Methodologies for Data-Poor Fisheries Assessment in the Mediterranean Basin: Status, Challenges, and Future Directions

Abstract

Fisheries management in the Mediterranean Sea faces persistent challenges due to the prevalence of data-poor and data-limited stocks, small-scale multi-species fisheries, and limited long-term monitoring programs. Effective assessment methodologies are critical to ensuring sustainable exploitation, yet traditional data-rich stock assessment models remain infeasible for many Mediterranean fisheries. This review provides a comprehensive synthesis of current methodologies developed and applied to assess data-poor fisheries in the Mediterranean context. We examine catch-only approaches, length-based methods, empirical indicators, and multi-indicator frameworks increasingly adopted by the General Fisheries Commission for the Mediterranean (GFCM) and the EU’s Data Collection Framework (DCF). Special attention is given to case studies from the western, central, and eastern Mediterranean that demonstrate the opportunities and limitations of these approaches. We further explore emerging tools, including integrated modeling frameworks, simulation-based harvest control rules, and participatory approaches involving fishers’ local knowledge, to highlight innovations suited to mixed, small-scale Mediterranean fisheries. The review concludes by identifying key gaps in data collection, assessment capacity, and institutional coordination, and proposes a roadmap for improving data-poor fisheries management under Mediterranean-specific ecological, socio-economic, and governance constraints. By consolidating methodological advances and practical lessons, this review aims to provide a reference framework for researchers, managers, and policymakers seeking to design robust, adaptive strategies for sustainable fisheries management in data-limited Mediterranean contexts.

1. Introduction

1.1. Socio-Economic and Ecological Importance of Mediterranean Fisheries

The Mediterranean Sea, though covering less than 1% of the global ocean surface, is a critical hotspot of marine biodiversity and a cornerstone of the cultural and economic fabric of surrounding nations [1]. Its fisheries, highly diverse and dominated by small-scale, artisanal operations, form a deeply embedded socio-ecological system that supports food security, coastal livelihoods, and cultural heritage within a fragile marine environment. Hosting 7–10% of the world’s marine species in a small area and exhibiting high endemism, it is recognized as a global biodiversity hotspot [2,3]. This diversity stems from its complex oceanography, varied habitats, such as Posidonia oceanica meadows, coralligenous assemblages, and submarine canyons, and its long evolutionary history [4]. Fisheries here are highly multi-specific, with over 100 species landed, though a few dominate catches [5]. Key commercial species include small pelagics like European anchovy (Engraulis encrasicolus) and sardine (Sardina pilchardus), as well as demersal species such as European hake (Merluccius merluccius), red mullet (Mullus barbatus), and deep-water rose shrimp (Parapenaeus longirostris) [6]. This multi-species nature means fishing activities can alter trophic structures and community composition [7], making sustainable fisheries management essential for ecosystem health.

Socio-economically, Mediterranean fisheries are vital for employment, livelihoods, and nutrition. Small-scale fisheries (SSF) dominate, comprising about 80% of the fleet and 60% of employment [8]. Typically, family-run and using vessels under 12 m with low-impact gears such as trammel nets, gillnets, and handlines [9], they sustain hundreds of coastal communities across the region [10]. The total first-sale value of Mediterranean and Black Sea catches is around USD 3.5 billion annually, with the Mediterranean contributing substantially [11]. Beyond monetary value, SSF enhance local food security [12] and preserve maritime traditions and cultural identity [11], dimensions often overlooked by bio-economic models.

However, Mediterranean fisheries face mounting pressures from climate change, habitat degradation, and overfishing. Rapid warming, acidification, and non-indigenous species are reshaping ecosystems [13,14], while urbanization and pollution further stress habitats. Around 75% of assessed stocks are exploited beyond sustainable limits [15], threatening both biodiversity and the socio-economic stability of the sector. Declining stocks like hake jeopardize small-scale fishing livelihoods and coastal community resilience [16]. Sustainable management is thus essential not only for environmental preservation but also for safeguarding livelihoods, cultural heritage, and one of the world’s most valuable marine ecosystems.

1.2. The “Data-Poor” Paradigm and Its Consequences

Mediterranean fisheries are predominantly managed under a “data-poor” paradigm, as detailed catch-at-age and effort data, standard in data-rich industrial regions like the North Atlantic, are typically insufficient or unreliable [17,18]. In contrast, data-rich fisheries rely on precise analytical models to estimate biomass and mortality relative to MSY and implement explicit harvest control rules [19]. Following the distinction of Caddy & Gulland [20] and Costello et al. [21], most Mediterranean stocks are “data-limited,” possessing some information like total catch or length frequencies but insufficient for full assessments [6,22], while many others are truly “data-poor,” lacking even basic biological or catch records [23].

This state stems from the dominance of small-scale, multi-species fisheries that employ a variety of gears to target numerous species simultaneously, making standardized monitoring exceptionally difficult [24]. Logbooks are often impractical for small vessels, and mixed landings are typically sold in local markets without species-level resolution [25]. These operational constraints are compounded by limited budgets, fragmented coastal infrastructures, and disparities in technical capacity, particularly in southern and eastern Mediterranean states [9,26,27,28]. Governance and institutional fragmentation further exacerbate the challenge, with more than twenty bordering countries differing in systems, capacity, and priorities [9]. Although the DCF provides structure for member states, harmonization across the basin remains incomplete [29]. The lack of consistent, long-term datasets for abundance or age-structured information prevents the application of advanced assessment models [30].

The data-poor reality has created a substantial “assessment gap,” where only a small fraction of exploited species are regularly assessed [6,15]. While hundreds of species are landed, quantitative assessments exist for only a few dozen, leaving most stock statuses unknown. Consequently, management has often relied on indirect measures such as mesh size regulations, effort limits, or spatial closures, applied with limited understanding of their biological impact [31]. The data-poor paradigm generates cascading consequences that undermine sustainability. Ecologically, it obscures overfishing, allowing “silent” depletion to proceed undetected [32]. Many unassessed stocks are likely exploited beyond sustainable limits [33,34]. The absence of reliable reference points forces reliance on generic assumptions mismatched to species-specific dynamics [35], leading to ineffective measures and failed recovery efforts [31]. In multi-species fisheries, depletion of predators like hake can trigger trophic cascades [7,36].

Socio-economically, declining stock abundance reduces catch-per-unit-effort (CPUE), increasing operational costs and eroding profitability for small-scale fishers [10,37]. Management responses are often reactive and politically driven, leading to the “ratchet effect” (describes how politically driven management tends to allow incremental increases in capacity or effort, while reductions occur slowly or only after crises, locking systems into progressively unsustainable baselines) [38]. Subsidies intended to support fleets can entrench overcapacity [39]. Over time, this undermines intergenerational continuity and the loss of traditional knowledge [11,40]. Institutionally, data poverty traps management in inertia. The high cost and complexity of conventional assessments delay action, a “paralysis of analysis” that permits continued overfishing [41]. The absence of transparent, science-based advice fuels stakeholder conflict [9,28]. Moreover, data scarcity undermines the Precautionary Approach, as uncertainty is used to justify inaction rather than conservative management [42,43]. It also constrains advanced governance tools such as ITQs or TURFs [44].

Ultimately, the data-poor paradigm perpetuates a vicious cycle: insufficient data leads to weak management, resulting in ecological decline, social conflict, and further barriers to effective data collection. Breaking this cycle requires embracing pragmatic, data-limited assessment methods capable of delivering precautionary, risk-based advice under uncertainty [45,46].

1.3. Objectives and Scope

Given the profound socio-economic and ecological importance of Mediterranean fisheries, the pervasive data-poor paradigm has long constrained sustainable management. This review evaluates how science and management are responding, synthesizing recent methodological advances that offer new opportunities for progress [21,31].

This review provides the first comprehensive, Mediterranean-specific synthesis of Data-Poor Assessment Methods (DPAMs), critically examining their principles, applications, and integration into management frameworks. Building on global advances [18,23,47], it systematically catalogs the main classes of methods, catch-only, length-based, life-history, and qualitative approaches, and evaluates their performance through regional case studies [30,48].

Importantly, the review aims to assess the institutional uptake of these approaches within regional governance structures such as the GFCM and the Scientific, Technical and Economic Committee for Fisheries (STECF) [15,49], identifying the barriers that continue to limit their operational use. By consolidating fragmented knowledge and contextualizing it within the Mediterranean’s unique ecological and governance landscape [25,50], this work seeks to provide a strategic roadmap for improving data-limited management capacity.

Ultimately, the novelty of this study lies in bridging methodological development with real-world implementation (Figure 1), offering both a critical diagnosis of current practice and a forward-looking vision for sustainable, equitable fisheries management in the Mediterranean Basin [35,51,52].

2. The Data Landscape of Mediterranean Fisheries

2.1. Primary Data Sources and Their Characteristics

The Mediterranean fisheries data ecosystem comprises multiple complementary streams, each with distinct strengths, limitations, and spatial-temporal coverage (Figure 2). These sources collectively enable stock assessments despite pervasive data limitations.

The MEDITS (International Bottom Trawl Survey in the Mediterranean) program, initiated in the mid-1990s, provides standardized, fishery-independent abundance indices, length-frequency distributions, and community metrics across continental shelves and upper slopes [53]. This harmonized survey series is a cornerstone for regional stock assessments and ecological indicators. However, MEDITS has inherent limitations that must be accounted for. Spatially, it is restricted to depths down to ~800 m, omitting deeper assemblages, coastal nurseries, and structured habitats like seagrass meadows. Temporally, its annual spring/summer sampling may miss seasonal dynamics and episodic recruitment pulses. Furthermore, the fixed GOC-73 trawl gear has known selectivity biases [54], underrepresenting smaller size classes and species with low catchability, which can skew abundance indices and size structure metrics used in assessments.

National reporting systems supply catch and effort data critical for reconstructing exploitation trajectories and supporting management benchmarks. Yet, comparative analyses reveal that official landings often understate total fishing pressure, particularly when unreported components from SSF, discards, and illegal, unreported, and unregulated (IUU) fishing are omitted [55,56]. Catch reconstructions integrating household consumption, local sales, observer data, and expert knowledge produce more comprehensive removal series that offer counterfactuals to official statistics for retrospective policy evaluation [24,55].

Observer programs and targeted sampling reveal substantial and heterogeneous discard ratios in Mediterranean fisheries, particularly in bottom trawl operations. Basin-wide reviews document species- and size-classes disproportionately affected, guiding mitigation strategies, selectivity improvements, and the design of discard sampling schemes [56,57]. However, onboard observation remains uneven across fleets and subregions, leading to uncertain mortality estimates and weak baselines for evaluating bycatch mitigation efficacy.

