Abstract Sivel et al. (2025) provided a review of peer-reviewed literature focused on food-web modeling and highlighted the many important contributions from research focused on the Northeastern US Shelf. Link (2025) raised several points related to our criteria of model evaluation and exclusive use of peer-reviewed literature, among others. We defend our use of peer-reviewed literature, clear search methodology, and our definition of model evaluation in determining whether a model application was validated. Our team that was led by early-career scientists found areas of agreement with Link (2025) but strongly disagree that the involvement of senior scientists or “experts” from outside the author team is needed in conducting analyses such as those presented in Sivel et al. (2025).

Abstract This essay outlines a personal scientific journey in molecular ecology and evolution, with particular focus on marine fish and fisheries genetics. Along the way, opportunities were taken, challenges navigated, and various lessons learnt. From early studies of counting protein variants on gels (allozymes) to the adoption of high throughput genomics, personal research highlights are described, with some reference to wider advances in the field. Developments in high throughput sequencing technology enabled progressively more detailed analysis of population structure, connectivity, temporal and spatial trends in marine biodiversity, and exploration of gene function. While the focus here is on marine fishes, emphasis is placed on the value of a comparative evolutionary approach, with study of taxa differing in lifestyle, distribution and ecology. An optimistic tone is now possible linking science to policy, including for example, contributions of genetic tools to mixed-stock fisheries and real time management, forecasting the impacts of environmental change on marine biodiversity and resilience, and traceability of fish populations and products. Importantly, the benefits of genomic tools are quantifiable, both in terms of promoting sustainability of exploited resources, and impacts on economic gain from catches. Various underlying drivers underpinning translation of genetic and genomic outputs into policy are identified, including salient technological advances, increased communication between scientists and policy makers, and the escalating urgent need to monitor and forecast impacts of environmental change on marine biodiversity. Based on personal experience the aim here is to highlight facilitators and strategies to navigate the often-conflicting demands of academia and scientific research.

Abstract Regenerative food systems, such as restorative aquaculture, are an approach to food production that can support ecosystem function and repair while meeting demands for protein and nutrition. However, it is not well understood whether these approaches are preferred by industry and supporting organizations, and, if so, whether they could become more widespread in the future. Using the United States of America marine aquaculture industry as a case study, we undertook a collaborative, mixed-methods scenario analysis to understand two potential futures for restorative aquaculture in 2035: the perceived business-as-usual scenario and a preferred-and-plausible scenario. We conducted an online survey, collecting 100 individual responses from industry, academia, Indigenous community organizations, government, and extension services. We found many respondents expected restorative aquaculture to provide co-benefits to a range of user groups, and the industry was positive about prioritizing restorative approaches in future expansion; 67.3% of industry respondents indicated they would give preference to restorative species and practices if expanding operations. To develop the two scenarios, we analysed these survey results through three regional workshops with stakeholders. The workshops confirmed interest and optimism for a future restorative aquaculture industry, but that constraints to this future exist and the preferred state is unlikely to be achieved without effective policies and supporting mechanisms. Active intervention across investment, research, regulation, market development, training, and Indigenous knowledge and leadership, is needed. Promising interventions include strengthened partnerships with government agencies to build more enabling conditions, science partnerships to build evidence of ecological and social co-benefits, and greater industry coordination to harness existing leadership and pro-environmental commitments in the early stages of development.

Abstract Sivel et al. reviewed ecosystem models applied in the Northeast US (NEUS). Their conclusions and recommendations have significant implications for the use and adoption of ecosystem modeling in the NEUS, elsewhere in the USA, and around the world, especially in the context of operational management of living marine resources. Sivel et al. claim that NEUS ecosystem modeling (1) has no standards for evaluation, (2) largely has not considered climate change, (3) has not largely considered or been used in management strategy evaluations, (4) has had modelers in the NEUS that mostly do not communicate with ecosystem modelers in other regions, (5) has not updated or been cognizant of ecological theory as seen in outputted indicators, and (6) has not been used to advance ecosystem-based (fisheries) management nor that any advancement of EBM is occurring. Their conclusions are based on incomplete information and, as a result, are inaccurate. Their recommendations are general or self-evident enough such that no one would disagree with them, but are also dated in that they miss that such recommendations have already been made (often decades earlier) and are already being adopted. I refute all of this with factual evidence that was not included in their review. Because these ecosystem models are used in an operational context, the implications that these models were not adequately built, reviewed, or used undermines confidence in their application. I conclude with a few other lessons for the broader marine science community. I share the authors’ goal to increase the potential move of NEUS, US, and global fisheries management towards ecosystem-based management frameworks, and trust that this commentary provides a more complete status of the state of NEUS (and broader) ecosystem modeling employed operationally for living marine resource management. I provide the comments herein not to discourage Sivel et al., but rather (1) to encourage them and the entire community to consider some of the broader things that can be commonly missed in reviews like what they presented, and (2) to ensure that misleading messages that can be damaging to the ecosystem modeling enterprise are corrected.

