Multi-Level Driving Mechanisms: Cascading Relationships Among Physical Factors, Nutrient Cycling, and Biological Responses in the Yangtze River–Lake Ecosystems

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript examines the spatial and temporal dynamics of river–lake ecosystems in the Yangtze Basin, focusing on how physical factors, nutrient cycling, and biological responses interact. Although the topic is ecologically relevant, the study’s design and analysis contain serious flaws that weaken the validity and generalizability of its conclusions. The connection to sustainability is only superficial: the work emphasizes ecological pattern description rather than management or policy applications. To align with the aims of Sustainability, the authors must explicitly relate their findings to the UN 2030 Agenda and propose actionable strategies for sustainable river–lake governance. The current version lacks this transdisciplinary perspective, offering little integration between ecological results and broader sustainability goals.

The study’s sampling and analytical framework are its main weaknesses. The single-year data collection prevents distinguishing long-term ecological trends from short-term variability, a critical issue in systems influenced by multi-year hydrological cycles such as those created by the Three Gorges Dam. Missing December data undermine the analysis of seasonal patterns, while inadequate spatial replication, absence of depth stratification, and lack of reference sites further limit interpretive power. The identification of phytoplankton only to the genus level also reduces ecological precision, and several statistical methods (including redundancy and structural equation modeling) are applied without sufficient rigor or justification. Model evaluation lacks standard fit indices, and the interpretation of correlations as causal relationships is methodologically unsound. These issues collectively undermine the manuscript’s main claims about cascade mechanisms and ecological drivers.

To reach publishable standards, the authors must redesign their inferential framework and substantially expand their analysis. This includes multi-year monitoring to capture hydrological and climatic variability, proper replication and stratification, and species-level taxonomic identification. The statistical modeling should incorporate multiple fit metrics, alternative hypotheses, and explicit controls for multicollinearity and Type I error. Integrating land use variables, dam operation data, and sustainability-oriented recommendations would greatly enhance the manuscript’s relevance and rigor. With these major revisions, the study could provide valuable insight into the functioning and sustainable management of river–lake systems in the Yangtze Basin, but in its current state, it remains an exploratory effort with limited inferential depth.

Comments on the Quality of English Language

The manuscript is generally understandable and employs appropriate scientific terminology, but numerous grammatical errors, inconsistent terminology, awkward phrasing, and stylistic inconsistencies reduce its clarity and professional tone. The writing reflects non-native English authorship with solid technical vocabulary yet uneven grammatical control and idiomatic accuracy. Professional language editing or review by a native English-speaking co-author is strongly recommended. Specific issues include typographical errors (“Furthemore” instead of “Furthermore,” “Abstact” instead of “Abstract”), verb tense inconsistencies, and spacing problems. Style can be improved by reducing passive constructions, removing redundant or clichéd expressions, simplifying complex sentences in the Results section, and standardizing article use and acronym formatting.

Author Response

Comment 1: This manuscript examines the spatial and temporal dynamics of river–lake ecosystems in the Yangtze Basin, focusing on how physical factors, nutrient cycling, and biological responses interact. Although the topic is ecologically relevant, the study’s design and analysis contain serious flaws that weaken the validity and generalizability of its conclusions. The connection to sustainability is only superficial: the work emphasizes ecological pattern description rather than management or policy applications. To align with the aims of Sustainability, the authors must explicitly relate their findings to the UN 2030 Agenda and propose actionable strategies for sustainable river–lake governance. The current version lacks this transdisciplinary perspective, offering little integration between ecological results and broader sustainability goals.

Response 1: We thank the reviewer for this constructive suggestion. We have extensively revised the manuscript to explicitly link our ecological findings to the UN 2030 Agenda and propose actionable strategies for sustainable governance. In the Introduction, we have added two paragraphs (the sixth and seventh paragraphs) to frame our research within the context of the Sustainable Development Goals (SDGs) and China's national policies for the Yangtze River Basin. Furthermore, we have added a new section, "4.4 Implications for sustainable management" to the Discussion. This section details how our results provide a scientific basis for managing the Three Gorges Dam and agricultural pollution, and offers five specific policy recommendations directly linked to achieving SDGs 6, 13, 14, 15, and 17.

 

Comment 2: The single-year data collection prevents distinguishing long-term ecological trends from short-term variability, a critical issue in systems influenced by multi-year hydrological cycles such as those created by the Three Gorges Dam.

Response 2: We appreciate the reviewer's concern but respectfully note that our study objectives differ fundamentally from long-term trend analysis. Our research is specifically designed as a mechanism-identification study rather than a trend-prediction study, focusing on quantifying the cascading relationships among physical drivers, nutrient cycling, and biological responses.

