the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 May 2023
Coral Bleaching Induced Mortality Transforms Local and Global Carbon Cycles: An Unrecognized Feedback Loop That May Accelerate Reef Decline
Abstract. Increases in atmospheric carbon have led to widespread, frequent, and severe coral bleaching, resulting in global coral reef decline. Here, we show that bleaching corals severely impact the local carbon cycle by releasing significant amounts of dissolved organic carbon (DOC), which further stresses the local reef community and may trigger coral mortality. During a severe bleaching event in Mo’orea, French Polynesia, we measured DOC concentrations 37 % greater than Total Organic Carbon (DOC and Particulate Carbon combined), compared to non-bleaching conditions. In addition, this DOC was highest immediately adjacent to the reef, indicating that the corals were the source of this carbon. Further, when exposed to bleaching-derived DOC-rich exudate (~2 mM DOC), otherwise healthy corals experienced bleaching, tissue loss, and mortality within 48 hours. While this is an extreme amount of DOC to be found on a reef in situ, it identifies a potential mechanistic impacts of coral-derived DOC on healthy corals. By extending our findings to regional scales, we estimate that large scale bleaching events can significantly alter the marine carbon cycle. For example, a single bleaching event on the Great Barrier Reef could have released an estimated 150 Gmol C as DOC alone. Our research identifies a previously unrecognized mechanism of coral mortality during bleaching events, in which biogeochemical shifts across the reefscape that may result in a positive feedback loop that accelerates coral reef loss.
- Preprint
(3969 KB) - Metadata XML
-
Supplement
(3011 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2023-779', Anonymous Referee #1, 20 Jun 2023
Summary:
The manuscript by Seabrook et al. investigates the release of Dissolved Organic Carbon (DOC) by bleached corals and the impacts of this DOC on apparently healthy corals. This is a very important research topic, given the increase in frequency and extent of bleaching events and the lack of knowledge on how bleaching affects reef biogeochemistry. However, there are significant gaps in the description of the methods and results and a lack of statistics reporting (details below). These gaps prevent making conclusions about whether DOC release by corals increases during bleaching and, more importantly, if the regional and global scalability of this observation is valid.
Specific comments:
The methods section describes one set of mesocosm experiments, while there seem to be three individual experiments: the first, where DOC exudates were produced from two species of corals, the second, where a new set of corals were exposed to the exudates from the first group, plus temperature stress, and a third experiment where other three species of coral were bleached for quantification of DOC release rates. If that is the case, these experiments need to be explained individually in detail. In general, the methods do not describe the number of replicate coral fragments used in each incubation, what was defined as light stress, and the filter type/model used to prepare the seawater.
Second, the results do not provide individual data points (only color gradients from averages) or report statistics. For example, in the results section 3.1 DOC enrichment, are the variance values representing standard error or deviation? Were these values statistically tested against the baseline? From Figures 1 and S2, it is unclear if there is a real difference in DOC values over the reef. The reef sites are situated in the orange zone of the contour map, and the variance is very large in FigS2. It would be more helpful to plot individual data points for each site against time. Given the lack of strong support, assertive passages like “dramatic shift” should be avoided.
In the results section 3.2, the text mentions coral mortality without reporting how many fragments died. More importantly, no individual data points are provided, only color codes from means (no variance or statistics reported). From the fluorescence data, it looks like Acropora did not bleach with the temperature stress. And it looks like DOC goes down at later time points for Pocillopora, raising the question of whether the DOC increase is significant or if the high DOC jar is an outlier.
In the experiment where corals were subject to exudates from bleached colonies, it is unclear if exudates were first pooled from several bleached colonies before exposure. More importantly, the T0 values seem to have low oxygen concentration (pink half of the jars in Figure 2). If that is the case, it is expected that the corals will suffer from anoxia, making it challenging to disentangle the effects of DOC and anoxia.
The third experiment provides release rates that are used to calculate regional and global DOC release rates. However, there is no information about how these experiments were done: how many coral fragments, of which size, incubation volume, time, light intensity, temperature, sampling strategy, etc. Incubation time is given as a range in the supplementary information. Also in the supp, the units for control and bleached are different. Is that a typo, or the values for bleached were somehow normalized by volume? There is no statistical reporting for these results, presumably because of low replication.
Minor comments:
Line 20: “DOC concentrations greater than TOC”: I believe this is a typo, as TOC = DOC + POC
Lines 118-123: Were these images taken from benthic transects, quadrats? How many images per site, and how were sites chosen?
Line 154: What water chemistry tests were performed?
Citation: https://doi.org/10.5194/egusphere-2023-779-RC1 -
AC2: 'Reply on RC1', Andrew Thurber, 04 Oct 2023
We would like to thank both reviewers for providing areas where we can refine our manuscript to increase its clarity and better communicate the results. Both reviewers also highlighted important concerns that resulted in changes throughout the manuscript, and we point out how they will be addressed if BGS allows us to submit a revised version. In particular, we acknowledge that our experimental design has limitations, which are now more explicitly stated, however we point out that this is a new line of inquiry and the resulting hypothesis (as posed) provides a novel understanding of how shifting biogeochemistry can alter coral health.
Reviewer 1 - Review comments in plain text - responses in bold.
Summary:
The manuscript by Seabrook et al. investigates the release of Dissolved Organic Carbon (DOC) by bleached corals and the impacts of this DOC on apparently healthy corals. This is a very important research topic, given the increase in frequency and extent of bleaching events and the lack of knowledge on how bleaching affects reef biogeochemistry. However, there are significant gaps in the description of the methods and results and a lack of statistics reporting (details below). These gaps prevent making conclusions about whether DOC release by corals increases during bleaching and, more importantly, if the regional and global scalability of this observation is valid.
We thank the reviewer for highlighting the importance of this work. As specified below, we address each aspect of the concerns both within this response as well as in the revised manuscript. Included in the revised manuscript is a far more expansive and explicit methods section.
Specific comments:
The methods section describes one set of mesocosm experiments, while there seem to be three individual experiments: the first, where DOC exudates were produced from two species of corals, the second, where a new set of corals were exposed to the exudates from the first group, plus temperature stress, and a third experiment where other three species of coral were bleached for quantification of DOC release rates. If that is the case, these experiments need to be explained individually in detail. In general, the methods do not describe the number of replicate coral fragments used in each incubation, what was defined as light stress, and the filter type/model used to prepare the seawater.
In our revised manuscript, we have described in detail how the experiments were carried out. In all cases, we have 3 replicates of separate individual coral colonies from the reef that were both the source of the DOC and the individual nubbins that were then exposed. There were 2 experiments - (1) the DOC exposure and (2) the DOC quantification. In the DOC exposure experiment, we took a single coral head from the reef, fragmented it into many nubbins and let them acclimate in a water table. We then took a portion of these and heat and light stressed them by removing them from the seawater table and letting them bleach in the sun in a beaker. This exudate was filtered through a non-quantitative filter to remove the large particulate (Melitta filter - none of the scientific filters we brought - which were many - would not clog immediately). This filtrate was then diluted volumetrically and nubbins of the coral were put in experimental treatments. All of the corals across all of the treatments within each replicate came from the same coral head. All replicates were different coral heads. Each time point of each treatment was an individual container that was sacrificed so we could quantify its health and not open the additional containers - meaning this is not a repeated measures experiment and time points are independent of each other within replicates. This was replicated 3 times for 2 species. The DOC quantification experiment (2 - above) was simply to quantify how much DOC and POC was released and that was done by placing corals in our closed filtration systems and bleaching them in the water tables to quantify how much DOC came out. This was done by adding full light to them - which rapidly bleaches these corals as they were collected at 5 m water depth and full sunlight is a significant stresser. This is all now in the revised manuscript.