Automatic Identification System (AIS) and Vessel Monitoring System (VMS) datasets have transformed spatial fisheries science by enabling inference on trawling corridors, intensity gradients, and transboundary effort distributions [58,59]. Recent advances integrate AIS with Synthetic Aperture Radar (SAR) and predictive modeling to address coverage gaps in nearshore areas and areas with imperfect AIS reception, strengthening evidence for spatial planning and cumulative impact assessment [60,61]. For instance, studies demonstrate that combining SAR with AIS enhances fishing vessel detection in the Mediterranean [62], while Generalized Additive Models (GAMs) applied to AIS data identify fishing corridors crucial for spatial management.

Most Mediterranean small-scale vessels fall below electronic tracking thresholds, creating significant data gaps in nearshore ecosystems. Coordinated participatory mapping with fishers has emerged as a critical alternative, documenting gear use, seasonal patterns, and fishing grounds to improve representativity in data-poor coastal zones [63,64]. Recreational fisheries, frequently overlooked, can locally rival commercial catches for specific taxa, necessitating structured survey designs to systematically capture participation, effort, catch, and economic expenditure [65].

Divers, anglers, and coastal communities contribute timely, georeferenced observations of rare and non-indigenous species, extending the reach of formal monitoring. When validated by experts, these inputs enhance early-warning systems for biosecurity and support ecological risk assessments amid rapid environmental change [66]. Participatory approaches have proven especially valuable for reconstructing historical changes that predate formal monitoring and for mapping essential habitats absent from official records [67].

To address the limitations of individual data sources, catch reconstructions synthesize disparate evidence, including gray literature, local knowledge, and consumption data, into comprehensive removal time series. Recent methodological advances stress the importance of integrating catch-based methods (CMSY) with length-based and abundance-based approaches to improve the robustness of stock status evaluations in data-limited settings [68]. Furthermore, hybrid frameworks that combine empirical data with model-based simulations offer a promising pathway for assessing the mixed-species fisheries that dominate the Mediterranean [69].

2.2. Gaps and Biases in Existing Data Systems

A consistent finding across data reconstructions is that official statistics systematically underreport total removals, especially from small-scale fisheries, discards, and IUU activities, which biases stock status indicators and hinders the evaluation of rebuilding potential in data-limited contexts [24,55]. The practical invisibility of small-scale fleets in electronic tracking systems, due to vessel size thresholds and frequent gear switching, creates major spatial and temporal blind spots in nearshore ecosystems, precisely where interactions with sensitive habitats and juvenile life stages are most intense [63]. Recreational fisheries introduce additional uncertainty, as participation and effort vary seasonally and geographically; without sustained, standardized monitoring, their catches are often omitted despite evidence of locally significant impacts [65].

Onboard observation of discards and bycatch remains uneven across fleets and subregions, leading to unreliable mortality estimates and inadequate baselines for assessing the effectiveness of bycatch mitigation and discard management measures in mixed demersal fisheries [56,57]. Survey-based indicators are constrained by the MEDITS program’s focus on shelf and upper slope habitats and by the selectivity of the GOC 73 bottom trawl gear, which may undersample deeper assemblages, fragile habitats, and certain size classes [53,54]. Complementary observation systems, such as deep-water surveys, baited remote underwater video systems, and ROV-based monitoring, are needed to fill these gaps, particularly in meso- and bathyal zones.

Data collection across Mediterranean countries often involves heterogeneous fleets employing diverse gear types under varying operational and environmental conditions. This methodological heterogeneity can introduce significant bias in catch composition, size selectivity, and abundance indices, complicating the interpretation of stock structure and trends. Without systematic inter-calibration, both across gears and between national programs, the comparability and consistency of data are jeopardized, undermining regional assessments and transboundary stock advice. Calibration exercises, such as parallel fishing trials and gear efficiency studies, are therefore essential to harmonize data streams and ensure that national statistics contribute reliably to basin-wide management frameworks.

In the EU Mediterranean, VMS is mandatory for vessels ≥ 12 m in length, and AIS for vessels ≥ 15 m. However, small-scale vessels (SSVs) under these thresholds, which constitute approximately 80–90% of the Mediterranean fleet by vessel count, are largely invisible to these electronic monitoring systems. This creates a significant spatial and temporal blind spot, particularly in coastal zones and nearshore habitats where SSVs operate intensively. Consequently, while AIS and VMS offer high-resolution data for larger commercial vessels, their representativeness of total fishing pressure remains partial.

AIS/VMS-based estimates of fishing effort are sensitive to classification algorithms, signal reception gaps, and vessel non-compliance, which can introduce spatial bias into fishing footprint and effort maps [58,59,60]. Without careful correction, these biases may lead to misidentification of fishing hotspots or underestimation of effort in nearshore and high-traffic areas. Integrating complementary remote sensing data, such as SAR, with statistical modeling approaches like GAMs has been shown to mitigate some of these gaps and improve the spatial resolution of fishing intensity maps.

Temporal interpretation of trends is further complicated by presentist bias and shifting baselines: historical removals, changes in gear selectivity, and the truncation of size distributions are often incompletely represented in contemporary datasets, potentially leading to overestimates of current stock status relative to historical baselines [25,70]. In a jurisdictionally complex basin, differences in reporting practices, taxonomic resolution, and sampling intensity across countries hinder harmonization and introduce comparability issues that affect subregional assessments and transboundary stock advice [25,71]. Collaborative initiatives, data-sharing platforms, and standardized protocols under the GFCM and EU DCF are essential to reconcile these differences and support integrated, ecosystem-based management.

Despite these persistent gaps, methodological and data innovations are emerging: integration of AIS with SAR and predictive models enhances the robustness of spatial effort maps; participatory mapping improves coverage of small-scale fleets; and targeted discards research refines species- and size-specific mortality estimates critical for mixed-fishery management and policy evaluation [60,63,72]. Together, these developments point toward a progressively more complete, multi-source data landscape capable of capturing the Mediterranean’s intrinsic multi-gear, multi-species complexity, while underscoring the continued need for sustained observation, harmonization, and validation to support science-based rebuilding and ecosystem-based management [25,46,73]. Figure 3 depicts a conceptual pipeline from primary data through integration layers to management-relevant outputs, emphasizing how tracking data feed spatial analysis, survey and market data underpin stock assessments, and citizen science contributes to early warning systems alongside qualitative knowledge channels.

3. Review of Data-Poor Assessment Methods (DPAMs)

The assessment of data-poor and data-limited fisheries has advanced considerably over the past two decades, driven by the global recognition that traditional, data-intensive methods are infeasible for the majority of exploited stocks worldwide [18,21,23]. A diverse suite of methods has been developed and tested across varied geographical and ecological contexts, offering pathways to management even when information is scarce.

Globally, DPAMs are broadly categorized by their data requirements. Catch-only models estimate biomass and fishing mortality from historical catch series using surplus production theory and Bayesian frameworks [47,74]. Length-based methods (LB-SPR, LBB) use size-frequency data to infer exploitation and spawning potential, bypassing the need for age information [22]. Abundance-based approaches (SPiCT, trend analyses) incorporate fishery-independent survey indices or standardized CPUE to track biomass trends [75]. Life-history invariant methods leverage empirical trait relationships to derive productivity and vulnerability proxies when fisheries data are absent [45]. To aid in selecting and applying the most appropriate method from these assessment methods, decision-support platforms have been developed. Tools such as FishPath [76] and the Stock Assessment Toolbox [77] provide structured, interactive frameworks that guide practitioners based on data availability, life-history traits, and management context. These platforms offer a wider suite of assessment options and are particularly valuable for navigating complex, multi-species fisheries.

Beyond purely quantitative methods, qualitative and integrative frameworks are increasingly used, especially in complex, multi-species systems. These include risk assessment tools like Productivity–Susceptibility Analysis (PSA) [78], multi-criteria indicator dashboards [79], and the systematic integration of fishers’ ecological knowledge [80]. Emerging hybrid platforms and Management Strategy Evaluation (MSE) frameworks, such as DLMtool, enable the synthesis of multiple data streams and the testing of robust harvest control rules under uncertainty [17,35].

3.1. Overview of Core Quantitative DPAMs

DPAMs applied in Mediterranean fisheries are broadly categorized into four groups: catch-only, length-based, abundance-based and risk assessment approaches [18,22,47]. As summarized in Table 1, each category differs in data requirements, underlying assumptions, and sensitivity to common regional biases such as unreported catches, hyperstability, and gear selectivity.

Moreover, the application of length-based and abundance-based methods is highly sensitive to gear-specific selectivity and catchability. Nationwide sampling programs that employ multiple gear types without calibration can produce biased length-frequency distributions and CPUE trends, leading to erroneous stock status diagnoses. Regular inter-calibration campaigns, comparing catches across gears, seasons, and regions, are crucial to adjust assessment inputs and enhance the reliability of data-limited methods.

In practice, catch-only models offer the broadest applicability where only catch time series exist, though their outputs are sensitive to prior assumptions about stock resilience and historical depletion [24,81]. Length-based methods provide direct proxies for exploitation and spawning potential where age data are unavailable but require representative length samples and reliable life-history parameters [82,83]. Abundance-based approaches (AMSY, SPiCT, state-space models) deliver higher precision when fishery-independent survey indices exist but remain vulnerable to hyperstability in CPUE data and effort creep [84,85]. Consequently, Mediterranean best practice increasingly emphasizes multi-method consensus and hybrid approaches (see Section 3.4), prioritizing catch reconstruction, fishery-independent data, and the communication of uncertainty [6,15,46,49].

Indicators such as shifting length-frequency distributions (reductions in mean length or L_max), or negative survey trends can provide compelling, though not always unambiguous, signals for precautionary management action, particularly when formal stock status estimates are uncertain [86,87]. However, interpreting a decline in mean length requires careful examination of the full length-frequency structure, as it may result from both the loss of larger adults due to fishing pressure or an influx of recruits from episodic strong year-classes.

Table 1. Summary of data-limited assessment methods applied in Mediterranean fisheries. The table compares three major categories of stock-assessment approaches, catch-only, length-based, and abundance-based, outlining the primary data requirements, key analytical models, representative Mediterranean applications, strengths, limitations, and core literature references (CMSY: Catch-Maximum Sustainable Yield; BSM: Bayesian State-Space Model; DB-SRA: Depletion-Based Stock Reduction Analysis; SSP: State-space Production model; AMSY: Abundance-Maximum Sustainable Yield; LB-SPR: Length-Based Spawning Potential Ratio; LBRPs: Length-Based Reference Points; ELEFAN-GA: Electronic Length Frequency Analysis—Genetic Algorithm; SPiCT: State-Space Production Model in Continuous Time; CPUE: Catch Per Unit Effort; IUU: Illegal, Unreported, and Unregulated fishing; MEDITS: Mediterranean International Trawl Survey; MEDIAS: Mediterranean Acoustic Survey; SSF: Small-Scale Fisheries, PSA (Productivity-Susceptibility Analysis), UAIDR (Uncertainty-Adjusted Index of Depletion Risk), SRA (Stock Reduction Analysis), L∞ (Asymptotic length, von Bertalanffy growth function), K (Growth coefficient, von Bertalanffy model), M/K (ratio of natural mortality to growth coefficient), and Lm (length at maturity).