Abstract A complete bomb radiocarbon (14C) chronology — covering an 80-year period (1938–2018) and reflecting thermonuclear testing in the 1950s and 1960s — was established using known-date otolith material to trace the signal in the Baltic Sea and to establish a tool in validating the age of regional fishes. Of particular interest were Baltic herring (Clupea harengus) and sprat (Sprattus sprattus) that are estimated to live >20 years in the Baltic Sea, a longevity that is ∼2× greater than determined for other locations. The new 14C chronology was used to validate age estimates up to 22 years for herring and 18 years for sprat, although there was evidence of 14C emissions from nuclear facilities, as well as deficiencies from other environmental factors. This work confirmed the accuracy of high-confidence age reading protocols used for decades and is a first step in establishing a valid basis for observed growth differences between northern-southern Baltic Sea populations. An exploration of 14C in otoliths of coastal Baltic pike (Esox lucius), pikeperch (Sander lucioperca), and Baltic cod (Gadus morhua) lend support for broader use of bomb 14C dating on marine and freshwater organisms of the Baltic Sea basin.

Abstract The size selectivity of gillnets can be estimated using the catch data from experimental fishing of gangs of net panels having different mesh sizes. Gillnet selectivity curves can take a wide variety of shapes, and currently their estimation requires consideration of several different parametric forms, with right-skewed or bimodal curves typically being preferred. Here it is shown that the generalized additive model (GAM) framework provides a convenient and more flexible alternative. The GAM approach also generalizes the scope of analysis by permitting the population length frequencies of encounter to be jointly estimated as a smooth function of length. Moreover, the GAM framework allows for the inclusion of covariates such as sex or a condition index, and accommodates hierarchical sampling designs and spatial or temporal effects. Relative fishing power of the different sized meshes can also be included, notwithstanding that care with non-identifiability is required. The ease of use of the GAM approach is demonstrated on previously published lake trout data.

Abstract Offshore wind energy production is on the rise globally, projected to occupy significant areas in shallow shelf seas and moving into deeper waters as floating turbine technology is becoming more mature. However, knowledge about the potential impact of wind farms on the physical oceanography and lower trophic organisms is still severely limited. In this review, we assess the current state of knowledge on the effects and impacts of offshore wind farms on regional and local hydrography and circulation, nutrient distribution, phytoplankton and primary production, and sediment load in the water column during the operational phase of the wind farms and identify critical knowledge gaps. The body of literature on the topic has grown rapidly over the last years, but most studies focus on wind farms in relatively shallow (<60 m water depth) and mainly unstratified or seasonally stratified shelf seas, predominantly on the northern European shelf and around China. In situ observations are scarce, leading to heavy reliance on numerical models. As floating wind farms have become operational only very recently, few studies focus directly on their specific impacts. There is general understanding of local impacts on ocean physics, e.g. on turbulence, mixing and stratification due to flow past turbine foundations, or the potential of wind wake impacts on surface currents, up- and downwelling. Consequences for phytoplankton and primary production are much less clear and both physical and biogeochemical impacts on regional scales remain uncertain. There is a critical need for observational data for validation and targeted impact studies. Particularly characteristics and temporal and spatial scales of circulation and hydrographic changes and their effects and impacts on primary producers, vertical flux, and pelagic-benthic coupling are little understood especially in stratified and deep shelf regions. Given the rapidly accelerating growth of the offshore wind farm industry and expansion into deeper seas using floating technology, addressing these knowledge gaps is crucial for reliable environmental impact assessments and sustainable development of this still relatively new energy sector.

