This manuscript examines coral reef production rates (Net Ecoystem Production) in Hong Kong across a gradient of coral cover under described “marginal conditions”. The methods are sound and the introduction illustrates the need for these types of studies. However, the conclusions are not supported by the results. Firstly, the description of marginal conditions is too broad. Of the proposed factors that contribute to marginality, such as nutrients or rainfall, many were either not measured or did not significantly differ between sites or previous studies. The paper needs to be reframed as a case study on reef NEP specifically in low light conditions (due to turbidity and latitude).

Secondly, the paper is missing estimates of reef accretion or calcification which are critical to making conclusions about overall community metabolism and reef health. We simply do not know if changes in NEP have anything to do with the corals without simultaneous measurements of calcification. Coral cover of 30 – 60% is above average to high. 0% algae cover is very rare. There is an interesting story here why these reefs are so healthy (high coral cover, no algae) yet are net respiring and persist under low light conditions. This is briefly alluded to on L496 and should be the focus of the paper.  

I encourage the authors to re-write this manuscript and avoid broad brushing terms like “marginal conditions” and “community metabolism”. The authors measured NEP on a low light reef. There needs to be a stronger tie-in to other literature values of NEP, light, and coral cover. There are multiple studies provided below that have these values as well as insight to how reef NEP may function under bleaching, a co-occuring stressor during the study. The values of NEP are extremely low relative to other reefs, this also needs to be addressed. 

Overall I do not support the publication of this manuscript until the authors are able to focus the conclusions specifically on reef production and low light. Statements cannot be made about reef health or function without calcification rates.

L7 – Are these anthropogenic stressors or local climactic variability?

L14 – mmol missing “l”

L15 – There are multiple studies that look at metabolic variability during bleaching periods, on degraded reefs, and in subtropical conditions, all arguably “marginal” conditions. You go on to cite several studies that work in these locations.

L21- too many “ands” in this sentence

L23 – remove changeable

L39 – The linkage obstacles between benthic metabolism and overlying water chemistry referred to in these citations are due to water volume and residence time. In-situ measurements of DO or pH to calculate reef metabolism are restricted to reefs where there is a measurable change in these parameters. Your study does not address this, as it relies on chemical anomaly methods. Remove this line

L41- unclear. You are measuring NEP with two different approaches? Methods indicate you are using one approach, chemical changes in the overlying water column.

Figure 1 – Higher resolution image please

L122 -125 – This belongs in the results

L189 – This does not need to be its own section. Consider adding to the statistical analyses section.

L308 – Your PAR units are missing a spatial value, usually m-2

L309 – The SD of these values overlap, was DO and salinity significantly different between seasons?

L313 – Remain consistent with your DO units, either mg L-1 or umol kg-1. Choose one.

Figure 3 – Text and frame size is too small. Consider placing the wet season plot below to increase size to see trends and values

L322 -327 – 36 and 65% coral cover with no algae is a very healthy reef and compares more to usual studies in optimal conditions, not marginal conditions.

L329 – Unclear why the lowest coral cover makes it a “control” site. It seems instead to be your “treatment” site since you are trying to examine how marginal conditioned reefs compare to your standard reef, which, in the literature, would be closer to Tung Ping Chau at 36% coral. In any case you aren’t performing experimental manipulations, so you don’t have a control. You have an environmental gradient with Tung Ping Chau being most comparable to average coral reefs.

L341- Here and elsewhere you need a space between mmol and O2

L342 -343 – These are extremely low rates of production and respiration, especially for a reef with 30 60% coral cover. Generally shallow benthic environments are closer to the order of 10 – 30 mmol O2 m-2 h-1. What were your changes in DO for these calculations?

L392- The majority of studies have not used autonomous methods, this is only a relatively recent development. The majority

L406 – Please use PAR units here that align with your measurements of umol m-2 s-1

L414 – These citations suggest that you are not the first study to look at reef production in marginal environments

L430 – You say above that rainfall did not significantly differ between seasons but here you are saying it does?

L437-L454 – Need a paragraph break in here.

