proliferation of harmful algal blooms (Richard et al., 2022). Krishna et al. 10.3389/fmars.2024.1481734Nutritional composition and thus the commercial value of oysters and mussels decrease under combined acidi?cation and warming stress (Tate et al., 2017). 4.4 Research gaps Effects of different stressor levels on populations in orthogonal experimental designs are dif?cult to examine, particularly when many stressors are involved Griffen et al. (2016). It is costly in terms of time and resources and the inclusion of different stressor intensities may complicate the experimental design (King et al., 2022). We identi?ed only a few studies (only 8%) which investigated the interactive effects of multiple stressors on ecosystem-level dynamics in coastal waters. This has been also noted by Crain et al. (2008) in their review of non-additive effects of human stressors in marine systems, pointing out the existing bias towards studying single species in ex-situ setups. Experimental determination of complex interactions in coastal environments is challenging (Carrier-Belleau et al., 2021), and it is dif?cult to measure stressor effects at the community or ecosystem level in natural settings (Adams, 2003; Elliott and Quintino, 2007; Borja, 2014; Wake, 2019). As the interactive effects at every trophic level vary, depending on factors such as stressor magnitude and exposure duration, measurements of multiple endpoints have to be considered while designing the experiments on an ecosystem scale (King et al., 2022). A ?rst step would start with individual stressor studies across a wide range of intensities to understand responses and then examine combinations of multiple stressors across a smaller range of stressor levels to explore interactions between them. This approach would helpVihtakari et al. (2013) suggested that warming affects larval development, whereas acidi?cation affects the reproductive capacity of adults. A moderate increase in water temperature can counteract the growth effects of reduced pH in M. galloprovincialis by allowing more active feeding time Kroeker et al. (2014), and thus constituting an antagonistic interaction. Mostly, however, the combined effect of OA and warming in bivalves has been observed as negatively synergistic: Thomsen et al. (2013), e.g., showed that heat shock proteins are downregulated under elevated pCO2, amplifying heat stress experienced by Mytilus edulis. Many traits may be affected by the combination of warming and acidi?cation: for both M. edulis and Mytilus galloprovincialis growth rate, shell size, and acid-base buffering capacity were found to decrease (Gazeau et al., 2014; Fitzer et al., 2015). The occurrence of further stressors along with warming and acidi?cation exacerbates the negative synergistic effects. Adding hypoxia impairs the ?tness of marine mussels by reducing the activity of digestive enzymes (Khan et al., 2020). Likewise, an increased frequency of extreme climatic events has been reported to impact bivalve species at 12 coastal regions around the Mediterranean (Rodrigues et al., 2015). Multi-stressor effects on bivalve species propagate to the entire coastal ecosystem via the ecosystem services provided, foremost the habitat creation and water quality improving ?ltration services, but also food provisioning. The presence of bivalves prevents theFrontiers in Marine Science 09to identify critical stressors for multiple stressor experiments. Another dif?culty in studying interactive effects across ecosystems is the variable response time of different taxa to a variety of stressors (Griffen et al., 2016; Turschwell et al., 2022). Furthermore, the time at which stressor response is measured can also affect the classi?cation of the interaction type (Garnier et al., 2017). While studying and quantifying the direct cumulative effects of multiple stressors themselves is challenging, their indirect effects further complicate the problem. This has implications for coastal management, as they do not follow typical cause-and-effect pathways but signi?cant modify ecosystem responses (Adams, 2005; Gladstone-Gallagher et al., 2023). Thus, it has been suggested to focus on both the direct and indirect effects of stressor interactions for effective stressor management (Gladstone-Gallagher et al., 2023). Despite the complications in measuring ecosystem responses to multiple stressors, some efforts have been made in this direction. With the advent of mesocosm experiments, it has become possible to study ecosystem-level responses and effects of climatic and anthropogenic stressors in quasi-natural habitats (Stewart et al., 2013). For example, the mesocosm experiments of Pansch and Hiebenthal (2019) facilitated assessment for a whole range of effects of multiple stressors (temperature, salinity, pH, light) on benthic ecosystems and communities. Given the large number of stressor combinations and species in coastal ecosystems, it will remain unfeasible to fully understand multi-factorial stressor effects by means of observational experiments. Therefore, other approaches such as modelling and expert opinions have been proposed by Halpern et al. (2007); Griffen et al. (2016); Stelzenmüller et al. (2024). For modelling, a 3-tiered approach has been proposed. First, understandmechanistically the stressor effects at the individual level Griffen et al. (2016), and then scale these to population-level responses and, ?nally, assess the risks for communities across ecosystems. This practice has been adopted by some modelling studies (Pörtner, 2012; Cornwall and Eddy, 2015; Queiro?s et al., 2015). Other researchers propose that combining methods can improve understanding such as aligning experiments from the beginning with the models they aim to inform (Hodgson and Halpern, 2019). Although we found that the majority of multi-stressor studies focus on species-level effects, these are largely restricted to phytoplankton and bivalves. Other taxonomic groups, such as zooplankton, ?sh and benthic organisms, which constitute important trophic linkages in the coastal food webs, are underrepresented (Figure 8). It is dif?cult to manage physiologically complex and larger organisms in manipulative experiments and to track variable responses to the same stressors by different species of the same group, e.g. in ?sh. Most of the multi-stressor studies performed on seagrass are in-situ experiments where co-variations in environmental boundary conditions make it hard to unravel non- additive interactions (Stockbridge et al., 2020). As a consequence, we could not identify a trend in reported interactive effects for seagrass or ?sh. Intra-speci?c responses in higher organisms (such as ?sh) vary depending on stressor magnitude and duration (Barton, 2002). Stressor-driven changes at higher trophic levels trigger cascading effects that impact the food web dynamics in coastal ecosystems (Pinnegar et al., 2000; Scheffer et al., 2005; Murphy et al., 2020). The same applies to the coupling of pelagic and benthic ecosystems. In our analysis, we found that only a few studies focused on benthicfrontiersin.org