Fisheries are increasingly needing to grapple with the need to consider human dimensions and external drivers such as fuel prices and climate change. In response, explicit adaptive approaches such as harvest strategies are being redesigned to strengthen the resilience of feedback management approaches and reduce the need to evoke exceptional circumstances. This can be achieved using Management Strategy Evaluation (MSE) to forecast alternative future scenarios and bound the range of plausible uncertainties. However, this process is also challenged by heterogeneous preferences and aspirations of stakeholders, especially for cross-jurisdictional multi-sector fisheries that are shared between Indigenous owners and non-Indigenous stakeholders. Here, we considered the Torres Strait tropical rock lobster Panulirus ornatus fishery, for which the empirical harvest control rule (eHCR) needed revision to improve resilience to extreme climate events, geopolitical influences on the supply chain and rising fuel prices. Based on consultation with the ‘Group of Rights holders, Advisors, Managers, Partners and Stakeholders’ (GRAMPS), a range of specially-tailored eHCR candidates were developed and tested using MSE, resulting in the successful adoption of a revised rapidly-adaptive eHCR in 2025. The preferred decision rule outputs annual recommended biological catch limits based on medium-term trends in fishery-independent and fishery-dependent data, adjusted by a catch multiplier that is weighted further by the most recent survey index and its associated precision. The success of this process was attributable in part to use of communication strategies to explain highly technical concepts to the GRAMPS to support their evaluation of trade-offs between competing management goals.
Small-scale demersal fisheries in tropical territories support local economies and food security but are difficult to assess because of multispecies catches, diverse fishing practices, and limited biological and fisheries data. As a result, most stocks fall within data-limited categories for which conventional analytical assessments are not feasible. This study applied a productivity–Susceptibility Analysis (PSA) to evaluate the relative vulnerability of 172 demersal stocks (234 species) exploited by artisanal fisheries in five tropical territories: French Guiana, French Antilles, Mayotte, Réunion Island, and French Polynesia. To improve the interpretability of the results, vulnerability thresholds were calibrated using categories from the IUCN Red List as an independent reference for biological risk. The analysis revealed distinct regional differences in the vulnerability of the stocks. Mayotte exhibited the highest vulnerability, driven by both low productivity and high susceptibility, which is potentially linked to limited management measures regulating catch and harvest size. In French Guiana, vulnerability was also high but was mainly associated with low productivity, whereas susceptibility remained close to the regional average. Conversely, the French Antilles showed the lowest susceptibility scores, contributing to an overall lower vulnerability. These results highlight strong regional and taxonomic differences in vulnerability and reveal substantial knowledge gaps in life history and fisheries data. This study demonstrates that context-specific calibration of PSA thresholds can improve vulnerability screening and help prioritise research and precautionary management in tropical, data-limited fisheries.
Yellowfin tuna (Thunnus albacares) are a highly mobile species of commercial importance distributed across tropical to temperate waters. However, most research has focused around their tropical to subtropical extents, with comparatively little known about habitat use and movement patterns at their temperate range extent. Here, we deployed pop-up satellite archival tags on yellowfin tuna off eastern Australia, resulting in 378 data days between 15°S and 40°S. Core thermal habitat was narrow, with the interquartile range spanning only 1.5 °C (18.4–19.9 °C). One tuna was tracked in the tropics (15°S–18°S) and inhabited thermal habitats consistent with previous studies (12.2–29.4 °C; median ± sd. = 26.6 ± 1.9 °C). In contrast, tuna tagged towards their southern latitudinal extent (29.5°S–40°S) occupied substantially cooler temperate habitats (4.6–23.2 °C; median ± sd. = 18.9 ± 1.1 °C), with southern distribution appearing constrained by the extent of the East Australian Current. Alongside thermal tolerance, variable vertical distributions were observed. Tuna mostly occupied the upper ocean (median depth = 25.5 m; 72% of time above 50 m), with only 1% of depth measurements occurring beyond 200 m. Despite this, one tuna was observed to dive to 1252 m, and ‘extreme’ deep dives were observed from 8 of the 13 tagged tuna (62%). These, along with crepuscular movements, were interpreted as potential forage or navigation-driven movements. As Western Boundary Currents strengthen and oceans warm, the temporal persistence of tuna towards their latitudinal extents may increase. Contrasting habitat use across ecological contexts informs adaptive fisheries management to better anticipate how tuna respond to changing environmental conditions.
