Fisheries are complex systems that extend beyond fish populations and those who harvest them. Whether situated in freshwater habitats or the deep sea, fisheries function as intricate networks tied to aquatic ecosystems, global economies, and the societies in which they are situated (Garcia and Charles, 2008). Natural resource sustainability and economic viability is often the overall objective when managing fisheries. This begins from a human-based communication process involving various stakeholders, such as communities reliant on fishing-related livelihoods, scientists studying the human impact on marine ecosystems, and regulatory bodies responsible for sustainable management (Dybas, 2002; Armada et al., 2018; Bentley et al., 2019b).

Incorporating the social and economic components of fisheries is key for developing management strategies and fostering informed policy decisions (Costello et al., 2016; Stephenson et al., 2017; Bennett et al., 2021; Ouréns et al., 2022). Thus, explicitly addressing ecological, economic, and social trade-offs is necessary when drafting fisheries regulations, or when assessing their performance. This need is based on the interconnectedness of humans and the environment. On one hand, the fisheries sector is the main source of income for approximately 8% of the global population (Cheung and Sumaila, 2015). Seafood consumption is projected to increase to 182 million tons by 2030 (FAO, 2022). Hence, the increased pressure over these renewable resources and their effects on people’s livelihoods must be properly accounted for policies to be successful at protecting nature, the economy and global food security. On the other hand, fishers respond to economic and behavioral drivers (Russo et al., 2015; Wang et al., 2024). These, however, are dependent on how target stocks: (i) respond to harvesting, or (ii) are influenced by external drivers, such as environmental change or economic shocks, or (iii) react to altered food webs (Heymans et al., 2011; Kaplan and Leonard, 2012; Blenckner et al., 2015; Hodgson et al., 2018; Weijerman et al., 2018; Agnetta et al., 2022; Dowd et al., 2022).

The adoption of a systems, or holistic, approach to managing fisheries is now part of global legislation (Ramírez-Monsalve et al., 2016; Marshall et al., 2018; Haugen et al., 2024). Yet, the breadth of the system differs. In some cases, it is composed of a single target stock and its immediate environment (i.e., ecosystem approach to fisheries, EAF), while others include additional elements, such as multiple target stocks, trophic dynamics, species interactions (i.e., ecosystem-based fisheries management, EBFM), and others even include additional human activities affecting and being affected by fisheries (i.e., ecosystem-based management, EBM) (Patrick and Link, 2015). However, all frameworks recognize the need to comprehensively include socio-economic factors to inform decision-making processes (Stephenson et al., 2017; Marshall et al., 2018; Craig and Link, 2023).

Food web models are key tools for assessing the consequences of environmental and policy changes on fisheries systems (Craig and Link, 2023). They are representations of the complex network of feeding relationships and energy flows among species in an ecological community (Pimm et al., 1991; Plagányi, 2007). Models can be further categorized depending on the level of complexity that is explicitly addressed. End-to-end models aim to simulate the entire ecosystem, from primary producers to top predators, to provide comprehensive insights into ecosystem dynamics and the effects of fisheries management scenarios. These include Ecopath with Ecosim (EwE) (Christensen and Walters, 2004), Atlantis (Fulton et al., 2011; Audzijonyte et al., 2019), Object-oriented Simulator of Marine ecOSystem Exploitation (OSMOSE), Integrated Generic Bay Ecosystem Model (IGBEM), Dynamic Multi-Species Models or Minimum Realistic Models (Plagányi, 2007). Simple trophic models, on the other hand, focus on specific segments of the food web, analyzing only a few predator-prey relationships or trophic levels of interest - such as Models of Intermediate Complexity for Ecosystem assessments (MICE) (Thorson et al., 2019).

