three latent components of vulnerability, and double-headed arrows represent their standardized covariances. Blue and red arrows indicate positive and negative associations, respectively. Asterisks denote statistical significance of regression path coefficients: *P < 0.05; ***P < 0.001. Sensitivity was significantly negatively associated with adaptive capacity (r = −0.731, 95% CI: −0.873 to −0.588; P < 0.001). Exposure was positively associated with sensitivity (r = 0.360, 95% CI: −0.003 to 0.722; P = 0.052), which we considered supported despite narrowly missing the α = 0.05 threshold. When considering all 3 vulnerability components, exposure was not significantly associated with adaptive capacity (r = −0.256, 95% CI: −0.562 to 0.051; P = 0.102; see Extended Data Table 1 for model fit statistics).

Scores of the SIFI index of vulnerability show uneven spatial distribution (Fig. 2a). Countries in Africa exhibit the highest overall vulnerability, followed by Oceania, and approximately half of the analysed countries in Asia and South and Central America show similarly high vulnerability (SIFI index > 0.5), whereas no European or North American countries exhibit high vulnerability (Fig. 2e). The most vulnerable subregions within these continents were western Africa and eastern Asia (Extended Data Fig. 1). The three most vulnerable countries globally were Guinea, Sierra Leone and Mozambique (Extended Data Fig. 2).

Extended Data Fig. 1. Distribution of the SIFI Index of vulnerability scores, grouped by region and subregion.

Panels show a) and b) overall SIFI Index of vulnerability scores and scores for the three components of vulnerability: c) and d) exposure, e) and f) sensitivity, and h)-g) adaptive capacity, as produced by a confirmatory factor analysis for 126 countries in Africa (n = 36), Asia (n = 29), South and Central America (n = 27), Europe (n = 24), Oceania (n = 8), and North America (n = 2). Points represent individual countries with jitter was added for visibility. Diamonds represent mean values and horizontal line indicate median values. Boxes show the 25th–75th percentiles, with whiskers extending to 1.5 × the interquartile range (IQR); observations beyond the whiskers remain visible as individual points. Minimum and maximum values are represented by the lowest and highest points, respectively.

Extended Data Fig. 2. National scores (n = 126) grouped by region for each of the three components of vulnerability, and for the SIFI Index of vulnerability.

Scores are the result of a confirmatory factor analysis and are scaled from 0 to 1.

Global patterns in SIFI index components

To explore the underlying factors of the SIFI index scores, each of the components of drivers of vulnerability—exposure, sensitivity and adaptive capacity—was examined. Exposure (nearshore LSF activity and SSF production) was highest in Asia, particularly eastern and southeastern Asia, followed closely by Africa, particularly West Africa (Fig. 2b and Extended Data Fig. 1c,d). Of the automatic identification system (AIS)-broadcasting vessels that could be identified, 90% were identified as trawlers (Extended Data Fig. 3).

Extended Data Fig. 3. Components of SSF-LSF exposure for 126 countries.

(a) Composite index of nearshore LSF (<25 km from coast), which reflects both detection density of AIS-broadcasting LSF and of non-broadcasting (detected by SAR and unmatched to AIS) per sq. km. (b) Total SSF catch scaled to nearshore area. (c) LSF density and SSF catch were combined to produce a composite value of exposure to SSF-LSF interactions. Basemap data from Natural Earth (www.naturalearthdata.com).

Sensitivity (coastal economic and nutrition reliance on SSFs) was highest in Africa, Oceania and Asia, particularly in southern and eastern Asia, followed by West and East Africa (Fig. 2c and Extended Data Fig. 1e,f).

Adaptive capacity (development and governance) was highest overall for countries in North America and lowest overall for countries in Africa (Fig. 2d and Extended Data Fig. 1g,h).

Outlier countries with low exposure to SSF–LSF interactions

The SIFI index of vulnerability offers a tool to identify countries with unexpectedly low exposure, which may serve as bright spots and/or precursors to identifying factors (for example, policy, geography or external influences like political instability) that explain variation in exposure. Countries with low present exposure may become more vulnerable if LSF activity expands in nearshore zones, making proactive identification important. We identified 15 low-exposure outlier countries, including Comoros, Libya, Yemen, Greece and the Solomon Islands, that exhibited relatively low exposure and adaptive capacity (lower than median) and high sensitivity (higher than median) compared with counterparts in the same region (Fig. 3a).

