This review addresses the critical interactions between small-scale fisheries (SSFs) and vulnerable marine megafauna—specifically marine mammals, sea turtles, and chondrichthyans (sharks, rays, skates)—in the southwestern Indian Ocean (SWIO). SSFs are vital for food security and livelihoods in coastal communities across developing regions. However, their operations often lead to significant bycatch and direct impacts on megafauna populations, which are characterized by K-selected life histories (slow growth, late maturity, low fecundity), making them exceptionally susceptible to population declines from even low levels of anthropogenic mortality.
The paper synthesizes existing knowledge, highlights severe data and monitoring deficiencies, and argues for urgent, collaborative, and evidence-based management strategies to ensure the sustainability of both fisheries and the marine ecosystems they depend on.
The review encompasses peer-reviewed literature, grey literature (e.g., NGO reports, government documents), and expert knowledge from multiple SWIO nations, including Kenya, Tanzania (including Zanzibar), Mozambique, South Africa, and Madagascar. The methodology involved a systematic collation of data on:
A key finding is the fragmented and often anecdotal nature of the data, preventing robust regional assessments.
Poor / Anecdotal
Landings & bycatch data are unreliable.
Exceptionally High
Due to K-selected life histories.
Weak Evidence
Strategies often lack scientific foundation.
Catch and landings data are universally described as poor in quality, resolution, and consistency. Compositional data is biased towards easily identifiable species, and cryptic mortality (e.g., animals that die after release) is largely unquantified. Understanding of fishing effort is limited, often relying on unsuitable proxies like number of boats, which fails to capture effective fishing power or spatial-temporal effort distribution.
All three megafauna groups show signs of overexploitation and population decline in several SWIO areas.
SSFs are not just an economic activity but a critical component of food security and cultural identity. Management interventions that fail to understand the degree of fishers' reliance on these resources—including the sale of bycatch for income—are likely to be ineffective or face strong resistance. Livelihood diversification is often limited.
A central flaw is the formulation of management strategies without a strong evidence base. This leads to regulations that may be biologically inappropriate, socio-economically unsustainable, or unenforceable. The lack of baseline data and ongoing monitoring makes it impossible to assess the status of stocks or the impact of management measures.
Governance is often top-down, with limited meaningful engagement of fishers and local communities. There is a lack of regionally collaborative frameworks to address transboundary stocks and shared challenges. The paper calls for a coalition of government agencies, NGOs, researchers, and fishers.
The review concludes with a call for action centered on:
Core Insight: This review exposes a fundamental systemic failure in the SWIO: the management of a complex socio-ecological system is being attempted with pre-industrial data infrastructure and governance models. The paper correctly diagnoses the problem—a severe evidence deficit—but the proposed solution hinges on a level of regional cooperation and institutional capacity that currently does not exist at the required scale.
Logical Flow: The argument is logically sound: poor data → poor understanding → ineffective management → unsustainable outcomes. The paper effectively traces this causal chain, using the vulnerability of K-selected species as the biological amplifier of the problem.
Strengths & Flaws: Its major strength is its comprehensive, regional scope and its clear-eyed assessment of data poverty. A significant flaw, however, is its relatively light treatment of implementation pathways and political economy barriers. It advocates for "good governance" and collaboration but offers few concrete strategies for overcoming the entrenched interests, funding shortfalls, and political inertia that characterize fisheries management in many developing regions. Compared to technological leaps in other fields (e.g., the use of Generative Adversarial Networks (GANs) like CycleGAN for image-based species identification and monitoring as discussed in Zhu et al., 2017), the solutions proposed here feel incremental.
Actionable Insights: For practitioners and funders, the immediate priority should be investing in lean, technology-enabled data pipelines. Instead of waiting for perfect, government-led monitoring, support should go to pilot projects using:
1. Drones and satellite imagery (inspired by applications from organizations like Global Fishing Watch) to map effort independently.
2. AI-assisted image recognition on smartphones for fishers to log bycatch, reducing reliance on taxonomic expertise.
3. Blockchain or secure ledger systems for catch documentation to improve traceability and combat IUU fishing, a related issue. The goal must be to generate "good enough" data rapidly to inform crisis-level decisions, while building the longer-term institutional framework the paper envisions.
To move from anecdotal to quantitative assessment, a standardized analytical framework is needed. A core component is modeling population vulnerability. This often uses a Potential Biological Removal (PBR) framework, adapted for bycatch. PBR estimates the maximum number of animals that can be removed from a population without causing decline:
$\text{PBR} = N_{min} \times \frac{1}{2} R_{max} \times F_r$
Where:
$N_{min}$ = Minimum population estimate
$R_{max}$ = Maximum theoretical growth rate
$F_r$ = Recovery factor (typically 0.1-1.0)
However, in the SWIO context, $N_{min}$ is usually unknown. Therefore, a priority-setting framework based on relative risk is more practical. This can use a semi-quantitative Ecological Risk Assessment (ERA) approach:
$\text{Risk Score}_{species, fishery} = \text{Exposure} \times \text{Consequence}$
Exposure is a function of spatial/temporal overlap and gear susceptibility. Consequence is a function of species' biological productivity (inversely related to K-selection) and current population status.
Scenario: Assessing bycatch risk for the dugong (Dugong dugon) in gillnet fisheries off northern Mozambique.
Step 1 - Data Collation: Gather fragmented data: (a) Fisher interviews suggesting occasional captures. (b) Historical sighting maps from aerial surveys (WCS, 2010). (c) GIS layers of reported gillnet fishing zones.
Step 2 - Exposure Index: Calculate spatial overlap between dugong habitat (seagrass beds) and gillnet effort. Use a simple scoring: 3 (High Overlap), 2 (Medium), 1 (Low), 0 (None). Assume score = 2.
Step 3 - Consequence Index: Dugongs have very low $R_{max}$ (~5% per year). IUCN status is Vulnerable. Assign a high consequence score: 3.
Step 4 - Risk Score: $\text{Risk Score} = 2 \times 3 = 6$ (on a scale of 0-9). This flags a High Priority for research and mitigation (e.g., testing acoustic pingers or modifying net profiles).
This framework allows managers to triage actions despite imperfect data.
Conceptual Chart: Data Fidelity vs. Management Action Timeline
A hypothetical chart would show two curves. Curve A (Current Paradigm): Shows a long, flat period of "Data Collection" with low fidelity (high uncertainty), followed by a delayed and often ineffective "Management Action." Curve B (Proposed Agile Paradigm): Shows rapid iterations. It starts with "Rapid Risk Assessment" (moderate fidelity), leading to a "Pilot Mitigation Measure" (e.g., community-led time closure), which then generates "Local Compliance & Bycatch Data," feeding back to refine the assessment in a continuous loop. The key insight is that action cannot wait for perfect data; management must become a learning process.
The future of sustainable SSF management in the SWIO lies in the convergence of participatory governance, appropriate technology, and adaptive science.