Second, the results do not provide individual data points (only color gradients from averages) or report statistics. For example, in the results section 3.1 DOC enrichment, are the variance values representing standard error or deviation? Were these values statistically tested against the baseline? From Figures 1 and S2, it is unclear if there is a real difference in DOC values over the reef. The reef sites are situated in the orange zone of the contour map, and the variance is very large in FigS2. It would be more helpful to plot individual data points for each site against time. Given the lack of strong support, assertive passages like “dramatic shift” should be avoided.
We have added figures to the supplemental to compliment the color gradient figures presented. Error bars are standard error. Yes, the variance among replicates was high on the reef and that has impacted the ability to have conclusive statistical support. We now include those data. We note that the replicates were taken along transects that followed depth contours along the reef, and the variance among replicates reflects natural variability along the reef which is expected during a spatially variable bleaching event. In addition, previous years' data are single data points twice a year making it a challenge to have a statistically resolved shift. As pointed out by the other reviewer, we also cannot directly compare our data with the long term data as they measured TOC and we measured DOC, and on different instruments without intercalibration. We highlight that it is the pattern that is key for showing the impact of the bleaching on the corals and suggest that more quantitative data over bleaching events focusing on DOC is needed to understand how bleaching impacts reef biogeochemistry. That is a key take home point of this manuscript – DOC should be measured across bleaching events to better understand the ecosystem role of coral bleaching.
In the results section 3.2, the text mentions coral mortality without reporting how many fragments died. More importantly, no individual data points are provided, only color codes from means (no variance or statistics reported). From the fluorescence data, it looks like Acropora did not bleach with the temperature stress. And it looks like DOC goes down at later time points for Pocillopora, raising the question of whether the DOC increase is significant or if the high DOC jar is an outlier.
Each of the jars/ time points are independent so there was not a decrease in the jar that had high DOC but there was a fragment that bleached in its mesocsosm that we sampled at that point and when we sampled the next time point in a separate jar that fragment has not bleached. The lower value after it is from a separate sample container that did not have a bleaching coral fragment. We did this to avoid a repeated measures design but we understand how this needs to be better clarified in the methods section, which it now is. In all cases, when bleaching was observed there was a significant increase in DOC. We now include figures in the Supplemental that have the individual data points for each treatment so the among replicate variance is more transparent to the reader.
Coral mortality is difficult to definitively measure after bleaching. In certain cases tissue sloughing was evident but that is a later stage of death and there can be a few polyps alive (the Phoenix effect). We have err’d on the side of using PAM and Chl concentration to indicate coral health rather than mortality, although we agree that this would be a valuable metric.
We also now provide plots that provide the among replicate variance as requested.
In the experiment where corals were subject to exudates from bleached colonies, it is unclear if exudates were first pooled from several bleached colonies before exposure. More importantly, the T0 values seem to have low oxygen concentration (pink half of the jars in Figure 2). If that is the case, it is expected that the corals will suffer from anoxia, making it challenging to disentangle the effects of DOC and anoxia.
Each bleaching exudate was generated from a separate fragment of the genets that were also used for the experiments. The T0 value provided was in response to the addition of the DOC that rapidly drew down the oxygen. Yes, hypoxia is a stressor that co-occurs with carbon stress.
The third experiment provides release rates that are used to calculate regional and global DOC release rates. However, there is no information about how these experiments were done: how many coral fragments, of which size, incubation volume, time, light intensity, temperature, sampling strategy, etc. Incubation time is given as a range in the supplementary information. Also in the supp, the units for control and bleached are different. Is that a typo, or the values for bleached were somehow normalized by volume? There is no statistical reporting for these results, presumably because of low replication.
Again, we see that there was a lack of clarity in our methods that led to this misunderstanding in the methodology that we have adjusted in the revised version thanks to the input of this reviewer. We provide (in the Supplemental) a table with all of these data including the number of replicates. We did standardize the surface area of the corals using photogrammetry but there were no corals in the controls hence the different units as there is no cm-2 surface area of coral to correct for.
Minor comments:
Line 20: “DOC concentrations greater than TOC”: I believe this is a typo, as TOC = DOC + POC
Fixed.
Lines 118-123: Were these images taken from benthic transects, quadrats? How many images per site, and how were sites chosen?
> 40 photos were stitched using photogrammetry together into a map and then used that to come up with the total surface area sampled per site. Together this allows us to quantify a much larger area than quadrats would have done. These were done at the three long term data sites on this reef (LTER O, LTER 1, and LTER 2 from the Moorea LTER datasets), although this methodology does differ from those employed by the LTER. We have provided those images now in supplemental as well as the number of photos for each one of the stitched images.
Line 154: What water chemistry tests were performed?
We have removed this line since we mainly focused on DOC values for this.
Citation: https://doi.org/10.5194/egusphere-2023-779-AC2
-
AC2: 'Reply on RC1', Andrew Thurber, 04 Oct 2023
-
RC2: 'Comment on egusphere-2023-779', Anonymous Referee #2, 27 Jul 2023
General comments:
The manuscript addresses a highly valuable topic concerning coral bleaching and its impact on dissolved organic carbon (DOC) release. The approach to test the hypothesis of a positive feedback loop that may accelerate reef decline was divided into three parts: 1) DOC measurements in the environment, 2) exposing corals to released DOC due to bleaching in experiments, and 3) modeling possible global impacts of coral bleaching events on DOC release based on a different bleaching experiment.
The study’s approach is interesting and holds great potential for contributing valuable insights into DOC release during bleaching events. However, the experimental set ups and sampling techniques as well as the quality and interpretation of the data are concerning. The number of samples are low and the manuscript is missing statistics to justify their findings. The exercise of calculating the net DOC that could potentially be released during a bleaching event is intriguing. However, it is unclear whether the results from the experiments (considering the coral species and number of replicates) provide a sufficient basis for any global calculation.
Specific comments:
Field observations:
The observed increase in DOC concentration compared to previous data (Nelson et al., 2011 and LTER) raises concerns about potential methodological artifacts. Proper sampling of seawater DOC is essential but challenging and contamination during sampling can significantly impact results (e.g. leaching plasticisers, fumes from the boat engine, contamination by touching the sample). Therefore, consistency in methodology in both sampling and analysis of samples are essential for the direct comparison of DOC datasets. According to the methods, the DOC samples were taken using whirl bags, which are leaching plasticizers and are not an acceptable method to sample for DOC, especially for samples with low DOC concentrations such as seawater. It is highly possible that the observed increase in DOC (Nelson et al., 2011 and LTER data vs. data of this study) is simply caused by the handling of the seawater samples. The inclusion of procedure blanks (e.g., MilliQ treated the same as the samples, including storage in whirl bags and filtering through the same setup) would be essential for evaluating potential artifacts and systematic biases between the two collection methods. Additionally, the fact that offshore concentrations sampled by the authors, which should not be affected by bleaching, are also elevated, further highlights contamination issues.