Our primary goal is to quantify the cascading mechanisms linking physical drivers, nutrient cycling, and biological responses. Such mechanistic understanding relies fundamentally on spatial covariation and cross-system comparisons rather than multi-year temporal replication. This approach is well-established in ecological literature (Soranno et al., 2014). Our design encompassed 36 sites across 4 rivers and 5 lakes, spanning substantial environmental gradients in hydrodynamic conditions (rivers vs. lakes), trophic states, and anthropogenic disturbance levels. This spatial heterogeneity provides a "space-for-time" substitution that is particularly valuable for identifying ecological drivers (Blois et al., 2013). Moreover, Our bimonthly sampling from January to November captured a complete hydrological year in this subtropical monsoon region, including all critical seasonal events: winter low-productivity period (January), spring phytoplankton bloom (March, May), summer high-temperature period (July), and autumn organic matter accumulation (September-November). With 36 sites × 6 temporal replicates = 216 spatiotemporal samples, our dataset substantially exceeds the recommended sample size for structural equation modeling (Kline, 2015). This sample size is also comparable to or exceeds those in published studies from similar regions (e.g., Gao et al., 2024).

Reference:

Soranno P A, Cheruvelil K S, Bissell E G, et al. Cross‐scale interactions: quantifying multi‐scaled cause–effect relationships in macrosystems[J]. Frontiers in Ecology and the Environment, 2014, 12(1): 65-73.

Blois J L, Williams J W, Fitzpatrick M C, et al. Space can substitute for time in predicting climate-change effects on biodiversity[J]. Proceedings of the national academy of sciences, 2013, 110(23): 9374-9379.

Kline R B. Principles and practice of structural equation modeling[M]. Guilford publications, 2023.

Gao W, Xiong F, Lu Y, et al. Water quality and habitat drive phytoplankton taxonomic and functional group patterns in the Yangtze River[J]. Ecological Processes, 2024, 13(1): 11.

 

Comment 3: Missing December data undermine the analysis of seasonal patterns.

Response 3: The absence of December data was due to practical constraints (extreme low temperatures and ice formation preventing safe field operations). However, this does not compromise our seasonal pattern analysis. Our sampling covered four complete seasons with all major seasonal transitions: winter (January), spring (March, May), summer (July), and autumn (September, November). All critical ecological events were well-represented, including spring bloom (May), summer thermal stratification (July), autumn organic matter accumulation (September-November), and winter low-productivity baseline (January). In the middle-lower Yangtze region, December typically exhibits environmental conditions highly similar to January (low temperature, minimal primary productivity), making the missing single month inconsequential for characterizing winter conditions or annual patterns. Sampling frequency for season is widely accepted for seasonal pattern analysis in subtropical lake systems (Gao et al., 2024; Yan et al., 2023).

Reference:

Gao W, Xiong F, Lu Y, et al. Water quality and habitat drive phytoplankton taxonomic and functional group patterns in the Yangtze River[J]. Ecological Processes, 2024, 13(1): 11.

Yan G, Yin X, Wang X, et al. Effects of summer and autumn drought on eutrophication and the phytoplankton community in Dongting Lake in 2022[J]. Toxics, 2023, 11(10): 822.

 

Comment 4: Inadequate spatial replication, absence of depth stratification, and lack of reference sites further limit interpretive power.

Response 4: We respectfully but firmly disagree with these characterizations, which appear to reflect misunderstanding of our study design and system characteristics.

Regarding spatial replication: Our design encompasses 36 sites across 9 water bodies (4 rivers, 5 lakes), yielding 216 spatiotemporal samples. This sampling intensity exceeds published studies in similar regions: Yan et al. (2022) used 47 sites in Dongting Lake basin, while Gao et al. (2024) employed 90 sites in the middle-lower Yangtze.

Regarding depth stratification: The studied river-lake systems are predominantly shallow and polymictic, with most lakes having average depths < 5 m and experiencing frequent vertical mixing due to wind action and water flow in this subtropical monsoon climate. Our sampling strategy specifically targeted the benthic-pelagic interface by collecting surface water samples (0.5 m depth, representing the euphotic zone), sediment samples (0-5 cm), and benthic macroinvertebrates from sediments. This design is optimal for investigating benthic-pelagic coupling, as the interface between pelagic and benthic zones is more critical than mid-water column stratification for understanding coupling processes.

Regarding reference sites: Our study employs a gradient analysis approach rather than reference-impact comparison, which is neither feasible nor appropriate for our research questions. The 36 sites inherently span environmental gradients from relatively pristine to moderately disturbed conditions, from oligotrophic to eutrophic states, and from lotic to lentic systems. In the middle-lower Yangtze River Basin, truly "undisturbed reference sites" are essentially nonexistent due to millennia of human activity and recent hydrological modifications. Contemporary macrosystem ecology increasingly favors gradient-based designs over reference-site approaches, particularly in human-dominated landscapes.

 

Comment 5: The identification of phytoplankton only to the genus level also reduces ecological precision.

Response 5: We appreciate the reviewer's concern regarding taxonomic resolution. We would like to clarify that phytoplankton were identified to the species level whenever possible. However, for certain taxa where morphological characteristics were ambiguous or diagnostic features were not clearly visible under light microscopy, identification was retained at the genus level to avoid misidentification errors. This approach prioritizes taxonomic accuracy over artificial precision. Moreover, our research focuses on identifying ecosystem-level drivers of biodiversity and benthic-pelagic coupling, not species-specific physiological tolerances.