Method	Data Required	Key Models	Mediterranean Applications	Strengths	Limitations	References
Catch-only	Annual catch series	CMSY, BSM, DB-SRA, CMSY++, SSP	Hake (Merluccius merluccius), red mullet (Mullus barbatus), deep-water rose shrimp, Aegean sardine	Minimal data; scalable to 100+ stocks	Sensitive to priors (r, K); assumes equilibrium; biased by IUU	[17,18,24,25,30,46,47,48,74,81,88,89,90,91,92,93,94]
Length-based	Length-frequency + life-history (L∞, K, M/K, Lm)	LB-SPR, LBRPs, ELEFAN-GA	Red mullet (Adriatic/Aegean), annular sea bream, sparids, crustaceans	No age data needed; direct F/M estimate	Steady-state assumption; sensitive to L∞, M/K; gear selectivity bias	[22,82,83,95,96,97,98,99,100,101,102,103]
Abundance-based	CPUE or survey index (MEDITS, MEDIAS)	SPiCT, trend rules, state-space models, AMSY	Sardine/anchovy (Adriatic acoustics), demersal stocks (MEDITS)	Tracks trends; less prior-dependent	Hyperstability in SSF CPUE; effort creep; survey selectivity; <10% stocks	[15,51,75,84,85,86,87,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122]
Risk Assessment	Life-history traits, gear exposure, qualitative risk criteria	PSA, UAIDR, SRA, Risk-based frameworks	Elasmobranchs, bycatch species, deep-water stocks	Low data needs; prioritizes management actions; good for multi-species	Qualitative/semi-quantitative; depends on expert judgment; hard to calibrate	[23,34,45,78,123,124,125]

Table 1. Summary of data-limited assessment methods applied in Mediterranean fisheries. The table compares three major categories of stock-assessment approaches, catch-only, length-based, and abundance-based, outlining the primary data requirements, key analytical models, representative Mediterranean applications, strengths, limitations, and core literature references (CMSY: Catch-Maximum Sustainable Yield; BSM: Bayesian State-Space Model; DB-SRA: Depletion-Based Stock Reduction Analysis; SSP: State-space Production model; AMSY: Abundance-Maximum Sustainable Yield; LB-SPR: Length-Based Spawning Potential Ratio; LBRPs: Length-Based Reference Points; ELEFAN-GA: Electronic Length Frequency Analysis—Genetic Algorithm; SPiCT: State-Space Production Model in Continuous Time; CPUE: Catch Per Unit Effort; IUU: Illegal, Unreported, and Unregulated fishing; MEDITS: Mediterranean International Trawl Survey; MEDIAS: Mediterranean Acoustic Survey; SSF: Small-Scale Fisheries, PSA (Productivity-Susceptibility Analysis), UAIDR (Uncertainty-Adjusted Index of Depletion Risk), SRA (Stock Reduction Analysis), L∞ (Asymptotic length, von Bertalanffy growth function), K (Growth coefficient, von Bertalanffy model), M/K (ratio of natural mortality to growth coefficient), and Lm (length at maturity).

Method	Data Required	Key Models	Mediterranean Applications	Strengths	Limitations	References
Catch-only	Annual catch series	CMSY, BSM, DB-SRA, CMSY++, SSP	Hake (Merluccius merluccius), red mullet (Mullus barbatus), deep-water rose shrimp, Aegean sardine	Minimal data; scalable to 100+ stocks	Sensitive to priors (r, K); assumes equilibrium; biased by IUU	[17,18,24,25,30,46,47,48,74,81,88,89,90,91,92,93,94]
Length-based	Length-frequency + life-history (L∞, K, M/K, Lm)	LB-SPR, LBRPs, ELEFAN-GA	Red mullet (Adriatic/Aegean), annular sea bream, sparids, crustaceans	No age data needed; direct F/M estimate	Steady-state assumption; sensitive to L∞, M/K; gear selectivity bias	[22,82,83,95,96,97,98,99,100,101,102,103]
Abundance-based	CPUE or survey index (MEDITS, MEDIAS)	SPiCT, trend rules, state-space models, AMSY	Sardine/anchovy (Adriatic acoustics), demersal stocks (MEDITS)	Tracks trends; less prior-dependent	Hyperstability in SSF CPUE; effort creep; survey selectivity; <10% stocks	[15,51,75,84,85,86,87,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122]
Risk Assessment	Life-history traits, gear exposure, qualitative risk criteria	PSA, UAIDR, SRA, Risk-based frameworks	Elasmobranchs, bycatch species, deep-water stocks	Low data needs; prioritizes management actions; good for multi-species	Qualitative/semi-quantitative; depends on expert judgment; hard to calibrate	[23,34,45,78,123,124,125]

3.2. Length and Age-Based Assessment Methods

Life-history-based methods are grounded in theoretical ecology and empirical relationships among biological traits, allowing inference of stock productivity and vulnerability when fisheries data are scarce. These approaches are particularly valuable for newly exploited stocks, bycatch species, or regions lacking historical monitoring, providing precautionary reference points in truly “data-less” situations [23,45].

3.2.1. Theoretical and Methodological Foundations of Life-History-Based Assessments

Life-history-based methods are founded on theoretical principles of energy allocation trade-offs [126,127] and provide practical, low-data tools for estimating stock productivity and vulnerability [23,45]. In the Mediterranean, where biological data are often fragmentary and age-structured information is rare, these approaches offer a critical pathway for generating precautionary reference points.

Key invariants such as the M/K ratio enable estimation of natural mortality (M) when only growth parameters are available [126,127]. This is particularly valuable for deep-water and elasmobranch species, where direct mortality estimates are seldom feasible [97,128]. Empirical equations for M, such as those based on gonadosomatic index [129], temperature-dependent growth [130], or maximum age [131], provide flexible, regionally adaptable inputs for data-limited stock assessments.

Simplified per-recruit models (Yield Per Recruit; YPR, SBPR) derive biologically grounded reference points without the need for full age-structured data [132,133]. In the Mediterranean, these proxies support the evaluation of exploitation levels and the identification of growth overfishing, especially in mixed trawl fisheries [83]. The Spawning Potential Ratio (SPR) proxy method, based on the ratio of age at maturity to age at full selectivity ($A_{\mathrm{mat}}/A_{\mathrm{sel}}$), offers a tractable indicator of reproductive status [132,134]. Its application in Mediterranean trawl fisheries has highlighted risks of growth overfishing where mesh sizes result in high captures of immature individuals [83].

PSA provides a semi-quantitative, risk-based framework that is well-suited to Mediterranean multi-species and bycatch assessments [78,125]. By integrating life-history traits (growth rate, fecundity) with susceptibility factors (gear overlap, habitat), PSA allows for the prioritization of vulnerable species, such as elasmobranchs and deep-water crustaceans, even in the absence of time-series data [34]. GFCM has endorsed PSA as a standardized tool for evaluating fishery impacts on data-poor and sensitive species [15].

The selection and performance of DPAMs in the Mediterranean are dictated by the type and quality of available data. Catch-based methods are the most widely applicable, requiring only a time series of total removals. This aligns with the Mediterranean context where annual catch statistics are often the sole consistent data stream, despite issues of under-reporting, especially in small-scale fisheries [24]. Their utility is highest for stocks with catch series spanning >20 years and informed priors for resilience, but they are sensitive to historical depletion assumptions and catch completeness [81].

Length-based methods require representative length-frequency samples, which are increasingly available through national port-sampling programs under the EU’s DCF. Their suitability for Mediterranean stocks is high for demersal species like red mullet (Mullus barbatus) and sparids, where age data are rare but length sampling is routine. In recent years, the Length-Based Integrated Mixed-Effects (LIME) model has emerged as a significant advancement over static length-based approaches. Unlike equilibrium-based methods, LIME integrates length-frequency data through time, allowing for the estimation of time-varying fishing mortality (F) and recruitment, and relaxing the steady-state assumption that can be problematic in rapidly changing or highly exploited fisheries [135]. By fitting to a time series of length compositions, LIME can estimate annual biomass trends and reference points like FMSY and BMSY, providing a more dynamic and data-adaptive assessment framework. This is particularly valuable for Mediterranean stocks subject to episodic recruitment pulses, interannual environmental variability, or non-equilibrium dynamics due to historical overfishing. The model’s ability to incorporate process error in recruitment and observation error in length samples makes it a robust tool for data-limited contexts where time-series of length data exist but age information remains scarce. Its application in mixed-fishery contexts also allows for the inclusion of multiple fleets with different selectivity patterns, enhancing its suitability for the multi-gear Mediterranean setting [49,136]. However, performance is contingent on accurate life-history parameters and can be biased by gear selectivity and non-representative sampling of market sizes [102,137].

Age-based or pseudo-age methods are the most data-demanding and are feasible for only a handful of Mediterranean stocks. Simplified age-structured require at least some age-length keys and catch-at-age information. Their application is currently limited to key commercial species like European hake in sub-regions with dedicated otolith collections [99] or small pelagics with age-reading programs. The data requirements often exceed the capacity for most Mediterranean stocks, reinforcing the reliance on length-based and catch-only approaches. While life-history methods are invaluable under data scarcity, their application in the Mediterranean must account for regional variability in life-history parameters, which can differ from global averages due to distinct environmental conditions and evolutionary histories [83]. Furthermore, these methods provide theoretical reference points rather than diagnostic assessments of current stock status; thus, they are best used to inform precautionary management targets, especially for low-resilience species, and should be cross-validated with other data-limited approaches where possible.

3.2.2. Performance, Limitations, and Best Practices in the Mediterranean Context

Life-history-based methods offer an indispensable safety net for fisheries management in data-limited contexts such as the Mediterranean, where conventional stock assessment inputs are often unavailable. However, while invaluable for providing first-order management guidance, their outputs are inherently generalized and must be interpreted with caution.

A major limitation stems from interspecific and intraspecific variation in life-history traits. Life-history invariants represent broad, central tendencies across taxa, yet significant variability exists within and among species. Mediterranean sub-populations of widely distributed species often display distinct growth rates or maturity schedules compared to their Atlantic counterparts, reflecting differences in temperature, productivity, and habitat conditions [82,138]. Applying global parameter estimates to these regional populations can therefore lead to biased inferences about productivity or sustainable mortality.

