Explicitly Closing and Quantifying Feedback Loops in Aquatic Environments Should Be a Priority

The Earth system is complex, with reciprocal interactions among its chemical, physical, and biological components which create cycles of cause and effect. While aquatic ecologists have long focused on how the environment shapes organisms, ecosystems, and biological function, less attention has historically been paid to how biology shapes its environment across local, regional, and global scales (Pausas and Bond 2022). Yet, in this modern moment of global change, it is imperative that we work to understand how these causal cycles, termed “feedbacks,” impact both aquatic ecosystems and Earth's climate. Ecological processes that have a clear unidirectionality—for example, solar radiation increases surface water temperature, or heat loss of warm-blooded animals to their surrounding environment—are the exceptions. Rather, ecological processes in which causes and consequences influence each other, that is, feedback loops, are the norm (Pichon et al. 2024); it is thus critical to study them as circular, interconnected processes. Feedbacks arise when the effect of a change in one system component causes an effect in another component, which in turn either dampens (stabilizes) or reinforces the initial change (Selden et al. 2024). While stabilizing feedback loops contribute to resilience (Mastrangelo and Cumming 2024), reinforcing feedback loops sustain processes within a system and can erode resilience. For example, the biological carbon pump modulates the amount of carbon in the atmosphere which stabilizes the climate on biological timescales (Henson et al. 2022), thus favoring stable conditions under which phyto- and zooplankton thrive. On the other hand, anoxia in lake bottom waters creates self-sustaining conditions (Carey et al. 2022; Lewis et al. 2024) that can modify food webs by altering biota behavior and increasing mortality (Townsend and Edwards 2003; Doubek et al. 2018). Biogeochemical feedbacks are a subset of feedbacks that act specifically on biogeochemical cycles. For example, sudden loss of transparency in lakes through heavy rainfall can create anoxic conditions that drive the release of dissolved organic matter from sediments, which in turn stimulate respiration rates of heterotrophic bacteria, thus exerting internal pressure to maintain the lake ecosystem under anoxic conditions (Brothers et al. 2014). In this example, the event-driven sudden loss of transparency represents the external stimulus or “forcing” which initiates the reinforcing feedback loop driving anoxia. Some feedback loops, such as the impacts of the biological carbon pump on carbon sequestration, are well-studied and consequently represented in conceptual and computational models, though often with significant uncertainties caused by eco-evolutionary feedbacks that interact with biogeochemistry (e.g., Tagliabue et al. 2021; Henson et al. 2022). However, many, if not most, feedbacks remain poorly understood and thus unconstrained (Selden et al. 2024). Given the role of feedbacks in promoting or eroding ecosystem resilience (see biological carbon pump and lake anoxia examples above), with potential consequences on local to global scales, we believe that better understanding interactive effects between distinct and nested feedback loops (Fig. 1) should be a priority for limnologists and oceanographers alike. Briefly, we define nested loops as feedbacks that act on one or multiple nodes of the main feedback loop (Fig. 1B orange arrows) and distinct loops as feedbacks that share the initial change but in which ecological mechanisms are different (Fig. 1C,E, dark cyan and blue arrows respectively). Indeed, interactive effects modulate the magnitude of the main feedback and represent large sources of uncertainties in models (Henson et al. 2022). Quantitatively addressing interactive effects in feedback loops is thus required to provide meaningful climate and biogeochemical projections under changing environmental conditions. For example, microbial methane production and consumption are both temperature dependent, but production increases faster than consumption (Bastviken 2009; Thottathil et al. 2019). Yet, how the methane budget will change with global warming is still unconstrained due to interactive effects, such as substrate availability, that can alter microbial production and consumption of methane differently (Selden et al. 2024). As such, we encourage our community to adopt a systems-thinking approach and to consider how feedbacks may play a role in their own research. Using a systems-thinking approach aims to contextualize a particular biogeochemical process within the broader system. In doing so, it can bring insights into how seemingly disparate phenomena may influence each other and constitute unforeseen feedback loops. Indeed, feedback loops often link components within a system that are traditionally studied in different scientific fields, such as ecology, biogeochemistry, eco-evolutionary, and social sciences. Interdisciplinary teams may thus help study feedbacks in a cohesive rather than fragmented way, providing critical insights into how feedback loops operate within the broader system. Feedback loops are generally complicated to formalize as there are many interacting components that may change both spatially and temporally. However, considering the spatial and temporal scales of feedback loops is important for understanding the processes at work. For example, on relatively short time scales, permafrost thaw creates conditions that further amplify thawing through increased temperature absorption in dark-colored ponds and by increasing emissions of potent greenhouse gases which contribute to global warming (Schuur et al. 2015; Leal Filho et al. 2023). By contrast, enhanced carbon sequestration in the deep ocean through the biological carbon pump stabilizes the climate over centennial timescales (Henson et al. 2022), which prevents a direct experience of