In addition, the lack of intercalibration or references hinders comparison across datasets (commonly used for oceanic DOC is Hansell CRMs at <45uM). The measured DOC values of this study exhibit significant variation, and this variation seems to increase with the number of samples (see Fig S2). While this could potentially be attributed to heterogeneity in DOC concentration in the environment, it is highly probable that inconsistent contamination and inaccurate measurement of low DOC are the underlying factors. In the manuscript's methods, it is mentioned that the minimum reproducible and reportable value from UMCES was 13.3 umol C/L. However, the UMCES protocol (link provided in the SI) indicates that the lowest value of the calibration curve is 0.5 mg/L, which suggests that the concentrations of the samples may fall outside the calibration curve window. Additionally, the methods do not specify whether the instrument used was equipped with a magnesium perchlorate water-trap, essential for accurate measurements of low DOC samples, such as oceanic seawater.
Lastly, the presentation of the DOC data raises concerns. The limited number of samples and their spacing do not allow for interpolation, as seen in Fig.2. A more appropriate representation would be Fig S2, which provides a more realistic picture. However, the absence of statistical analysis to support the claims in the paper is a major issue.Lab experiments:
As the authors acknowledge themselves (Line 148), the addition of 2mM DOC in the experiment is an order of magnitude higher than what is typically found in nature. Consequently, it is not surprising that the mesocosms turned hypoxic and led to coral death. At such high concentrations, the source and composition of dissolved organic matter are likely irrelevant. Therefore, drawing the conclusion that the bleaching-derived DOC-rich exudates directly cause bleaching or death to healthy corals is not supported.
Model:
Making global calculations based on an experiment with an unrealistic time frame (bleaching within 48h) is somewhat speculative. But it is an interesting thought experiment. If the aim is to determine the maximum DOC that a coral reef could release, it would be insightful to provide context. For instance, it would be beneficial to consider the total carbon or total organic carbon stored in the Great Barrier Reef and assess whether the calculated numbers are realistic.
Overall, although the proposed mechanism of the positive feedback loop is an interesting hypothesis, unfortunately this study lacks sufficient data to support the existence of this mechanism.
Technical corrections:
Line 108: SI Figure 1 does not present Nelson et al., 2011 data but rather showcases photos of corals.
Line 144: Please specify the material and brand of the filter used.
Line 147: Clarify which DOC samples from the field are being referenced here.
Line 165: Remove the underscore.
Line 186 & 201: Unfortunately, comparing the data of this study to previous data is not feasible due to the issues discussed in the Specific comment section.
Line 199: While comparing trends within this study's data set is more appropriate, the significant variations caused by flaws in sample preparation and measurement hinder the interpretation of the data. Simply comparing averages is not sufficient; statistical analysis would be necessary to support the author's argument that the reef switches from sink to source. Given the methodological concerns and the limited number of measurements, the authors should be cautious when drawing conclusions from only a few samples taken on a single day.
Line 206: Please provide additional evidence to support this claim. According to the methods in the SI, the applied calibration curve appears not to cover the range of seawater samples (with the lowest concentration being 0.5mg/L). The TOC analyzer does not seem to be equipped with a magnesium perchlorate water-trap, which is essential for accurate measurements of low DOC samples, such as seawater. Clarifying these points will enhance the validity of the study's results.
Line 209 & 295: This conclusion cannot be drawn from the experiment presented here. The substantial amount of DOC added to the mesocosm would likely cause hypoxia and coral death regardless of the source or composition of the added DOC. This outcome is not unexpected and cannot be used as an argument to support the notion that DOC-rich coral exudates in particluar cause bleaching and coral mortality.
Line 305: Please clarify whether oxygen levels were directly measured on the reef during the time of bleaching in Moorea to validate the statements made, or if the following conclusions are solely based on the DOC concentration.
Fig. S2: The fonts used in the figure for units and coordinates are too small to read, and the captions need to provide clarification on the meaning of "combined data set."
Fig. S3: Nelson et al 2011 and LTER samples have a higher number of samples than indicated in the figure. Please explain what subset was chosen to create the figure.
Citation: https://doi.org/10.5194/egusphere-2023-779-RC2 -
AC1: 'Reply on RC2', Andrew Thurber, 04 Oct 2023
We would like to thank both reviewers for providing areas where we can refine our manuscript to increase its clarity and better communicate the results. Both reviewers also highlighted important concerns that resulted in changes throughout the manuscript, and we point out how they will be addressed if BGS allows us to submit a revised version. In particular, we acknowledge that our experimental design has limitations, which are now more explicitly stated, however we point out that this is a new line of inquiry and the resulting hypothesis (as posed) provides a novel understanding of how shifting biogeochemistry can alter coral health
Review comments in plain text - responses in bold
The manuscript addresses a highly valuable topic concerning coral bleaching and its impact on dissolved organic carbon (DOC) release. The approach to test the hypothesis of a positive feedback loop that may accelerate reef decline was divided into three parts: 1) DOC measurements in the environment, 2) exposing corals to released DOC due to bleaching in experiments, and 3) modeling possible global impacts of coral bleaching events on DOC release based on a different bleaching experiment.
The study’s approach is interesting and holds great potential for contributing valuable insights into DOC release during bleaching events. However, the experimental set ups and sampling techniques as well as the quality and interpretation of the data are concerning. The number of samples are low and the manuscript is missing statistics to justify their findings. The exercise of calculating the net DOC that could potentially be released during a bleaching event is intriguing. However, it is unclear whether the results from the experiments (considering the coral species and number of replicates) provide a sufficient basis for any global calculation.
We thank this reviewer for these valuable comments and the insight provided throughout. We would like to emphasize both in this response, and we do this explicitly, that the role of DOC release as a causative agent of reef decline is proposed as a hypothesis and we agree that our data do not test this result without caveats (as all experiential work in a new direction will have) - for many of the reasons the reviewer provides (note we are adding “may” in the title). These results will stimulate research and increased understanding the role of bleaching corals in a reef ecosystem. We provide lines of evidence that support a new hypothesis and more research is needed to identify the magnitude of this. This work is the outcome of opportunistic alignment of experimental study coinciding with a major bleaching event on the reef that was our study site. We did not expect such a shift in DOC across the reefscape, or during bleaching in our experimental work, but upon seeing this consistently across both studies considered these results to be important and necessary to communicate to the scientific community to fuel future work to test our resultant hypothesis. Regardless, we do demonstrate an increase of DOC from the reef vs above it during a bleaching event - something that has never been quantified previously. While the magnitude may be impacted by our sampling method, the pattern is not.
Specific comments:
Field observations:
The observed increase in DOC concentration compared to previous data (Nelson et al., 2011 and LTER) raises concerns about potential methodological artifacts. Proper sampling of seawater DOC is essential but challenging and contamination during sampling can significantly impact results (e.g. leaching plasticisers, fumes from the boat engine, contamination by touching the sample). Therefore, consistency in methodology in both sampling and analysis of samples are essential for the direct comparison of DOC datasets. According to the methods, the DOC samples were taken using whirl bags, which are leaching plasticizers and are not an acceptable method to sample for DOC, especially for samples with low DOC concentrations such as seawater. It is highly possible that the observed increase in DOC (Nelson et al., 2011 and LTER data vs. data of this study) is simply caused by the handling of the seawater samples. The inclusion of procedure blanks (e.g., MilliQ treated the same as the samples, including storage in whirl bags and filtering through the same setup) would be essential for evaluating potential artifacts and systematic biases between the two collection methods. Additionally, the fact that offshore concentrations sampled by the authors, which should not be affected by bleaching, are also elevated, further highlights contamination issues.