The sensitivity of these methods to parameter estimates further compounds this issue. Risk-based approaches rely on accurate life-history data to classify species vulnerability. Errors in key parameters such as age at maturity, fecundity, or natural mortality can significantly distort risk scores and management priorities [124]. Moreover, the static nature of life-history frameworks limits their capacity to reflect ecological dynamics. Most models assume fixed trait values and do not incorporate density dependence or environmentally driven variability. In the rapidly warming Mediterranean, for instance, shifts in temperature regimes have already altered growth and mortality rates in several species [139], potentially invalidating static assumptions.

Equally important is the conceptual distinction between reference point estimation and stock status evaluation. Life-history-based methods typically estimate potential or theoretical benchmarks, such as FMSY or sustainable yield proxies, derived from species productivity characteristics. They inform “what fishing mortality should be sustainable,” but not “what fishing mortality is currently occurring.” As noted by Prince & Hordyk [23], these tools thus provide management targets rather than diagnostic assessments of present stock condition.

To ensure robust and context-appropriate application in the Mediterranean, several best practices are recommended. First, whenever possible, models should be parameterized using Mediterranean-specific life-history data. Regional syntheses, such as those by Tsikliras & Stergiou [83] on size at maturity, substantially reduce bias relative to global averages and better capture the ecological idiosyncrasies of the basin. Second, life-history estimates should be incorporated as priors within Bayesian assessment frameworks, such as DB-SRA or BSM, where subsequent fisheries data can iteratively update these prior beliefs [94]. This integration preserves the value of life-history knowledge while allowing empirical evidence to refine estimates. Third, tools like PSA should be used primarily for triaging and prioritization, not as substitutes for full stock assessments. High-risk species identified through PSA, such as many elasmobranchs, should be prioritized for enhanced monitoring and precautionary management [34]. Fourth, life-history-derived reference points should be cross-checked against outputs from complementary methods, including catch- or length-based approaches. For instance, an FMSY estimated from a catch-based model should remain biologically consistent with the species’ resilience classification; large discrepancies can signal data or model issues requiring further scrutiny [81]. Finally, all applications of life-history-based outputs should adopt a precautionary management stance. Given their generalized and assumption-driven nature, these methods are best suited for establishing conservative rather than optimal exploitation limits, especially for species with low resilience or limited data [45].

3.3. Integrative and Qualitative Approaches

As the limitations of relying on any single, quantitative data-poor method become apparent, a paradigm shift towards integration and qualitative synthesis has gained prominence. Integrative and qualitative approaches acknowledge that in complex, data-limited systems like Mediterranean fisheries, a more holistic understanding of stock status and fishery health can be achieved by combining disparate data sources, expert judgment, and local knowledge. These approaches do not seek to produce a single, precise estimate of fishing mortality or biomass, but rather to build a robust, consensus-based narrative of the fishery’s state and trajectory, which can provide critical, actionable guidance for management when data-intensive, age-structured assessments are not feasible [51,140].

3.3.1. The Model of Intermediate Complexity for Ecosystem Assessments (MICE)

An emerging and powerful approach integrates qualitative and historical data from fishers to inform and validate quantitative ecosystem models. Local Ecological Knowledge (LEK) on species interactions, historical ecosystem states, and major ecological shifts can be incorporated into MICE, which bridge the gap between single-species and full ecosystem models [108]. In the Mediterranean, such integration has helped explain trophic dynamics and ecosystem shifts; for example, LEK documenting historical shark declines has been used to model mesopredator release effects on species like hake and rays, providing a more ecologically grounded basis for management [34].

It is important to note that MICE represents a distinct assessment and management paradigm, shifting the focus from single-species stock status to ecosystem-level indicators and trade-offs. Consequently, this approach may not be directly applicable in management environments strictly oriented toward single-species assessment and catch advice. However, in the complex, multi-species context of Mediterranean fisheries, where data for individual stocks are often lacking and trophic interactions are significant, MICE offers a valuable complementary framework for strategic, ecosystem-based management.

3.3.2. Multi-Criteria and Indicator-Based Frameworks

Integrated assessment frameworks combine multiple indicators, biological, ecological, social, and economic, to provide a holistic evaluation of fishery and stock health. The Traffic Light Approach offers an intuitive means of visualization, presenting indicators such as CPUE trends, mean length, catch-to-maximum ratios, and economic performance, each assigned a color (red for poor, yellow for caution, green for good) based on predefined reference points. The resulting color pattern provides an at-a-glance synthesis of stock status and highlights priority areas for management attention [79]. A more formal alternative, the State-Space Data-Limited Method (SS-DLM) framework, treats the outputs of data-poor assessments as uncertain but complementary indicators of the true stock state. By applying state-space modeling or Bayesian averaging, these methods integrate disparate signals into a consolidated estimate of stock status with quantifiable uncertainty [141]. In the Mediterranean context, multi-indicator dashboards are increasingly employed by STECF and GFCM to guide management decisions. Stocks are often classified as overfished when several indicators consistently signal deterioration, even when no single method alone is definitive [15,142].

3.3.3. Expert Elicitation and the Delphi Method

In data-poor contexts, structured expert judgment offers a transparent and defensible foundation for management decisions. The Delphi Method is a widely used approach in which a panel of experts anonymously answers sequential questionnaires, receiving feedback after each round until a consensus or stable range of opinions is achieved [143]. In fisheries assessment, it is used to elicit probability distributions for key parameters, or to classify stock status categories (“healthy,” “overfished”) [144]. In the Mediterranean, the Delphi method has supported the prioritization of research needs for data-poor stocks and the evaluation of management tools such as Fisheries Restricted Areas (FRAs), facilitating the synthesis of knowledge from scientists, managers, and experienced fishers into coherent, consensus-based assessments [50].

3.3.4. The Integration of Fishers’ Knowledge (IFK)

The systematic integration of Fishers’ Ecological Knowledge (FEK) or LEK has greatly expanded the data available for fisheries assessment. Fishers hold a unique, long-term understanding of marine ecosystems and resource dynamics that is often missing from scientific surveys [80]. Knowledge is typically elicited through structured and semi-structured approaches, including interviews and questionnaires to collect quantitative or semi-quantitative data on catch rates, species composition, and spatial distributions [145] participatory mapping to identify fishing grounds, spawning sites, and bycatch hotspots [11] and historical reconstructions that capture long-term changes predating formal monitoring [146]. In the Mediterranean, FEK has proven especially valuable for small-scale fisheries, supporting the identification of historical stock declines, validation of survey data, and mapping of essential habitats critical for both management and conservation [147,148]. This approach is particularly useful for poorly monitored or low-value species that fall outside traditional scientific programs.

3.3.5. Performance, Limitations, and Best Practices

While integrative approaches provide a valuable pathway for assessing data-poor stocks, they require careful implementation to avoid introducing new errors. Expert judgment and fishers’ knowledge are subject to cognitive biases, such as recall and optimism/pessimism bias [149], and assigning quantifiable uncertainty to qualitative assessments can be challenging for management bodies accustomed to numeric advice [150]. These approaches are also resource- and time-intensive compared with quantitative DPAMs [149].

Best practices for Mediterranean application emphasize triangulation, seeking convergence of evidence from catch, length, life-history, and local knowledge [87,140]. Structured, transparent protocols for data collection and synthesis are critical [151], as is co-production with scientists, managers, and fishers to build trust and improve compliance [152]. Integrative assessments should act as a bridge, guiding future monitoring and data collection [136], and outputs must communicate clear, actionable advice [153].

Selecting the most appropriate integrative or qualitative method depends on the specific management question, data context, and available resources. Each approach offers distinct trade-offs between scientific rigor, stakeholder engagement, and operational feasibility.

Table 2. Comparative overview of integrative and qualitative assessment methods for Mediterranean data-poor fisheries. LEK: Local Ecological Knowledge; MPA: Marine Protected Area; CPUE: Catch Per Unit Effort; DCF: Data Collection Framework; GFCM: General Fisheries Commission for the Mediterranean; STECF: Scientific, Technical and Economic Committee for Fisheries; FRA: Fisheries Restricted Area.

Approach	Primary Inputs and Data Needs	Typical Outputs/Objectives	Key Strengths	Key Limitations and Biases	Best Suited For	References
Ecosystem/Trophic Models	Quantitative data; LEK on past states	Trophic impacts; management trade-offs	Links single-species and ecosystem models; integrates LEK	Data-heavy; needs skilled modelers; sensitive assumptions	Transboundary stocks; MPA evaluation	[7,34,36,108]
Indicator-Based	Multiple indicators (CPUE, length, surveys, economics)	Status dashboards; integrated classifications	Clear synthesis; good communication; widely used	Arbitrary reference points; can hide conflicts; frequent updates	Routine DCF screening; multispecies advice	[79,141]
Expert Elicitation	Expert questionnaires	Consensus estimates; risk/option ranking	Structured expert judgment; works in data-poor cases	Biases; resource-intensive; depends on panel	Research prioritization; assessing new tools	[50,143,144]
Local Ecological Knowledge	LEK interviews and mapping	Historical baselines; spatial/seasonal patterns	Fine-scale, long-term insight; fills gaps	Recall bias; hard-to-quantify uncertainty; localized	Small-scale fisheries; habitat mapping; trend reconstruction	[80,147]

summarizes the key characteristics, applications, and limitations of the four primary methods discussed, providing a guide for practitioners in Mediterranean data-limited fisheries. Best practices emphasize triangulation, combining multiple lines of evidence to compensate for individual method weaknesses, and co-production with stakeholders to ensure legitimacy and compliance [51,140,152]

3.4. Hybrid and Multi-Method Frameworks

While Section 3.3 described integrative approaches that combine qualitative and quantitative data into multi-indicator dashboards or narratives, this section focuses on hybrid frameworks that formally integrate multiple quantitative DPAMs within a single statistical or simulation platform. These frameworks aim not just to combine indicators, but to merge likelihoods, priors, or model outputs into unified stock assessments.

The diverse toolkit of DPAMs available for Mediterranean fisheries, ranging from catch-only to length-based and life-history approaches, has advanced considerably, yet their greatest potential lies in integration rather than isolation. The next logical evolution is the development of hybrid and multi-method frameworks that merge multiple lines of evidence to enhance the robustness and credibility of stock assessments. These approaches move beyond a “model competition” paradigm toward an “evidence-weighting” or “ensemble” philosophy, systematically combining data from catch, length, abundance, life-history traits, and qualitative sources to generate more defensible, precautionary management advice under pervasive uncertainty [136,154].

