the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
A proxy of subsurface Chlorophyll-a in shelf waters: use of density profiles and the below mixed layer depth (BMLD)
Arianna Zampollo
Thomas Cornulier
Rory O'Hara Murray
Jacqueline F. Tweddle
James Dunning
Beth E. Scott
Abstract. Primary production dynamics are strongly associated with vertical density profiles, which dictate the depth of stratification and mixed layers. Climate change and artificial structures (e.g. windfarms) are likely to modify the strength of stratification and vertical distribution of nutrient fluxes, especially in shelf seas where fine scale processes are important drivers, affecting the vertical distribution of phytoplankton. To understand the effect of physical changes on primary production, identifying the linkage between density and phytoplankton profiles is essential. Here, the ecological relevance of eight density layers (DLs) obtained by multiple methods that define three different portions of the pycnocline (above, centre, below) was evaluated to identify a valuable proxy for subsurface Chlorophyll-a (Chl-a mg m-3) concentrations. The associations of subsurface Chl-a with surface and deep mixing were investigated by hypothesizing the occurrence at the same depth of any DL and the maximum Chl-a layer (DMC) using Spearman correlation, linear regression, and a Major Axis analysis. Out of 1237 observations of the water column exhibiting a pycnocline, 78 % reported DMCs above the bottom mixed layer depth (BMLD). This suggests that the BMLD is a boundary trapping Chl-a in shallow waters (≤ 120 m). BMLD constantly described Chl-a vertical distribution despite surface mixing indicators, suggesting a significant contribution of deep mixing processes in supporting subsurface production under specific conditions (e.g. prolonged stratification, tidal cycle, and bathymetry). Using BMLD for defining subsurface Chl-a could be a valuable tool for understanding the spatiotemporal variability of Chl-a in shelf seas, representing a potential variable for ecological assessments.
Arianna Zampollo et al.
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2022-140', Anonymous Referee #1, 23 May 2022
General:
The paper presents a surprisingly complex statistical approach (GAM) to a large dataset of chlorophyll profiles, breaking the profiles down into key groups based on profile shapes and specifying a large range of parameters based on density and chl contrasts and gradients.
The methods section is incredibly dense and difficult to keep track of. By the start of the Results section, I am afraid I had got so confused with the details I decided to skip to the Discussion and Conclusions to see if I understood what the key, novel results/messages were. I still found it difficult to glean the important findings and their implications in the Discussion. I was left with the view that a surprisingly complex statistical model had been used to analyse the shapes of a lot of chlorophyll profiles (which implicitly have been assumed to be temporally static?) But I did not feel I had learnt anything useful about the general properties of sub-surface chlorophyll layers. There may well be something here that I have missed – but the challenge of deciphering the methods and holding a large number of acronyms in my mind defeated me.
I think the key, interesting point that is being made (though not clearly articulated) is that descriptions of ocean mixed layers are largely informed by starting with the surface ocean and working downwards from there. In shelf seas, particularly where tidal mixing plays an important role, working upwards from the seabed makes more sense. This is a nice way of thinking about the system. Phytoplankton will utilise the stability of the deepest pycnocline where there is still sufficient light, which will set the nitracline position. But pulling this out of the paper was exceptionally difficult – and I may have missed other key points. If you want to make some clear, useful points about chlorophyll layers in shelf seas in a way that a wide audience will be able to understand and use, then I think the paper needs to lift itself out of the statistics and focus much more on the resulting chl shapes and the processes underpinning them. These ideas may well be in there somewhere, but just too buried in the details for me to extract.
The writing overall tends to make things sound overly complicated. It would be useful to go through and simplify the prose and cull some of the repetition – the paper needs to be more concise and clearly argued if you want it to have some impact.
Specific:
Line 38: “Climate change is introducing….” You could also mention the increasing recognition of possible changes associated with large-scale roll-out of renewable energy in deep shelf seas (e.g. Dorrel et al., 2022: https://www.frontiersin.org/articles/10.3389/fmars.2022.830927/full).
Line 58: “…where the stratification is maintained by tidal cycles mixing the water column through horizontal circulation…” I think this needs rewording. Stratification is not maintained by tidal mixing – the existence and strength of stratification are controlled by a balance between mixing processes (which in NW European shelf seas are generally dominated by tidal mixing) and the source(s) of buoyancy (surface heating and estuarine inputs of low salinity water).
Overall, I get a little confused by the term “deep mixing processes”. Do you mean mixing at the pycnocline or mixing near the seabed?
Line 62: a general statement about ocean productivity and climate change should probably also reference something like Steinacher et al., Biogeosciences, 2010. Clarify that the canonical view is that at low and temperate latitudes in the open ocean productivity will decrease because of strengthening stratification inhibiting vertical mixing of nutrients.
Line 74: What is meant by the nutricline exhibiting positive correlations with MLD? What aspects of the nutricline? The depth, the strength?
Line 107: “Vertical samples….” Do you mean vertical profiles? “Samples” to me implies bottle samples rather than CTD.
Line 109: Is there a particular reason for the choice of 120 metres as the deepest? Is it simply forced by the data available, or do you have a different reason?