There are potential methodological challenges associated as highlighted above and we also agree that proper sampling of DOC is challenging. We agree that we should not compare our DOC data to the previous TOC data due to this and have addressed this explicitly and carefully throughout the revised manuscript, and modified it as such. What we do compare, and what we have kept in the manuscript, is the shift in on reef / off reef DOC balance within each data set. The DOC on reef was consistently depleted relative to offshore samples in all LTER data as well as our data at the onset of the bleaching event. During the peak of the bleaching event, we saw this balance shift in the data (coinciding with a shift in water quality along the reef, notably) with the reef becoming enriched when compared to offshore samples within the same dataset, from samples collected on the same day. However, we do argue that contamination - if present - would be uniform across our samples and the relative patterns are robust. This is also a unique time with very high water temperature and so there is no reason to dismiss the relatively high offshore value a priori, as the water was warmer than it has been previously.
In addition, the lack of intercalibration or references hinders comparison across datasets (commonly used for oceanic DOC is Hansell CRMs at <45uM). The measured DOC values of this study exhibit significant variation, and this variation seems to increase with the number of samples (see Fig S2). While this could potentially be attributed to heterogeneity in DOC concentration in the environment, it is highly probable that inconsistent contamination and inaccurate measurement of low DOC are the underlying factors. In the manuscript's methods, it is mentioned that the minimum reproducible and reportable value from UMCES was 13.3 umol C/L. However, the UMCES protocol (link provided in the SI) indicates that the lowest value of the calibration curve is 0.5 mg/L, which suggests that the concentrations of the samples may fall outside the calibration curve window. Additionally, the methods do not specify whether the instrument used was equipped with a magnesium perchlorate water-trap, essential for accurate measurements of low DOC samples, such as oceanic seawater.
This comment has a few points to address:
- DOC values of the site are not lower than 45uM. These are not oligotrophic open ocean waters from both our measurements as well as those from the LTER at this site are not less than 60 uM in concentration. We did make an error as the 13.3 umol is the detection limit and the reporting limit from this method is 41.6 umol - at the time when we analyzed our samples from the analytical lab. We have adjusted the text accordingly, however this is still below the minimum from the waters samples. There was no magnesium-perchlorate water-trap as these are within the reproducible result range of the instrument used.
- None of the samples ran had a value less than 0.5 mg/L - all samples fell within the calibration curve window of the instrument.
- Yes - our data was highly variable however we also were 1m off the seafloor in a highly dynamic location. We would argue that it would be heterogeneity in the environment that led to this high variability rather than differential contamination. In part this is because we sampled these at three different days and replicated across those sites, in each case there was increased DOC on the reef which would be a highly unlikely occurrence if there was stochastic contamination from the whirl packs.
- There are no existing data from the LTER from this time to directly compare. There was sampling earlier in the year before the onset of the main bleaching event so a true intercalibration is not possible with existing data.
Lastly, the presentation of the DOC data raises concerns. The limited number of samples and their spacing do not allow for interpolation, as seen in Fig.2. A more appropriate representation would be Fig S2, which provides a more realistic picture. However, the absence of statistical analysis to support the claims in the paper is a major issue.In the figure legend we highlight the faded section is for aesthetic and communication purposes as to why it did not look like figure S2. We feel this provides a more clear view of the patterns. This has no impact on the interpretation of the results. The statistical treatment was challenged by the high variability across the reef as discussed above (data are now more comprehensively presented in the supplemental.) We would also like to point out that our replication (3 across a transect at three sites) is in contrast to a single data point per depth collected via nisken for both this (LTER) and most long term data sets. So in contrast we have had sample numbers that can elucidate the natural variability across a reefscape that is not always observed.
Lab experiments:
As the authors acknowledge themselves (Line 148), the addition of 2mM DOC in the experiment is an order of magnitude higher than what is typically found in nature. Consequently, it is not surprising that the mesocosms turned hypoxic and led to coral death. At such high concentrations, the source and composition of dissolved organic matter are likely irrelevant. Therefore, drawing the conclusion that the bleaching-derived DOC-rich exudates directly cause bleaching or death to healthy corals is not supported.
We will better emphasize this point. While we had tried to specify this in the text (including calling out the order of magnitude increase) we have further specified this as a hypothesis, one that we did not test (yet). We also more clearly state the impact role of High DOC leading to the hypoxic results and its role in generating hypothesis and stimulating research.
Model:
Making global calculations based on an experiment with an unrealistic time frame (bleaching within 48h) is somewhat speculative. But it is an interesting thought experiment. If the aim is to determine the maximum DOC that a coral reef could release, it would be insightful to provide context. For instance, it would be beneficial to consider the total carbon or total organic carbon stored in the Great Barrier Reef and assess whether the calculated numbers are realistic.
We are unaware and have not been able to find a value for the total carbon pool stored in corals on the GBR. We also agree that this is a thought experiment and appreciate that it was interpreted as such.
Overall, although the proposed mechanism of the positive feedback loop is an interesting hypothesis, unfortunately this study lacks sufficient data to support the existence of this mechanism.
We feel that the real take home message of this is the potential for this mechanism and the significant DOC released from corals when they bleach - something that has not previously been shown during a bleaching event. We now specify this in the conclusion section.
Technical corrections:
Line 108: SI Figure 1 does not present Nelson et al., 2011 data but rather showcases photos of corals.
Fixed.
Line 144: Please specify the material and brand of the filter used.
Fixed and answered in other review.
Line 147: Clarify which DOC samples from the field are being referenced here.
Fixed.
Line 165: Remove the underscore.
Fixed.
Line 186 & 201: Unfortunately, comparing the data of this study to previous data is not feasible due to the issues discussed in the Specific comment section.
Agreed, and we have modified the text as a result of this.
Line 199: While comparing trends within this study's data set is more appropriate, the significant variations caused by flaws in sample preparation and measurement hinder the interpretation of the data. Simply comparing averages is not sufficient; statistical analysis would be necessary to support the author's argument that the reef switches from sink to source. Given the methodological concerns and the limited number of measurements, the authors should be cautious when drawing conclusions from only a few samples taken on a single day.
Data sets on DOC frequently are single points in time without replication. In fact our replication is greater than previous studies. We now have added figures and text to clarify this, including statistics to back it up. Again, we argue that our methodology for DOC measurement is appropriate for the concentrations at this location (they would not be in other regions).
Line 206: Please provide additional evidence to support this claim. According to the methods in the SI, the applied calibration curve appears not to cover the range of seawater samples (with the lowest concentration being 0.5mg/L). The TOC analyzer does not seem to be equipped with a magnesium perchlorate water-trap, which is essential for accurate measurements of low DOC samples, such as seawater. Clarifying these points will enhance the validity of the study's results.
This point is addressed above. These concerns are valid in other locations but not the location that this study was carried out.
Line 209 & 295: This conclusion cannot be drawn from the experiment presented here. The substantial amount of DOC added to the mesocosm would likely cause hypoxia and coral death regardless of the source or composition of the added DOC. This outcome is not unexpected and cannot be used as an argument to support the notion that DOC-rich coral exudates in particular cause bleaching and coral mortality.
We modify the sentences to more cleary state that our data are supportive of and suggestive of the hypothesis rather than conclusive.
Line 305: Please clarify whether oxygen levels were directly measured on the reef during the time of bleaching in Moorea to validate the statements made, or if the following conclusions are solely based on the DOC concentration.
Oxygen data are not available across a sufficient time frame to contrast the bleaching event with normal conditions. These statements were solely based on DOC concentrations.