The hybrid framework approach is particularly critical in the Mediterranean due to the patchy nature of data streams. A typical stock may have a catch time-series, intermittent length samples from port monitoring, and a sporadic MEDITS survey index. No single method can fully utilize this fragmented information. Hybrid frameworks, such as those implemented in DLMtool or through Bayesian model averaging, formally weight these disparate lines of evidence. For example, a state-space model can integrate a catch series with a biomass index from MEDITS, while simultaneously fitting to length-composition data to inform selectivity and growth, thereby providing a more robust assessment than any single-method application [17,19]. This multi-method triangulation is the most promising pathway for improving stock status diagnoses for the majority of data-limited Mediterranean stocks.

Within the landscape of integrated assessment platforms, the Stock Assessment Continuum (SAC) Tool [19,155] represents a significant advance toward unified, model-agnostic assessment in data-limited contexts. Built upon the Stock Synthesis (SS) framework, SAC allows users to fit a spectrum of models, from simple production models to length-based, age-structured, and state-space approaches, using whichever data types are available. This flexibility enables a continuous transition from data-poor to data-rich assessments within a single, consistent statistical platform, while also accommodating life-history parameters and recruitment estimation. Its modular design supports the principles of triangulation and evidence weighting emphasized in this section, offering a transparent and reproducible workflow for Mediterranean stocks where data types are often fragmented and heterogeneous [22,135,156].

At the heart of these hybrid systems is the principle of triangulation, using convergent evidence from independent methods to strengthen confidence in stock status estimates. Each DPAM carries unique sensitivities: catch-based models may be biased by underreporting, length-based analyses by sampling errors, and life-history proxies by spatial variability in parameters. When methods converge, the combined assessment gains credibility; when they diverge, the discrepancies highlight critical data gaps or flawed assumptions. For example, a high CMSY estimate paired with a low LB-SPR result might indicate hyperstability in catches or inaccuracies in biological parameters [81,87].

Operationally, hybrid frameworks have evolved into structured, probabilistic, and tiered systems capable of integrating multiple data streams. The Data-Limited Methods Toolkit (DLMtool) exemplifies this paradigm by functioning not as a single method but as an MSE platform [157]. Within DLMtool, an Operating Model (OM) simulates the “true” population dynamics conditioned on all available data-catch series, length compositions, abundance indices, and life-history priors. Various management procedures (MPs), are tested against the OM to evaluate their capacity to achieve objectives like maintaining biomass above BMSY or preventing collapse, thus shifting the emphasis from static assessment to adaptive management [17]. Similarly, the Stock Synthesis (SS) framework, though originally developed for data-rich scenarios, can be adapted to data-limited contexts by integrating diverse data types through likelihood-based weighting [19]. Simplified SS models can fit catch, abundance, and length data simultaneously, estimating trajectories while accounting for the relative reliability of each input [19]. In the Mediterranean, pilot studies combining MEDITS survey indices, standardized CPUE, and port sampling length-frequency data have shown that such integration produces more comprehensive and transparent assessments than single-method approaches [15].

For the most data-limited situations, the Tier-5 precautionary decision tree offers a pragmatic sequence for structured assessment [158]. Starting with catch-only models like CMSY, assessors can incrementally add data-length information for LB-SPR, abundance indices for state-space models like AMSY, and resort to life-history proxies or qualitative triangulation when data are sparse [136]. This hierarchical workflow maximizes data use, provides a transparent audit trail, and supports regional standardization under frameworks such as the GFCM [49,142]. Despite their promise, hybrid frameworks face notable challenges in the Mediterranean. Their computational and technical complexity can strain the capacity of regional institutions [159], and combining low-quality datasets may propagate rather than reduce uncertainty, a manifestation of the “garbage in, garbage out” problem [160]. Moreover, overly complex or opaque models risk eroding stakeholder trust if assumptions are not communicated transparently [151]. To mitigate these risks, best practices emphasize rigorous data vetting (“data archaeology”) before modeling, ensuring consistency, completeness, and bias detection [24]. The adoption of standardized multi-method workflows, such as the Tier-5 structure, can ensure reproducibility and comparability across stocks, supported by GFCM-led guidance and capacity building [31].

MSE remains the gold standard for testing hybrid frameworks, as it explicitly evaluates performance under uncertainty by simulating ecological, data, and implementation variability. MSE applications to Mediterranean small pelagics are already demonstrating the approach’s potential to strengthen adaptive, evidence-based management [117,161]. Equally vital is effective communication of results: visual tools like Kobe plots or traffic-light summaries can distill complex model outputs into intuitive formats, enhancing transparency and stakeholder engagement [162]. Ultimately, the goal of hybrid assessments is to produce actionable advice, for example, when consensus among CMSY, LB-SPR, and fishers’ knowledge indicates growth overfishing, managers might raise mesh size so that the length at first capture (Lc) exceeds the length at maturity (Lm), thereby increasing spawning biomass by a measurable margin within a defined timeframe [23].

4. Applications in the Mediterranean: Case Studies and Adoption

The application of data-limited assessment methods in the Mediterranean has expanded markedly over the last decade, reflecting the structural realities of the region’s fisheries-multi-species, multi-gear fleets dominated by small-scale operations, uneven monitoring across GSAs, and fragmented time series, while responding to management imperatives set by regional bodies such as the GFCM and EU advisory frameworks. Within this context, catch-only surplus production approaches and length-based methods have provided tractable pathways to quantify status and explore rebuilding under uncertainty, particularly where age-structured assessments remain infeasible or impractical for most stocks.

Stock-Level Data Inventory and Assessment Suitability

Given the pervasive under-reporting and data gaps in Mediterranean small-scale fisheries (Section 2.2), complete, stock-specific catch series are often unavailable or biased. In this context, catch-only models should be used cautiously: their responsible application requires (i) prior catch reconstructions, (ii) carefully specified resilience and depletion priors based on Mediterranean life-history information, and (iii) cross-validation against independent indicators such as length compositions or survey trends where available.

When these steps are followed, catch-only methods can provide preliminary, albeit uncertain, estimates of stock status and productivity that are otherwise absent, as demonstrated in large-scale assessments of Turkish stocks [163] and in regional screening studies [48]. Nevertheless, in many Mediterranean fisheries, length-based methods often represent a better-matched alternative because they rely on size-frequency data, which are increasingly collected through standardized port-sampling programs under the EU Data Collection Framework and are less susceptible to the systematic under-reporting that plagues catch records [49,102]. The key lesson from recent applications is that no single data-poor method is universally appropriate; rather, a diagnostic, multi-method approach, tailored to data availability and transparent about uncertainties, is essential to generate credible advice for Mediterranean fisheries.

Length-based methods are increasingly applied where port or market sampling provides size-frequency data, offering direct proxies for exploitation and biomass while maintaining transparent links to size-selective regulations [164,165,166]. The Length-Based Bayesian estimator infers relative biomass and exploitation from length distributions and has been benchmarked against independent assessments, while LBSPR estimates equilibrium spawning potential ratio (SPR), enabling managers to align technical measures with reproductive targets such as SPR40% [22]. Methodological guidance now emphasizes responses to recruitment pulses, binning effects, and the stabilization of priors for L∞ and M/K using local life-history data, increasingly available from Mediterranean studies [22,167]. For example, recent work in the Aegean Sea on common pandora provides age, growth, and mortality parameters directly usable as priors in length-based and surplus production assessments, reducing structural uncertainty and reconciling reference point diagnostics in Greek waters [168].

Structured catch-composition programs in the southeastern Aegean reveal seasonal mosaics of species and sizes across gears and habitats, enabling empirical selectivity profiles and improved realism in length-based diagnostics [169]. These detailed catch records also provide the foundation for empirical indicator approaches based on shifts in species composition, a valuable tool for assessing small-scale, multi-species fisheries where formal logbooks are absent and stock-specific data are limited. For example, the dominance of non-indigenous species in specific gears (>90% in trammel-net catches) not only requires careful treatment to avoid bias in LBB/LBSPR outputs [170] but can itself serve as an indicator of ecological change and fishing pressure. Because length-based estimators rely on representative samples and valid selectivity/life-history priors, violations can bias outputs [22,47]. Consequently, studies recommend stratified spatial–temporal sampling, separate treatment or exclusion of non-target taxa, and explicit accounting for gear- and species-specific selectivity or life-history priors when preparing inputs for LBSPR/LBB. In data-poorest settings, monitoring changes in species composition, such as the rise in non-indigenous species or the loss of historically abundant taxa, can offer a pragmatic, early-warning signal of ecosystem-level impacts, even when formal stock assessment is not feasible.

For deep-dwelling species like red coral, slow growth, late reproduction, and depth-dependent densities complicate survey indices [171]. Here, state-space surplus production models absorb noisy abundance data, while LBSPR translates size structures into reproductive capacity proxies [172]. This pairing allows managers to compare outcomes of technical measures under uncertainty while accounting for life history and spatial heterogeneity. Regulatory contexts, including EU Habitats Directive and Barcelona Convention listings, further motivate conservative targets, and combining dynamic abundance models with length-based reproductive metrics provides a defensible management pathway [171,173].

Additional regional case studies expand the evidence base and demonstrate portability across taxa and geographies [165,174,175]. For instance, CMSY and length-based indicators applied to Atlantic bonito in the eastern Atlantic and Aegean have estimated stock status and sustainability metrics, providing templates for small tunas in data-limited contexts [176]. Comparative evaluations in island systems, such as the Azores, benchmarked LBB, LBSPR, and indicator suites against independent data, revealing mixed performance and emphasizing the importance of aligning model choice with data availability, selectivity realism, and life-history knowledge, lessons directly relevant to heterogeneous Mediterranean GSAs [165]. Method portfolios for Macaronesian resources position LBSPR, LBB, length-based indicators, and SPiCT as complementary tools matched to data conditions, length-frequency time series, growth and maturity inputs, and intermittent abundance indices, offering structured pathways to triage stocks and prioritize data collection [165,177,178,179]. Collectively, these studies demonstrate that catch-only and length-based approaches can yield coherent, actionable advice in Mediterranean fisheries, provided that local ecological knowledge and life-history information stabilize priors and selectivity, and diagnostics are interpreted in the context of fleet behavior and environmental change [57,163,180].

The comparative performance of data-poor assessment methods in the Mediterranean is informed by regional applications and insights from simulation studies tailored to local conditions, where short time-series, unknown selectivities, and mixed-gear fleets are pervasive [64,106,174,181]. Applications of catch-only models to Mediterranean stocks, for example, the assessment of 54 Turkish stocks, which estimated 94% below B_msγ and 85% subject to overfishing, demonstrate their utility for rapid, large-scale stock triage [163,182]. However, these analyses also reveal the difficulty of independent validation in the absence of age-structured benchmarks, and regional evaluations note that CMSY tends to overestimate F/F_msγ and underestimate B/B_msγ when catch series are short or incompletely reconstructed, underscoring the need for careful catch reconstruction and informed priors [154]. Length-based methods have been extensively applied and tested in Mediterranean settings, with the GFCM Scientific Advisory Committee highlighting length-cohort analysis (LCA) as a particularly appropriate short-term tool under data-poor conditions, while stressing that multi-method testing is essential due to inherent biases [8]. Performance of length-based approaches is highly sensitive to life-history parameters and sampling representativeness [102,137,183], reinforcing the limitations outlined in Section 3.1.