Countries that have already established SSF preferential access were more vulnerable to SSF–LSF conflict, indicated by the SIFI index (mean = 0.59 versus 0.42, P < 0.001; Extended Data Table 2), suggesting that these tools have been used in contexts where conflict is already elevated. Countries with preferential access areas exhibited significantly higher sensitivity (P < 0.01) and exposure (P < 0.05), and lower adaptive capacity (P < 0.001) than countries without them (Fig. 3b).

Extended Data Table 2.

Comparison of vulnerability to conflict between SSF and LSF for countries with and without SSF preferential access areas

Component	Mean (with PAA)	sd (with PAA)	Mean (without PAA)	sd (without PAA)	t	df	p-value
SIFI Index	0.685	0.215	0.486	0.255	4.409	68.802	0.000
Exposure	0.350	0.178	0.290	0.153	1.750	51.740	0.086
Sensitivity	0.600	0.159	0.494	0.178	3.223	64.778	0.002
Adaptive Capacity	0.411	0.164	0.552	0.251	−3.691	89.391	0.000

Summary statistics and two-sample t-tests comparing SIFI Index of vulnerability components (exposure, sensitivity, and adaptive capacity) between countries with and without SSF preferential access areas (PAAs). Differences were assessed using two-sided Welch’s t-tests.

Drivers of vulnerability

Understanding drivers of vulnerability is essential for interpreting the conditions that drive high-vulnerability countries as well as potential bright spots of low exposure. We used a confirmatory factor analysis to examine key drivers of vulnerability. The SIFI index of vulnerability increased most strongly in places with high marine SSF nutrition contribution and economic value (P < 0.001; Fig. 4). In contrast, countries with low measures of development and governance (both drivers of adaptive capacity) were most vulnerable to SSF–LSF interactions (P < 0.001).

SSF reliance shapes vulnerability

We hypothesized that countries with higher reliance on SSFs (greater sensitivity) would experience higher exposure to LSFs, and that lower governance and management capacity (lower adaptive capacity) could exacerbate this exposure (Fig. 1c). We used linear models to test these relationships individually and then applied a mediation analysis using a structural equation model to evaluate whether these relationships persist when modelled simultaneously within an integrated framework.

As hypothesized, bivariate linear models indicated that countries with increased SSF reliance (high sensitivity) experience greater exposure to SSF–LSF interactions (Fig. 5a; P < 0.001). Furthermore, low adaptive capacity is a strong predictor of high sensitivity (Fig. 5b; P < 0.001) and high exposure, confirming our hypotheses (Fig. 5c; P < 0.001). In contrast, structural equation modelling revealed that these relationships do not all persist when modelled simultaneously. Instead, this model indicates that sensitivity mediates the relationship between exposure and adaptive capacity (Fig. 5d), as the direct path between exposure and adaptive capacity is weakened once sensitivity is included in the model.

Discussion

The SIFI index of vulnerability represents an estimate of the risk of interactions between fisheries sectors to assess potential drivers of fisheries conflict. We find that vulnerability remains highest where SSFs are most critical, especially for nutrition, and where national governance is weakest, highlighting the need for targeted interventions.

Risks and opportunities presented by low-exposure outlier countries

The low-exposure countries identified here illustrate that the nature and distribution of exposure to SSF–LSF interactions can vary widely even between adjacent countries, requiring tailored policy responses. For example, while most countries in West Africa exhibit some of the highest exposure measured in this study, Guinea demonstrates relatively lower exposure despite very high sensitivity and low adaptive capacity. Low exposure in these countries may be attributed to several factors, such as policies that limit nearshore industrial fishing, previous overexploitation of fish populations, robust historic or current protections for SSF access, biophysical features that reduce LSF accessibility to nearshore zones, or persistent political instability and governance failure, which could constrain industrial expansion. One key benefit of using the SIFI index to identify potential bright spots is the ability to explore explanatory factors that may have helped reduce exposure in these areas. Future analyses could investigate differences in exposure in neighbouring countries to show that exposure to LSFs in coastal waters is not an inevitable outcome of geography or governance limitations and offer strategies that reduce exposure.