Fig. S2: The fonts used in the figure for units and coordinates are too small to read, and the captions need to provide clarification on the meaning of "combined data set."
Fixed throughout.
Fig. S3: Nelson et al 2011 and LTER samples have a higher number of samples than indicated in the figure. Please explain what subset was chosen to create the figure.
The LTER collects a single data point at each depth and location and we chose those points that matched up with our site due to the high level of bleaching. We also did not subsample Nelson et al. 2011. This highlights how our data has more replication than the landmark papers that are used to track changes in DOC at these sites.
Citation: https://doi.org/10.5194/egusphere-2023-779-AC1
-
AC1: 'Reply on RC2', Andrew Thurber, 04 Oct 2023
Status: closed
-
RC1: 'Comment on egusphere-2023-779', Anonymous Referee #1, 20 Jun 2023
Summary:
The manuscript by Seabrook et al. investigates the release of Dissolved Organic Carbon (DOC) by bleached corals and the impacts of this DOC on apparently healthy corals. This is a very important research topic, given the increase in frequency and extent of bleaching events and the lack of knowledge on how bleaching affects reef biogeochemistry. However, there are significant gaps in the description of the methods and results and a lack of statistics reporting (details below). These gaps prevent making conclusions about whether DOC release by corals increases during bleaching and, more importantly, if the regional and global scalability of this observation is valid.
Specific comments:
The methods section describes one set of mesocosm experiments, while there seem to be three individual experiments: the first, where DOC exudates were produced from two species of corals, the second, where a new set of corals were exposed to the exudates from the first group, plus temperature stress, and a third experiment where other three species of coral were bleached for quantification of DOC release rates. If that is the case, these experiments need to be explained individually in detail. In general, the methods do not describe the number of replicate coral fragments used in each incubation, what was defined as light stress, and the filter type/model used to prepare the seawater.
Second, the results do not provide individual data points (only color gradients from averages) or report statistics. For example, in the results section 3.1 DOC enrichment, are the variance values representing standard error or deviation? Were these values statistically tested against the baseline? From Figures 1 and S2, it is unclear if there is a real difference in DOC values over the reef. The reef sites are situated in the orange zone of the contour map, and the variance is very large in FigS2. It would be more helpful to plot individual data points for each site against time. Given the lack of strong support, assertive passages like “dramatic shift” should be avoided.
In the results section 3.2, the text mentions coral mortality without reporting how many fragments died. More importantly, no individual data points are provided, only color codes from means (no variance or statistics reported). From the fluorescence data, it looks like Acropora did not bleach with the temperature stress. And it looks like DOC goes down at later time points for Pocillopora, raising the question of whether the DOC increase is significant or if the high DOC jar is an outlier.
In the experiment where corals were subject to exudates from bleached colonies, it is unclear if exudates were first pooled from several bleached colonies before exposure. More importantly, the T0 values seem to have low oxygen concentration (pink half of the jars in Figure 2). If that is the case, it is expected that the corals will suffer from anoxia, making it challenging to disentangle the effects of DOC and anoxia.
The third experiment provides release rates that are used to calculate regional and global DOC release rates. However, there is no information about how these experiments were done: how many coral fragments, of which size, incubation volume, time, light intensity, temperature, sampling strategy, etc. Incubation time is given as a range in the supplementary information. Also in the supp, the units for control and bleached are different. Is that a typo, or the values for bleached were somehow normalized by volume? There is no statistical reporting for these results, presumably because of low replication.
Minor comments:
Line 20: “DOC concentrations greater than TOC”: I believe this is a typo, as TOC = DOC + POC
Lines 118-123: Were these images taken from benthic transects, quadrats? How many images per site, and how were sites chosen?
Line 154: What water chemistry tests were performed?
Citation: https://doi.org/10.5194/egusphere-2023-779-RC1 -
AC2: 'Reply on RC1', Andrew Thurber, 04 Oct 2023
We would like to thank both reviewers for providing areas where we can refine our manuscript to increase its clarity and better communicate the results. Both reviewers also highlighted important concerns that resulted in changes throughout the manuscript, and we point out how they will be addressed if BGS allows us to submit a revised version. In particular, we acknowledge that our experimental design has limitations, which are now more explicitly stated, however we point out that this is a new line of inquiry and the resulting hypothesis (as posed) provides a novel understanding of how shifting biogeochemistry can alter coral health.
Reviewer 1 - Review comments in plain text - responses in bold.
Summary:
The manuscript by Seabrook et al. investigates the release of Dissolved Organic Carbon (DOC) by bleached corals and the impacts of this DOC on apparently healthy corals. This is a very important research topic, given the increase in frequency and extent of bleaching events and the lack of knowledge on how bleaching affects reef biogeochemistry. However, there are significant gaps in the description of the methods and results and a lack of statistics reporting (details below). These gaps prevent making conclusions about whether DOC release by corals increases during bleaching and, more importantly, if the regional and global scalability of this observation is valid.
We thank the reviewer for highlighting the importance of this work. As specified below, we address each aspect of the concerns both within this response as well as in the revised manuscript. Included in the revised manuscript is a far more expansive and explicit methods section.
Specific comments:
The methods section describes one set of mesocosm experiments, while there seem to be three individual experiments: the first, where DOC exudates were produced from two species of corals, the second, where a new set of corals were exposed to the exudates from the first group, plus temperature stress, and a third experiment where other three species of coral were bleached for quantification of DOC release rates. If that is the case, these experiments need to be explained individually in detail. In general, the methods do not describe the number of replicate coral fragments used in each incubation, what was defined as light stress, and the filter type/model used to prepare the seawater.
In our revised manuscript, we have described in detail how the experiments were carried out. In all cases, we have 3 replicates of separate individual coral colonies from the reef that were both the source of the DOC and the individual nubbins that were then exposed. There were 2 experiments - (1) the DOC exposure and (2) the DOC quantification. In the DOC exposure experiment, we took a single coral head from the reef, fragmented it into many nubbins and let them acclimate in a water table. We then took a portion of these and heat and light stressed them by removing them from the seawater table and letting them bleach in the sun in a beaker. This exudate was filtered through a non-quantitative filter to remove the large particulate (Melitta filter - none of the scientific filters we brought - which were many - would not clog immediately). This filtrate was then diluted volumetrically and nubbins of the coral were put in experimental treatments. All of the corals across all of the treatments within each replicate came from the same coral head. All replicates were different coral heads. Each time point of each treatment was an individual container that was sacrificed so we could quantify its health and not open the additional containers - meaning this is not a repeated measures experiment and time points are independent of each other within replicates. This was replicated 3 times for 2 species. The DOC quantification experiment (2 - above) was simply to quantify how much DOC and POC was released and that was done by placing corals in our closed filtration systems and bleaching them in the water tables to quantify how much DOC came out. This was done by adding full light to them - which rapidly bleaches these corals as they were collected at 5 m water depth and full sunlight is a significant stresser. This is all now in the revised manuscript.
Second, the results do not provide individual data points (only color gradients from averages) or report statistics. For example, in the results section 3.1 DOC enrichment, are the variance values representing standard error or deviation? Were these values statistically tested against the baseline? From Figures 1 and S2, it is unclear if there is a real difference in DOC values over the reef. The reef sites are situated in the orange zone of the contour map, and the variance is very large in FigS2. It would be more helpful to plot individual data points for each site against time. Given the lack of strong support, assertive passages like “dramatic shift” should be avoided.