Food web models, however, are seldom applied to address environmental, ecological, and socio-economic issues together (Memmott, 2009; Wang et al., 2016; Bossier et al., 2018; Heijboer et al., 2018; Nielsen et al., 2018; Bauer et al., 2019). The literature using EwE, for example, the most widely used food web model in fisheries research (Plagányi, 2007; Bentley et al., 2024), has primarily addressed concerns regarding how ecosystems are affected by fisheries and vice versa. This has been done by simulating changes in fisheries policy and environmental drivers (e.g., sea surface temperature) and estimating their effects on target stocks, their predators, their prey, and the broader ecological community (Colléter et al., 2015; Keramidas et al., 2023). For example, these models have been successful at predicting: the consequences of harvesting prey on predators and non-target groups (Scotti et al., 2022), climate change impacts on fish biomass and energy transfer across the food web (Hodgson et al., 2018; Bauer et al., 2019; Agnetta et al., 2022), and the effects of implementing Marine Protected Areas (MPAs) on protected species (Ramirez et al., 2015). While these models conceptualize fisheries as agents of removal and sources of fishing mortality, a notable gap exists in the literature, with limited exploration into the implications of fisheries on human systems, such as how environmental changes may impact fleet revenue or a nation’s Gross Domestic Product (GDP). Moreover, as models are developed to address a particular question, their representation of the human subsystem is sometimes oversimplified (e.g., using a single fishing fleet as the source of fishing mortality in areas known to have multiple fisheries fleets operating simultaneously; Ramirez et al., 2015). While these models might be addressing important ecological questions, their usefulness for operationalizing EAF, EBFM and EBM require paying even more attention to the interplay between the biophysical and human components of these systems (Stephenson et al., 2017; Fulton, 2021; Craig and Link, 2023; Haugen et al., 2024).

The use of food web models to provide tactical advice for fisheries policy requires broadening the scope of research and how this is communicated to decision-makers and fisheries sector stakeholders (Fulton, 2021; Craig and Link, 2023). The end-users of the models are quite diverse and include government officials, fishers and their direct representatives, local businesses, seafood processors, distributors, wholesalers, and retailers, coastal communities, and seafood consumers – all key stakeholders whose choices and wellbeing are directly influenced by the information provided by ecological models (Schwermer et al., 2020). However, these stakeholders seek insights into the broader consequences of environmental change or fisheries policies on employment, salaries, fish prices, food security, availability of culturally significant resources, broader economic considerations, and gender issues. Addressing these objectives require adopting a language that resonates with their concerns and ensuring a close collaboration between policymakers and scientists (Stephenson et al., 2017; Galland et al., 2018; Craig and Link, 2023; Haugen et al., 2024).

This systematic review focuses on characterizing how food web models are being used to address the social and economic consequences of fisheries policies and environmental change. We first focus on characterizing the development of the field over time. Next, we focus on the questions addressed by the research and the methodological approaches undertaken by modelers. Finally, we discuss our findings within the broader literature, using examples from the reviewed literature to highlight interesting and novel approaches, while also drawing attention to topic areas that need greater consideration in future research.

The current review followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) methodology (Moher et al., 2015). First, we developed a comprehensive research plan to set the boundaries of the analysis. Then we developed a multi-layer search string ( Table 1 ) that was used on both Scopus (www.scopus.com) and Web of Science (www.webofknowledge.com) databases. A preliminary search of these databases resulted in very few papers meeting our criteria before the year 2011. Thus, we decided to focus only on peer-reviewed literature published since 2010. The search strategy (i.e., Scopus, Web of Science, and personal archive) resulted in 506 articles after duplicates were removed.

Next, we proceeded to exclude articles based on the information included in their abstracts. The abstract screening process was implemented using the web-based systematic review platform HubMeta (Steel et al., 2023). All abstracts were screened by at least two researchers. A total of 406 publications were excluded in this stage ( Figure 1 ). The main reasons for the exclusion of articles were that they: (i) did not use socio-economic drivers (model inputs) or indicators (model outputs) (44%), (ii) were not using food web models (19%), or (iii) were not primary sources of information (e.g., literature reviews, book chapters, synthesis reports) (5%). However, almost one third of the excluded articles were false positives (i.e., abstracts using keywords that matched our search string but whose content was not relevant for our research). As some abstracts were not informative, a second level of screening was required. This screening focused on the papers’ methods and results sections. After which only 47 publications were selected for data extraction and analysis ( Figure 1 ).