Importantly, genus-level classification does not compromise ecological precision in our study. Sodré et al. (2020) conducted a comprehensive analysis across 1,010 lakes in the contiguous USA and demonstrated that genus-level classifications have good concordance with species responses to environmental variation and are particularly effective at detecting differences in plankton assemblages among ecoregions. Their study further revealed that genus-level, taxonomic, and functional analyses yielded very similar results in terms of responses to environmental variables, suggesting that genus-level resolution adequately captures major ecological patterns. Moreover, in large-scale datasets, genus-level records provide greater consistency and are more robust against human identification errors compared to forced species-level determinations. The concept of "taxonomic sufficiency" supports that organisms should be identified to a level matching the information needs of the study, and for community-level analyses examining responses to environmental drivers, coarse classifications at the genus level provide reliable predictive power while conserving resources.

Therefore, our mixed species- and genus-level approach represents a balanced strategy that maximizes both taxonomic accuracy and ecological interpretability.

Sodré E D O, Langlais-Bourassa A, Pollard A I, et al. Functional and taxonomic biogeography of phytoplankton and zooplankton communities in relation to environmental variation across the contiguous USA[J]. Journal of plankton research, 2020, 42(2): 141-157.

 

Comment 6: Several statistical methods (including redundancy analysis and structural equation modeling) are applied without sufficient rigor or justification.

Response 6: We thank the reviewer for this constructive suggestion. However, our statistical analyses followed rigorous, well-established protocols with appropriate preliminary tests. The perceived lack of rigor stems from insufficient detail in our Methods description rather than actual methodological deficiencies.

For Redundancy Analysis (RDA): We explicitly conducted Detrended Correspondence Analysis (DCA) prior to RDA to determine the appropriate ordination method, selecting linear RDA because all gradient lengths were <3. This is the recommended procedure (Braak & Šmilauer, 2012; Legendre & Legendre, 2012). We employed forward selection with Monte Carlo permutation tests (999 permutations), retaining only variables with P < 0.05, which prevents overfitting and controls Type I error through permutation-based significance testing following the protocol. Environmental variables were standardized to zero mean and unit variance prior to analysis. These procedures constitute the standard for RDA in ecological studies (Borcard et al., 2018, Numerical Ecology with R).

For Structural Equation Modeling (SEM): Our Methods section 2.3 explicitly states that multicollinearity was assessed using Variance Inflation Factors (VIF), with variables having VIF > 10 removed to minimize redundancy—the widely accepted standard threshold. Our model was theory-driven with variables organized into three conceptually distinct hierarchical layers reflecting known ecological processes. We incorporated random effects to account for the nested data structure (multiple samples within sites), which is a key advantage of piecewise SEM over traditional covariance-based SEM.

 

Comment 7: Model evaluation lacks standard fit indices, and the interpretation of correlations as causal relationships is methodologically unsound.

Response 7: These concerns reflect fundamental misunderstanding of piecewise structural equation modeling methodology and the theoretical basis for causal inference in ecological systems. We address each point directly.

Regarding model fit evaluation: The piecewise SEM uses a fundamentally different approach than traditional covariance-based SEM. Traditional SEM fits a global covariance matrix and uses goodness-of-fit indices such as CFI, RMSEA, and TLI. In contrast, piecewise SEM decomposes the path model into a series of local regression models (lm, glm, lmer) and evaluates fit through tests of directed separation (d-separation), which examine whether variables assumed to be conditionally independent in the model are indeed statistically independent in the data. The appropriate fit metric for piecewise SEM is Fisher's C statistic, which aggregates all d-separation tests. A non-significant P-value (P > 0.05) indicates good model fit, meaning no significant missing paths exist. Our model achieved Fisher's C = 43.132, P = 0.261, indicating excellent fit. Traditional fit indices are mathematically inapplicable to piecewise SEM because it does not estimate a global covariance matrix. This is not a weakness but a fundamental characteristic of the methodology. The theoretical development and state that Fisher's C is the appropriate and sufficient measure of overall model fit for piecewise structural equation models. We chose piecewise SEM because it accommodates complex hierarchical data structures, incorporates random effects, handles non-normal response variables, and is more flexible and robust for ecological data. This approach is widely used in high-impact ecological research (Lefcheck, 2016).

Regarding causal inference: We appreciate the reviewer's attention to the rigor of causal inference in structural equation modeling. However, we respectfully disagree with the assertion that our causal interpretations are methodologically unsound. This criticism appears to reflect a common misconception about SEM that has been extensively addressed in the methodological literature.

As Pearl and Bollen (2013) clarify in their influential paper "Eight Myths About Causality and Structural Equation Models," SEM has never claimed to derive causality solely from correlations. Rather, SEM is a "reasoning engine" that requires three essential inputs: (1) qualitative causal assumptions based on theory, (2) specific research questions, and (3) empirical data for parameter estimation. The critical distinction, often missed by critics, is that SEM does not discover causal relationships from data; it tests whether data are consistent with theoretically-specified causal hypotheses. Duncan (1966) and Goldberger (1973) established early on that structural equations represent "causal linkages rather than mere empirical associations," but these linkages are researcher-specified assumptions that must be justified independently of the model fitting process.