The SIFI index provides a broad snapshot of SSF–LSF interactions but does not fully capture their dynamic nature. For example, Madagascar is classified as ‘low exposure’, yet recent documented SSF–LSF conflicts suggest otherwise³⁵. This probably reflects the fact that the most recent LSF vessel data used in this study are from 2021, whereas there has been a recent increase in China-flagged trawlers in Madagascar’s waters (including in areas farther offshore than would be captured by the SIFI index), underscoring the need for continual localized updating of data that inform the SIFI index. This temporal mismatch underscores that while the index can identify areas generally at risk, localized and rapidly changing dynamics, such as new industrial fleets or recent policy changes, may generate conflicts not yet detected by the SIFI index, emphasizing the need for continual updating with the most recent LSF and SSF data. In fact, a version of the index that integrates dynamic time series data, for example, in the form of annual averages, would be a major area for future research. Such an approach would require yearly updates of LSF activity from AIS and synthetic aperture radar (SAR) data, along with repeated national-level estimates of SSF catch, employment and the other variables that inform the SIFI index. This would enable retrospective analyses to identify the temporal drivers of increased or reduced exposure, track responses to fisheries policy and support proactive management interventions.

While the SIFI index provides a robust measure of relative vulnerability to SSF–LSF interactions, it is not currently validated against observed conflict events due to the lack of comprehensive data. Future work could integrate case study-based observations, fisheries monitoring data and fisher-reported conflict incidents to empirically test the index’s predictive accuracy. Until such validation is possible, the index should be interpreted as an indicator of potential risk rather than as a direct measure of realized conflict.

Furthermore, the presence of LSFs does not necessarily equate to exposure³⁶, defined simply here as SSF–LSF co-occurrence. Our estimate of exposure does not capture differences in fishing practices or target species, and as a result does not encompass resource competition, which is a potential major driver of SSF–LSF conflict. For example, SSFs often depend heavily on nearshore pelagic species, which may be simultaneously targeted by LSF midwater trawlers. In contrast, LSF bottom trawl fleets targeting demersal species often operate at depths and distances beyond those most accessible to SSFs, resulting in comparatively lower conflict risk than midwater gear. Exposure depends on multiple complex factors, and in some cases LSFs may operate near SSFs without meaningful interaction.

Nevertheless, given that 90% of the AIS-broadcasting LSF vessel density in this study is attributed to trawlers—vessels that may use bottom trawl gear, which is known to exert substantial damage on seafloors and coastal ecosystems³⁴—future iterations of the SIFI index could focus on trawlers’ overlap with mutually targeted species or direct reports of SSF–LSF conflicts to produce a more nuanced representation of exposure. This is particularly relevant in the context of ongoing debates around restricting or banning destructive trawling in certain coastal areas³⁷, as such management measures could substantially reduce SSF–LSF interactions.

SSF nutritional dependence as a key driver of vulnerability

The overall vulnerability of coastal populations to SSF–LSF interactions is driven strongly by their reliance on SSFs, especially for nutrition (Fig. 4). This pattern is further supported by our structural equation model, which indicates that the relationship between exposure to SSF–LSF interactions and adaptive capacity is mediated by sensitivity (for example, nutritional and economic reliance on SSFs), pointing to SSF dependence as a key pathway through which LSF pressure translates into vulnerability. Evidence has shown that LSFs can drive reduced access to SSF fishing grounds and competition over fish populations targeted by SSFs^9,38, and drive up the price of fish at local markets³⁹. This is likely to reduce local food availability, as high retention of SSF catch helps to secure local protein and micronutrients to coastal populations, particularly in food-deficient countries^1,40,41. In contrast, although some regions consume at least portions of LSF catch locally, most LSF catch is exported via complex supply chains facilitated by international fishing agreements, with a large proportion destined to become fishmeal to feed livestock^31,42. These patterns highlight the need to minimize harmful SSF–LSF interactions to protect local food availability. Empowering SSF organizations, unions and cooperatives can further enhance resilience by ensuring their participation in decision-making^7,43.

Towards proactive policy solutions

For all these reasons, countries considered vulnerable by the SIFI index could consider policy strategies that limit SSF–LSF interaction exposure, such as: (1) the implementation of the Food and Agriculture Organization (FAO) SSF Guidelines, especially via spatial access rights; (2) the restriction of LSF overcapacity; and (3) the negotiation of fair fishing agreements. It is important to note that the indicators used in this study reflect proximate factors associated with SSF–LSF exposure rather than root-cause drivers. Thus, while our analysis identifies where interactions are likely to occur, it does not fully explain why. Identifying ultimate drivers, such as historical legacies, market dynamics and governance structures, requires qualitative, context-specific research. Accordingly, our policy recommendations are broad starting points rather than prescriptive solutions.