For short-lived species or those with high recruitment variability, length-based estimates exhibit wide uncertainty, limiting their diagnostic power [136]. Ultimately, data-poor assessments in the Mediterranean should be viewed as triage tools that prioritize stocks for enhanced monitoring or precautionary management, rather than as sources of definitive reference points, a perspective reinforced by regional advisory frameworks and case-study evidence [15,142].

European hake benefits from the long-term MEDITS trawl survey (since 1994), which provides standardized abundance indices across continental shelves [184]. State-space models, such as SPiCT and AMSY, are increasingly applied where MEDITS data exist, delivering robust estimates of exploitation status, fishing mortality, and biomass relative to MSY [75,106,161,185,186]. However, discard estimates remain uncertain, commercial CPUE often exhibits hyperstability, and age data are sparse, limiting the applicability of age-structured assessments [187,188]. In data-poorer areas of the southern Mediterranean, CMSY has been applied but tends to underestimate historical depletion, inflating relative biomass estimates [154].

Table 3. Data availability, current assessment methods, and recommended data-poor assessment methods (DPAMs) for selected Mediterranean fisheries. DPAMs are highlighted as either primary (recommended) or exploratory/supplementary where noted. MEDITS: Mediterranean International Trawl Survey; MEDIAS: Mediterranean Acoustic Survey; CPUE: Catch Per Unit Effort; SPiCT: Stochastic Production Model in Continuous Time; LB-SPR: Length-Based Spawning Potential Ratio; LBB: Length-Based Bayesian Biomass; CMSY/CMSY++: Catch-MSY (catch-only stock assessment model); AMSY: Adapted Catch-MSY; VPA: Virtual Population Analysis; SSF: Small-Scale Fisheries.

Species	European Hake (M. merluccius)
Primary Data Available	MEDITS survey indices, catch time series, length samples (port), CPUE available but often biased/spatially limited (hyperstability concerns)
Current Assessment Methods	SPiCT, age-structured (few stocks), CMSY/CMSY++ are exploratory/complementary for hake
Key Data/Knowledge Gaps	Age data sparse, CPUE hyperstability, discards poorly quantified
Recommended DPAMs	SPiCT (with MEDITS), LB-SPR, CMSY/CMSY++ (require careful priors and LB-SPR needs representative length data)
Key References/Applications	[25,56,142,161,163,189,190]
Species	Sardine (S. pilchardus)
Primary Data Available	MEDIAS surveys, catch, length-frequency, age–length keys when available
Current Assessment Methods	SPiCT, surplus production, age-based/catch-at-age/VPA where sufficient data exist, catch-only/exploratory methods occasionally used
Key Data/Knowledge Gaps	Spatio-temporal variability and environmental/climate-driven shifts, SSF catches, shrinkage of length-at-age, under-reporting/discard uncertainty
Recommended DPAMs	SPiCT, LB-SPR (if lengths representative), CMSY only as supplementary or exploratory
Key References/Applications	[15,191,192,193,194]
Species	Deep-water rose shrimp (P. longirostris)
Primary Data Available	Catch, length-frequency/carapace-length samples, some trawl-survey data,
Current Assessment Methods	Quantitative stock assessment under GFCM, regional catch-based demographic studies where data exist
Key Data/Knowledge Gaps	Uncertainty in stock structure, spatial coverage gaps, high and poorly quantified discard/bycatch, variable life-history parameters, environmental effects on abundance and distribution, lack of fishery-independent survey
Recommended DPAMs	LB-SPR or life-history/demographic methods in data-poor areas and catch-only approaches (cautiously due to discard and environmental variability) [15,195,196,197]
Species	European anchovy (E. encrasicolus)
Primary Data Available	MEDIAS acoustic surveys, catch data, port/landing length samples
Current Assessment Methods	SPiCT, biomass dynamic models, CMSY/AMSY (exploratory)
Key Data/Knowledge Gaps	Recruitment variability, seasonal availability, SSF under-reporting, limited age data
Recommended DPAMs	SPiCT (acoustic), LB-SPR if length samples representative, CMSY/AMSY (with catch reconstruction; exploratory)
Key References/Applications	[15,198,199,200,201]
Species	Red mullet (M. barbatus)
Primary Data Available	Length-frequency (port), catch data, some MEDITS indices
Current Assessment Methods	LB-SPR, CMSY, empirical indicators
Key Data/Knowledge Gaps	Age data are rare and difficult to validate, selectivity varies by gear, discards poorly quantified; growth and maturity vary regionally
Recommended DPAMs	LB-SPR, LBB, CMSY, traffic-light indicators
Key References/Applications	[182,202,203,204,205]

presents some of the most common caught fish in the Mediterranean Sea and their assessment methods.

Length-based methods are used where representative port samples exist, offering insights into exploitation status and growth overfishing; their performance is sensitive to natural-mortality assumptions and the representativeness of length samples [136]. The GFCM has adopted size-based regulations informed by LB-SPR results and uses SPiCT advice for western-Mediterranean rebuilding plans, highlighting the importance of combining complementary data streams to address hyperstability, discard uncertainty, and limited age information [206].

Sardine status is primarily informed by MEDIAS acoustic surveys in the Adriatic, Aegean, and Alboran Sea, with SPiCT as the preferred assessment tool where survey indices are available [207]. In regions lacking regular acoustic coverage, such as parts of the Ionian Sea, CMSY++ has been applied; however, this method is highly sensitive to priors for intrinsic growth and should be used only as an exploratory or supplementary tool, as it may mask overfishing [208]. The 2023 benchmark for the Adriatic sardine stock indicates a depleted spawning stock relative to MSY (SSB/B_msγ ≈ 0.65) and a fishing mortality well above sustainable levels (F/F_msγ ≈ 1.52), underscoring over-exploitation [15]. Age data are limited, and small-scale fishery catches are often under-reported, further increasing uncertainty. Length-based indicators, such as mean length and Lmax, provide auxiliary signals, with declining mean lengths serving as early warnings of growth overfishing in several GSAs [209,210]. The GFCM’s multiannual management plan for Adriatic small pelagics relies on SPiCT outputs to set catch ceilings, while in the Aegean a traffic-light approach integrates CMSY results, length trends, and expert judgment [15].

Deep-Water Rose Shrimp exemplifies a truly data-poor stock. Fishery-independent surveys are absent, as MEDITS does not sample its deep-water habitat (>800 m), catch data are often aggregated, and life-history parameters are poorly. Consequently, formal stock assessments are rare [211]. Catch-only models have been tentatively applied but require strong, uncertain priors and assume equilibrium conditions that are unlikely in expanding deep-water trawl fisheries [154]. Length-based or life-history proxy approaches remain the only feasible quantitative tools in these areas. Management has therefore relied primarily on input controls, such as depth closures and effort restrictions [212].

European anchovy assessments are most advanced in the Adriatic Sea (GSAs 17–18), where annual MEDIAS acoustic surveys have provided robust biomass indices since the early 2000s [198,207]. In the latest assessment for the Adriatic, spawning-stock biomass remains above reference levels with B/B_pa = 1.23 and B/B_lim = 1.62, while fishing mortality is near the target (F/F_msγ ≈ 1.04) [213]. These survey data are integrated with catch data in state-space models (SPiCT) to estimate fishing mortality and biomass relative to MSY [214]. SPiCT performs reliably, capturing recruitment pulses and supporting interannual catch limits adopted by the GFCM. However, age data are limited, and small-scale fisheries are often under-reported, which increases uncertainty [215]. In contrast, assessments in the Aegean and eastern Mediterranean rely primarily on catch-only methods, applied after careful catch reconstruction [106]. These approaches are exploratory and can produce optimistic biomass estimates if priors or catch completeness are inadequate. Key lessons include the indispensability of acoustic surveys for anchovy assessment and the necessity of catch reconstruction before applying data-poor methods in areas dominated by small-scale fisheries. Length-based indicators can be applied where representative size data are available, providing supplementary insight into exploitation status and growth overfishing [154,216].

Red Mullet is well-suited for length-based assessment, with extensive port-sampling length-frequency data available in the Adriatic, Aegean, and Ionian seas [202]. LB-SPR is routinely applied and indicates widespread growth overfishing (SPR < 20%) in several GSAs. The method’s performance depends critically on accurate life-history priors and representative length samples; violations, such as oversampling marketable sizes, can bias results [136]. Figure 4 shows the decision-making framework for assigning Mediterranean fish stocks to assessment methods based on data availability and quality.

The process begins with catch data and proceeds through a series of data-availability questions, including length-frequency data, survey or CPUE indices, whether CPUE can be interpreted as an abundance index, age-structured data, and life-history parameters. Stocks are allocated to one of five methodological tiers, catch-only, length-based, abundance-based, age-structured, or life-history proxies, with abundance-based models preferred over catch-only approaches whenever a reliable abundance index is available, and data quality assessments and insufficient-data branches can redirect stocks or recommend additional data collection.

Age data are rare and difficult to validate, which reinforces the reliance on length-based methods. Catch-only approaches have been used for cross-validation but often disagree with LB-SPR outcomes, highlighting the sensitivity of CMSY to catch completeness and priors [136,154,210].

In summary, the heterogeneous data landscape of Mediterranean fisheries precludes a one-size-fits-all assessment solution. A tiered, hybrid strategy, employing catch-only models for rapid triage, length-based methods where samples exist, and state-space models where surveys are available, provides the most feasible and robust pathway to stock status evaluation within the region’s governance frameworks.

5. Challenges and Limitations

The application of DPAMs in the Mediterranean has expanded the scope of assessable stocks, moving management from a state of near-total ignorance to one of informed, albeit uncertain, guidance. However, significant methodological, implementation, and governance challenges persist, which can undermine the reliability of assessments and lead to misguided management advice if not explicitly acknowledged and addressed.