Line 118: What does “426 profiles” mean in the context of a mix of towed and vertical-profiling CTD data? Are the individual undulations of the towed systems each counted as a single profile? Is is clearer later inn the paragraph – so maybe the full 1273 profiles needs noting here?
Line 136: “samples’ distance” I think should be “sample vertical resolution”.
Section 2.2.1: Why was a GAM/spline used instead of a simple spline (or an even simpler moving average)? Some justification/explanation of this choice would be useful Also, a couple of example profiles in a Fig would help – e.g. one profile where the GAM worked well and another where the visual fixing was required.
Line 194: not strictly “density gradient” – the values you state are densities.
In two of the Methods sections (2.3 – 2.4) I had to work inordinately hard to see what was going on. I think these sections could be clarified with some better ordering. For instance, AMLD is talked about in section 2.3, but the full description of what it is does not occur until 2.4. There is a raft full of HPDs that pops up line 195-200, but it is unclear what they all mean. If you find yourself having to refer to a section further on in the paper (e.g. line 198 you refer to section 2.4 for the explanation of adjusted AMLD) then you need to rethink how you are structuring the material. You need a clear, logical progression of explanations that does not leapfrog – this is really important, as the reader needs to keep track of a large number of different acronyms and their meanings.
Line 213: “transient” – do you mean “transition”? Unclear what you are trying to say.
Line 215: delta-rho is a density difference, not a density gradient. This occurs a few times.
Lines 216 – 227: It is really hard to understand what is meant here (partially, but not wholly, because when you say “this paper” I cannot work out if you mean your paper or the Chu & Fan paper cited in the previous sentence). Clarification needed.
Around this stage I just got very confused with the methods. They appear rather complicated and dense, and I found them difficult to follow. To me this difficulty began to detract from what I thought the paper was aiming to demonstrate. Perhaps consider a Supplementary Material section to deal with the details of the methods (though they would still need to be clarified) and focus the main paper on the results and implications?
Section 3.1 starts by repeating a lot of the methods. No real results appear until 3.2 and Fig. 4.
I stopped dealing with specific points at this stage - focusing instead on trying to pull out what the key points might be.
Citation: https://doi.org/10.5194/egusphere-2022-140-RC1 -
AC1: 'Reply on RC1', Arianna Zampollo, 08 Jul 2022
Dear reviewer,
We are very thankful for your time and your comments on the paper. According to all the reviewers, we identified some common issues that came across, and we have planned to improve the manuscript following all your advice.
The main points we want to work on are: i) better defining the scope of the paper by deleting the Chl-a shapes from the analyses, ii) simplifying the methods, and iii) providing the code to let users trying with the proposed algorithm.
Below, we describe the main changes we are going to introduce into the paper to address the above points.
The scope of the paper will be clarified by focusing on the BMLD (base of the pycnocline) and its use as a proxy for the depth of maximum Chl-a (DMC) in shelf waters. To date, the paper is packed with many details regarding the co-occurrence at the same depth of any density layer (that we will rename as “level”) (e.g. AMLD, BMLD, DHP and Max N2) and DMC. The current structure of the paper reports first the comparison for all the profiles together (section 3.2) and then the comparison for each Chl-a shape (section 3.3). However, the length of the paper and the amount of information has increased the confusion among all the reviewers, who struggled to identify the main scope of the paper and often focused mainly on issues referred to Chl-a shapes. On the contrary, we have written this paper to promote a different point of view in investigating subsurface Chl-a by using density profiles. Hence, the main aim of the paper is to highlight the BMLD as a useful tool to predict and investigate DMCs in shelf waters. The vertical distribution of DMCs nearby BMLDs suggests that this variable has an ecological relevance when we investigate the vertical distribution of Chl-a subsurface patches, and we suggest its use in further research (enlarging these applications in the Discussion). However, this point does not come across easily, and we decided to delete all the analyses related to Chl-a shapes to focus mainly on the use of the BMLD and its potential. The following paragraphs will be deleted: 2.2 in the methods will not include Chl-a shape identification, 3.3 in the results, 4.1 and 4.2 in the discussion. However, understanding the physical processes underpinning the vertical distribution of each Chl-a shape is an open question, and the presented results showed how each shape exhibits a different association of DMCs with the pycnocline. Hence, we are interested in detailing this question in another paper, to avoid hiding the main scopes of this paper, which are i) proposing a method to extrapolate the base of the pycnocline from density profiles and ii) evaluating its association with the vertical distribution of Chl-a (regardless the Chl-a shape).
The second and third points (“simplifying the methods” and “providing the code to let users trying with the proposed algorithm”) are ensuring that the reader fully understands the method and its potentialities. For this reason, we will reduce the number of details regarding the algorithm in paragraph 2.4 and we will focus on the requirements, limitations, and circumstances in which the method can be used. Since paragraph 3.1 describes what is considered a correct or wrong identification, and is a repetition of the methods, we decided to integrate it into the methods together with figure A1. Moreover, we will upload the code of the function on GitHub, where an example will be also provided. The details regarding the structure of the function will be reported in the supplementary material to allow people to replicate, improve and use the code. Therefore, Figure 3 and part of the methods will be moved to supplementary materials.