Securing spatial access rights for SSFs

The SIFI index confirms that countries with both high reliance on SSFs and limited governance capacity are especially vulnerable to interactions with LSFs, illustrating the multifaceted and deeply rooted nature of challenges faced by SSFs. SSFs globally are often overlooked in funding and policy decisions, in part because their labour and outputs are outside the traditional structure of productivist and capitalist economies⁴⁴. A transformative approach to limit SSF–LSF exposure would be for governments to value SSFs as engines of self-sufficiency and food sovereignty rather than solely economic productivity⁴⁵. Operationally, this could be achieved by implementing the Voluntary Guidelines for Securing Sustainable Small-Scale Fisheries, principles ratified by the UN Committee on Fisheries in 2014⁴⁶.

One focus of the FAO SSF Guidelines is responsible governance of tenure, including the establishment of preferential access areas where SSFs can operate without competition from LSFs⁴⁷. At least 44 countries have enacted preferential access areas, including Nicaragua, Equatorial Guinea, Suriname and Guinea, four of the low-exposure outliers identified in this study²⁶. Interestingly, exposure was moderately but significantly higher in countries with preferential access areas. In some countries, historic or ongoing exposure to LSF may have triggered the selective adoption of preferential access in the highest-conflict contexts. In others, preferential access areas may coexist with high LSF pressure, as seen in countries such as Senegal, Sierra Leone and Bangladesh, where preferential access areas have been established and SSF–LSF conflict exposure remains substantial. At the same time, weak enforcement may allow high exposure to persist despite formal SSF access protections. Our comparative analysis shows that countries with preferential access areas tend to have higher reliance on SSFs and lower adaptive capacity, indicating that these tools are an important first step in contexts where SSFs are socially and economically central and without sufficient governance resources to manage fisheries sustainably. Taken together, these findings suggest that while these areas are probably an important action for reducing spatial overlap, they alone are insufficient to eliminate competitive exclusion of resources, particularly where LSFs remain poorly regulated^31,33. Further investigation is needed to assess implementation and enforcement to confirm their role in mitigating SSF–LSF conflict.

Limiting LSF overcapacity

Beyond spatial access policies, reducing SSF–LSF exposure will also require the conditions that enable LSF pressure in vulnerable regions to be addressed. One of the most powerful of these drivers is LSF overcapacity, which is often maintained through harmful fisheries subsidies. These subsidies have long been a focus of international policy⁴⁸. Per capita, LSF fishers receive 3.5 times more subsidies than SSFs⁴⁹. Reducing these subsidies in highly exposed areas could curb artificially sustained fishing pressure and reduce poverty³⁸. Further research should identify how subsidies enable SSF–LSF interactions, especially in vulnerable countries.

At the same time, exposure may be underestimated in many countries due to industrial activity that is not captured by broadcast tracking systems. Future research using the SIFI index could focus and refine estimates of non-broadcasting LSF activity, which probably captures a mix of illegal, unreported or weakly regulated fishing⁵⁰, and which may disproportionately affect countries with limited enforcement capacity and amplify vulnerability.

Negotiating fair fishing agreements

Economic and geopolitical interests are likely to drive SSF–LSF exposure, particularly through fishing access agreements, bi- or multilateral agreements used to allocate rights, define fishing zones and establish other rules between governments¹⁵. In some lower-income countries, distant-water fishing nations represent or replace substantial proportions of the fishing sector, generally exporting their catch elsewhere²⁸. However, historically, fishing agreements have often not resulted in long-term benefits for host countries due to inequitable allocation of rights, inadequate fees and failure of the income to reach fishing communities impacted by increased LSFs⁵¹. For example, in West Africa, one of subregions most exposed to SSF–LSF interactions, conflicts with China- and European Union-affiliated vessels are well-documented, including in SSF-preferential zones^52,53. Addressing harmful distant-water LSF exposure requires negotiating (or renegotiating) agreements to ensure they support SSF livelihoods and food security while delivering genuine economic benefits to coastal states⁵¹.