5.1. Methodological Challenges

5.1.1. Sensitivity to Priors and Life-History Parameters

The well-documented sensitivity of DPAMs to prior distributions and life-history parameters (Section 3.1) transitions from a technical concern to a fundamental management challenge in the Mediterranean. When priors for resilience, natural mortality, or growth are drawn from global databases [47], that poorly represent regional populations [82,83], the resulting stock status estimates carry a hidden, unquantifiable risk of severe bias [34,140,154]. This creates a dilemma for managers: scientifically defensible advice requires locally derived biological parameters, yet the data-poor condition itself precludes their collection. Consequently, management bodies are forced to base potentially consequential decisions on estimates whose accuracy hinges on often-unverified assumptions. This dilemma can erode stakeholder trust, fuel contention over assessment results, and in the worst case, lead to the approval of unsustainable exploitation levels masked by overly optimistic priors. Overcoming this challenge requires institutionalizing protocols for sourcing and vetting life-history priors within regional assessment frameworks, and adopting a precautionary stance that explicitly discounts advice derived from unvalidated generic parameters.

5.1.2. Violation of Equilibrium and Steady-State Assumptions

The equilibrium assumptions underpinning many DPAMs, particularly surplus production models and length-based methods like LB-SPR, are frequently violated in Mediterranean fisheries due to the region’s distinct ecological and exploitation history [22,23]. Unlike theoretical scenarios where stocks begin assessment periods near carrying capacity, most Mediterranean populations have experienced prolonged, intensive exploitation, meaning catch series often start from already-depleted states. This historical depletion leads to systematic underestimation of carrying capacity and unexploited biomass, consequently inflating estimates of current relative biomass [24,217].

Beyond historical overfishing, contemporary Mediterranean dynamics further disrupt equilibrium conditions. Episodic recruitment pulses, common in small pelagics like anchovy and sardine, create transient population structures that conflict with steady-state assumptions [13]. Simultaneously, rapid warming, acidification, and the proliferation of non-indigenous species are fundamentally altering growth rates, mortality schedules, and species interactions across the basin [13,101]. For length-based methods like LB-SPR, these environmental shifts can decouple the assumed relationships between size, age, and reproductive output, producing biased estimates of spawning potential unless explicitly accounted for. This confluence of historical depletion and ongoing, rapid environmental change creates a fundamental mismatch with the steady-state and constant-productivity assumptions underpinning many DPAMs, such as surplus production models and equilibrium length-based methods. In a basin experiencing regime shifts and non-stationary dynamics, these assumptions are not merely violated but are increasingly untenable, posing one of the most significant challenges for generating reliable stock status advice. Future methodological development must prioritize frameworks that can accommodate time-varying productivity and non-equilibrium starting conditions.

These violations necessitate careful interpretation of DPAM outputs in Mediterranean contexts. Where possible, assessments should incorporate historical catch reconstructions to better approximate initial biomass, employ state-space formulations that accommodate non-equilibrium dynamics, and apply environmental covariates to account for climate-driven changes in life-history parameters.

5.1.3. Pervasive Issues with Data Quality and Biases

The data quality limitations detailed in Section 2.2 do not merely create uncertainty; they directly and disproportionately bias the outputs of DPAMs due to these models’ simplified structures and reliance on often-flawed input data. The consequences of core data problems, systematic under-reporting, hyperstability in fishery-dependent indices, and non-representative length sampling, for each class of DPAM were detailed in Section 3.1 and Stock-Level Data Inventory and Assessment Suitability. For example, under-reporting leads catch-only models to produce falsely optimistic stock status estimates [89,154,218], while CPUE hyperstability (Figure 5) can mask severe biomass declines in abundance-based approaches [16,84,107]. Similarly, unrepresentative length sampling inflates mean length and spawning potential estimates in length-based assessments [49,142,219].

The key challenge in the Mediterranean context is that these biases are often systemic and difficult to quantify, making standard correction techniques hard to implement at scale. Therefore, a defensive analytical posture is essential: DPAM applications must prioritize sensitivity testing across plausible bias scenarios, explicitly favor fishery-independent data where available, and transparently communicate how unquantifiable data limitations fundamentally constrain the inference of stock status.

5.1.4. Challenges of Model Structure, Multi-Species Contexts, and Uncertainty Communication

As assessment methods become more complex and integrated, challenges related to model structure and transparency arise. Surplus production models assume a specific functional form for the relationship between biomass and production, so if the stock’s true productivity curve differs, estimates of MSY and related reference points can be biased [220]. Similarly, sophisticated hybrid or Bayesian frameworks are often perceived as “black boxes” by managers and stakeholders, and lack of transparency can undermine trust and hinder adoption, an important concern in Mediterranean multi-gear, small-scale fisheries where stakeholder buy-in is essential [9,151].

Most DPAMs are single-species models applied to individual stocks, which contrasts with the inherently multi-species nature of Mediterranean fisheries. Management measures based on one stock can have unintended consequences for co-caught species with different selectivity and life-history traits [31]. Additionally, single-species models ignore trophic interactions; for instance, the recovery of a predator stock could increase predation on prey species, complicating their management [7]. While ecosystem models could address these dynamics, they are data-intensive and generally infeasible in data-poor Mediterranean contexts, leaving a gap in ecosystem-based management.

Data-poor assessments are characterized by high uncertainty, yet quantifying and communicating this uncertainty to decision-makers remains challenging. Bayesian methods can provide credible intervals, but these are conditional on the model structure and priors and may not capture structural uncertainties, such as hyperstability or abrupt environmental shifts [117]. Managers, seeking clear guidance, may struggle with probabilistic outputs, for example, a statement that there is a 60% chance a stock is overfished can lead to indecision or overly optimistic interpretations [116]. Effectively framing uncertainty within a precautionary management context is therefore a persistent methodological and communication challenge.

5.2. Implementation Challenges

While the methodological hurdles of DPAMs are significant, they represent only one facet of the challenge. The transition from a scientifically validated method to a routinely applied tool that generates actionable management advice is fraught with implementation barriers. These challenges are often rooted in socio-economic, institutional, and human resource constraints, and they frequently pose a greater obstacle to sustainable management than the development of the methods themselves [136,140]. In the fragmented and diverse context of the Mediterranean, these implementation challenges are particularly acute.

5.2.1. Technical, Institutional, and Socio-Political Barriers

The uneven distribution of technical expertise and financial resources, a foundational feature of the Mediterranean data-poor context (Section 1.2), represents a primary barrier to implementing DPAMs. Many advanced methods (SPiCT and DLMtool) demand proficiency in statistical programming and population dynamics [17,136]. Consequently, the shortage of trained scientists in southern and eastern Mediterranean states [6,26] means that even validated DPAMs cannot be applied consistently across Geographical Sub-Areas (GSAs). This creates assessment blind spots, undermines regional management coherence, and forces reliance on a small cadre of overburdened experts, slowing the entire advisory process. Furthermore, the long-term funding required to sustain the data that feed these assessments is often unstable [49], jeopardizing the continuity of even the most basic monitoring programs like MEDITS [9,110].

The utility of any DPAM is limited by the quality of the underlying data pipeline, which in many Mediterranean contexts is fragmented or inefficient. Data collection is often dispersed across multiple agencies with poor coordination and inconsistent sharing protocols [221], and methods for recording basic metrics like length frequencies or effort are rarely standardized, hindering regional aggregation and creating a “Tower of Babel” effect [25,222]. Monitoring small-scale, multi-gear, multi-species fleets remains particularly challenging (Section 2.2), and systematic gaps in nearshore and informal landing sites continue to bias assessments toward better-monitored industrial fleets (Figure 6) [10,11,24,25].

Effective management advice requires legitimacy and credibility among fishers, yet achieving buy-in remains a major challenge. Assessments based on methods that stakeholders do not understand are often met with skepticism, particularly when results conflict with fishers’ experience, a perception reinforced by phenomena like hyperstability [51,84]. Recommendations to reduce effort or change practices can face strong resistance, especially where livelihoods are threatened and alternatives are limited, creating a political economy that may delay or dilute science-based measures [28,31,37,40].

Fisheries management in the Mediterranean is constrained by complex institutional structures that are slow to integrate new scientific approaches. Established bodies like the GFCM and EU STECF are oriented toward data-rich assessments, making the adoption of probabilistic DPAM advice gradual and sometimes resisted as “second-best” or experimental [15,142,151]. Legal frameworks often lack clear mandates for precautionary action in data-poor contexts, allowing inertia to persist despite indicative evidence, which can contravene the Precautionary Approach [41,42,43]. Cross-jurisdictional coordination further complicates matters, as shared stocks require agreement on methods, data harmonization, and joint interpretation, challenges that often result in conflicting assessments or management deadlock [30,223].

A major challenge in implementing DPAMs is the communication gap between scientists and policymakers. Managers often seek simple, definitive answers, whereas DPAM outputs are inherently uncertain and probabilistic, making it difficult to convey risk without misinterpretation or misuse by stakeholders to justify minimal action [38,162]. Political cycles also operate on short-term timelines, whereas assessments, stakeholder consultations, and stock responses require long-term perspectives, often favoring quick fixes over adaptive management [224]. Additionally, fishery-dependent data frequently undersample the stock’s full spatial range, further complicating accurate assessment and sustainable management.

5.2.2. Policy Integration and Fragmented Governance

The successful application of DPAMs in the Mediterranean ultimately depends on governance systems capable of translating scientific advice into effective management. Yet, the region’s complex, multi-jurisdictional nature, entrenched political economies, and fragmented policy frameworks create a governance trap that often neutralizes even robust assessments [28,40]. The Precautionary Approach, meant to ensure that lack of full certainty does not delay conservation measures [42], is frequently distorted in practice. In the data-poor Mediterranean, uncertainty is often invoked to justify inaction, with managers claiming DPAM results are not “conclusive enough,” leading to endless calls for further research and continued overfishing [38,41,225]. Most governance bodies have yet to operationalize precaution through pre-agreed harvest control rules or a formal Tier 5 system, leaving management decisions subject to political negotiation rather than objective criteria [111,136,151].

Governance challenges are intensified by the region’s jurisdictional fragmentation: over 20 bordering states share numerous transboundary and straddling stocks that demand harmonized data, coordination, and compromise [6,8]. Divergent assessment methods, uneven institutional capacity, and conflicting national priorities often produce incompatible advice and management deadlock [25,30]. Disparities between northern EU states operating under the CFP and DCF and non-EU countries with differing legal and economic frameworks further exacerbate imbalance, frequently reducing regional management to the lowest common denominator [26,31].

These structural issues are reinforced by the political economy of fisheries. Management is shaped less by science than by short-term political and economic interests. Organized industry groups often resist restrictions that threaten immediate profits, while uncertainty in data-poor assessments is exploited to argue for leniency and sustain overcapacity [9,226,227]. Decision-makers focused on short electoral cycles tend to postpone difficult reforms, perpetuating harmful subsidies and weak precautionary measures [224,228,229].

Even when scientific assessments are available, institutional and political delays can hinder timely management responses, as exemplified by the slow policy reaction to declining Adriatic sardine stocks (Figure 7) [15,142].