Data was systematically extracted from each paper covering a series of standardized variables. These included general information about the articles (e.g., the year of publication, the research questions it addressed), the study area (e.g., the countries whose waters were included under the modeled area, the ecosystem type), the model (e.g., the modeling framework being used, the number of functional groups, the number of fishing fleets, the type of model integration procedures), the methodological approach used (e.g., if the study was simulation-based: what were the drivers used? If they were policy driven, was this achieved through input controls? Were fleet-dynamics explicitly addressed by the models? How did they measure the performance of the simulation outputs)?, and whether trade-offs and uncertainty were explicitly addressed and how. This information was the basis for assessing how food web models were being used to address the social and economic consequences of fisheries policies and environmental change. Variable descriptions and the resulting database are available for download ( Supplementary Tables 1, 2 ).

Food web models are great tools for integrating information, scenario testing and trade-off analyses (Fulton, 2021; Craig and Link, 2023; Rodriguez-Perez et al., 2023). These tools are being increasingly used for providing strategic and tactical advice to fisheries managers (Howell et al., 2021; Craig and Link, 2023). In this systematic review, we have found evidence that they were being used for characterizing the socio-economic importance of fisheries systems and simulating the socio-economic consequences of implementing fisheries policies and/or environmental change (including climate change) ( Figure 5 ). Although great papers have been written in this field, food web models in general are not properly integrating the human dimensions of aquatic ecosystems.

Hundreds, if not thousands, of papers using food web models were published across the studied period (2010-2023). Yet only in 47 papers, models included socio-economic drivers or expressed their outputs using socio-economic indicators. This finding is not entirely surprising, as less than 5% of all ocean science literature involves social sciences (Partelow et al., 2023). However, the publication rate remained constant over time. This is a feature not shared by other emerging fields of study using food web models (e.g., offshore renewables), which shows that the one which we have reviewed is somewhat stagnant. As a modeling community it is important to change this.

The human dimensions of aquatic systems really matter. This is not something new, and an issue mentioned by both scientists and decision makers alike (Marshall et al., 2018; Barreto et al., 2020; Bennett et al., 2021; Ramírez-Monsalve et al., 2021; Tittensor et al., 2021; Melbourne-Thomas et al., 2022; Partelow et al., 2023; Rodriguez-Perez et al., 2023). Decision makers, resource users and stakeholders of global aquatic systems want to understand, and anticipate, the consequences of environmental or policy change not only on ecosystem functioning, but also on human well-being (Marshall et al., 2018; Fulton, 2021; Weiskopf et al., 2022; Craig and Link, 2023; Rodriguez-Perez et al., 2023; Thébaud et al., 2023). This requires the explicit consideration of the costs, benefits, and trade-offs over a wide array of ecological, social, economic, cultural, and political objectives at play. Failure to do so can lead to unsuccessful policies. For example, management decisions that do not account for the social and economic costs of their implementation can result in public dissatisfaction, lack of societal acceptance and reduced compliance (Stephenson et al., 2017; Weber et al., 2019; Craig and Link, 2023; Haugen et al., 2024). These consequences further limit management success while generating mistrust and hampering future collaboration and data sharing (among stakeholders and with regulators), as well as diminishing the legitimacy of future interventions. Moreover, these pathways can lead to pervasive, albeit resilient, cycles that foster controversial decisions, reduce policy options, hasten environmental degradation, and perpetuate social injustices, loss of cultural identity, and human rights abuses (Barreto et al., 2020; Bennett et al., 2022; Melbourne-Thomas et al., 2022; Villasante et al., 2022; Craig and Link, 2023; Partelow et al., 2023; Haugen et al., 2024).

Additionally, given that sustainability is the overarching goal of management interventions, it is important to remember that it operates under a triple bottom line (i.e., environmental, financial, and social) that permeates across both the biophysical and human dimensions of aquatic systems (Marshall et al., 2018; Rodriguez-Perez et al., 2023; Thébaud et al., 2023). If we work under the assumption that achieving sustainable fisheries will inevitably lead to social and economic well-being, then we will certainly fail. Positive socio-economic outcomes are not inevitable features of sustainable fisheries (Barclay et al., 2023; Partelow et al., 2023). Trade-offs between fisheries and ecosystem objectives, and conflict between resource users, may still occur even when fishing sustainably (Stephenson et al., 2017; Barclay et al., 2023; Craig and Link, 2023; Haugen et al., 2024). Nonetheless, only two thirds of the papers included in this review explicitly addressed trade-offs among management objectives and outputs.