SSF–LSF interactions will increase under a blue economy logic

While this study examines industrial activity in coastal areas through the lens of marine capture fisheries, other sectors such as energy, mining, shopping, tourism, aquaculture and fishery development are predicted to expand in the blue economy agenda⁵⁴. The prioritization of large-scale economic activities over localized small-scale use could result in a coastal squeeze likely to drive novel spatial competition in areas critical to SSFs^55,56. The vulnerability of SSFs to conflict with LSFs is an important obstacle to blue justice, a social movement and area of scholarship focused on securing equity and rights for coastal communities including small-scale fishers^57–59. Furthermore, SSF–LSF conflict vulnerability, combined with climate change and continued overfishing, is likely to increase vulnerability for aquatic food systems⁶⁰. Stronger fishery management tools in coastal zones, particularly to delineate and enforce areas designated for SSFs, are necessary to avert future conflict and support sustainable and equitable coastal food systems.

Methods

We provide a global assessment of the vulnerability of SSFs to interactions with LSFs. To do this, we combined newly available SSF and LSF data to operationalize a well-tested vulnerability index based on exposure, sensitivity and adaptive capacity (Tables 1 and 2 and Supplementary Fig. 1). Primary data sources for SSF catch, employment and subsistence were extracted from ref. ¹, which highlights the contributions of SSFs to sustainable development using official catch statistics, unofficial catch estimates, local and sub-national data, and expert insights to produce national-level SSF catch estimates for 58 countries, and household surveys from 78 countries to estimate SSF employment and subsistence. These data were then extrapolated to sub-global and national estimates for additional countries^1,2. Primary data for LSFs were extracted from Global Fishing Watch, which uses AIS signals from fishing vessels to track vessel movements, which are then processed using machine-learning algorithms to classify fishing activity by gear type and flag state⁶¹. These data were complemented with SAR to estimate the density of vessels not broadcasting AIS, providing a more complete picture of fishing vessel density nearshore⁵⁰. Here we briefly summarize data sources and methods; more detailed data sources, estimation methods and modelling approach are contained in Supplementary Methods.

Statistics and reproducibility

The country was the unit of analysis, and the sample size was determined by the availability of national-level data and the prespecified inclusion criteria (final analytical sample: n = 126 countries). Missing-data patterns and imputation procedures are described in Supplementary Methods; no countries meeting the stated inclusion criteria were excluded from the analyses. Because this was an observational study using existing country-level datasets, randomization and blinding were not applicable. All statistical tests were two-sided, with statistical significance assessed at α = 0.05. All analyses were conducted in R version 4.4.3.

Components of vulnerability

Exposure was represented by the estimated degree to which LSF activities co-occur spatially with SSFs, and was calculated as a composite driver of nearshore LSF vessel presence and SSF production (Supplementary Information, section 1 and Supplementary Fig. 1). Higher likelihood of spatial co-occurrence was assumed if the country exhibited the following: (1) relatively dense detections of LSF vessels broadcasting AIS signals, measured as the annual average number of AIS-broadcasting industrial fishing vessels detected operating within 25 km from shore during 2017–2021 and scaled to the country’s available nearshore area⁶¹; (2) relatively dense detections of LSF vessels not broadcasting AIS, measured as the annual average number of non-broadcasting industrial fishing vessels detected using SAR imagery within 25 km from shore during 2017–2021 and also scaled to nearshore area⁶¹; and (3) relatively high marine SSF production (metric tons per square kilometre of nearshore area) representing the intensity of small-scale fishing within a country’s nearshore zone^1,2 (Supplementary Fig. 2).

Sensitivity was defined as a nation’s dependence on SSFs in importance to economy, employment and food security, following ref. ²⁹. Two composite drivers were used to measure the relative sensitivity of countries’ marine SSFs: economic dependence and nutritional dependence (Table 2).

Adaptive capacity was represented by a nation’s ability to effectively and resiliently manage the impacts of SSF–LSF interactions through policy and effective governance. Following ref. ²⁹, two drivers were used to produce a measure of adaptive capacity: the Human Development Index (HDI) and governance (Table 2).

Constructing the SIFI index

The construction of a vulnerability index is a function of exposure, sensitivity and adaptative capacity; however, there is no universally accepted framework for how to combine these three components to estimate vulnerability⁶². To address this uncertainty, we compared three methods for index construction (composite mean, principal component analysis, confirmatory factor analysis estimation) and selected confirmatory factor analysis because it best captured the relationships among components and explained the most variance (Extended Data Fig. 4 and Supplementary Methods). In the model, adaptive capacity indicators were inverted before inclusion so that higher adaptive capacity counterbalances vulnerability drivers, consistent with other studies that have quantified vulnerability²⁹.