the impact of the feedback loop. While both examples contain a long-term element (changes in climate), permafrost thaw also contains short-term dynamics that are rapidly observable and quantifiable. Thus, short-term processes may have more direct and observable impacts on the ecosystem than do long-term processes, which in turn may help identify how long-term processes impact the feedback loop. Additionally, anthropogenic or natural pressures happening in proximity to aquatic ecosystems can rapidly alter the ecosystem structure and chemistry, with direct consequences on biogeochemistry that can be amplified or stabilized through feedback loops. If the ecosystem is also influenced by large-scale feedbacks, such as ocean acidification, the additional pressure can be integrated in the general feedback loop impacting the system. As such, recognizing how the temporal and spatial scales of feedback loops influence the ecosystem may bring more attention to the system as a whole, shedding light on how longer-term feedback loops can also impact the system. Then, each short- and long-term feedback can be quantified and their relative importance assessed. Identifying interacting feedback loops is primarily done conceptually where biogeochemical and ecological processes are considered as potential elements of the feedback loop; interdisciplinary teams may thus help identify potential links that could be missed otherwise (Mastrangelo and Cumming 2024). Such a deductive approach provides a robust way to test conceptual models (Livingstone and Imboden 1996), and many statistical methods can then be used to assess the magnitude or relative importance of each component of the loop (e.g., variance partitioning, structural equation modeling, sensitivity analyses, etc.). We argue that explicitly closing feedback loops, a necessary first step to constrain their impacts quantitatively, is essential for understanding biogeochemical cycles, ecosystem functioning, and climate change, especially when the impacts of these interactions are not immediately observable or easily quantifiable. Explicitly closing feedback loops—linking the final consequence back to the initial stimulus—eases recognizing whether the feedback is destabilizing or stabilizing the system, thus deepening our understanding of the system as a whole. As feedback loops are often intricate with many interacting components (Fig. 1), we argue that using a reductionist approach first to simplify the whole feedback loop to its core components may help better identify the importance of the studied feedback. Reducing the complexity of the feedback loop in our representations provides a way to help close it by focusing on the most important aspects. Nested and interacting components can then be reintroduced to modulate the magnitude of the feedback. This approach ensures that the directionality (reinforcing or stabilizing) and impacts of the general feedback loop are well understood. The examples given earlier were short feedback loops about highly influential components (e.g., warmer climate creates permafrost thaw lakes, this increases the production of greenhouse gases which further warms the climate, Fig. 1A, or increased atmospheric CO2 increases temperature, stimulating bacterial respiration and reducing particulate organic matter flux to the deep ocean leading to more CO2 in the atmosphere, Fig. 1D), but the complexity of feedback loops can increase by including other components (Fig. 1B,C,E). For example, including the temperature dependency of microbial metabolic pathways related to greenhouse gas budget (production versus consumption), or how gas solubility decreases with rising temperature, will influence the overall feedback (Fig. 1C). In the example above, although using a reductionist approach may appear as introducing additional unnecessary steps if the studied topic is related to a nested loop, for example, the impact of reduced albedo on permafrost thaw (Fig. 1B), rather than the general loop (Fig. 1A), we argue that it provides critical information to better contextualize the research within the broader biogeochemical process. Putting the studied feedback into perspective may also reveal how it interacts with other biogeochemical processes, thus increasing our general understanding of ecosystems and creating novel research possibilities. The knowledge gained through this reductionist approach offers valuable insights for quantifying feedbacks and integrating them into biogeochemical models and integrating them into biogeochemical models. Lastly, closing feedback loops also provides actionable pathways to address situations where amplifying feedbacks are detrimental to the whole system or support efforts that enable natural communities to stabilize their ecosystems. For example, understanding the impact of self-sustaining anoxic conditions can help design solutions to remedy the problem, and understanding the role of aquatic plants on water quality can support conservation efforts to maintain ecosystem services. Explicitly closing feedback loops is thus a critical first step to recognize what creates or erodes resilience in aquatic ecosystems. Whether stabilizing or reinforcing, feedbacks shape the dynamics of aquatic ecosystems and influence broader processes such as biogeochemical cycles and ultimately the global climate—through silicate weathering (Penman et al. 2020) or the biological carbon pump (Henson et al. 2022) for example. By recognizing, closing, and quantifying these loops, we refine our understanding of natural processes, identify gaps in our knowledge to open up new research directions, and create actionable pathways for resilience. In turn, we hope that the thoughts included here provide new ideas on how feedbacks may impact your research and instill systems-thinking to better understand our complex world. Part of this work was supported by the McGill Wares and the Trottier Space Institute postdoctoral fellowships to RL.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.