To better illustrate this point, perhaps we can draw some examples from the reviewed literature. In Peru, research by Christensen et al. (2014) highlighted differences in the per ton contributions of functional group and fishing fleets to the country’s GDP and total employment. Hence, management measures seeking to improve stock status such as reductions in the total allowable catch of resources caught by multiple fleets (e.g., anchoveta or mackerels caught by both industrial and small-scale purse seiners) would impact resource users differently. Thus, it is likely that some stakeholder groups would oppose changes in management targets, that could be beneficial to them in the medium to long-term, if these did not include additional short-term assurances, such as socialized quota allocation among fleets (Walters and Martell, 2004). Similarly, research by Fay et al. (2019) or showed that effort reductions on the Northeast US Shelf caused greater change in employment than in sales, and that the consequences of management actions for individual sectors may be felt disproportionally through the region and across sectors. Finally, research by Celic et al. (2018) showed that the introduction of landing obligations, to reduce discards, in the Mediterranean Sea would have negative consequences on the ecosystem biomass, as well as on the fleets’ marketable landings and revenue. This work highlights that regulations with “good intentions” need to be validated and quantitatively assessed before being implemented as they could lead to unforeseen and unwanted consequences.

More work is required within the modeling community to frame and connect management actions with strategic goals related to social and economic objectives (Stephenson et al., 2017; Barclay et al., 2023; Partelow et al., 2023; Rodriguez-Perez et al., 2023; Thébaud et al., 2023; Haugen et al., 2024). Moreover, model outputs simulating the consequences of environmental or policy change also require to be expressed in meaningful socio-economic indicators for stakeholders and decision makers (e.g., fisher’s income, employment, access to seafood or recreation; Abedin et al., 2019; Bennett et al., 2021, 2022; Melbourne-Thomas et al., 2022; Barclay et al., 2023; Partelow et al., 2023; Thébaud et al., 2023; Haugen et al., 2024). These indicators, although likely sharing many similarities, should be context-specific (Bennett et al., 2021; Melbourne-Thomas et al., 2022; Partelow et al., 2023) and covering more than the most common attributes described in the reviewed papers (i.e., changes in biomass of target and not-target species, and changes in the average long-term yields by functional group and sector). Increased indicator diversity is needed particularly for the economic and social domains ( Figure 6 , Table 3 ). Many important, and viable, indicator candidates for these dimensions are available in the literature and could be linked to model outputs with relative ease (Barreto et al., 2020; Bennett et al., 2021, 2022; Barclay et al., 2023; Partelow et al., 2023; Thébaud et al., 2023). This is a task that requires modelers leaving their comfort zone and embracing interdisciplinarity (Weber et al., 2019; Fulton, 2021; Bennett et al., 2022; Weiskopf et al., 2022; Barclay et al., 2023; Craig and Link, 2023; Partelow et al., 2023; Rodriguez-Perez et al., 2023; Haugen et al., 2024).

Collecting comprehensive socio-economic data is crucial for developing standardized indicators that enable policymakers to monitor fishing activities effectively in both ecological and socio-economic contexts. Standardization can aid informed decision-making (Stephenson et al., 2017; Bennett et al., 2021), however reliance on a standardized set of indicators should not constrain the creativity and flexibility of existing modeling tools. It is important to mainstream their use while adapting models and indicators to specific local and stakeholder needs and contexts, ensuring that they remain relevant to diverse management objectives.

The inclusion of the human dimensions of aquatic systems will not only allow for more transparent debates and management advise, but it will also likely increase the uptake and usage of food web models in decision making processes (Weiskopf et al., 2022; Partelow et al., 2023; Haugen et al., 2024). Yet, this challenge is not homogenous across the globe. Greater efforts are required to increase representation of models for the Global South. Most publications featured in this review covered North American, European, and Australian case studies ( Figure 3 ). Although this is also the case for Atlantis (CSIRO, 2024) and EwE (Ecopath Research and Development Consortium, 2024) models in general, the picture is less skewed towards the Global North when no socio-economic filters are applied. Raising the profile of case studies in the Global South is key, granted that these areas are already disproportionally affected by environmental injustices (Bennett et al., 2023). Moreover, given current needs, limitations, and trends, it is likely that the negative cumulative environmental, social, and economic effects of the blue acceleration, climate change, and local blue growth policies will continue to perpetuate existing injustices in the Global South (Jouffray et al., 2020; Melbourne-Thomas et al., 2022; Bennett et al., 2023).