Our study adheres strictly to the requirements for valid causal inference from SEM as outlined in contemporary methodological guidance. Specifically:

1. Theory-Driven A Priori Specification: Our model structure was specified before data analysis based on well-established ecological mechanisms documented in the literature (physical drivers → nutrient dynamics → biological responses). This follows Pearl's requirement that causal assumptions (A) must be input into the model rather than inferred from it.
2. Clear Temporal and Logical Priority: We established causal ordering through multiple lines of evidence. Physical variables (e.g., turbidity, water flow) temporally and logically precede chemical changes (nutrient concentrations), which in turn precede biological responses (community composition). This temporal sequencing provides strong support for the assumed causal direction, as later events cannot cause earlier ones—a fundamental principle of causal inference.
3. Statistical Validation via d-Separation Tests: For piecewise SEM, the appropriate validation method is Shipley's d-separation test rather than global fit indices. Our Fisher's C statistic (p > 0.05) confirms that no theoretically-unjustified pathways were omitted from the model, supporting the structural validity of our causal assumptions. Recent methodological work has established d-sep tests as the gold standard for validating structural causal models in ecological contexts.
4. Mechanistic Plausibility: Each pathway in our model corresponds to known ecological mechanisms. For example: turbidity → total phosphorus (TP) operates through sediment resuspension releasing particle-bound nutrients; TP → phytoplankton richness through limitation relief in phosphorus-limited systems; phytoplankton → benthic richness through settling organic matter affecting substrate quality and food availability. These are not post-hoc interpretations but established processes in aquatic ecology.

As Bollen and Pearl note, if the criticism that "SEM cannot support causal inference from observational data" were accepted, it would invalidate not only decades of SEM research in top-tier journals, but also much of regression-based causal inference in epidemiology, econometrics, and social sciences. The appropriate question is not whether SEM can support causal inference, but whether the specific causal assumptions in a given study are adequately justified.

Our causal interpretations rest on the same logical foundation as accepted causal research in ecology and other observational sciences: explicit theoretical assumptions, validated through statistical testing and evaluated against empirical data. We have clearly stated these assumptions, provided theoretical and empirical justification for them, and validated the resulting model structure using appropriate statistical methods. This approach aligns with current best practices for causal inference in ecological SEM.

References:

Lefcheck, J.S. piecewiseSEM: Piecewise Structural Equation Modelling in r for Ecology, Evolution, and Systematics. Methods Ecol. Evol. 2016, 7, 573–579, doi:10.1111/2041-210X.12512.

Bollen K A, Pearl J. Eight myths about causality and structural equation models[M]//Handbook of causal analysis for social research. Dordrecht: Springer Netherlands, 2013: 301-328.

Pearl J. The causal foundations of structural equation modeling[J]. Handbook of structural equation modeling, 2012: 68-91.

Shipley B. Cause and correlation in biology: A user's guide to path analysis, structural equations and causal inference with R[M]. Cambridge university press, 2016.

 

Comment 8: The manuscript is generally understandable and employs appropriate scientific terminology, but numerous grammatical errors, inconsistent terminology, awkward phrasing, and stylistic inconsistencies reduce its clarity and professional tone. The writing reflects non-native English authorship with solid technical vocabulary yet uneven grammatical control and idiomatic accuracy. Professional language editing or review by a native English-speaking co-author is strongly recommended. Specific issues include typographical errors (“Furthemore” instead of “Furthermore,” “Abstact” instead of “Abstract”), verb tense inconsistencies, and spacing problems. Style can be improved by reducing passive constructions, removing redundant or clichéd expressions, simplifying complex sentences in the Results section, and standardizing article use and acronym formatting.

Response 8: We sincerely thank the reviewer for the detailed feedback on the manuscript's language. We have corrected all the specific errors noted by the reviewer, including typos and grammatical inconsistencies. As recommended, the entire revised manuscript has been professionally polished by a language editing service to improve its clarity, style, and overall readability. We are confident that the language now meets the high standards required for publication.

 

We believe these revisions will substantially improve the manuscript's clarity while maintaining the scientific rigor and validity of our original findings.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript focuses on the river-lake ecosystems in the middle and lower reaches of the Yangtze River. Through multi-site, multi-temporal monitoring, multivariate statistics (PCA, RDA), and structural equation modeling (SEM), it systematically explores the cascading relationships among physical factors, nutrient cycling, and biological responses, and identifies the "phosphorus paradox" and benthic-pelagic coupling characteristics. The research topic aligns with the ecological conservation needs of the Yangtze River Basin, holding clear scientific value and application potential. The study design is relatively rigorous, with sufficient data volume, and the conclusions demonstrate certain innovation at the regional scale. However, there is room for optimization in terms of research limitations, model integrity, depth of result interpretation, and detail presentation, requiring further supplementation and improvement.

