How does thermal stratification restructure zooplankton ecological functions in reservoirs: a hypothesis based on self-organised criticality theory

To clarify the mechanism by which reservoir thermal stratification influences the succession of zooplankton functional groups, this paper systematically reviews the national and international research progress and future development trends concerning both reservoir thermal stratification and zooplankton functional groups. Drawing on a systematic literature review and the theory of Self-organised Criticality (SOC), it proposes a novel quantitative research framework and explicates the logical chain through which the dynamic process of thermal stratification drives habitat differentiation, reorganises the topological structure of zooplankton functional-group interaction networks and consequently directs the community towards a self-organised critical state. The paper identifies the key scientific question that urgently requires resolution in this field as “How does the reservoir thermal stratification process drive zooplankton functional groups towards a self-organised critical state?” In response, three research priorities are established: (1) the driving effect of reservoir thermal stratification on zooplankton habitat differentiation; (2) the effect of habitat differentiation on the topological structure of zooplankton functional-group interaction networks; (3) the role of zooplankton functional-group interaction networks in determining the probability of occurrence of SOC states. It is worth emphasising that the SOC state is here construed as a probabilistic rather than an inevitable condition, contingent upon site-specific environmental circumstances. This research may offer scientific support for refining the theoretical framework of reservoir ecology and for strengthening adaptive reservoir management under extreme high-temperature conditions.

There are approximately 95,000 reservoirs in China, with a normal water storage capacity of about 600 billion m³, among which 729 large-scale reservoirs contribute more than 75% of the total storage capacity. These reservoirs constitute critical infrastructure for China to achieve sustainable development goals such as water resource security (SDG 6), aquatic biodiversity conservation (SDG 14) and climate change response (SDG 13) (Cao et al., 2022; Ministry of Water Resources of the People's Republic of China, 2025). The frequent occurrence of global extreme high-temperature events has significantly intensified the reservoir thermal stratification phenomenon, triggering water quality deterioration and ecological degradation, which directly threatens the water supply security and ecological service functions of reservoirs (Woolway et al., 2021; Cao et al., 2022; Gai et al., 2025; Wen et al., 2025). The thermal stratification (ThS) phenomenon exerts multi-dimensional stress effects on carbon and nitrogen cycles and biodiversity maintenance in large reservoirs by regulating the vertical distribution of dissolved oxygen, nutrient migration and biological community succession (Shi et al., 2023b; Jia et al., 2025; Wen et al., 2025). The multi-scale ecological and environmental effects and driving mechanisms of thermal stratification phenomenon in large reservoirs have become a cutting-edge hotspot in the field of water science.

Reservoir zooplankton communities are generally dominated by taxa exhibiting r-strategist life-history characteristics, having developed through natural succession from original river zooplankton communities under reservoir conditions (Han, 2010; Pociecha et al., 2018). They have the characteristics of short life history and rapid reproduction, respond quickly to changes in habitat conditions, and are core groups for material cycling, energy flow and stability maintenance in reservoir aquatic ecosystems. They can effectively indicate water quality status and regulate the risk of algal bloom outbreaks (Han, 2010; Cai et al., 2020; Baldwin et al., 2022). A functional group-based approach can provide greater ecological insight than a purely taxonomic approach, as functional traits—such as body size, feeding mode, and migration behaviour—directly capture organismal responses to environmental gradients and ecosystem processes (Fintelman-Oliveira et al., 2023; Obertegger and Flaim, 2015). By grouping species with similar ecological roles, functional group analysis reveals how environmental changes (e.g., thermal stratification) cascade through community structure and ecosystem function, offering a more mechanistic understanding than taxonomic identity alone (Mwagona et al., 2018; Venello et al., 2021). Zooplankton of different functional groups have differentiated feedback regulation capabilities to environmental stress, and the cascading effects of environmental changes on ecosystems can be more clearly revealed from the perspective of functional groups (Yang et al., 2020; Fintelman-Oliveira et al., 2023; Zheng et al., 2023).

At present, there are still obvious deficiencies in research on the impact of reservoir thermal stratification on zooplankton communities, especially the key processes and mechanisms by which the dynamic process of thermal stratification influences the succession of zooplankton functional groups remain unclear (Lv et al., 2024a; Chi et al., 2024). While several well-established mechanisms have been described—including habitat compression resulting from hypoxia, altered predator-prey interactions, changes in resource distribution, and shifts in diel vertical migration behaviour (Winton et al., 2019; Matthews and Ohman, 2023; Gao et al., 2019)—the complete mechanistic chain linking thermal stratification dynamics to the self-organised critical state of zooplankton functional groups remains unresolved. Based on relevant domestic and international research progress, this paper discusses the potential influence mechanism of reservoir thermal stratification on the succession of zooplankton functional groups, clarifies the key scientific problems to be solved urgently, and analyses future development trends, with a view to providing useful references for comprehensively revealing the eco-environmental effects of reservoir thermal stratification and improving the theoretical and methodological system of reservoir ecology.

This paper is a narrative review that aims to synthesise diverse lines of evidence—spanning reservoir thermal stratification dynamics, zooplankton functional ecology, and self-organised criticality theory—into a coherent conceptual framework. Unlike a meta-analysis, which requires quantitative synthesis of effect sizes from homogeneous studies, a narrative review is particularly suited to the present task because the relevant literature spans multiple disciplines (limnology, ecology, complex systems theory) and employs diverse methodologies (field observations, laboratory experiments, computational modelling). The review seeks not merely to summarise existing findings, but to integrate them critically, identify gaps and inconsistencies, and propose a novel theoretical synthesis—namely, the SOC-based mechanistic hypothesis chain linking thermal stratification to zooplankton functional degradation. In line with the standards for high-quality narrative reviews (Siddaway et al., 2019), the present review explicitly states its research questions, search strategy, and inclusion criteria.

As a narrative review, this study adopts a systematic and transparent approach to literature searching and synthesis, following the best practice principles outlined for systematic reviews (Siddaway et al., 2019) and the reporting guidance of the PRISMA 2020 statement (Page et al., 2021).