Models that capture socio-economic dimensions provide policymakers with a more holistic understanding of the potential outcomes of management actions, allowing them to anticipate and mitigate negative impacts on communities reliant on fisheries (Fulton, 2021; Craig and Link, 2023). Trade-offs between ecological and socio-economic objectives can be assessed, supporting the development of inclusive policies objectives that strive for both healthy ecosystems and human communities. Food web modeling can thus play a pivotal role in providing information to strengthen local and global decision-making processes in favor of just, profitable, and sustainable futures (i.e., combating path dependency).

Another issue worth highlighting is the need to enhance the characterization of economic actors in the food web models. The food web modeling community has learnt from experience that there is no one-size-fits-all solution for representing food webs. Some questions require more complex models than others (Fulton, 2021; Craig and Link, 2023). Yet, food web models tend to oversimplify humans within aquatic systems. Our review found that functional groups were consistently represented at a much finer resolution than fishing fleets ( Figure 4 ). In many cases fishing fleets were only seen as sources of fishing mortality, and in the most extreme situations all types of vessels and fishing operations were grouped under a single fishing fleet. This is not a serious representation of fishers or fisheries. Fleets group people whose decisions affect the system and who are affected by changes in the system. Thus, they need to be better characterized in the models so that drivers, trade-offs, and conflicts among them can be better explored (Ulrich et al., 2012; Weber et al., 2019; Schadeberg et al., 2021; Thébaud et al., 2023).

Additionally, human processes should be better represented in food web models – including the feedback between ecosystem dynamics and peoples’ choices or behavior (Tittensor et al., 2021; Melbourne-Thomas et al., 2022; Thébaud et al., 2023). Fishers’ behavior is not only dependent on regulations or profit (Girardin et al., 2017; Schadeberg et al., 2021; Thébaud et al., 2023). However, understanding how changes in profit affect effort dynamics is a key issue which is often overlooked. Very few of the reviewed papers accounted for fishing effort dynamics in their models ( Figure 5 ). Changes in fishing effort drove most models, yet not many considered how the outcomes of a level of effort on a given time step (e.g., the profitability of the fleet on year t) influenced fishing effort (or affected target species selection) in the next time step. Atlantis models can capture some aspects of fleet dynamics (e.g., Ainsworth et al., 2012), yet for food web models developed on EwE, capturing fleet dynamics seriously relies on ensemble modeling efforts (e.g., Dichmont et al., 2013; Lee et al., 2021; Apriesnig et al., 2022; Failler et al., 2022). In both cases, this task requires multiple additional steps for food web modelers, but it adds realism to predictions.

The ensemble modeling approaches (i.e., modeling chains) identified in the present review (those assessing fleet dynamics and beyond) are promising examples of how modeling efforts can help advance interdisciplinarity. Given that economists generally use the types of outputs these models generate, albeit with simpler biophysical model components, their mainstreaming can be a good path to increasing model development and uptake beyond the food web modeling community. Another promising example is EwE’s value chain plugin (Christensen et al., 2011), which is the tool that provides the most diverse set of socio-economic indicators among the works reviewed for this study, while also allowing for feedback between the biophysical and human components of the fisheries system. More case studies involving ensemble bio-economic modeling approaches using food web models at their cores are certainly needed and should be encouraged.

It is important to note, however, that modelers should be cautious of whether they are representing fleets (i.e., groups of vessels of similar characteristics) or métiers (i.e., groups of fishing operations with similar target species, fishing gears, areas, and seasonality) in their models. Although, fleets and métiers can be the same in some situations, the kind of questions that models can address, and the validity of the advice provided based on them, will vary depending on this distinction (Ulrich et al., 2012).