Extended Data Fig. 4. Comparison of statistical methods used to construct the SIFI Index of vulnerability for 126 countries.

Methods are in order of ascending complexity: a) composite mean of the three components of vulnerability (exposure, sensitivity, and adaptive capacity) and principal components analysis for the same three components; b) confirmatory factor analysis compared to composite mean; and c) principal components analysis compared to confirmatory factor analysis. Top panels show linear model results for the comparisons and bar plots show variation in country scores assigned by the different methods.

We used a multi-step confirmatory factor analysis to estimate the SIFI index of vulnerability, validate the consistency of our measurement theory, and evaluate the reliability and validity of its constructs. Confirmatory factor analysis is useful to assess how well measurable variables (for example, LSF density, SSF catch, SSF employment and so on) explain unobservable latent factors (for example, vulnerability). We ran a model with vulnerability as the latent factor using the following equation:

$$V_i={\lambda }_1{E}_i+{\lambda }_2{S}_i+{\lambda }_3\mathrm{A}{\mathrm{C}}_i+{ϵ}_i$$

where V_i represents vulnerability, E_i exposure, S_i sensitivity, AC_i adaptive capacity and ε_i the residual error term. Factor loadings (λ₁, λ₂, λ₃) indicate the strength of each component’s contribution to vulnerability. The model was estimated using maximum likelihood with mean and variance adjustment and 1,000 bootstrap iterations (comparative fit index (CFI) = 0.985, Tucker–Lewis index (TLI) = 0.955; Extended Data Table 1). New SIFI scores were derived from the confirmatory factor analysis model by extracting factor scores, which estimate each country’s vulnerability based on observed variables weighted by their predicted factor loadings. This process was repeated for three additional single-factor models for each of the three framework components (exposure, sensitivity, adaptive capacity) as latent factors.

Extended Data Table 1.

Model results for confirmatory factor analysis and structural equation model to estimate the SIFI Index (vulnerability) and evaluate relationships between its components

Fit measure	Structural equation model	Confirmatory factor analysis
CFI	0.996	0.985
TLI	0.985	0.955
ChiSq	5.214	9.622
p-value	0.266	0.087
RMSEA	0.049	0.086
SRMR	0.036	0.048
AIC	−689.337	−686.929
BIC	−641.120	−641.549

Model fit indices included in the table are: Comparative Fit Index (CFI), where values above 0.9 indicate good fit, Tucker-Lewis Index (TLI), with values above 0.9 indicate good fit, Root Mean Square Error of Approximation (RMSEA), where values below 0.08 are considered acceptable, and Standardized Root Mean Square Residual (SRMR), where values below 0.08 are considered acceptable.

We identified low-exposure outlier countries to highlight heterogeneity in exposure. Outlier countries were considered those with low exposure (lower than the median) and high sensitivity and adaptive capacity (higher than the median) within geographic regions defined by the FAO. In addition, we compared vulnerability components between countries with and without SSF preferential access areas identified by ref. ²⁶ using two-sample t-tests to assess differences in exposure, sensitivity and adaptive capacity.

Linear models were fitted to test hypotheses among the three components of vulnerability. To test for moderating effects between the three components, a structural equation model was used to estimate both direct and indirect effects among the three components and their observed variables (CFI = 0.996, TLI = 0.985; Extended Data Table 1).

Reporting summary

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

Author contributions

M.R.C. and X.B. conceived of the study design. M.R.C. led the data collection and analysis, with contributions from X.B., E.F., J.V. and G.C. M.R.C. wrote the first draft. All authors contributed to revisions and approved the final version.

Peer review

Peer review information

Nature Food thanks Gavin McDonald and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Funding

Funding for this study was provided by the Cedar Tree Foundation and the Smith Conservation Research Postdoctoral Fellowship Program. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

Data availability

The datasets generated and analysed during the current study are available via GitHub at https://github.com/melissarcronin/SSF-LSF-interactions.

Code availability

All code used for statistical analysis and visualization is available at https://github.com/melissarcronin/SSF-LSF-interactions.

Competing interests

The authors declare no competing interests.

Extended data

is available for this paper at https://doi.org/10.1038/s43016-026-01431-z.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s43016-026-01431-z.