Study design

Given the theoretical and integrative nature of this review, we did not apply the traditional PICOS (Participants, Interventions, Comparators, Outcomes) framework (Page et al., 2021) for clinical trials. Instead, we structured our conceptual framework around four core ecological dimensions aligned with our research objectives. Subjects (equivalent to Participants) was defined as zooplankton communities and their functional groups within freshwater reservoir ecosystems; Driving Process (equivalent to Interventions) was defined as the dynamic process of reservoir thermal stratification (including formation, development, and breakdown phases); Comparative States (equivalent to Comparators) was assigned as different thermal stratification regimes (e.g., stable stratified period vs. transitional mixing period vs. isothermal period) to enable comparative analysis of community responses; And Outcome Measures (equivalent to Outcomes) was appointed as habitat differentiation patterns, topological structural changes in species interaction networks, and the SOC states.

In view of the conceptual and integrative scope of this review, we did not apply a standardised risk-of-bias tool (e.g., Cochrane RoB 2 or ROBINS-I). Instead, methodological quality and internal validity were assessed qualitatively through (Siddaway et al., 2019): (a) critical evaluation of the study design, data collection methods, and analytical approaches reported in each included paper; (b) inclusion of peer-reviewed sources where theoretical claims were made; and (c) cross-verification of key findings across multiple independent studies to reduce the influence of single-study artifacts. Any limitations in study quality were narratively addressed during the synthesis and discussed explicitly in the final interpretation.

Information sources and search strategy

A systematic literature search was performed across four electronic databases—Web of Science, Scopus, Google Scholar, and CNKI—to ensure comprehensive coverage of peer-reviewed literature. The search was conducted up to December 2025, with no restriction on publication year to capture historical depth. Both English- and Chinese-language publications were included to minimise language bias. Search terms were operationalised from the core research questions, combining controlled vocabulary and free-text terms using Boolean operators (AND, OR) and truncation symbols (*). The primary search string included combinations of “reservoir” (or “dam”, “lake”), “thermal stratification” (or “thermal layering”), “zooplankton functional group” (or “zooplankton functional traits”, “zooplankton community”), “self-organised criticality” (or “self-organisation”, “critical state”), and “habitat differentiation” (or “habitat degeneration”).

Studies were considered eligible if they met the following inclusion criteria: (1) addressed reservoir or lake thermal stratification, zooplankton functional groups or community dynamics, or self-organised criticality in ecological systems; (2) were peer-reviewed journal articles, book chapters, or some important grey literature (e.g., doctoral dissertations, conference proceedings, technical reports); and (3) provided empirical data, theoretical frameworks, or methodological advances relevant to our research questions. Studies were excluded if they focused solely on marine systems without relevance to freshwater reservoirs, or if they addressed thermal stratification or zooplankton only as peripheral topics without substantive contribution to the central mechanisms.

Data extraction and analysis

All retrieved records were collated and deduplicated. The selection process was conducted in two stages by two independent reviewers. First, titles and abstracts were screened against the eligibility criteria. Subsequently, full texts of potentially relevant articles were retrieved and assessed for final inclusion. Disagreements at either stage were resolved through discussion or consultation with a third reviewer. Additionally, we performed backward snowballing (i.e., manual screening of reference lists from all included studies) to identify any further relevant work not captured by the primary database searches. The entire screening workflow is documented in a PRISMA-style flow diagram (Page et al., 2021).

We adopted a theory-driven narrative synthesis structured around the SOC conceptual framework, proceeding through four integrated stages. First, thematic categorisation and data extraction: all included studies were fully read by two reviewers, and key information (thermal stratification metrics, zooplankton functional group classifications, network indices, and evidence of critical transitions) was extracted into standardised forms and organised into three thematic clusters aligned with our research priorities. Second, causal pathway construction: we mapped a directed logic chain—thermal stratification dynamics → habitat differentiation → topological changes in interaction networks → SOC emergence. For each link, evidence across studies was cross-compared to assess consistency, plausibility, and modifying factors (e.g., reservoir morphology, trophic state). Third, cross-study comparison and contradiction handling: we used a “lines of evidence” approach. Convergent findings from ≥3 independent studies across different regions were deemed robust. Divergent results were scrutinised for methodological or contextual sources and narratively discussed to define boundary conditions. Fourth, graphical and framework integration, we synthesised the extracted evidence into a unifying SOC-based quantitative research framework, supplemented by a conceptual diagram and summary matrices to illustrate stepwise driving mechanisms across thermal stratification phases (stable vs. mixing).

During synthesis, we assigned greater inferential weight to studies with high-frequency, long-term monitoring, robust statistical modelling, and explicit threshold testing. Gray literature was used primarily as corroborative support, not as primary evidence. All interpretive decisions were reached through iterative team discussions to minimise individual bias. Where quantitative indicators (e.g., correlation coefficients, network connectivity) were reported, we qualitatively benchmarked relative link strengths, though formal meta-analysis was precluded by substantial heterogeneity in study designs and reporting formats.

Overview of existing related studies

The systematic search across Web of Science, Scopus, Google Scholar, and CNKI identified 95 records. After removing 52 pre-screening records (50 duplicates, 2 retracted), 43 records were screened, excluding 3 at title/abstract stage. Among 40 reports sought for retrieval, 5 could not be obtained. Fulltext assessment of the remaining 35 articles excluded 6 (3 with no usable information, 2 nonstandard literature, 1 erroneous results), yielding a final corpus of 29 studies. The screening process is shown in Figure 1.

These 29 studies cover three thematic clusters: (1) thermal stratification dynamics and habitat differentiation in reservoirs; (2) zooplankton functional group responses and species interaction network topology; and (3) evidence of self-organised criticality (SOC) and critical transitions in aquatic ecosystems. Geographically, the included studies span temperate and subtropical reservoirs, with the majority from Asia and Europe. The evidence base is dominated by field observational studies, supplemented by a smaller number of mesocosm experiments and modelling analyses. Methodologically, most studies employ seasonal or monthly sampling, though only a few utilize high-frequency monitoring or threshold-testing approaches, which limits the capacity to detect rapid transitions during stratification breakdown. Cross-study comparison reveals a general consensus that thermal stratification drives vertical habitat segregation and alters zooplankton interaction networks, but substantial disagreement persists regarding the mechanisms—whether stratification acts directly via physicochemical stress or indirectly through bottom-up resource control. The SOC framework, while conceptually promising, remains empirically under-tested in reservoir zooplankton systems, with only 4 of the 29 studies explicitly addressing criticality or regime shift indicators.