Subsidies for fuel, vessel construction, and modernization lower operational costs and maintain excessive fishing pressure even as stocks decline [39,49], creating a self-reinforcing cycle of overfishing and political dependency [230]. The reliance on input controls, coupled with poor data, impedes the introduction of output-based or rights-based systems such as ITQs or TURFs, producing a self-reinforcing governance trap (a ‘Catch-22’) that operationalizes the broader vicious cycle described in Section 1.2: poor data prevent sophisticated management, and weak management sustains poor data [10,44]. Further complications arise from weak policy integration and conflicting objectives across governance domains. Fisheries agencies often operate independently from environmental authorities responsible for implementing frameworks such as the EU Marine Strategy Framework Directive or the Barcelona Convention [50,231]. When stock assessments reveal the need to curb fishing, socio-economic priorities often override ecological imperatives [232]. The absence of a strong Ecosystem-Based Management approach means DPAM insights into broader ecosystem risks seldom drive action [34,233]. Even when precautionary measures are adopted, enforcement remains a chronic weakness. Widespread IUU fishing, combined with the small-scale, multi-gear character of Mediterranean fisheries, makes compliance costly and inconsistent [24,234]. Weak legal accountability allows authorities to ignore scientific advice with impunity, and because the Precautionary Approach is rarely legally binding, its practical influence in data-poor contexts remains limited [43,235].

6. A Roadmap for 2030

The persistent challenges of assessing and managing data-poor fisheries in the Mediterranean Basin demand a forward-looking roadmap that aligns scientific innovation, participatory governance, and socio-economic resilience. Building on recent advances in DPAMs, this vision emphasizes the integration of emerging technologies, interdisciplinary collaborations, adaptive management, and robust institutional support to shift from reactive to proactive fisheries governance [236,237]. This roadmap is conceptualized as a progression from enhanced data collection to robust management strategies (Figure 8).

Methodological innovation remains central to this transformation. Advanced computational techniques, such as machine learning algorithms integrated into catch-only and length-based models, can extract stronger inferences from limited datasets [238,239,240,241,242,243]. Neural networks and random forests have predicted life-history parameters and stock status in data-poor Indo-Pacific fisheries, offering transferable insights for Mediterranean applications [244]. Hybrid approaches like the SPiCT model, which couples catch data with biomass indices in a state-space framework, show promise for Mediterranean small pelagics [75], while size-spectrum models have proven effective in capturing community-level shifts in overfished systems [245]. Integrating genetic stock identification into enhanced Catch-MSY+ frameworks could further refine assessments for transboundary species such as anchovy [246]. Open-source platforms and electronic monitoring (EM) systems for fully documented fisheries (FDF) will be crucial to ensure transparency and accountability, allowing management to focus on outcomes rather than prescriptive inputs [247,248,249,250].

Furthermore, there is an urgent need to develop and validate non-equilibrium DPAMs that explicitly account for climate-driven shifts in productivity and life-history parameters. Integrating environmental covariates into state-space models (SPiCT) and using MSE to test the robustness of harvest control rules under various climate scenarios are critical steps to ensure assessments remain relevant in a non-stationary Mediterranean.

In addition, while long-standing programs such as MEDITS provide invaluable fishery-independent data, their spatial and temporal limitations, including restricted depth coverage, seasonal sampling, and gear selectivity biases, highlight the need for more adaptive and complementary monitoring frameworks. Future monitoring systems should integrate multi-gear surveys, year-round sensor networks, and targeted deep-water and coastal sampling to capture the full spatiotemporal dynamics of Mediterranean stocks. Coupled with emerging tools like environmental DNA (eDNA) metabarcoding and participatory mapping, such an integrated approach will help overcome the inherent constraints of single-survey programs and support more responsive, ecosystem-based management.

Reducing data poverty also requires investment in scalable data collection infrastructures. EM systems and onboard sensors, successfully applied in New Zealand, can improve compliance and data accuracy for Mediterranean fleets [251]. Adopting FDF approaches with electronic logbooks, as in Danish demersal fisheries, could enable verifiable discard estimates [248]. Complementary citizen science initiatives using mobile applications for length-frequency reporting, modeled on Great Barrier Reef programs, offer cost-effective engagement with small-scale and recreational fishers [252]. Remote sensing, AIS, and VMS integration, supported by machine learning to detect illegal activity, can strengthen transboundary oversight [253], while eDNA sampling provides non-invasive monitoring of abundance and biodiversity [254]. Together, these innovations expand spatial and temporal data coverage, improving assessment precision across jurisdictions. Additionally, the harmonization of multi-gear, multi-national sampling efforts must be prioritized. Inter-calibration initiatives, such as standardized comparative fishing experiments, gear efficiency trials, and cross-border data validation protocols, should be institutionalized under regional bodies like the GFCM. These efforts will ensure that data from diverse national monitoring programs are comparable, scalable, and suitable for integrated stock assessments, ultimately strengthening the scientific basis for Mediterranean fisheries management.

Adaptive evaluation through MSE can ensure that emerging assessment methods translate into effective, risk-based decision-making [35]. MSE frameworks simulate ecological and socio-economic trade-offs, testing the robustness of harvest control rules under uncertainty. Lessons from Australia, the North Pacific, and the Indian Ocean demonstrate their utility in building consensus among stakeholders and identifying resilient management strategies [17,255,256,257,258,259]. Adopting open-source MSE tools such as DLMtool could enable regional bodies like the GFCM to evaluate hybrid approaches combining catch-only assessments, spatial closures, and ecosystem indicators. A structured decision framework should further guide the translation of DPAM outputs into policy. Tiered systems, such as those pioneered by the U.S. Pacific Fishery Management Council, link data availability to graduated precaution, ensuring defensible decisions for all stock categories [260]. Integrating GFCM benchmarks, automated decision trees based on CMSY uncertainty, and risk assessment tools like PSA can help prioritize species for enhanced monitoring [78]. Decision support systems such as FishPath can be localized for Mediterranean case studies, incorporating adaptive harvest rules and real-time EM data [76,261]. To ensure fairness, socio-economic dimensions must be embedded into this framework, drawing on Pacific Island models that balance conservation goals with community livelihoods [262].

To translate the methodological insights and case-based lessons from this review into a coherent and actionable future, it is essential to present a structured framework that connects improved data streams with optimized assessment selection and, ultimately, effective management decisions. The path forward must move from generalized recommendations to a specific, iterative workflow that aligns data collection, analytical capacity, and governance needs.

First, the foundation lies in systematizing and expanding data collection to address the critical gaps identified in Section 2. This involves a tiered monitoring strategy: (1) strengthening and harmonizing core programs like MEDITS and national DCF reporting to ensure consistent, fishery-independent time series for key commercial stocks; (2) deploying targeted, cost-effective methods, such as EM for larger vessels, participatory mapping and logbooks for small-scale fleets, and eDNA metabarcoding for biodiversity and range shifts, to fill spatial and taxonomic blind spots; and (3) institutionalizing the systematic collection of essential life-history parameters through port-sampling programs and research surveys. Data reconstruction efforts should be formally integrated to account for historical depletion and present-day under-reporting, creating more accurate input time series for assessment models.

Second, the assessment selection process must be formalized and matched to data availability within a transparent, tiered decision tree. Building on the principles outlined in Section 3 and the comparative performance reviewed in Section 4, a standardized workflow can guide practitioners. For instance, where only catch time-series exist (Tier 1), catch-only methods with carefully chosen, region-specific priors are the entry point. The addition of representative length-frequency data (Tier 2) triggers the application of length-based methods, while the availability of reliable abundance indices (Tier 3) enables state-space models (SPiCT, AMSY). Crucially, this tiered system should mandate a multi-method cross-validation step wherever multiple data types are available, using consensus-building approaches or model-averaging techniques to reconcile estimates and quantify uncertainty. This workflow ensures that the simplest sufficient method is used, but that additional data are leveraged to increase diagnostic robustness.

Third, the output of assessments must be directly tailored to management information requirements. Management bodies typically require clear advice on stock status (overfished, subject to overfishing) and prescribe specific measures such as catch limits, size regulations, or effort controls. Therefore, assessment outputs should be translated into management-ready formats. This includes generating probabilistic statements about biomass relative to reference points, estimates of sustainable catch and indicators of growth overfishing (SPR). Tools like the Traffic Light Approach or Kobe plots can visually synthesize multi-indicator assessments for diverse stakeholders. Structured decision-support tools like FishPath and the FRDC Stock Assessment Toolbox [77] exemplify how assessment selection can be systematized and tailored to local data conditions. Integrating such platforms into Mediterranean assessment workflows could streamline method choice, improve transparency, and ensure that the selected DPAMs align with both data availability and regional management requirements. Furthermore, to be decision-relevant, assessments must be embedded within an MSE framework. MSE allows for the pre-testing of harvest control rules derived from data-poor assessments against a range of uncertainties, including data quality, model misspecification, and climate-driven non-stationarity, ensuring that management procedures are robust before implementation.

Finally, closing the loop requires strengthening the science-policy interface and institutional capacity. This entails (1) co-developing assessment protocols and management procedures with fishers, managers, and scientists to build trust and legitimacy; (2) investing in regional training and open-source toolkits (DLMtool, TropFishR) to build analytical capacity, particularly in southern and eastern Mediterranean states; and (3) fostering regional data-sharing platforms and standardized reporting under the auspices of the GFCM and EU DCF to ensure assessments for transboundary stocks are consistent and collaborative. Governance structures must formally adopt precautionary, adaptive decision rules that are triggered by assessment outcomes, moving away from negotiation-based decisions and towards evidence-based, pre-agreed responses.

Mediterranean fisheries remain constrained by multi-species complexity, small-scale operations, transboundary governance challenges, and pervasive data limitations. Yet, the convergence of digital monitoring, open data platforms, and participatory MSE offers an unprecedented opportunity to overcome these barriers. Achieving this vision by 2030 will require (1) investing in affordable, interoperable data infrastructures such as EM, eDNA, and citizen science; (2) strengthening regional capacity through GFCM-led training, open-source modeling, and south–north knowledge exchange; (3) embedding socio-economic evaluations into adaptive management design; and (4) institutionalizing MSE-based governance that continually tests and refines management strategies under uncertainty. By following this roadmap, connecting targeted data collection, a transparent and tiered assessment pathway, and decision-making processes designed for uncertainty, Mediterranean fisheries can transition toward resilient, ecosystem-based management that reconciles ecological integrity with socio-economic viability. This approach transforms data poverty from a barrier into a catalyst for innovation, transparency, and long-term sustainability, ultimately enabling the region to implement tested, science-driven, and participatory fisheries governance [236,237].