We have added figures to the supplemental to compliment the color gradient figures presented. Error bars are standard error. Yes, the variance among replicates was high on the reef and that has impacted the ability to have conclusive statistical support. We now include those data. We note that the replicates were taken along transects that followed depth contours along the reef, and the variance among replicates reflects natural variability along the reef which is expected during a spatially variable bleaching event. In addition, previous years' data are single data points twice a year making it a challenge to have a statistically resolved shift. As pointed out by the other reviewer, we also cannot directly compare our data with the long term data as they measured TOC and we measured DOC, and on different instruments without intercalibration. We highlight that it is the pattern that is key for showing the impact of the bleaching on the corals and suggest that more quantitative data over bleaching events focusing on DOC is needed to understand how bleaching impacts reef biogeochemistry. That is a key take home point of this manuscript – DOC should be measured across bleaching events to better understand the ecosystem role of coral bleaching.
In the results section 3.2, the text mentions coral mortality without reporting how many fragments died. More importantly, no individual data points are provided, only color codes from means (no variance or statistics reported). From the fluorescence data, it looks like Acropora did not bleach with the temperature stress. And it looks like DOC goes down at later time points for Pocillopora, raising the question of whether the DOC increase is significant or if the high DOC jar is an outlier.
Each of the jars/ time points are independent so there was not a decrease in the jar that had high DOC but there was a fragment that bleached in its mesocsosm that we sampled at that point and when we sampled the next time point in a separate jar that fragment has not bleached. The lower value after it is from a separate sample container that did not have a bleaching coral fragment. We did this to avoid a repeated measures design but we understand how this needs to be better clarified in the methods section, which it now is. In all cases, when bleaching was observed there was a significant increase in DOC. We now include figures in the Supplemental that have the individual data points for each treatment so the among replicate variance is more transparent to the reader.
Coral mortality is difficult to definitively measure after bleaching. In certain cases tissue sloughing was evident but that is a later stage of death and there can be a few polyps alive (the Phoenix effect). We have err’d on the side of using PAM and Chl concentration to indicate coral health rather than mortality, although we agree that this would be a valuable metric.
We also now provide plots that provide the among replicate variance as requested.
In the experiment where corals were subject to exudates from bleached colonies, it is unclear if exudates were first pooled from several bleached colonies before exposure. More importantly, the T0 values seem to have low oxygen concentration (pink half of the jars in Figure 2). If that is the case, it is expected that the corals will suffer from anoxia, making it challenging to disentangle the effects of DOC and anoxia.
Each bleaching exudate was generated from a separate fragment of the genets that were also used for the experiments. The T0 value provided was in response to the addition of the DOC that rapidly drew down the oxygen. Yes, hypoxia is a stressor that co-occurs with carbon stress.
The third experiment provides release rates that are used to calculate regional and global DOC release rates. However, there is no information about how these experiments were done: how many coral fragments, of which size, incubation volume, time, light intensity, temperature, sampling strategy, etc. Incubation time is given as a range in the supplementary information. Also in the supp, the units for control and bleached are different. Is that a typo, or the values for bleached were somehow normalized by volume? There is no statistical reporting for these results, presumably because of low replication.
Again, we see that there was a lack of clarity in our methods that led to this misunderstanding in the methodology that we have adjusted in the revised version thanks to the input of this reviewer. We provide (in the Supplemental) a table with all of these data including the number of replicates. We did standardize the surface area of the corals using photogrammetry but there were no corals in the controls hence the different units as there is no cm-2 surface area of coral to correct for.
Minor comments:
Line 20: “DOC concentrations greater than TOC”: I believe this is a typo, as TOC = DOC + POC
Fixed.
Lines 118-123: Were these images taken from benthic transects, quadrats? How many images per site, and how were sites chosen?
> 40 photos were stitched using photogrammetry together into a map and then used that to come up with the total surface area sampled per site. Together this allows us to quantify a much larger area than quadrats would have done. These were done at the three long term data sites on this reef (LTER O, LTER 1, and LTER 2 from the Moorea LTER datasets), although this methodology does differ from those employed by the LTER. We have provided those images now in supplemental as well as the number of photos for each one of the stitched images.
Line 154: What water chemistry tests were performed?
We have removed this line since we mainly focused on DOC values for this.
Citation: https://doi.org/10.5194/egusphere-2023-779-AC2
-
AC2: 'Reply on RC1', Andrew Thurber, 04 Oct 2023
-
RC2: 'Comment on egusphere-2023-779', Anonymous Referee #2, 27 Jul 2023
General comments:
The manuscript addresses a highly valuable topic concerning coral bleaching and its impact on dissolved organic carbon (DOC) release. The approach to test the hypothesis of a positive feedback loop that may accelerate reef decline was divided into three parts: 1) DOC measurements in the environment, 2) exposing corals to released DOC due to bleaching in experiments, and 3) modeling possible global impacts of coral bleaching events on DOC release based on a different bleaching experiment.
The study’s approach is interesting and holds great potential for contributing valuable insights into DOC release during bleaching events. However, the experimental set ups and sampling techniques as well as the quality and interpretation of the data are concerning. The number of samples are low and the manuscript is missing statistics to justify their findings. The exercise of calculating the net DOC that could potentially be released during a bleaching event is intriguing. However, it is unclear whether the results from the experiments (considering the coral species and number of replicates) provide a sufficient basis for any global calculation.
Specific comments:
Field observations:
The observed increase in DOC concentration compared to previous data (Nelson et al., 2011 and LTER) raises concerns about potential methodological artifacts. Proper sampling of seawater DOC is essential but challenging and contamination during sampling can significantly impact results (e.g. leaching plasticisers, fumes from the boat engine, contamination by touching the sample). Therefore, consistency in methodology in both sampling and analysis of samples are essential for the direct comparison of DOC datasets. According to the methods, the DOC samples were taken using whirl bags, which are leaching plasticizers and are not an acceptable method to sample for DOC, especially for samples with low DOC concentrations such as seawater. It is highly possible that the observed increase in DOC (Nelson et al., 2011 and LTER data vs. data of this study) is simply caused by the handling of the seawater samples. The inclusion of procedure blanks (e.g., MilliQ treated the same as the samples, including storage in whirl bags and filtering through the same setup) would be essential for evaluating potential artifacts and systematic biases between the two collection methods. Additionally, the fact that offshore concentrations sampled by the authors, which should not be affected by bleaching, are also elevated, further highlights contamination issues.
In addition, the lack of intercalibration or references hinders comparison across datasets (commonly used for oceanic DOC is Hansell CRMs at <45uM). The measured DOC values of this study exhibit significant variation, and this variation seems to increase with the number of samples (see Fig S2). While this could potentially be attributed to heterogeneity in DOC concentration in the environment, it is highly probable that inconsistent contamination and inaccurate measurement of low DOC are the underlying factors. In the manuscript's methods, it is mentioned that the minimum reproducible and reportable value from UMCES was 13.3 umol C/L. However, the UMCES protocol (link provided in the SI) indicates that the lowest value of the calibration curve is 0.5 mg/L, which suggests that the concentrations of the samples may fall outside the calibration curve window. Additionally, the methods do not specify whether the instrument used was equipped with a magnesium perchlorate water-trap, essential for accurate measurements of low DOC samples, such as oceanic seawater.