L471 – Bleaching events have been shown to have no effect on NEP. Please review the following papers and include with your discussion of how anthropogenic stressors have been measured to affect reef metabolism:

Lantz, C. A., Leggat, W., Bergman, J. L., Fordyce, A., Page, C., Mesaglio, T., & Ainsworth, T. D. (2022). Will daytime community calcification reflect reef accretion on future, degraded coral reefs?. Biogeosciences, 19(3), 891-906.

McMahon, A., Santos, I. R., Schulz, K. G., Scott, A., Silverman, J., Davis, K. L., and Maher, D. T.: Coral Reef Calcification and Production After the 2016 Bleaching Event at Lizard Island, Great Barrier Reef, J. Geophys. Res.-Ocean., 124, 4003–4016, https://doi.org/10.1029/2018JC014698, 2019. 

Pisapia, C., Hochberg, E. J., and Carpenter, R.: Multi-Decadal Change in Reef-Scale Production and Calcification Associated With Recent Disturbances on a Lizard Island Reef Flat, Front. Mar. Sci., 6, p. 575, https://doi.org/10.3389/fmars.2019.00575, 2019. 

L530 – You did not measure calcification, remove.

L536 – We don’t know this because you did not measure calcification or net accretion. As you state in L496, its possible that coral accretion remained unaffected due to enhanced heterotrophy. It is very difficult to make assertions about reef health, function, or “community metabolism” with only estimates of NEP.

L543 – You cannot make assertions about the opportunity for coral growth because you did not measure reef accretion rates. With coral coverage in 30 – 60% range, these are healthy reefs obviously capable of growing. There is a missing piece here why a reef would have such high coral cover yet be net heterotrophic.

Reviewer Response: This study presents in situ measurements of net ecosystem production (NEP) at three coral reef sites in Hong Kong. The goal of the study was to characterize spatial and seasonal variability in NEP trends across gradients in environmental conditions and benthic community composition. Although the datasets provide valuable insight into spatiotemporal variability of NEP within a region that remains relatively understudied, I have several concerns regarding the interpretation and framing of the results.

In particular, the manuscript tends to overextend the interpretation of NEP as a proxy for coral community metabolism and ecosystem functioning without adequately acknowledging the need for complementary metabolic measurements (net ecosystem calcification; NEC) to fully capture ecosystem-scale processes. Additionally, several aspects of the manuscript would benefit from further development, particularly the discussion section where many interpretations are presented without sufficient mechanistic explanation or supporting context. Finally, the reported NEP rates are considerably lower than those in prior studies, and are surprisingly similar between the 66% and 4% coral cover sites. This is a key result that warrants more interpretation and discussion than is currently provided in the manuscript.

Overall, I think this study has merit but I recommend major revisions to address these conceptual, methodological, and interpretive concerns. Please see all the specific comments below.

Specific Comments

Abstract

General: I appreciate the concise abstract but if word count allows for it, I suggest including some additional details about the three sites to support the characterization that the coral communities are “persisting under marginal environmental conditions”. For example, briefly mentioning coral cover trends, community stability, and/or other long-term metrics would help contextualize this claim alongside the short-term metabolic measurements presented.

(Lines 8-9 and Lines 15-17): The use of “adapt” may overstate the biological interpretation of the results. While I suspect the authors likely intended it to mean an adjustment to marginal environmental conditions, it can also imply improved fitness or performance. The NEP measurements identified net heterotrophy at the three sites and the absence of NEC data to accompany these findings makes it difficult to assess overall ecosystem functioning. I therefore recommend more neutral phrasing such as “how coral ecosystems may respond to changing ocean conditions” or “how coral ecosystems may cope with marginal ocean conditions in the future.”

Introduction

General: This section would benefit from additional citations and further development of the conceptual framework. Many statements are currently supported by a single reference despite a large body of literature addressing the topics. Expanding the in-text citations and more holistically synthesizing existing work would strengthen the introduction and overall framing of the study. This is particularly noticeable in the second and seventh paragraphs, both of which are central to the study’s motivation and should be expanded upon.