The known fact about the reservoir thermal stratification

Reservoir thermal stratification is a stratification phenomenon formed by temperature variation with depth, which can profoundly alter water physical, chemical and ecosystem processes. It is jointly controlled by climate change, hydrological processes and artificial reservoir regulation, and has a unique spatio-temporal evolution pattern compared with thermal stratification in lakes and oceans (Shi et al., 2023a; Jia et al., 2025). Existing research is mainly carried out around the directions of formation mechanisms and driving factors of reservoir thermal stratification, eco-environmental effects, monitoring and numerical simulation technologies, and adaptive management and regulation strategies.

The general pattern of reservoir thermal stratification is shown in Figure 2. In terms of temporal dimension, thermal stratification mainly occurs in summer, with the formation period in late spring and early summer, and the dissipation period in autumn. The formation of reservoir thermal stratification is jointly regulated by heat flux (balance between heat absorption and dissipation) and mixing energy (wind-wave disturbance). In summer, solar radiation dominates surface warming: cold water with higher density sinks, while warm water with lower density rises, and vertical heat exchange is blocked to form stable stratification. In autumn and winter, surface cooling triggers vertical convection, and increased wind speeds enhance mechanical mixing of the water column, which now has a smaller density gradient compared with the warmer months, collectively leading to stratification collapse (Li et al., 2022; Shi et al., 2023b; Wang et al., 2023). Longitudinally, the thermal stratification intensity is highest in the lacustrine zone and lowest in the riverine zone; along the water depth direction, the depth, thickness and temperature difference gradient of the thermocline determine the thermal stratification intensity. Current research has confirmed that the spatio-temporal dynamic evolution of reservoir thermal stratification is mainly driven by factors including climate (solar radiation, air temperature, wind speed, rainfall, etc.), hydrology and topography (inflow, reservoir longitudinal morphology and underwater topography), and human activities (reservoir regulation, stratified water intake, shipping, fishery operations, etc.) (Li et al., 2022; Shi et al., 2023a; Wang et al., 2024; Jia et al., 2025). The main research methods include long-term observation, remote sensing inversion, and climate-watershed-reservoir coupling models (such as the combination of GOTM and mHM) (Gai et al., 2025). The development trend of research in this direction is to quantify the coupling effect of extreme climate events (especially extreme high-temperature events) and complex terrain, and further clarify the superimposed influence mechanism of long-term climate change and reservoir operation (Chen et al., 2024; Gai et al., 2025).

The current understanding of the zooplankton functional groups

The classification of zooplankton functional groups is centred on functional traits, aiming to simplify complex community structures and analyse ecological functions and environmental response mechanisms. Existing research is mainly carried out from the perspectives of body size, feeding habit, reproduction mode, environmental tolerance and migration ability, and the characteristics of various methods are shown in Table 1. According to body size (or biomass level), they can be divided into large, medium and small groups. Body size is a functional trait that strongly correlates with metabolic rate, predation risk, and trophic position; size-based categories are therefore treated as functional groups because they reflect organismal function in energy flow and food web dynamics, though it is acknowledged that this approach aggregates species with potentially different ecological roles within the same size class. This classification can intuitively explain the ability of zooplankton to control algal blooms (small filter feeders regulate microalgae, large herbivores inhibit macroalgae) and the efficiency of food supply to fish (medium and large groups are the core food resources of commercial fish). Meanwhile, body size is also related to metabolic rate and predation risk, and is the most commonly used basic classification method for marine and freshwater zooplankton functional groups (Obertegger and Flaim, 2015; Fintelman-Oliveira et al., 2023; Zheng et al., 2023). From the perspective of feeding habits, according to feeding mode and trophic level, they can be divided into filter-feeding, predatory and detritivorous groups. This method can analyse the niche breadth of zooplankton and accurately characterise the driving effect of zooplankton communities on the biogeochemical cycles of important elements such as carbon, nitrogen and phosphorus (Venello et al., 2021; Fintelman-Oliveira et al., 2023; Matthews and Ohman, 2023; Zheng et al., 2023). From the perspective of reproduction mode, they are mainly divided into parthenogenetic and sexual reproduction groups, which can explain genetic diversity and population resilience (Fintelman-Oliveira et al., 2023; Matthews and Ohman, 2023; Zheng et al., 2023). According to the adaptation strategies of zooplankton to environmental stress, they can be divided into hypoxia-tolerant, eurythermal, resting egg strategy and sensitive groups. This classification is mainly applicable to waters with obvious environmental stress, and can explain the steady-state maintenance mechanism of zooplankton communities, quantify environmental stress effects and environmental quality indication value (Venello et al., 2021; Fintelman-Oliveira et al., 2023). According to the migration behaviour of zooplankton in vertical or horizontal directions, they can be divided into collective migration, individual migration and non-migration groups. This classification is applicable to waters with significant environmental gradient changes (such as marine cross-shelf gradient zones, deep-water reservoirs, waters with significant thermocline/stratification), and can analyse the response ability of zooplankton to environmental gradients such as temperature, light and nutrients, as well as the resource utilisation efficiency of populations, and is a key classification dimension for water space-for-time prediction and hydrological response research (Fintelman-Oliveira et al., 2023; Matthews and Ohman, 2023). At present, the development trend of research on zooplankton functional group classification mainly lies in addressing the following challenges: missing trait data, inconsistent standards, high redundancy, mismatch between molecular sequencing and trait databases, non-linear environmental driving, multi-factor interaction, strong regional specificity, decoupling between functional group structure and real ecological functions, as well as insufficient classification resolution, spatio-temporal scale and model interpretability (Fintelman-Oliveira et al., 2023; Matthews and Ohman, 2023). Importantly, we distinguish among taxonomic diversity (the richness of species), functional-group structure (how species are clustered in trait space), functional diversity (the distribution and spread of functional traits within a community) and ecosystem functioning (the aggregate ecological processes, such as grazing, nutrient recycling and energy transfer, sustained by the community). The traits selected for classification—body size, feeding mode, reproduction mode, environmental tolerance and migration behaviour—are ecologically meaningful because they directly link organismal responses to environmental gradients with ecosystem-level processes.