Lastly, the presentation of the DOC data raises concerns. The limited number of samples and their spacing do not allow for interpolation, as seen in Fig.2. A more appropriate representation would be Fig S2, which provides a more realistic picture. However, the absence of statistical analysis to support the claims in the paper is a major issue.Lab experiments:
As the authors acknowledge themselves (Line 148), the addition of 2mM DOC in the experiment is an order of magnitude higher than what is typically found in nature. Consequently, it is not surprising that the mesocosms turned hypoxic and led to coral death. At such high concentrations, the source and composition of dissolved organic matter are likely irrelevant. Therefore, drawing the conclusion that the bleaching-derived DOC-rich exudates directly cause bleaching or death to healthy corals is not supported.
Model:
Making global calculations based on an experiment with an unrealistic time frame (bleaching within 48h) is somewhat speculative. But it is an interesting thought experiment. If the aim is to determine the maximum DOC that a coral reef could release, it would be insightful to provide context. For instance, it would be beneficial to consider the total carbon or total organic carbon stored in the Great Barrier Reef and assess whether the calculated numbers are realistic.
Overall, although the proposed mechanism of the positive feedback loop is an interesting hypothesis, unfortunately this study lacks sufficient data to support the existence of this mechanism.
Technical corrections:
Line 108: SI Figure 1 does not present Nelson et al., 2011 data but rather showcases photos of corals.
Line 144: Please specify the material and brand of the filter used.
Line 147: Clarify which DOC samples from the field are being referenced here.
Line 165: Remove the underscore.
Line 186 & 201: Unfortunately, comparing the data of this study to previous data is not feasible due to the issues discussed in the Specific comment section.
Line 199: While comparing trends within this study's data set is more appropriate, the significant variations caused by flaws in sample preparation and measurement hinder the interpretation of the data. Simply comparing averages is not sufficient; statistical analysis would be necessary to support the author's argument that the reef switches from sink to source. Given the methodological concerns and the limited number of measurements, the authors should be cautious when drawing conclusions from only a few samples taken on a single day.
Line 206: Please provide additional evidence to support this claim. According to the methods in the SI, the applied calibration curve appears not to cover the range of seawater samples (with the lowest concentration being 0.5mg/L). The TOC analyzer does not seem to be equipped with a magnesium perchlorate water-trap, which is essential for accurate measurements of low DOC samples, such as seawater. Clarifying these points will enhance the validity of the study's results.
Line 209 & 295: This conclusion cannot be drawn from the experiment presented here. The substantial amount of DOC added to the mesocosm would likely cause hypoxia and coral death regardless of the source or composition of the added DOC. This outcome is not unexpected and cannot be used as an argument to support the notion that DOC-rich coral exudates in particluar cause bleaching and coral mortality.
Line 305: Please clarify whether oxygen levels were directly measured on the reef during the time of bleaching in Moorea to validate the statements made, or if the following conclusions are solely based on the DOC concentration.
Fig. S2: The fonts used in the figure for units and coordinates are too small to read, and the captions need to provide clarification on the meaning of "combined data set."
Fig. S3: Nelson et al 2011 and LTER samples have a higher number of samples than indicated in the figure. Please explain what subset was chosen to create the figure.
Citation: https://doi.org/10.5194/egusphere-2023-779-RC2 -
AC1: 'Reply on RC2', Andrew Thurber, 04 Oct 2023
We would like to thank both reviewers for providing areas where we can refine our manuscript to increase its clarity and better communicate the results. Both reviewers also highlighted important concerns that resulted in changes throughout the manuscript, and we point out how they will be addressed if BGS allows us to submit a revised version. In particular, we acknowledge that our experimental design has limitations, which are now more explicitly stated, however we point out that this is a new line of inquiry and the resulting hypothesis (as posed) provides a novel understanding of how shifting biogeochemistry can alter coral health
Review comments in plain text - responses in bold
The manuscript addresses a highly valuable topic concerning coral bleaching and its impact on dissolved organic carbon (DOC) release. The approach to test the hypothesis of a positive feedback loop that may accelerate reef decline was divided into three parts: 1) DOC measurements in the environment, 2) exposing corals to released DOC due to bleaching in experiments, and 3) modeling possible global impacts of coral bleaching events on DOC release based on a different bleaching experiment.
The study’s approach is interesting and holds great potential for contributing valuable insights into DOC release during bleaching events. However, the experimental set ups and sampling techniques as well as the quality and interpretation of the data are concerning. The number of samples are low and the manuscript is missing statistics to justify their findings. The exercise of calculating the net DOC that could potentially be released during a bleaching event is intriguing. However, it is unclear whether the results from the experiments (considering the coral species and number of replicates) provide a sufficient basis for any global calculation.
We thank this reviewer for these valuable comments and the insight provided throughout. We would like to emphasize both in this response, and we do this explicitly, that the role of DOC release as a causative agent of reef decline is proposed as a hypothesis and we agree that our data do not test this result without caveats (as all experiential work in a new direction will have) - for many of the reasons the reviewer provides (note we are adding “may” in the title). These results will stimulate research and increased understanding the role of bleaching corals in a reef ecosystem. We provide lines of evidence that support a new hypothesis and more research is needed to identify the magnitude of this. This work is the outcome of opportunistic alignment of experimental study coinciding with a major bleaching event on the reef that was our study site. We did not expect such a shift in DOC across the reefscape, or during bleaching in our experimental work, but upon seeing this consistently across both studies considered these results to be important and necessary to communicate to the scientific community to fuel future work to test our resultant hypothesis. Regardless, we do demonstrate an increase of DOC from the reef vs above it during a bleaching event - something that has never been quantified previously. While the magnitude may be impacted by our sampling method, the pattern is not.
Specific comments:
Field observations:
The observed increase in DOC concentration compared to previous data (Nelson et al., 2011 and LTER) raises concerns about potential methodological artifacts. Proper sampling of seawater DOC is essential but challenging and contamination during sampling can significantly impact results (e.g. leaching plasticisers, fumes from the boat engine, contamination by touching the sample). Therefore, consistency in methodology in both sampling and analysis of samples are essential for the direct comparison of DOC datasets. According to the methods, the DOC samples were taken using whirl bags, which are leaching plasticizers and are not an acceptable method to sample for DOC, especially for samples with low DOC concentrations such as seawater. It is highly possible that the observed increase in DOC (Nelson et al., 2011 and LTER data vs. data of this study) is simply caused by the handling of the seawater samples. The inclusion of procedure blanks (e.g., MilliQ treated the same as the samples, including storage in whirl bags and filtering through the same setup) would be essential for evaluating potential artifacts and systematic biases between the two collection methods. Additionally, the fact that offshore concentrations sampled by the authors, which should not be affected by bleaching, are also elevated, further highlights contamination issues.
There are potential methodological challenges associated as highlighted above and we also agree that proper sampling of DOC is challenging. We agree that we should not compare our DOC data to the previous TOC data due to this and have addressed this explicitly and carefully throughout the revised manuscript, and modified it as such. What we do compare, and what we have kept in the manuscript, is the shift in on reef / off reef DOC balance within each data set. The DOC on reef was consistently depleted relative to offshore samples in all LTER data as well as our data at the onset of the bleaching event. During the peak of the bleaching event, we saw this balance shift in the data (coinciding with a shift in water quality along the reef, notably) with the reef becoming enriched when compared to offshore samples within the same dataset, from samples collected on the same day. However, we do argue that contamination - if present - would be uniform across our samples and the relative patterns are robust. This is also a unique time with very high water temperature and so there is no reason to dismiss the relatively high offshore value a priori, as the water was warmer than it has been previously.