(Lines 27-30): I suggest rewording this to retain the key concepts while avoiding a reliance on definitions from a single study that are not universally defined across the broader literature.

(Lines 30-31): This sentence could be interpreted as directly equating reef metabolism to ecosystem functioning. While reef metabolism reflects underlying ecosystem processes, it is more of an indicator of functionality rather than functionality in itself.

(Lines 31-33): The statement “marginal conditions, by definition, limit functionality” is overly broad. While marginal environments are often associated with reduced diversity or altered ecosystem functioning, there are also examples of systems maintaining ecosystem functioning under certain conditions. Consider clarifying or rewording this sentence since “by definition” comes across too strong.

(Lines 38-40): This is an example where including additional citations would strengthen the statement. There are many studies that touch on this point, such as the Takeshita et al. publications that are currently in the reference section but not included here. I also recommend rephrasing “lack of understanding” to instead focus on the need for further characterization of NEP across various reef types and environmental conditions. There isn’t a lack of understanding, rather a need for broader understanding.

(Lines 42-44): NEP is also an important indicator of biomass assimilation. Consider incorporating this aspect to provide a more complete description.

(Lines 45-50): This paragraph mentions that the physical environment impacts NEP but does not expand on that point. Suggest briefly including how key physical drivers, such as residence times, hydrodynamics and tidal variability, can influence metabolic processes.

(Lines 51-52): This comment relates to my earlier point about the definitions of “extremeness” and “marginality” (see Lines 27-30). In those earlier definitions, marginality appears to refer to community characteristics, whereas here, both terms describe environmental conditions. Clarifying whether these terms are distinct or synonymous would improve readability.

(Lines 56-59): Consider providing additional details on the dominant bioeroders at the sites (e.g., parrotfish, urchins, sponges), as this information can provide important context for the measured NEP rates.

(Lines 61-64): This statement would benefit from additional references, if available.

(Lines 65-70): Please expand this paragraph since it is highly important to your study. What type of adjacent anthropogenic activities and why would those impact susceptibility? Additionally, the current wording seems to imply that turbid environments are favorable reef habitat. While turbidity can reduce heat and light stress, those are typically acute benefits that occur during thermal stress events. This context is needed since oligotrophic conditions are generally considered favorable habitats, so clarifying this would strengthen the introduction.

(Lines 79-81): Please be careful when equating NEP with coral community metabolism. While NEP is an important indicator of organic carbon cycling, the absence of NEC limits the ability to fully characterize community metabolism and ecosystem functioning. The current wording is not incorrect, but should be clarified to reflect this distinction.

Methods

General: The methods section would benefit from additional details throughout, particularly regarding site-specific benthic community composition, instrument configuration, and data processing. Clarifying these components would improve reproducibility and confidence in the results. Additionally, the direct comparison between wet and dry season needs further justification, as the sampling periods are separated by ~10 months and three of the four wet season time points could represent seasonal transition periods.

(Lines 91-94): Consider adding mean values (+/- SEM) alongside the reported ranges to provide a clearer comparison across seasons. SEM is more meaningful for data interpretation, so alternatively that could replace the ranges entirely.

(Lines 102-104): This statement would benefit from including quantitative coral cover values for each site.

(Lines 109-110): The use of “most marginal” is too qualitative for the point being made. Suggest specifying the environmental and/or ecological characteristics that support this terminology.

(Table 1 and Lines 120-125): The distinction between wet and dry season requires further support. Several wet season time points could be classified as transitional periods between wet and dry season. The “prolonged wet season in 2022” should be justified with observational data and quantitative values that support this claim, rather than inferred through “prolonged elevated temperatures.” Providing this type of evidence would strengthen confidence in the seasonal comparison.

(Lines 126-135): The use of monthly averaged data may not capture conditions during short deployment periods so this limitation should be acknowledged and carefully considered throughout the manuscript.

(Lines 160-161): The position of the lower sensor height (Z₂ = 8cm) is closer to the benthos than prior gradient-flux studies. Since this can lead to different boundary layer assumptions, consider adding justification for this configuration and clarify how this might influence these assumptions and subsequent flux calculations.