The controversies regarding the relationship between reservoir thermal stratification and zooplankton community

Reservoir thermal stratification significantly alters the physical and chemical characteristics of water bodies, yet the extent to which these changes drive the structure and dynamics of zooplankton communities remains a subject of considerable debate. When the temperature difference between surface water temperature and bottom water temperature reaches above 10 °C, the reshaping effect on the water environment is particularly pronounced (Li et al., 2022; Shi et al., 2023b; Dai et al., 2023; Wang et al., 2023). A thermocline with a temperature gradient of 1–3 °C/m can substantially hinder the vertical exchange of water masses and enhance stratification stability (Zhang et al., 2022; Castrillo et al., 2024). Light intensity may attenuate by more than 90% upon reaching the thermocline, rendering the bottom layer nearly aphotic (Shi et al., 2023b). Dissolved oxygen in the surface layer tends to be saturated (8–12 mg/L), but can drop sharply to below 3 mg/L on the lower side of the thermocline, with the bottom layer potentially forming an anaerobic zone due to prolonged hypoxia (<1 mg/L), occasionally triggering “black water events” (Li et al., 2021; Mercado-Bettín et al., 2021; Chen et al., 2024; Zhang et al., 2024). Photosynthesis of surface algae absorbs CO 2, raising pH to 8.5–9.0, while decomposition of organic matter in the bottom layer releases CO 2, lowering pH to 7.0–7.5 and forming an acid-base cline (Liu et al., 2019, 2020; Mercado-Bettín et al., 2021; Zhang et al., 2024). Nutrients in the surface layer are heavily consumed by phytoplankton, whereas the bottom layer becomes enriched through sediment release and anaerobic decomposition; however, the thermocline blocks the vertical circulation of nutrients (Winton et al., 2019; Liu et al., 2020b; Shi et al., 2023b; Gao et al., 2024). Particulate organic carbon (POC) is delayed by the thermocline during settlement, easily forming a POC enrichment zone (Dong et al., 2022). While these physicochemical gradients are well documented, their cascading effects on zooplankton—and the causal mechanisms involved—are far from settled.

The relationship between reservoir thermal stratification and zooplankton community dynamics is marked by conflicting findings and interpretive disagreements. On one hand, a number of studies suggest that thermal stratification exerts a strong, deterministic control over zooplankton. The thermocline creates vertical gradients of temperature, oxygen, and dissolved substances that drive the stratified distribution of plankton communities, and hypoxia in the deep water layer during the stratification period can enhance anaerobic bacterial activity, aggravating the risk of eutrophication (Liu et al., 2019). Some researchers have reported that when thermal stratification reaches a certain intensity, it is highly likely to cause degradation or even collapse of zooplankton communities, thereby weakening the ability to control algal blooms (Gao et al., 2019; Rizo et al., 2020; Banerjee et al., 2022). On the other hand, an opposing body of evidence indicates that zooplankton communities exhibit considerable resilience and adaptive capacity, with vertical migration behaviour and species-specific physiological tolerances allowing them to persist under stratified conditions. The extent to which stratification acts as a direct driver versus a modulating factor remains contested, and the relative importance of stochastic versus deterministic processes in shaping community responses is still under active investigation (Lv et al., 2024a, 2024). Furthermore, the lack of long-term, high-resolution monitoring datasets makes it difficult to disentangle stratification-induced effects from natural seasonal variability and other confounding factors such as nutrient loading, predation pressure, and hydrological regime (Chi et al., 2024).

Methodological discrepancies further complicate the picture. Different studies employ varying definitions of stratification intensity, use disparate sampling frequencies and depths, and apply different analytical frameworks to assess zooplankton community responses. These methodological inconsistencies make cross-study comparisons challenging and may partly explain the divergent conclusions reported in the literature. Some researchers argue that the apparent correlation between stratification and zooplankton decline is spurious, driven instead by concurrent changes in trophic state or fish predation (Gao et al., 2019; Banerjee et al., 2022). Others contend that the effect is real but indirect, mediated through stratification-induced changes in food availability (phytoplankton composition and biomass) rather than through direct physicochemical stress (Rizo et al., 2020).

The accuracy of monitoring and simulation of reservoir thermal stratification is the primary limitation

The accuracy of monitoring and simulation of reservoir thermal stratification has become the primary limitation constraining the current field of thermal stratification research; although a series of breakthroughs have been made in recent years, the core bottleneck still lies in the insufficient accuracy of both monitoring and simulation.

In terms of monitoring technology, Zhang et al. (2022) established accurate thermal stratification thresholds based on hourly temperature profile data generated by high-frequency monitoring technology, and captured the abnormal phenomenon of thermocline intensity in the tailwater area caused by cascade operation. However, the accuracy of such monitoring remains insufficient. The current primary limitation is that high monitoring costs restrict the deployment and maintenance of long-term, high-frequency observation networks, and the difficulty in standardised integration of multi-source data (remote sensing, buoys, model outputs) further reduces the accuracy and consistency of the monitoring network.

In terms of simulation technology, a cross-scale method integrating the advantages of multiple models has been formed. For example, Gai et al. (2025) comprehensively applied the lake model GOTM, watershed model mHM and river temperature model Air2stream, effectively integrating meteorological data, reservoir topography and inflow characteristics, and realised predictions of the seasonal dynamics of the stratification structure; however, the accuracy of such predictions remains limited for extreme conditions. The primary limitation in simulation is that there are uncertainties and deviations in the responses to strong inflow disturbances (such as storm runoff processes) and extreme climate events (such as extreme drought), leading to significant prediction errors in the cold water retention time in the deep water layer (Shi et al., 2023a). In the future, it is necessary to further develop coupling models adapted to complex hydrological conditions so as to improve simulation accuracy, and optimise low-cost sensors and edge computing technologies to improve the spatio-temporal resolution and efficient applicability of thermal stratification monitoring and assessment.