This study evaluated the relative size selectivity of four trammel nets for black bream (Megalobrama skolkovii) using the Share Each LEngth's Catch Total (SELECT) method and a generalized additive model (GAM) framework. The mesh sizes (knot-to-knot, mean ± SE) of the inner layers were 2.85 ± 0.002, 6.27 ± 0.008, 9.79 ± 0.018, and 14.06 ± 0.034 cm, corresponding to hanging ratios of 0.49, 0.33, 0.23 and 0.22, respectively. Initial SELECT analysis identified the bilognormal model as the best relative fit among candidates but showed lack of fit and overdispersion. Consequently, GAM models were adopted as the primary analysis tool. Maximum retention occurred at a length-to-mesh ratio (l/m) of ∼ 3.5. The curve similarity of GAM models before and after inclusion of zero-catch length classes suggested that the estimated selectivity pattern was stable. The relative fishing power had little effect on the overall shape of the selectivity curve. Because mesh size and hanging ratio were perfectly confounded in this study, the estimated selectivity curves represented specific gear configurations (fixed mesh-size/hanging-ratio packages) rather than mesh size alone. The population distribution demonstrated consistent bimodal structure (modes at ∼20 cm and ∼48 cm) under different fishing power assumptions. The encountered length structure should be interpreted as a combined outcome of gear selectivity and fishing power, and these factors should be disentangled to avoid systematic bias when doing population assessments.
Achieving ecosystem-based fisheries management (EBFM), which seeks to optimise fisheries catches whilst maintaining the structure and functioning of ecosystems, is becoming more challenging in the current context associated with budget cuts and the restructuring of many research and management institutions. As such, more consideration could be given to the ecosystem approach to fisheries management (EAFM) which focuses on individual fisheries, to build strong foundations for EBFM. Here, we report an approach we developed to identify relevant tools to assist EAFM in New Zealand, towards EBFM. We present it and demonstrate it through an application to the west coast South Island (WCSI). Our application includes principal component analyses (PCAs) and heatmaps on landing compositions from the WCSI mixed bottom trawl fishery and research survey biomasses. Analyses indicated that the last 34 years can be divided into four clusters, starting with a period where the biomasses and landings of flatfishes and associated bycatch species were particularly high and ending with a period where the biomasses and landings of the commercially valuable snapper (Chrysophrys auratus) and red gurnard (Chelidonichthys kumu) increased markedly. Using the literature and knowledge from resource managers and stakeholders, we were able to interpret the clusters of years in light of changes from natural or anthropogenic causes. For example, the increased snapper biomass and landings in the most recent years may be partly explained by warming and more regular marine heatwaves that benefited the snapper population. The tools we identified will help progress EAFM in New Zealand and can support current management in several ways. In particular, user-friendly visual analyses based on PCAs and heatmaps and métier analysis can contribute to a better framework for defining fisheries, by delivering an understanding of catch composition as an emergent property of fisheries. We recommend further applications of the tools we identified in New Zealand, and our approach can also be adapted in other world regions.