(Lines 178-188): Additional details on instrument specifications (e.g., accuracy, calibration procedures, and relevant settings) would improve reproducibility of the methods.

(Line 184): Clarify whether the sensors recorded continuous measurements, discrete pulse measurements, or a single measurement within each 60-second interval. This distinction is important for evaluating data quality and the filtering steps applied.

(Lines 195-196): Please provide additional detail on the velocity threshold used. This would enable direct comparisons between the thresholds used in prior studies (e.g., <0.5 cm s^-1 in Khrizman et al. 2025).

(Lines 203-206): A substantial proportion of the data was removed during the filtering step. Given this, additional detail on the filtering criteria and justification for data removal would improve confidence in the remaining dataset.

Results

General: The NEP rates are consistently low across all sites, and it is surprising that values are broadly similar between the two with markedly different coral cover (Sharp Island at ~66% and Sham Wan at ~4%). Additional context or discussion would help explain why NEP does not appear to drastically change with coral cover or overall benthic composition, as might be expected. Clarifying whether this reflects methodological constraints would strengthen interpretation of the results.

(Lines 254-255): Please clarify whether the reported values are mean +/- SD or mean +/- SEM. This distinction should be stated clearly the first time it is used in the results section. 

(Lines 281-293): The EPD monitoring data provides useful environmental context, but the limitations should be made clearer. These values are based on periodic surface measurements and monthly averages, which may not fully capture the variability experienced during the shorter 7-14 day deployments.

(Lines 298-299): The lack of significant rainfall differences between the wet and dry season deployment periods suggests that the seasonal distinction may be less clear-cut than presented. I recommend providing strong clarification for how these deployment periods properly represent wet versus dry season conditions.

(Figure 3): For clarity, consider indicating directly in the figure the time periods of each deployment. This could allow for easier interpretation and replace the (# days) sub-text.

(Lines 327-329): This provides important support for why Sham Wan is treated as the marginal site. Suggest introducing some of this context earlier in the manuscript so that readers understand this distinction before reaching the results.

Discussion

General: The interpretation of NEP as a proxy for coral community metabolism and ecosystem functioning should be treated with more caution. While NEP provides important insight into organic carbon cycling, it does not capture inorganic carbon cycling (calcification and dissolution) and therefore does not fully characterize community metabolism. This distinction is important to address within the discussion section.

(Lines 403-406): Suggest broadening this interpretation since these patterns could also reflect differences in benthic biomass and community composition, rather than physiological performance alone.

(Lines 412-415): This is an incomplete characterization of the relationship between environmental drivers of NEP. While increased PAR can enhance photosynthesis, elevated temperatures during the wet season can also increase respiration. Given that all sites are largely net heterotrophic, this interaction introduces additional complexity that should be acknowledged.

(Lines 427-429): This is an important point but is underdeveloped in its current state. Consider expanding on the underlying mechanisms or providing supporting context to strengthen this interpretation.

(Lines 437-439): This sentence is difficult to follow and the underlying concept would benefit from clarification.

(Lines 449-452): There appears to be a notation and unit inconsistency here. In the methods, z₁ and z₂ are defined as the sensor heights (z₁ = 124 cm and z₂ = 8 cm), whereas here they are presented with incorrect values, reversed notation, and are described in the context of bottom water velocity.

(Lines 456-458): Please develop this point further, as benthic community composition is likely a stronger driver of metabolic activity than diversity. Several parts of this sub-section put too much emphasis on diversity being a main contributor.

(Lines 471-474): This statement is not well supported by prior studies and would therefore benefit from additional context and further development.

(Lines 487-489): Please avoid generalizing the results to all coral reef habitats, as this is not broadly supported by the literature. Consider rephrasing to focus specifically on the three study sites.

(Lines 528-530 and Lines 536-537): Please keep the discussion focused on production and respiration since NEC was not measured in this study. Statements referring to “calcification” or “growth” should be avoided.