The self-organisation of zooplankton functional group succession is widespread

Aggregated spatial distribution is a pattern commonly observed in zooplankton, and self-organisation represents one possible mechanism that can generate such patterns. Studies have found that zooplankton usually present aggregated distribution in water, with the spatial range of a single cluster ranging from tens of centimetres to hundreds of metres (Rothenberger and Calomeni, 2016; Maznah et al., 2018), and the interaction distance between individuals ranges from several times to more than a hundred times the body length (several millimetres to several centimetres) (Hart and Bychek, 2011). Such scale-invariant clustering patterns recur across different water bodies and taxonomic groups, indicating that aggregation itself is a common property rather than a species- or habitat-specific phenomenon. Importantly, the internal interactions of zooplankton clusters have been shown to be spontaneous and localised, belonging to self-organising systems with self-replication ability (derived from reproductive ability). Because these interactions are driven by local rules among individuals rather than by external top-down control, they exhibit the characteristic signatures of self-organisation. Furthermore, the event frequency distribution of cluster succession processes follows a power-law distribution, and the key stage of the zooplankton community degradation process is entering the unstable Self-organised Criticality (SOC) state (Huang et al., 2024). The recurrence of power-law dynamics and SOC transitions across diverse observational systems provides considerable evidence for the widespread nature of self-organisation in zooplankton succession. It is important to note, however, that aggregated spatial distributions are common in planktonic organisms, and not all aggregation implies self-organisation. Self-organisation requires that the observed patterns arise from local, decentralized interactions among individuals rather than from external forcing alone. Therefore, aggregation is treated here as a pattern potentially compatible with self-organisation, rather than as direct evidence of it; the stronger evidence for self-organisation comes from the power-law dynamics and SOC transitions observed in these systems.

The SOC state is the critical state of the self-organisation of biological communities, which has been proposed as a potential transitional state through which a community may reorganise as it shifts from an old steady state to a new steady state, and is characterised by instability (Guill et al., 2021). If this critical transition is indeed governed by universal dynamic laws rather than by any particular species or ecosystem, it could constitutes a common pathway through which zooplankton communities reorganise, thereby underscoring the potential generality of the self-organisation process. The self-organisation of zooplankton community succession is a core characteristic of spontaneously forming stable structures and functions through interspecific interactions, environment-trait feedback and ecological process coupling, and is the key to analysing community assembly mechanisms and ecosystem stability (Liu et al., 2021a; Huang et al., 2023). Its core principle is that interactions such as interspecific predation and competition form density-dependent feedback to regulate cluster structure, two-way adaptation of functional traits and environmental factors realises adaptive regulation, and coupling of energy flow and material cycling forms internal closed-loop regulation (Mwagona et al., 2018; Mesquita et al., 2024). These mechanisms, namely density-dependent feedback, adaptive regulation and closed-loop regulation, are generic in nature and operate irrespective of the specific taxa or habitats involved, which is why self-organisation manifests consistently across zooplankton functional groups.

The research paradigm of plankton community self-organisation is “field observation - experimental analysis - model prediction”, that is, based on long-term field monitoring and spatio-temporal statistical analysis, combined with microcosm and mesocosm scale controlled experiments to verify causal mechanisms, and supplemented by relevant mathematical models or computer models for simulation (Chakraborty and Manthena, 2015; Liu et al., 2021a; Huang et al., 2024). The fact that such a uniform paradigm can be applied across disparate systems itself reflects the generality of the underlying self-organisation phenomenon. The development trends of research on the self-organisation mechanism of zooplankton functional groups are: (1) establishing a unified quantitative index system for self-organisation processes; (2) determining the environmental stress intensity threshold that leads to the collapse of self-organisation processes; (3) constructing a multi-scale fusion model for self-organisation processes. These trends ultimately converge on the goal of further testing and understanding the extent of self-organisation across zooplankton functional group succession.

Self-organised criticality theory can effectively bridge the gap of differences

In the study of the influence of reservoir thermal stratification on zooplankton functional group succession, there are long-standing research divergences: different studies reach inconsistent or even conflicting conclusions on how stratification regulates functional group succession, owing to differences in reservoirs, seasons, spatial scales and statistical methods. Self-organised criticality theory is often used to explain the mechanisms of phenomena such as species extinction, wildfire spread and population fluctuation in the field of ecology (Lian et al., 2016; Kato et al., 2020; Walter and Hinterberger, 2022). Because it reveals the common self-organised critical dynamics underlying superficially different ecological processes (Turcotte, 1999; Watkins et al., 2016), SOC theory can serve as a unified framework to reconcile the divergent findings on the thermal-stratification-driven succession of zooplankton functional groups. However, it is important to acknowledge that other theoretical frameworks — such as resilience theory, critical transitions theory, ecological network theory, and regime-shift frameworks — may provide alternative or complementary explanations for the observed ecological patterns. These perspectives, while differing in their mechanistic assumptions and methodological approaches, share the common goal of understanding how ecological systems respond to external perturbations and reorganise across scales. SOC theory offers several distinctive advantages for the specific question addressed in this paper: (1) SOC provides a quantitative, physically grounded framework for predicting the probability and magnitude of abrupt community transitions, rather than merely describing them post hoc; (2) SOC explicitly links micro-scale interactions (e.g., local predation, competition) to macro-scale emergent properties (e.g., power-law distributions, avalanche dynamics), offering a mechanistic bridge between individual-level processes and community-level outcomes; (3) the power-law scaling relationship inherent to SOC systems allows for probabilistic prediction of the likelihood of large-scale functional degradation events, which is difficult to achieve with threshold-based or resilience-based approaches alone. These features make SOC particularly well-suited for analysing the cascading effects of thermal stratification on zooplankton functional group dynamics, where local habitat changes propagate through interaction networks to potentially trigger system-wide functional degradation.