Specific Revision Suggestions

• Supplement details and rationality explanation of sampling design: The selection of sampling sites only mentions "representativeness and accessibility" but fails to clarify whether different gradients of human disturbance are covered. Human activities are key influencing factors for the river-lake ecology in the middle and lower reaches of the Yangtze River. It is recommended to supplement the land use types of each sampling site and the main surrounding disturbance sources, and explain how this design supports the research objective of "environmental heterogeneity"
• Supplement the rationality basis for sediment sampling depth
• Optimize data presentation and supplement statistical indicators:.
• Supplement key classic literatures and regional studies

Comments for author File: Comments.pdf

Author Response

Comment 1: Supplement details and rationality explanation of sampling design: The selection of sampling sites only mentions "representativeness and accessibility" but fails to clarify whether different gradients of human disturbance are covered. Human activities are key influencing factors for the river-lake ecology in the middle and lower reaches of the Yangtze River. It is recommended to supplement the land use types of each sampling site and the main surrounding disturbance sources, and explain how this design supports the research objective of "environmental heterogeneity".

Response 1: We thank the reviewer for this valuable comment, which has prompted us to provide a more comprehensive description of our sampling design rationale. There may be certain ambiguities in the description of the methodology section in our original text. We have substantially added the Methods section to explicitly address the spatial representativeness of our sampling framework and the anthropogenic disturbance gradients captured across the 36 sampling sites.

Regarding the spatial and ecological representativeness of the sampling design, we have added detailed descriptions explaining how site selection explicitly considered the morphological and ecological characteristics of each water body.

To address the reviewer's concern regarding anthropogenic disturbance gradients, We used remote sensing and geographic data to find that the sampling actually exhibits a certain environmental gradient, which had already been taken into account at the initial stage of selecting sampling points.

Modified sections: Methods 2.1 (Paragraph 3-5), it was confirmed that all sampling points cover various gradients from low anthropogenic disturbance to high anthropogenic disturbance, exhibiting a certain degree of environmental heterogeneity.

 

Comment 2: Supplement the rationality basis for sediment sampling depth.

Response 2: We appreciate this comment and have added a detailed explanation of the scientific rationale for selecting the 0-5 cm sediment sampling depth. The revised Methods section now explicitly states that this surface layer was selected based on several important considerations. First, the 0-5 cm surface layer represents the biogeochemically active zone where benthic-pelagic coupling processes are most intensive, with target benthic macroinvertebrates primarily residing at depths of a few centimeters in shallow lake systems. Second, the 0-1 cm interval is recognized as critical for capturing sediment-water interface dynamics, and the 0-5 cm depth encompasses the primary zone of recent organic matter deposition and biogeochemical transformations. Third, preliminary field observations confirmed that most benthic macroinvertebrates collected in our study system inhabited this surface layer. These rationales are supported by established protocols for benthic sampling in shallow lake and river systems, and the chosen depth is consistent with standard methodologies employed in similar benthic-pelagic coupling studies.

Modified section: Methods 2.2 (Paragraph 2), with supporting references [34,35] added to the reference list.

 

Comment 3: Optimize data presentation and supplement statistical indicators.

Response 3: We have carefully reviewed all figures and tables throughout the manuscript to optimize data presentation and ensure complete statistical reporting. Figure 1 has been revised to improve image clarity and resolution. Figure 2 has been updated with a more comprehensive figure caption.

Modified sections: Figure captions for Figures 1-7.

 

Comment 4: Supplement key classic literatures and regional studies.

Response 4: We have substantially strengthened the literature review by incorporating seminal works on benthic-pelagic coupling and recent regional studies specific to the Yangtze River Basin. These additions appear primarily in the sixth and seventh paragraphs of the Introduction, where we contextualize our study within the broader framework of Yangtze Basin environmental challenges and management priorities. Furthermore, we have added a new Discussion section 4.4 titled "Implications for sustainable management" that explicitly connects our findings to Sustainable Development Goals and China's national policies for Yangtze River protection.

Modified sections: Introduction paragraphs 6-7, Discussion section 4.4, and References (15 new citations added).

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript is based on the theory of 'multi-scale driving mechanisms', the research object is divided into three levels: physical factors, nutrient cycling , and biological response . Their interactions are analyzed through cascade path analysis, with particular attention to benthic-phytoplankton coupling and the 'phosphorus paradox' phenomenon. Measuring 20 environmental parameters and biological community data, with multivariate statistics (PCA/RDA) and structural equation modeling (SEM) to quantify causal relationships between variables; Kruskal-Wallis test was used to analyze spatial differences, and time series analysis was employed to reveal seasonal patterns. The results elucidate the cascade mechanisms of physical factors, nutrient cycling, and biological response in the river-lake system of the middle and lower reaches of the Yangtze River, providing a quantitative framework for the management of freshwater ecosystems. I think this is an excellent manuscript. Minor revisions  needed before publication.