Efficient fisheries management relies on understanding population structure and dynamics of the target species. This study investigated the population structure of the Atlantic chub mackerel (Scomber colias), an economically important pelagic species widely exploited across the northeastern Atlantic and Mediterranean, but with no formally defined stock units. The currently proposed fisheries management units suggest either a single unit or two units for the Bay of Biscay and Iberian waters, with a potential division between European and African waters, but lack formal comparison to assess biological consistency. Samples from eight Atlantic locations, ranging from the Cantabrian Sea to central Morocco, and three Mediterranean locations were compared through otolith shape analysis, body morphometry, and gill raker counts. The results from different methods broadly aligned with current recommendations for regional fisheries management, although each method captured population differences at distinct temporal scales. Body morphometric differences suggested higher spatial heterogeneity in Atlantic waters, distinguishing among Cantabrian, Iberian (west and south), African (central Morocco and Canary Islands), and Atlantic islands (Madeira and Azores) samples. Further, differences were found between Atlantic and Mediterranean waters, as well as between western and eastern Mediterranean samples. Gill raker counts further suggested potential distinct source populations in the Cantabrian Sea, Iberian waters (including the western Mediterranean), west African waters (extending to all Atlantic islands), and the eastern Mediterranean. High inter-individual variability in S. colias otolith shape limited the power to detect spatial differences and highlighted the need for increased sampling effort for future studies.
Many recreationally harvested demersal fishes inhabit high-relief, rocky habitat, and lack comprehensive fishery-independent surveys to inform stock assessments and management. In these situations, stock assessments often rely on fishery-dependent surveys to track patterns in abundance. To address these challenges and align methods more closely with the stock assessment process, we develop novel methods to account for changes in relative abundance, using vermilion rockfish (Sebastes miniatus) as a case study. We develop a habitat-weighted time series of relative abundance using maps of high-relief habitat, coupled with fine-scale catch-per-unit-effort (CPUE) data from a recreational fishery survey. We compare this approach to two alternative methods that implicitly assume equal area weights, or disregard regional abundance trends altogether, both commonly used in the absence of habitat information. We also describe a habitat-weighted, spatial index of abundance for use in allocating catch across management areas and compare it to three other methods of catch allocation. We propose that in the absence of fishery-independent surveys, habitat-weighted, fishery-dependent CPUE indices can better inform spatial and temporal indices of relative abundance, relative to methods in which habitat data are lacking. In our case study, differences among methods for catch allocation based on spatial indices of abundance were greater than differences observed for temporal indices. This result may differ among species and regions. Our work demonstrates that when only fishery-dependent data are available, the use of CPUE as a proxy for density (with habitat area weights) is more appropriate than the assumption that CPUE alone is proportional to abundance.
Estuaries and adjacent marine habitats play an important role in the migrations of Pacific salmon (Oncorhynchus spp.), but these ecosystems are increasingly affected by coastal development, potentially altering salmon behaviour. The behaviour of adult salmon in these habitats prior to upriver migrations remains understudied and represents a key knowledge gap for conservation. Cowichan Bay, British Columbia, Canada is a critical habitat for Chinook salmon (Oncorhynchus tshawytscha), with traditional knowledge suggesting that salmon stage in the bay during the fall. Adult Chinook salmon were acoustically tagged over three years, and an array of receivers was deployed in Cowichan Bay. Tag detections, depth, and activity levels of adult Chinook salmon were modelled against environmental conditions to understand behaviour and movement patterns of staging salmon before upriver migration. Chinook salmon exhibited strong diel behaviour patterns, with adult salmon moving into Cowichan Bay, staging near the estuary, remaining at shallower depths, and being less active during the day. At night, salmon moved out of the bay, occupied deeper depths, and became more active. This study revealed pronounced diel and tidal patterns in Chinook salmon behaviour which may be influenced by predation avoidance. Understanding such patterns is essential for effective conservation and could provide insights into bottlenecks to the survival of Chinook salmon.