Abstract Accurate age estimation is an important factor in understanding the temporal ecology, life history, and population structure of marine mammal species. The presence of annual growth layer groups (GLGs), bands of differing densities within teeth, has been documented and studied for decades as a method of age estimation. However, the methodology is still associated with complexities and uncertainties that challenge practical applications. Characteristics of GLGs can vary by species, and interpretation of GLG boundaries can vary by person. These complications are especially notable for elusive species, which are underrepresented in dentin studies. Here, we examine a dataset of ∼80 killer whale (Orcinus orca) teeth within an established framework to improve empirical understanding of GLG growth and identification. This is, to our knowledge, the most comprehensive dataset for this species and the first to reference known-age individuals. We describe characteristics of killer whale GLGs, including GLG width, presence of accessory lines, and possible distortion from pulp stones (i.e. calcified masses). We also compared our own age/GLG estimations with those of numerous alternate readers who have studied these teeth in the past, finding potentially large discrepancies (0 to 38 years). To achieve more accurate age estimations and better understand the discrepancies, we were able to examine teeth from wild, age-referenced killer whales. In addition to analysing the GLGs, we also calculated the pulp:tooth area ratio as another measure through which to understand age, as the pulp cavity gradually fills with dentin as the animal ages. We found support for an exponential relationship, and we found that killer whale ecotype is a factor for consideration, since pulp:tooth area ratios were consistently lower for residents than for transients. Our study provides a modern foundation for age estimation of killer whales, an elusive species of increasing conservation concern in some regions. We also provide a structured and empirical framework for better understanding age estimation via dentinal GLGs in any marine mammal species.

Abstract DNA metabarcoding of plankton samples is a cost-effective approach to analyze the richness and composition of zooplankton in coastal waters. Plankton surveys using metabarcoding can also be used to monitor species of concern, both for planktonic species and benthic species with planktonic larvae. We used DNA metabarcoding with the mitochondrial cytochrome oxidase I gene to: (1) evaluate time of sampling (year, month, week), tidal cycle, time of day, and sampling location as potential sources of variation for zooplankton community richness and composition, and (2) as an early detection tool for species of concern in Prince William Sound. We found that seasonality of sampling had the strongest impact on species richness and community composition. Differences in community composition were mostly driven by differences among meroplankton groups, with the peak in meroplankton abundance occurring in April and May, depending on the location. We detected eight potentially introduced species, three holoplanktonic (Limnoithona tetraspina, Oithona davisae, Pseudodiaptomus marinus), one benthic (Monocorophium acherusicum), and four meroplanktonic species (Amphibalanus improvisus, Philine auriformis, Alitta succinea, Marenzelleria neglecta). Most of these potentially introduced species detections occurred in a single sample with low read counts; but Oithona davisae was detected in three locations over two years. We discuss confidence in the genetic ID and invasion history for each of the species of concern. We showed that (1) plankton surveys can be a broad monitoring tool for species of concern, (2) timing of sampling can be critical depending on the organisms’ life histories, and (3) sampling during the meroplankton peak concentration can increase the chance of detecting larval stages of introduced benthic species. Finally, we provide some bioinformatic recommendations to improve species detection and validate identifications.

Abstract In the Northeast Atlantic, the International Council for the Exploration of the Sea (ICES) provides scientific advice under the precautionary approach (PA) and maximum sustainable yield (MSY) principles. FMSY, the exploitation level that achieves MSY in the long term, is derived in ICES through stochastic simulations with the software EQSIM for most of the stocks. The computed FMSY is then conditioned on the PA such that fishing at FMSY does not exceed a 5% probability of the spawning stock biomass falling below the limit reference point Blim. We compared reference points estimated using EQSIM and a short-cut management strategy evaluation (MSE) tool (RPETool), which approximates a full-feedback control loop, and, differently from EQSIM, mimics the stock assessment advice process, including all data lags, and retains the original structure of the assessment model. Here, we showed that the simplifications of the management system and the assessment model, which are necessary for conducting simulations with EQSIM, result in the breaching of the PA for 3 of 4 recently benchmarked stocks. For the only stock for which the PA was not violated, the EQSIM estimate of FMSY was still larger than the actual FMSY estimated within the assessment model, and long-term yields were not maximized. Considering that EQSIM has been used since 2017 to derive reference points in ICES for more than 75% of the data-rich stocks, it is urgent that it is phased out and substituted by a more appropriate approach, like RPETool. Furthermore, the analysis highlights the asymmetry between increased risk of breaching Blim when fishing at or above FMSY and low risk in terms of the minimal loss in long-term yield when fishing below FMSY, with <5% yield loss even if F were reduced to around 60% of FMSY. Considering the wide ranges of additional uncertainties about the assessment model and the associated FMSY estimate, we propose that it is time for a paradigm shift within ICES to advocate for FMSY being the upper limit for advice on fishing opportunities rather than the target.