（1）Fig.1 needs to revise (It is recommended to enlarge the image and clearly display the sampling point locations and sample numbers，etc)

(2)references have many mistakes, need to be revised one by one(eg: 1 authors name; The full name of the journal must be written.).

Author Response

Comment 1:Fig.1 needs to revise (It is recommended to enlarge the image and clearly display the sampling point locations and sample numbers，etc)

Response 1: We would like to express our sincere gratitude to the reviewer for the suggestions on image revisions. At present, Figure 1 has been completely reconstructed to enhance the clarity of each location point.

 

Comment 2: references have many mistakes, need to be revised one by one(eg: 1 authors name; The full name of the journal must be written.).

Response 2: We would like to express our sincere gratitude to the reviewer for the meticulous observations. At present, we have reviewed and verified all the references throughout the manuscript to ensure they fully comply with the formatting requirements of the Sustainability journal.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

This second version of the manuscript, titled “Multi-level driving mechanisms: cascading relationships among physical factors, nutrient cycling, and biological responses in the Yangtze river-lake ecosystems,” offers an in-depth examination of how environmental conditions interact with biological communities in the middle and lower Yangtze River Basin. The authors have made substantial progress since earlier drafts, especially in clarifying the framework for sustainable development and expanding the methodological explanations. By explicitly linking their findings to the Sustainable Development Goals (particularly SDGs 6, 14, 15, and 17), the study gains clear policy relevance. The combination of multivariate analyses and structural equation modeling is well-suited to reveal the cascading mechanisms within this complex system, while the identification of the “phosphorus paradox” and the strong relationship between turbidity, phosphorus, and biological responses adds meaningful insight into shallow lake dynamics. The broad spatial coverage of 36 sites also gives the dataset solid representativeness.

Despite these advances, some issues still need refinement. The justification for selecting a sediment sampling depth of 0–5 cm remains weak and would benefit from system-specific references or data showing that this layer captures most benthic macroinvertebrate activity in the Yangtze. Likewise, the bimonthly sampling frequency may not capture all seasonal dynamics; the authors should acknowledge this limitation and discuss its potential influence on their interpretation of benthic–pelagic interactions. Given only six sampling events per year, it would also be useful to evaluate the statistical power of temporal analyses. The discussion of policy implications is well-conceived but could go further by identifying specific management thresholds (for example, turbidity limits that could inform Three Gorges Dam operations).

The methodological section has improved with the addition of land-use characterization, which strengthens the treatment of anthropogenic influences. Still, the presentation of structural equation modeling results would benefit from including detailed fit statistics and p-values in a supplementary table, and confidence intervals in Figure 7 would make the visual interpretation stronger. The section addressing the “phosphorus paradox” is clear and thoughtfully contextualized through comparison with other systems, such as Nanwan Reservoir. Minor issues persist, including some grammatical inconsistencies and figure resolution problems. The data availability statement should specify a concrete repository or access procedure. Overall, the manuscript makes a valuable contribution to understanding the cascading dynamics of river–lake ecosystems under a sustainability lens. With modest revisions to refine methodological justification and management guidance, it stands as a strong candidate for publication.

Comments on the Quality of English Language

The manuscript is written in generally good English, but several areas require refinement. Verb tenses are occasionally inconsistent, such as alternating between “exhibited” and “exhibits,” and a few missing articles and subject-verb agreement errors appear throughout. Some sentences, particularly in the introduction, are overly long and contain multiple subclauses, making them harder to follow. The phrasing could also be more direct in places, and hyphenation should be applied consistently (for instance, “river-lake” versus “river lake”). While the technical writing is solid, with accurate scientific terminology, clear statistical reporting, and correct abbreviation use, certain expressions should be revised for smoother readability, such as replacing “The top 0–5 cm layer of sediment was collected” with “The upper 5 cm of sediment was collected.” A final professional edit by a native English speaker experienced in scientific writing would help ensure stylistic consistency and overall clarity before publication.

Author Response

Comment 1: The justification for selecting a sediment sampling depth of 0–5 cm remains weak and would benefit from system-specific references or data showing that this layer captures most benthic macroinvertebrate activity in the Yangtze.

Response 1: We sincerely appreciate the reviewer's insightful comment regarding the need for system-specific justification of our 0-5 cm sediment sampling depth. We acknowledge that our original manuscript lacked sufficient regional context from Yangtze River Basin studies. We have substantially strengthened this section in section 2.2 Paragraph 2:

The selection of the upper 0-5 cm sediment layer for sampling in Yangtze River basin studies is scientifically justified based on biogeochemical and ecological considerations. Previous investigations in the Yangtze River system have consistently adopted this sampling depth for assessing organic carbon dynamics, with forest stream sediment studies in the upper Yangtze River successfully employing 0-5 cm samples to quantify organic carbon concentrations and stocks across seasonal variations [39]. Studies in similar Chinese shallow lakes have consistently demonstrated that benthic macroinvertebrates predominantly inhabit the surface sediment layer. In Lake Taihu, the third-largest freshwater lake in China with comparable depth and substrate characteristics to our study lakes, 39.25% and 24.87% of total benthic abundance occurred within the 0-5 cm and 5-10 cm sediment layers respectively, with density markedly declining with increasing depth [40]. In other study, additional Peterson grab sediment samples were collected at each sampling station in Lake Taihu to determine the total nitrogen and total phosphorus contents in the surficial sediments, with the sampling depth specified as approximately the upper 5 cm [41]. Moreover, .in coastal sediment research, the study of the La Paz lagoon in Northwest Mexico demonstrated that 0-5 cm surface samples effectively capture pharmaceutical residues associated with wastewater inputs, supporting this interval as an appropriate depth for assessing contaminant retention[42]. These findings collectively support the 0-5 cm sampling protocol as an ecologically meaningful and methodologically sound approach for characterizing active sediment biogeochemistry in river and lake systems.

Action taken: Method section 2.2 Environmental variables and biological community ( Paragraph 2 in the revised manuscript).

 

Comment 2: Likewise, the bimonthly sampling frequency may not capture all seasonal dynamics; the authors should acknowledge this limitation and discuss its potential influence on their interpretation of benthic–pelagic interactions.

Response 2: We appreciate this thoughtful comment. While our bimonthly design was appropriate for our research objectives, we acknowledge that higher temporal resolution could reveal additional dynamics, and it is important to discuss this transparently. We have revised in section 4.1 Heterogeneity in the river-lake systems Paragrapha 4 and was shown in below:

While our bimonthly sampling design (six events per year) successfully captured broad spatiotemporal patterns across multiple river-lake systems, we acknowledge that higher temporal resolution could have revealed additional ecological dynamics. Comparative studies have demonstrated that bimonthly sampling regimes (8 sampling events per year) may capture >75% of seasonal variance, whereas traditional seasonal sampling (4 events per year) typically explains <65% of total variance [54]. Moreover, increased sampling frequency has been shown to strengthen correlations between biotic data and seasonal variables such as temperature, salinity, and nutrients [54]. For rapidly growing planktonic taxa, population abundances can fluctuate by a factor of 10 or more over periods of 1-2 weeks, imposing important constraints on sampling design and indicating that monitoring at longer intervals may omit critical ecological information [55,56]. Additionally, episodic carbon flux events in benthic-pelagic systems, such as those associated with large diatom colonies, are temporally and spatially scarce and therefore difficult to capture through fixed-interval sampling [57]. Consequently, our bimonthly sampling frequency may not have fully captured rapid interactions between benthic and pelagic communities or their short-term responses to environmental fluctuations, potentially underestimating seasonal variations in benthic-pelagic coupling strength and missing transient but ecologically significant coupling events.

 

Comment 3: The discussion of policy implications is well-conceived but could go further by identifying specific management thresholds (for example, turbidity limits that could inform Three Gorges Dam operations).

Response 3:

We sincerely appreciate the reviewer's constructive suggestion regarding the provision of specific management thresholds. While our structural equation model quantifies the turbidity-phosphorus-biology cascade (standardized coefficient = 0.717, P < 0.001), we recognize that precise numerical thresholds require caution given the observational study design and system heterogeneity. We have revised the text to incorporate a focused, data-driven recommendation.

This approach adds a single, focused sentence that directly links our quantitative finding to operational guidance without overstating what the data can support. The modification is intentionally conservative because our observational study design does not include experimental manipulation of turbidity levels, and the specific turbidity values from our field measurements are system-specific. By referencing the strongest quantitative relationship from our SEM and connecting it to ecologically relevant timing considerations already evident in our seasonal analysis, we provide meaningful guidance while acknowledging that precise numerical thresholds would require additional experimental validation across controlled turbidity gradients. This balanced approach responds constructively to the reviewer's request while maintaining the scientific integrity appropriate for correlational field data.

Section 4.4, policy recommendation 1 have beend revised to: Develop ecosystem-based TGD operation protocols maintaining hydrological variability within ranges supporting native biodiversity while minimizing sediment resuspension, particularly avoiding prolonged high-turbidity discharge during ecologically sensitive periods given the strong turbidity-to-TP pathway (standardized coefficient = 0.717; SDG 15.1, 15.5)

 

Comment 4: Still, the presentation of structural equation modeling results would benefit from including detailed fit statistics and p-values in a supplementary table, and confidence intervals in Figure 7 would make the visual interpretation stronger.

Response 4: We thank the reviewer for this constructive suggestion. We have prepared a comprehensive table in supplementary information (Table S1) containing all SEM statistical details, including non-standardized estimates, standard errors, 95% confidence intervals, standardized estimates, P-values, and complete model fit statistics (Fisher's C, AIC, degrees of freedom, and marginal R² values).