Furthermore, we need to understand how context shapes fishers’ interactions (among them and with nature) (Partelow et al., 2023). This requires modelers to be critical and sincere about which aspects of such contexts are dependent on processes that cannot be adequately predicted (e.g., changes in seafood demand and consumer preferences, technological developments, population growth, future fuel costs, changes in employment across sectors, or collaboration among governments; Pinnegar et al., 2005; Melbourne-Thomas et al., 2022; Thébaud et al., 2023). A possible solution to this challenge can come from co-designing (with stakeholders and regulators) and operationalizing scenarios for standardizing model runs (e.g., based on shared socio-economic pathways; Pinnegar et al., 2005; Kreiss et al., 2020; Hamon et al., 2021). In these exercises model users can develop an explicit shared understanding of their assumptions regarding the interplay of future environmental, social, economic, technological, political, and cultural conditions. Scenario development of this sort has been promoted in the European context through projects such as CERES (2024) and FutureMARES (2024) – the latter directly applying such scenarios on EwE models. More efforts, and financial support, is needed for such undertakings across the globe.

Knowledge co-production is increasingly being promoted in the food web modeling arena given that it leads to better model fits and increased trust about model predictions (Bevilacqua et al., 2016; Bentley et al., 2019a; Fulton, 2021; Craig and Link, 2023). Yet, we found very few examples of such model development among the papers reviewed for this publication (e.g., Armada et al., 2018; Bevilacqua et al., 2019; Alms et al., 2022). However, we also champion model co-development as it can lead to additional benefits when seeking to include socio-economic drivers or expressing model outputs in socio-economic terms. First, it allows modelers to integrate the human dimensions of the system early on the modeling process, potentially reshaping how they represent the system’s ecological dimensions – so models are better at highlighting attributes that are important for a broader user base (Melbourne-Thomas et al., 2022; Barclay et al., 2023). Second, they improve our collective understanding of the socio-ecological system, particularly on how (i) different stakeholder groups operate, (ii) their management objectives, desires, and goals, (iii) the way they measure the success of management actions, and (iv) the thresholds for what they find acceptable or desirable, as well as (v) what information is available and what are the persistent data needs and data collection priorities (Galland et al., 2018; Weber et al., 2019; Barreto et al., 2020; Bennett et al., 2021; Ramírez-Monsalve et al., 2021; Bennett et al., 2022; Melbourne-Thomas et al., 2022; Partelow et al., 2023; Rodriguez-Perez et al., 2023). Third, they can help modelers, resource users and regulators to communicate better, over a shared language, (a) reducing chances of misinterpreting findings or advice (Galland et al., 2018; Weiskopf et al., 2022), (b) diminishing existing biases about the use of food web models for guiding management actions (Weijerman et al., 2018; Fulton, 2021; Craig and Link, 2023), and (c) collectively prioritizing issues that should be at the top of decision makers’ agendas (Galland et al., 2018; Weber et al., 2019; Ramírez-Monsalve et al., 2021; Bennett et al., 2022; Partelow et al., 2023; Rodriguez-Perez et al., 2023; Haugen et al., 2024). Although this is key for ecosystem-based fisheries management initiatives, its relevance might even be greater when using food web models to address ecosystem-based management pursuits – where fisheries (and fishers) are just some of many groups being affected and affecting the system, and where cumulative impact assessments need to be robust and comprehensive on the social and economic fronts (Weber et al., 2019; Bennett et al., 2022; Melbourne-Thomas et al., 2022; Weiskopf et al., 2022; Frazão Santos et al., 2023; Partelow et al., 2023; Haugen et al., 2024).

Finally, the elephant in the room: uncertainty. All models rely on assumptions and are based on incomplete knowledge and imperfect data. Yet, it is uncommon for integrated ecological-economic fisheries models (whether they have a food web model at their core or not) to produce confidence intervals or report uncertainty (Nielsen et al., 2018). In this review we found that more than half of the papers did not explicitly address uncertainty in model parametrization, structure, or outputs ( Figure 7 ).