Annual acoustic-trawl (AT) surveys are conducted to manage the harvest of small pelagic fishes (SPF) off Northwest Mexico. The AT method uses species proportions and fish-length distributions from trawl catches to convert echosounder data to estimates of SPF biomasses. Here, we consider the effects of SPF diel behaviors on the accuracy and efficiency of AT sampling. We analyze echosounder data and trawl catches collected irrespective of the time-of-day during four AT surveys between 2019 and 2021. We use a multi-frequency algorithm to identify acoustic backscatter from schooling fish with swimbladders and attribute it to SPF. We compare observations of the putative SPF backscatter versus depth and time-of-day. We assume that SPF backscatter is less biased by the sea-surface ‘blind zone’ if it is deeper, and less biased by non-SPF backscatter if nearby catches have less bycatch. We also assume that trawl catches are less biased if they have higher SPF-species richness and larger length ranges. We compare these metrics across diel periods. Furthermore, geostatistical methods were applied to the Nautical Area Scattering Coefficient in order to demonstrate that vertical migration constitutes a source of uncertainty. During daytime, SPF aggregate deeply enough to be sampled by the echosounders and, during nighttime, they ascend near the surface and disperse for efficient capture. During daytime, SPF and bycatch are separated, and fewer non-SPF contribute to the catch and presumably the SPF backscatter, compared to nighttime. However, at nighttime, SPF catches are generally larger and are inferred to more accurately represent the species composition and length distribution of SPF in the area, compared to daytime catches. We conclude that, at least for the studied area and times, echosounder transects should be conducted during daytime and trawling should be conducted at nighttime to improve AT-survey efficiency and reduce uncertainty in estimates of SPF biomasses. To confirm if these conclusions are generally valid, similar analyses should be conducted in other regions.
Monitoring programs are often faced with a decision to allocate resources into either robust spatiotemporal coverage to estimate a population index (e.g., not true abundance) or confirming closed sampling conditions (e.g., with block nets) for an unbiased population estimate at the cost of spatiotemporal coverage. However, making accurate and precise abundance estimates at robust spatiotemporal scales is possible when combining open and closed sampling designs with integrated modeling techniques. We used simulations and a case study of backpack electrofishing surveys in the Santa Ana River, California to test the efficacy of an integrated abundance model (temporary emigration model, TE) to estimate abundance of fishes using removal sampling methods with a hybrid sampling design (sampling with and without block nets during removal sampling). We found that the TE model performed well under most modeling scenarios (sample size, amount of closure violation, number of samples collected during closure), although at least a few samples with block nets were necessary for all parameters to be estimable. When applied to fish surveys in the Santa Ana River, we found that catch of the fishes fit to the TE model (Santa Ana Sucker, Arroyo Chub, Channel Catfish, Largemouth Bass, Yellow Bullhead) showed little evidence that the closure assumption was violated when block nets were not used. Additionally, we found that the abundance of non-native fishes negatively affected the abundance of the native Santa Ana Sucker, which was also found to adversely affect the native fish’s access to critical habitat consisting of gravel and cobble substrate. Our results indicate that the TE model presents a viable solution to common sampling problems that impact many monitoring programs, where precise and accurate population estimates can be made at large spatiotemporal scales even when most samples violate the closure assumption.
The Mediterranean Sea, one of the most sensitive regions to global climate change, has experienced notable fluctuations in the life history traits of marine fish, which in turn affect population dynamics. Among these, small pelagic fish such as the European sardine (Sardina pilchardus) are particularly responsive due to their short life cycles and sensitivity to environmental variability. This study investigates temporal variations in key life history traits of S. pilchardus in the Southern Alboran Sea, using data collected from four Moroccan ports over two periods (1999–2003 and 2012–2023). In total, 20,383 individuals were sampled, and analyses included length–weight relationships, gonadosomatic (GSI) and body condition (Kn) indices, size classes, maximum length, and length at first maturity (L₅₀). Length–weight relationships remained strongly correlated (r = 0.92–0.94), with males exhibiting slightly negative allometric growth in 1999–2003 (b = 2.96) and both sexes showing positive allometry in 2012–2023 (b > 3). Mean total length declined for females (158.23–154.16 mm) and males (158.23–154.15 mm), with maximum lengths decreasing from 199.56 mm to 190.93 mm in females and 187.61 mm to 183.55 mm in males. The population showed a persistent female bias, with annual F:M ratios increasing from 1.17 to 1.87, while small size classes (<100 mm) disappeared in the later period. Female L₅₀ declined significantly (143.1 ± 11.3 mm to 128.2 ± 15.3 mm, slope = –1.22 mm•year⁻¹, p = 0.006), whereas males showed no consistent trend. Reproductive activity intensified and slightly extended in Period 2, with spawning females present across most months, coinciding with decreased Kn in females and a minor improvement in males. These descriptive findings highlight substantial interannual, seasonal, and sex-specific variations in sardine life history traits, providing a baseline for understanding population resilience and informing future assessments of ecological and fisheries dynamics in the context of ongoing environmental change in the Mediterranean Sea.