According to this theory, the phase transition process of complex systems theoretically expected to pass through an unstable SOC state, in which small disturbances to the system may trigger chain reactions (such as the avalanche effect), and the relationship between event scale and frequency follows a power-law distribution (Watkins et al., 2016). The basic principle is shown in Figure 3. Applied to reservoir ecosystems, this scale-invariant power-law relationship implies that the responses of zooplankton functional group succession to thermal stratification, although differing in magnitude and threshold among reservoirs, follow the same critical dynamics, thereby bridging the divergences between micro- and macro-scale observations. The identification of the SOC state is the key to research, which is usually carried out based on core statistical characteristics such as power-law distribution of system events, long-range correlation, scale invariance and critical slowing down. These complementary methods collectively bridge the gap between observable data and the underlying critical state, each addressing a different aspect of the identification problem. The main methods include (Turcotte, 1999; Lian et al., 2016; Watkins et al., 2016; Kato et al., 2020; Walter and Hinterberger, 2022): (1) Statistical analysis to verify whether the event scale frequency of the system conforms to the power-law distribution, supplemented by the Kolmogorov-Smirnov test to exclude pseudo-power laws; (2) Quantifying long-range correlation through detrended fluctuation analysis (DFA) or power spectrum analysis, and calculating the fractal dimension of the system event distribution to verify self-similarity; (3) Dynamic simulation method, constructing sand pile models, forest fire models, etc., to observe whether the system spontaneously evolves to a critical state and reproduces the characteristics of actual data; (4) Machine learning method, using neural networks or support vector machines to extract critical state features (such as fluctuation patterns near the phase transition point) from high-dimensional data. In practical application of the above methods, it is necessary to be alert to the interference of non-stationary noise, and integrate multi-scale evidence (consistency between physical mechanism and statistical characteristics) to improve the reliability of identification, thereby providing a consistent and standardised basis for comparing results across different reservoirs and studies, which helps to resolve the research divergences in the influence of thermal stratification on zooplankton functional group succession. It must be emphasised, however, that power-law scaling, long-range correlation and related statistical signatures are necessary but not sufficient evidence for SOC. Robust identification additionally requires finite-size scaling, avalanche statistics, critical exponents and tests of universality, as well as explicit comparison against alternative mechanisms (e.g., stochastic resonance, multiplicative noise, threshold dynamics) that can generate similar statistical patterns. Accordingly, evidence of self-organisation should be distinguished from evidence specifically supporting SOC; where such diagnostics cannot be satisfied, the SOC interpretation should remain tentative.

These trends further reflect the potential bridging role of SOC theory, extending its possible unified framework to reconcile the remaining divergences in SOC state identification within living systems. The development trends of research on SOC state identification in living systems are (Lian et al., 2016; Kato et al., 2020; Liu et al., 2021b; Banerjee et al., 2022): (1) Effectively distinguishing the characteristics of SOC states in living systems from those of other mechanisms (such as stochastic resonance); (2) Realising effective observation of SOC states in living systems with short duration; (3) Constructing effective SOC state evolution models for scenarios with insufficient actual observation data.

Key scientific hypotheses and their logical chains

Research on the eco-environmental effects and driving mechanisms of thermal stratification in large reservoirs is still in the exploratory and development stage. As global warming intensifies extreme high-temperature weather, abnormal thermal stratification in reservoirs is becoming increasingly severe. Unlike tropical reservoirs (mostly in thermal mixing state) and temperate reservoirs (generally weak stratification), large reservoirs in subtropical monsoon regions undergo more intense dynamic changes in thermal stratification that generate more complex eco-environmental problems demanding focused study (Winton et al., 2019; Chen et al., 2024). The central scientific question is: “How does the reservoir thermal stratification process drive zooplankton functional groups towards a self-organised critical state?”

We hypothesise that in large reservoirs, the dynamic process of thermal stratification drives the differentiation of zooplankton habitat conditions, and different functional groups of zooplankton produce differentiated responses to habitat differentiation, which changes the topological structure of the functional group interaction network, leading to an increased diffusion rate of small disturbances in the cluster, thereby accelerating the evolution of the functional group self-organisation process towards the SOC state, and ultimately leading to the loss of functional group diversity, manifested as functional degradation of the zooplankton community.

To operationalise the concept of functional degradation, we draw on established trait-based approaches (Obertegger and Flaim, 2015; Mwagona et al., 2018; Venello et al., 2021; Fintelman-Oliveira et al., 2023; Zheng et al., 2023) in functional ecology rather than equating it with a loss of functional-group diversity. Functional degradation is here defined as a sustained decline in the ecosystem-level performance of a zooplankton community, as evidenced by measurable trait-based indicators such as functional richness (FRic), functional dispersion (FDis) and Rao’s quadratic entropy, together with reduced functional redundancy and shifts in response traits relative to effect traits (Amorim and Kainz, 2025; Josué et al., 2021; dos Santos et al., 2025). Ecosystem-function measures—such as grazing pressure on phytoplankton, nutrient recycling rates and energy transfer efficiency (Bluhm et al., 2015)—provide complementary indicators of whether reduced functional performance translates into ecosystem-level consequences. Because functional redundancy and compensatory dynamics can buffer ecosystem functioning, a decline in functional-group diversity is not necessarily equivalent to functional degradation; degradation is inferred only when trait-based and ecosystem-function indicators decline concurrently.

It is important to emphasize that this hypothesis chain posits that thermal stratification drives zooplankton functional groups towards an SOC state as a probabilistic event rather than a deterministic outcome. Although the SOC state is universal during the system phase transition process, the system is not always in the phase transition process. Therefore, the SOC state is a condition that may occur with a certain probability under specific environmental circumstances. In a stable community, functional redundancy and niche overlap allow the disappearance or fluctuation of certain non-critical species to be compensated by other species within the same functional group—this is the normal homeostatic mechanism of a stable system. The SOC state, by contrast, represents a critical transition phase where such compensatory mechanisms have been overwhelmed, and small disturbances can propagate through the network (Turcotte, 1999; Watkins et al., 2016). Furthermore, the driving mechanism of thermal stratification on zooplankton functional group SOC state evolution is likely to exhibit a degree of site-specificity, varying with reservoir characteristics, climatic region, and local biotic context. The main challenge is that zooplankton communities are limited self-organising systems, in which external environmental changes can easily render them completely controlled systems; hence the applicability of existing SOC identification methods and evolution-rate calculations must be verified.

A key question is whether changes in the zooplankton interaction network can be uniquely attributed to thermal stratification, given that communities are also shaped by fish predation and invasive species. The framework addresses this through: (1) statistical hypothesis testing to quantify the relative contribution of thermal stratification versus other factors; (2) causal inference methods (e.g., Difference-in-differences, Bayesian Structural Time Series) to isolate the effect of thermal stratification from confounding factors; (3) testing the mechanistic pathway—thermal stratification → habitat differentiation → network topology change → SOC state—rather than a simple direct causal link; and (4) segmented proof of each link, with overall verification through system dynamics modelling. This multi-faceted design identifies the specific role of thermal stratification while acknowledging concurrent drivers. The design explicitly distinguishes three levels of evidence, including correlational evidence, causal-inferential evidence that explicitly controls for confounding factors such as trophic state, nutrient loading, hydrology, fish predation and seasonal variability, and mechanistic evidence. The proposed chain should therefore be regarded as a hypothesis to be tested rather than an established mechanism.