Not talking about uncertainty is misleading, reduces model uptake in decision-making processes and creates mistrust about model outputs across their broader user base (Weiskopf et al., 2022; Craig and Link, 2023; Rodriguez-Perez et al., 2023; Haugen et al., 2024). Effectively communicating uncertainty is essential for informed decision-making. Transparency regarding the sources and magnitude of uncertainty allows policymakers to assess the risks associated with different management options (Craig and Link, 2023; Weiskopf et al., 2022). Utilizing clear visualizations and explanations can help convey complex information on uncertainty, improving stakeholders’ confidence in model predictions and ensuring decisions account for potential variability in outcomes. We acknowledge that communicating uncertainty effectively to decision-makers and resource users is challenging and requires creativity and out-of-the-box thinking (Galland et al., 2018; Fulton, 2021; Craig and Link, 2023; Rodriguez-Perez et al., 2023; Haugen et al., 2024). However, the papers included in this review have a very different target audience. For academia, being explicit about the sources of uncertainty in our work should be as common practice as the peer review process (Schindler and Hilborn, 2015). There are numerous tools available for addressing parameter uncertainty (e.g., running sensitivity analysis; Bracis et al., 2020; Rovellini et al., 2024), structural uncertainty (e.g., promoting model comparisons; Link et al., 2012; Forrest et al., 2015; Natugonza et al., 2020; Tittensor et al., 2021; Craig and Link, 2023), and implementation error (e.g., assessing the outcomes of proposed management interventions under different implementation success scenarios; Ainsworth et al., 2012; Kaplan et al., 2012; Natugonza et al., 2020). Yet, their use was limited and mostly focused on understanding how uncertainty on the biophysical dimensions of the models (e.g., diets, abundance, sensitivity to temperature) affected model outputs (e.g., non-target group biomass, fisheries yields and revenue). Uncertainty on socio-economic parameters was only addressed explicitly in six papers (Wang et al., 2015; Wiedenmann et al., 2016; Bauer et al., 2019; Sanchirico and Essington, 2021; Apriesnig et al., 2022; Dowd et al., 2022). The effects of misrepresenting the economic components of the models in these works were significant for both ecological and socio-economic outputs. This can be quite problematic for some ensemble modeling initiatives that assume static representations (or stationary parametrizations) of the socio-economic dimensions of their models. This a challenge faced by researchers working with food web models coupled or linked to value chains, input-output models, social accounting matrices or computable general equilibrium models ( Table 2 ). Their socio-economic components are dynamic over time and hence require that multiple models be developed (or that they are updated as new information becomes available). This is best practice for food web models (Fulton, 2021; Weiskopf et al., 2022; Craig and Link, 2023; Haugen et al., 2024) and should also be the case for all components of their ensemble modeling chain.

We have been quite critical on how food web models are being used to characterize and simulate the socio-economic consequences of change in aquatic ecosystems. However, the works reviewed for this contribution are truly great, denote that progress is taking place, and that there is room for continued improvement. We hope that this paper is a call for action. Food web models, and the community of researchers that develop them, must embrace interdisciplinarity to better represent aquatic systems and the uncertainty across all modelled domains and components. Yet, it is important to recognize that there is no “free lunch” when seeking to seriously address the human dimensions of aquatic systems in the modelling arena. Extending data gathering and monitoring programs, training personnel, co-developing models, and actively engaging in multiple stakeholder forums and policy workshops is certainly a costly endeavor. Thus, we are hopeful that this review will also be a call for action to funders to increase their support for marine social science and interdisciplinary research. We strongly believe that interdisciplinarity is the path forward for increasing trust in model outputs, uptake in decision-making processes, and using our capacity to move the dial towards more resilient marine ecosystems and just futures for those who depend on them for their livelihoods and wellbeing.

Ecocentric fisheries management (and ecosystem-based management) is crucial for ensuring healthy aquatic ecosystems. This also includes the humans that depend on the aquatic biota. Food web models are one of the main tools available for representing the consequences of environmental and policy changes on aquatic systems. Although research interest in these models is on the rise, as is their use in decision-making processes, models are still overlooking critical aspects of the human dimensions of systems they are seeking to represent. This systematic review provides an overview of how food web models have been used to date to characterize and simulate the socio-economic consequences of change in aquatic systems. There is a pressing need to improve our understanding of the interlinkages between the biophysical and anthropogenic components of aquatic systems. We hope this review is seen as a call for action and that the food web modeling community rises to the challenge of embracing interdisciplinarity to bridge existing knowledge silos and improve our ability to model aquatic systems across all their domains and components.