Improving fishery eco-efficiency is essential for balancing fishery production with resource conservation and environmental sustainability. Although China has made notable progress in promoting environmentally sustainable fisheries, significant regional disparities in fishery eco-efficiency persist. Taking technological heterogeneity into account, this study evaluates the eco-efficiency of fisheries across 27 Chinese provinces from 2011 to 2023 under both group-specific and national technology frontiers, and further examines the dynamic evolution and spatial disparities of eco-efficiency. The results show that fishery eco-efficiency in China improved steadily during the study period, although regional disparities remained evident. Under the meta-frontier framework, eco-efficiency generally exhibited a decreasing pattern from eastern to central and western China. Under the group-frontier framework, however, the western region showed relatively higher efficiency levels, while the gap between the eastern and central regions gradually narrowed. The technology gap ratio (TGR) results indicate that the eastern and central regions remained close to the national technology frontier, with the central region showing a particularly clear catch-up trend, whereas the western region continued to lag behind. Convergence analysis further reveals significant catch-up effects at both the national level and in the central region. By contrast, the western region only showed strong convergence under the group-frontier framework, while the efficiency pattern in the eastern region became relatively stable over time. In addition, the decomposition of regional disparities shows that intra-regional differences consistently accounted for more than 70% of the total variation, suggesting that internal imbalance has become a major constraint on the coordinated green development of fisheries. The findings suggest that improvements in fishery eco-efficiency depend not only on overall technological progress, but also on the extent to which regional production systems match local technological conditions. Accordingly, differentiated development strategies should be adopted based on regional resource endowments and technological foundations. In particular, greater attention should be given to strengthening green technology diffusion, improving regional coordination, and promoting low-carbon fishery development suited to local conditions, so as to support the sustainable transformation of China’s fishery sector.
The gillnet sector of the New South Wales Estuary General commercial fishery uses a variety of fishing methods and gear specifications. These details, as well as rates of discarding, are only recorded during occasional scientific observer surveys. Having incomplete information on fishery structure creates challenges to the monitoring and management of this fishery. In this study we present an analysis which expands the partial observer data to fleet-wide estimates for two key variables: fishing method and discard rate of a key target species (mulloway, Argyrosomus japonicus). This involved fitting a set of four related models and using sequential prediction of logbook data to create fleet-wide estimates. A Bayesian framework was used to ensure uncertainty was accurately propagated between models. We were able to provide useful fleet-wide estimates for fishing method with their expected seasonality, as well as a time series of total mulloway discards. Predicted discard rates were often large but with considerable uncertainty (mean 3.3 kg/trip, 0.07–13.3 95% CI) and varying by fishing method and mesh size. This discarding time series will be especially useful for management processes such as integrated stock assessment. Of key importance was the quantification of uncertainty, which was generally large due to natural variability and related parameter uncertainty. The implication of this uncertainty is that 1) more detailed self-reporting (especially of fishing method) will enhance model outcomes; and 2) regular monitoring (with representative estuary-level coverage) is likely essential to ensure changes in the fishery and environment can be accounted for in future model predictions.