The multi-link hypothesis chain is therefore tested through “segmented proof and overall verification”. If verified, it will not only reveal the eco-environmental effects of reservoir thermal stratification more comprehensively, but also extend SOC theory into reservoir ecology and promote interdisciplinary innovation. However, full verification is difficult, facing challenges in data quality, model deviation and multi-method coupling. Three lines of research are prioritised: Research Content 1—Thermal stratification drives habitat differentiation, invert the three-dimensional temperature field from long-term water temperature monitoring, characterise thermocline spatio-temporal patterns and stratification intensity evolution; quantify vertical habitat differentiation using key factors (dissolved oxygen, pH, light, nutrients, suspended particulate matter, chlorophyll-a); and statistically establish the driving function of stratification on habitat differentiation. Research Content 2—Habitat differentiation reshapes interaction networks, construct functional group classification from functional trait space; clarify inter-group interactions and analyse network topology and its spatio-temporal evolution from long-term data; assess disturbance diffusion under different habitat-differentiation scenarios; and statistically establish that habitat differentiation significantly alters network topology and enhances disturbance diffusion, incorporating spatio-temporal lag effects. Research Content 3—Interaction networks determine SOC occurrence probability: define and identify functional group SOC states based on self-organisation dynamics; develop and map the spatio-temporal distribution of SOC occurrence probability; and statistically establish that higher disturbance diffusion efficiency significantly increases SOC occurrence probability.

Finally, a modular system dynamics (SD) model—comprising thermal stratification, habitat differentiation, zooplankton interaction network and SOC evolution modules—integrates the above contents for overall verification. Taking thermally sensitive water areas (such as the water area in front of the reservoir dam) as the object, taking the measured high-frequency dynamic sequence thermal stratification data as the model input, and taking the occurrence probability of zooplankton functional group SOC states as the output, if the modelled spatio-temporal distribution of SOC occurrence probability matches the measured distribution of zooplankton functional group diversity—with increasing SOC occurrence probability decreasing mean diversity and increasing its coefficient of variation—then the scientific hypothesis as a whole is validated. In addition to functional diversity, additional validation metrics should be considered to strengthen the robustness of the framework. For instance, network structure metrics (e.g., connectance, modularity, and interaction strength distribution within the zooplankton interaction network) can reveal whether SOC-induced shifts reorganise trophic linkages in a manner consistent with critical dynamics. Ecosystem stability indicators (e.g., temporal variability of biomass, resistance and resilience to perturbations) can provide independent tests of whether the system approaches a critical state. Ecosystem functioning indicators (e.g., nutrient recycling rates, energy transfer efficiency across trophic levels, water quality regulation capacity) can further corroborate whether the functional consequences of SOC transitions extend beyond community composition to broader ecosystem-level processes. Incorporating these complementary metrics would not only reduce the reliance on a single endpoint but also provide a more comprehensive validation of the proposed SOC-based framework.

A feasible methodology framework

To achieve the aforementioned research objectives and content, the proposed methodology proceeds from actual observed phenomena, integrating relevant concepts and principles from fields including zooplankton biology, self-organisation theory and reservoir ecology, concisely put forward scientific questions, and propose scientific hypotheses combining the small-world network principle and existing research outcomes. The methodological framework presented here is intended as a feasible technical pathway for validating the proposed scientific hypotheses, rather than as a rigid analytical workflow. It serves as a conceptual model that can be adapted to specific study contexts and data availability. A typical large-scale reservoir in a subtropical climate zone with well-established long-term observation data is proposed as the study area. The methodological approaches are selected based on their specific suitability for testing each component of the hypothesis chain. Finally, the findings will be integrated into the systematic dynamic mechanism of the overall causal chain. The research technical route is shown in Figure 4 as follows. To enhance practical applicability, we distinguish between essential analyses required to test the central hypothesis and optional analyses that may be implemented when data availability permits.

Essential analyses: Principal Component Analysis (PCA) is used to reduce the dimensionality of multi-factorial habitat variables and identify the dominant environmental gradients driving habitat differentiation. K-means Clustering Analysis is applied to classify zooplankton functional groups based on trait space, providing an objective and replicable grouping scheme. Cellular Automaton (CA) models are used to simulate the local interaction rules that govern the self-organisation dynamics of zooplankton functional groups, as CA models are well-suited for capturing emergent spatial patterns from decentralized local interactions.

Optional analyses that can be incorporated when sufficient data are available include: Hurst exponent and Moran’s I are employed to quantify the long-range dependence and spatial autocorrelation of thermal stratification and habitat variables, respectively, which are essential for understanding the spatio-temporal structure of the system. Bayesian Structural Time Series (BSTS) model is applied to infer causal relationships in the “thermal stratification → habitat differentiation” link, as it can handle non-stationary time series and quantify the uncertainty of causal effects. Difference-in-differences (DID) method is used to test the “habitat differentiation → network topology change” link, by comparing network structure changes between periods or locations with and without strong thermal stratification. Grey Model GM(1,1) is employed for predicting the evolution trend of SOC state occurrence probability under data-limited conditions, as it requires only a small number of data points to generate reasonable predictions. Set-pair Analysis (SPA) is used to assess the degree of similarity and difference between the modelled and observed SOC state occurrence probabilities, providing a quantitative measure of hypothesis validation.

The practical feasibility of this framework depends on the availability of long-term, high-frequency monitoring data, which are indeed rare for most reservoir ecosystems. To address this challenge, the framework can be implemented incrementally, starting with the best-available datasets from well-studied reservoirs, and progressively incorporating data from additional sites as monitoring networks expand. The validation strategy involves comparing model predictions against independent empirical observations, if the spatio-temporal pattern of modelled SOC state occurrence probability matches independent observations of functional group diversity decline (from separate monitoring campaigns or time periods), the hypothesis is supported; if no such match is found, the hypothesis is refuted.