Regarding Figure 7, we note that standardized coefficients with P-values and non-standardized coefficients with confidence intervals represent two alternative approaches for presenting SEM results, not complementary ones. The former is the prevailing convention in ecological SEM literature because standardized coefficients (ranging -1 to 1) enable direct comparison of relative effect sizes across variables with different measurement units, which is essential for interpreting complex ecological relationships. Combining both approaches in a single figure would be redundant and potentially confusing, as they convey similar statistical information (significance) but serve different interpretive purposes. We believe this combination—visual emphasis on standardized effects in the main figure and comprehensive statistical details in the supplementary table—provides the most effective communication of our SEM findings while maintaining the interpretability that the reviewer appropriately emphasized.

Action taken: Supplementary Table

Table S1. Path coefficients from structural equation modeling of benthic, planktonic, and environmental variables in river-lake ecosystems

Response	Predictor	Estimate	SE	95% CI	P	Std.Estimate
TP	Turb	0.0026	0.0002	[0.0022, 0.0030]	0	0.7169 ***
TP	TOC	-0.0001	00.00003	[−0.00016, −0.00004]	0.0013	-0.1563  **
TN	TOC	-0.0007	0.0003	[−0.0013, −0.0001]	0.0156	-0.1558   *
TN	TP	1.2709	0.4015	[0.484, 2.058]	0.0018	0.2119  **
phyto_richness	DO	0.8602	0.1427	[0.580, 1.140]	0	0.3685 ***
phyto_richness	TP	13.1799	2.4442	[8.389, 17.971]	0	0.3393 ***
phyto_richness	TN	-0.5416	0.4051	[−1.336, 0.252]	0.1829	-0.0836
phyto_pielou	Turb	-0.001	0.0004	[−0.0018, −0.0002]	0.0202	-0.2221   *
phyto_pielou	TP	0.0924	0.1213	[−0.145, 0.330]	0.4474	0.0721
benthos_richness	SMC	0.0062	0.0026	[0.0011, 0.0113]	0.018	0.1588   *
benthos_richness	TOC	-0.0023	0.0007	[−0.0037, −0.0009]	0.0006	-0.2343 ***
benthos_richness	TP	-2.6511	0.9619	[−4.536, −0.766]	0.0065	-0.1941  **
benthos_richness	DO	-0.0761	0.0652	[−0.204, 0.052]	0.2452	-0.0927
benthos_richness	EC	-0.01	0.0064	[−0.023, 0.003]	0.1204	-0.1126
benthos_richness	phyto_richness	0.0815	0.0256	[0.031, 0.132]	0.0017	0.2317  **
benthos_richness	phyto_pielou	0.0445	0.6902	[−1.308, 1.397]	0.9486	0.0042

Note: *** P < 0.001; ** P < 0.01; * P < 0.05.

Model fit: Fisher's C = 43.132, P = 0.261, df = 38, AIC = 2057.848.

Marginal R²: TP = 0.52, TN = 0.08, Phyto richness = 0.22, Phyto Pielou = 0.03, Benthos richness = 0.12.

SE = standard error; CI = confidence interval; TP = total phosphorus; TN = total nitrogen; Turb = turbidity; TOC = total organic carbon; DO = dissolved oxygen; SMC = soil moisture content; EC = electrical conductivity; Phyto = phytoplankton.

 

Comment 5: The section addressing the “phosphorus paradox” is clear and thoughtfully contextualized through comparison with other systems, such as Nanwan Reservoir. Minor issues persist, including some grammatical inconsistencies and figure resolution problems.

Response 5: We have carefully reviewed the manuscript and identified the grammatical inconsistencies. We will engage the journal's professional English editing service to thoroughly address these language issues and ensure grammatical consistency throughout the manuscript after making sure there are no major modifications.

 

Comment 6: The data availability statement should specify a concrete repository or access procedure.

Response 6: We thank the reviewer for highlighting the need for a more specific data availability statement. We have changed the statement: All data and code supporting the findings of this study are available upon reasonable request from the corresponding author by email.

 

Comment 7: The manuscript is written in generally good English, but several areas require refinement. Verb tenses are occasionally inconsistent, such as alternating between “exhibited” and “exhibits,” and a few missing articles and subject-verb agreement errors appear throughout. Some sentences, particularly in the introduction, are overly long and contain multiple subclauses, making them harder to follow. The phrasing could also be more direct in places, and hyphenation should be applied consistently (for instance, “river-lake” versus “river lake”). While the technical writing is solid, with accurate scientific terminology, clear statistical reporting, and correct abbreviation use, certain expressions should be revised for smoother readability, such as replacing “The top 0–5 cm layer of sediment was collected” with “The upper 5 cm of sediment was collected.” A final professional edit by a native English speaker experienced in scientific writing would help ensure stylistic consistency and overall clarity before publication.

Response 7: We have carefully reviewed the manuscript and acknowledge the identified issues and also note the reviewer's specific example regarding sediment collection description, which we will revise accordingly. Pending confirmation that no major revisions are required, we will engage this journal's professional English editing service to comprehensively address these language issues. This professional editing will ensure: consistent verb tense throughout the manuscript, appropriate article usage and subject-verb agreement, streamlined sentence structures for improved readability, standardized hyphenation (we will consistently use "river-lake" as a compound modifier), and refined phrasing for clarity and directness.

 

Author Response File: Author Response.pdf