In addition, the lack of intercalibration or references hinders comparison across datasets (commonly used for oceanic DOC is Hansell CRMs at <45uM). The measured DOC values of this study exhibit significant variation, and this variation seems to increase with the number of samples (see Fig S2). While this could potentially be attributed to heterogeneity in DOC concentration in the environment, it is highly probable that inconsistent contamination and inaccurate measurement of low DOC are the underlying factors. In the manuscript's methods, it is mentioned that the minimum reproducible and reportable value from UMCES was 13.3 umol C/L. However, the UMCES protocol (link provided in the SI) indicates that the lowest value of the calibration curve is 0.5 mg/L, which suggests that the concentrations of the samples may fall outside the calibration curve window. Additionally, the methods do not specify whether the instrument used was equipped with a magnesium perchlorate water-trap, essential for accurate measurements of low DOC samples, such as oceanic seawater.
This comment has a few points to address:
- DOC values of the site are not lower than 45uM. These are not oligotrophic open ocean waters from both our measurements as well as those from the LTER at this site are not less than 60 uM in concentration. We did make an error as the 13.3 umol is the detection limit and the reporting limit from this method is 41.6 umol - at the time when we analyzed our samples from the analytical lab. We have adjusted the text accordingly, however this is still below the minimum from the waters samples. There was no magnesium-perchlorate water-trap as these are within the reproducible result range of the instrument used.
- None of the samples ran had a value less than 0.5 mg/L - all samples fell within the calibration curve window of the instrument.
- Yes - our data was highly variable however we also were 1m off the seafloor in a highly dynamic location. We would argue that it would be heterogeneity in the environment that led to this high variability rather than differential contamination. In part this is because we sampled these at three different days and replicated across those sites, in each case there was increased DOC on the reef which would be a highly unlikely occurrence if there was stochastic contamination from the whirl packs.
- There are no existing data from the LTER from this time to directly compare. There was sampling earlier in the year before the onset of the main bleaching event so a true intercalibration is not possible with existing data.
Lastly, the presentation of the DOC data raises concerns. The limited number of samples and their spacing do not allow for interpolation, as seen in Fig.2. A more appropriate representation would be Fig S2, which provides a more realistic picture. However, the absence of statistical analysis to support the claims in the paper is a major issue.In the figure legend we highlight the faded section is for aesthetic and communication purposes as to why it did not look like figure S2. We feel this provides a more clear view of the patterns. This has no impact on the interpretation of the results. The statistical treatment was challenged by the high variability across the reef as discussed above (data are now more comprehensively presented in the supplemental.) We would also like to point out that our replication (3 across a transect at three sites) is in contrast to a single data point per depth collected via nisken for both this (LTER) and most long term data sets. So in contrast we have had sample numbers that can elucidate the natural variability across a reefscape that is not always observed.
Lab experiments:
As the authors acknowledge themselves (Line 148), the addition of 2mM DOC in the experiment is an order of magnitude higher than what is typically found in nature. Consequently, it is not surprising that the mesocosms turned hypoxic and led to coral death. At such high concentrations, the source and composition of dissolved organic matter are likely irrelevant. Therefore, drawing the conclusion that the bleaching-derived DOC-rich exudates directly cause bleaching or death to healthy corals is not supported.
We will better emphasize this point. While we had tried to specify this in the text (including calling out the order of magnitude increase) we have further specified this as a hypothesis, one that we did not test (yet). We also more clearly state the impact role of High DOC leading to the hypoxic results and its role in generating hypothesis and stimulating research.
Model:
Making global calculations based on an experiment with an unrealistic time frame (bleaching within 48h) is somewhat speculative. But it is an interesting thought experiment. If the aim is to determine the maximum DOC that a coral reef could release, it would be insightful to provide context. For instance, it would be beneficial to consider the total carbon or total organic carbon stored in the Great Barrier Reef and assess whether the calculated numbers are realistic.
We are unaware and have not been able to find a value for the total carbon pool stored in corals on the GBR. We also agree that this is a thought experiment and appreciate that it was interpreted as such.
Overall, although the proposed mechanism of the positive feedback loop is an interesting hypothesis, unfortunately this study lacks sufficient data to support the existence of this mechanism.
We feel that the real take home message of this is the potential for this mechanism and the significant DOC released from corals when they bleach - something that has not previously been shown during a bleaching event. We now specify this in the conclusion section.
Technical corrections:
Line 108: SI Figure 1 does not present Nelson et al., 2011 data but rather showcases photos of corals.
Fixed.
Line 144: Please specify the material and brand of the filter used.
Fixed and answered in other review.
Line 147: Clarify which DOC samples from the field are being referenced here.
Fixed.
Line 165: Remove the underscore.
Fixed.
Line 186 & 201: Unfortunately, comparing the data of this study to previous data is not feasible due to the issues discussed in the Specific comment section.
Agreed, and we have modified the text as a result of this.
Line 199: While comparing trends within this study's data set is more appropriate, the significant variations caused by flaws in sample preparation and measurement hinder the interpretation of the data. Simply comparing averages is not sufficient; statistical analysis would be necessary to support the author's argument that the reef switches from sink to source. Given the methodological concerns and the limited number of measurements, the authors should be cautious when drawing conclusions from only a few samples taken on a single day.
Data sets on DOC frequently are single points in time without replication. In fact our replication is greater than previous studies. We now have added figures and text to clarify this, including statistics to back it up. Again, we argue that our methodology for DOC measurement is appropriate for the concentrations at this location (they would not be in other regions).
Line 206: Please provide additional evidence to support this claim. According to the methods in the SI, the applied calibration curve appears not to cover the range of seawater samples (with the lowest concentration being 0.5mg/L). The TOC analyzer does not seem to be equipped with a magnesium perchlorate water-trap, which is essential for accurate measurements of low DOC samples, such as seawater. Clarifying these points will enhance the validity of the study's results.
This point is addressed above. These concerns are valid in other locations but not the location that this study was carried out.
Line 209 & 295: This conclusion cannot be drawn from the experiment presented here. The substantial amount of DOC added to the mesocosm would likely cause hypoxia and coral death regardless of the source or composition of the added DOC. This outcome is not unexpected and cannot be used as an argument to support the notion that DOC-rich coral exudates in particular cause bleaching and coral mortality.
We modify the sentences to more cleary state that our data are supportive of and suggestive of the hypothesis rather than conclusive.
Line 305: Please clarify whether oxygen levels were directly measured on the reef during the time of bleaching in Moorea to validate the statements made, or if the following conclusions are solely based on the DOC concentration.
Oxygen data are not available across a sufficient time frame to contrast the bleaching event with normal conditions. These statements were solely based on DOC concentrations.
Fig. S2: The fonts used in the figure for units and coordinates are too small to read, and the captions need to provide clarification on the meaning of "combined data set."
Fixed throughout.
Fig. S3: Nelson et al 2011 and LTER samples have a higher number of samples than indicated in the figure. Please explain what subset was chosen to create the figure.
The LTER collects a single data point at each depth and location and we chose those points that matched up with our site due to the high level of bleaching. We also did not subsample Nelson et al. 2011. This highlights how our data has more replication than the landmark papers that are used to track changes in DOC at these sites.
Citation: https://doi.org/10.5194/egusphere-2023-779-AC1
-
AC1: 'Reply on RC2', Andrew Thurber, 04 Oct 2023
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 602 | 152 | 26 | 780 | 50 | 9 | 10 |
- HTML: 602
- PDF: 152
- XML: 26
- Total: 780
- Supplement: 50
- BibTeX: 9
- EndNote: 10
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1