It is necessary to emphasize that the construction of ecological interaction networks is a critical methodological step. Several approaches are available: (1) correlation-based networks (e.g., Spearman rank correlations between functional group abundances), which are easy to compute but may include spurious associations; (2) co-occurrence networks (e.g., using probabilistic or graph-theoretical methods), which provide better statistical rigor but do not directly reflect mechanistic interactions; and (3) mechanistic interaction networks (e.g., based on known predator-prey relationships, feeding modes, and competitive interactions), which offer the most ecologically meaningful interpretation but require extensive prior knowledge. Given the functional group focus of this study, a hybrid approach is recommended: starting with co-occurrence network analysis to identify statistically robust associations, and then overlaying known functional trait-based interaction rules (e.g., filter-feeders compete for similar resources, predators consume filter-feeders) to refine the network into a more ecologically realistic interaction structure. The choice of network inference method should be explicitly justified based on the specific research question and data characteristics.

The limitations of this study and the corresponding coping strategies

As a narrative review of the eco-environmental effects of reservoir thermal stratification and the self-organised criticality of zooplankton functional groups, this study provides a comprehensive and critically evaluated synthesis of current knowledge, yet it is not without limitations. Rather than treating these limitations as passive caveats, we pair each with a corresponding coping strategy to chart a concrete path forward.

First, although a systematic search was conducted, the review did not follow a formal registered protocol (e.g., PROSPERO), which may introduce selection bias in the identification and inclusion of studies. To cope with this, future reviews should register a protocol and follow PRISMA guidelines, apply dual-reviewer independent screening, and explicitly document search strings and inclusion criteria to enhance transparency and reduce bias.

Second, the review is primarily qualitative in nature; quantitative synthesis (meta-analysis) was not possible due to the heterogeneity of study designs, outcomes, and reporting formats across the included literature. The corresponding coping strategy is to standardise outcome reporting and promote comparable metrics among studies, enabling future meta-analyses or subgroup analyses once sufficiently homogeneous data are accumulated.

Third, the search was limited to English- and Chinese-language publications, which may exclude relevant work published in other languages. This limitation can be addressed by extending retrieval to multilingual databases and grey literature, thereby broadening linguistic and geographical coverage.

Fourth, the proposed SOC-based framework, while grounded in existing theory and empirical observations, remains a hypothesis that requires rigorous empirical testing. This is the most consequential limitation, and the coping strategy centres on empirical validation through the three priority research contents—thermal stratification driving habitat differentiation, habitat differentiation reshaping functional group interaction networks, and interaction networks determining SOC occurrence probability—implemented under the principle of “segmented proof and overall verification”.

Taking large reservoirs in subtropical monsoon regions with sufficiently rich monitoring data as objects and accumulating reservoir-level case studies, the general driving mechanism can be summarised and integrated within a modular system dynamics model for overall verification. Once validated, the framework could inform the development of thermal stratification ecological risk early warning models and adaptive reservoir regulation technologies to support reservoir water supply security and ecological health. Such adaptive strategies may include selective withdrawal, artificial mixing or destratification, and forecast-based adaptive reservoir operation (Liu et al., 2022, 2023; Wang et al., 2024).

Although the framework is developed with reference to large reservoirs in subtropical monsoon regions, where intense stratification dynamics pose particularly pressing ecological challenges, its underlying principles are intended to be generalisable to reservoirs differing in morphometry, trophic status, climate, mixing regime, hydrological conditions and zooplankton community composition. The probabilistic formulation of SOC occurrence and the modular, trait-based design are deliberately system-flexible, allowing parameterisation for shallow or deep, oligotrophic or eutrophic, and temperate or tropical reservoirs, provided that site-specific stratification regimes and community trait distributions are characterised. The emphasis on subtropical large reservoirs is therefore a starting point for empirical validation rather than a claim of universal applicability.

In summary, although these limitations temper the strength of the present conclusions, each is matched by a feasible coping strategy. Systematically addressing them will enhance monitoring quality, enable cross-regional comparative research, and convert the identified gaps into operational tools, thereby advancing both the theory of thermal stratification ecological effects and its practical application.

This paper presents a conceptual framework, grounded in a narrative review of the literature, linking reservoir thermal stratification to the functional degradation of zooplankton communities via the lens of self-organised criticality (SOC) theory. The key contributions of this framework are as follows:

First, the paper identifies the key scientific question—”How does reservoir thermal stratification drive zooplankton functional groups towards a self-organised critical state?”—and proposes a mechanistic hypothesis chain: thermal stratification → habitat differentiation → altered topological structure of zooplankton functional group interaction networks → increased disturbance diffusion efficiency → accelerated evolution towards the SOC state → functional degradation of zooplankton communities. Each link in this chain is presented as a testable hypothesis, with the understanding that the SOC state is a probabilistic outcome rather than a deterministic consequence of thermal stratification.

Second, the paper synthesizes existing literature on reservoir thermal stratification (formation mechanisms, eco-environmental effects, monitoring and simulation, adaptive management) and zooplankton functional groups (classification schemes, self-organisation dynamics) to identify knowledge gaps and establish the empirical basis for the proposed SOC framework. The literature review reveals that while individual components of the hypothesized pathway have been studied in isolation, an integrated framework linking thermal stratification to zooplankton functional degradation through SOC theory has not been previously constructed.

Third, the paper proposes a methodological framework for testing the hypothesis chain, including: (1) specific topological parameters for quantifying interaction network changes (inter-module connections, average path length, modularity, connectivity density, clustering coefficient); (2) an operational definition of SOC state occurrence probability as a composite index integrating multiple SOC diagnostic metrics; (3) a suite of analytical methods selected for their specific suitability to each component of the hypothesis chain; and (4) an overall validation strategy using system dynamics modelling with independent data for calibration and validation.

Fourth, the paper discusses the limitations and challenges of the proposed framework, including: (1) the observational nature of much of the supporting evidence, which limits causal inference; (2) the difficulty of distinguishing SOC from alternative processes generating similar statistical patterns; (3) the substantial data requirements for implementing the full methodological framework; and (4) the site-specificity of thermal stratification-zooplankton interactions, which may limit the generality of findings.