The removal of Chl-a shapes from the paper will change the discussion section, which will be reduced and will focus on describing the relationship between density and Chl-a profiles. We will review the physical variables that are playing a role in the definition of BMLD and AMLD, and their association with the vertical distribution of maximum Chl-a in the water column. Figure A2 will be moved to the main text to better understand the vertical distribution of the depth-integrated Chl-a with regard to each density layer (AMLD, BMLD, DHP and Max N2).
Here we respond to your main specific comments:
“I was left with the view that a surprisingly complex statistical model had been used to analyse the shapes of a lot of chlorophyll profiles (which implicitly have been assumed to be temporally static?) But I did not feel I had learnt anything useful about the general properties of sub-surface chlorophyll layers.”
We think this is related to the high number of sub-analyses that were presented in the paper. We wanted to show the relationship between the different density layers and the vertical distribution of Chl-a, and we thought that reporting the information at the level of each Chl-a shape was actually helpful. On the contrary, it created more confusion, and we are considering now focusing on describing how AMLD and BMLD can be used to investigate Chl-a throughout the water column without considering each Chl-a shape (his subdivision can be part of a future analyses/paper). Your comment “I think the key, interesting point that is being made (though not clearly articulated) is that descriptions of ocean mixed layers are largely informed by starting with the surface ocean and working downwards from there. In shelf seas, particularly where tidal mixing plays an important role, working upwards from the seabed makes more sense” was summarising the scope of the paper, and we intend to make this message coming across easier.
“I think the paper needs to lift itself out of the statistics and focus much more on the resulting chl shapes and the processes underpinning them.”
We agree that there is a need of understanding the processes underpinning each Chl-a shape, although we think that this paper may be more suitable to describe the use of BMLD by comparing it to the other characteristics of the density profile (e.g., AMLD), and methods. The investigation of different processes underpinning each Chl-a shape can be expanded by involving further physical variables, which we think would be more suitable for another paper (research question). Hence, we suggest deleting all the sections referring to Chl-a shape and focusing on the different interpretations that can be obtained by investigating Chl-a in association with either AMLD or BMLD.
“the paper needs to be more concise and clearly argued if you want it to have some impact.”
We hope that reducing the number of analyses and details will improve the readability. The methods are going to be eased, and repetitions between methods and results will be solved into a unique section that describes the use of the algorithm.
Citation: https://doi.org/10.5194/egusphere-2022-140-AC1
-
AC1: 'Reply on RC1', Arianna Zampollo, 08 Jul 2022
-
RC2: 'Comment on egusphere-2022-140', Anonymous Referee #2, 09 Jun 2022
Dear Editor.
I have read the manuscript entitled A proxy of subsurface Chlorophyll-a in shelf waters: use of density profiles and the below mixed layer depth (BMLD) by Zampollo A. et. al.
The manuscript deals with the topology of water column profiles (hydrography and Chlorophyll) aiming to relate both in large dataset at a shelf sea region. The work develops methodology for the systematic characterization of the seasonal pycnocline, pursuing objective labelling of transitional depths separating the stratified water column from mixed layers above and below. This applies to the specific circumstance of shelf seas where, below the seasonal pycnocline, there is a bottom mixed layer instead a permanent pycnocline as in Open Ocean.
I consider the subject of the manuscript interesting and I appreciate the authors’ efforts of systematic characterization of water column structure, however I find issues with the methodology and also I find the overall scope unclear. Therefore I cannot recommend its publication at this stage. My main concerns are as follows.
Starting with the methodology, the authors develop an algorithm that delimitate the stratified portion of the water column with special focus on tagging the top of the bottom mixed layer. The algorithm is complex, as can be inferred from Fig.3 diagram. Appraising the need of such iterative clustering-based approach requires time I acknowledge I could not invest for this review, and probably some trials with the code. The authors argue that their method is more accurate than simpler systems as those based in thresholds or gradients. I feel that further comparisons with outcomes produced by simpler methods would provide a stronger case for the use of their complex method. Besides thresholds from top and bottom, approaches based on curve segmentation may provide accurate results. To be clear in this point, I think there is likely a detailed analysis of the large profiles dataset that support the development and use of their algorithm instead of others, but at this stage this is not easily assimilated by the reader.
The main focus seems to be classifying whether the Deep Chlorophyll Maximum (DCM) is located above or below reference pycnocline levels (roughly speaking top/middle/bottom). This is strongly dependent on distinct Chla profile shapes, which are classified in 6 types following literature. If I understand properly, Chla profiles have not been classified or clustered following any systematic objective method but manually. Therefore, one of the main strengths of the work (i.e. providing automatic algorithms to process large amounts of profiles) weakens. The analysis of the dataset is therefore mostly manual and the real advantage in objectively tagging density levels to draw scientific conclusions of their dataset is unclear. To improve the manuscript, I suggest applying an automated Chla profiles classification system so the processing gets fully objective and can be applied to much larger datasets.
I am also confused about the overall scope of the paper. Conclusions are e.g. that AMLD is only correlated to Chla for certain fluorescence shapes (HCU), that there is tendency for deep DCMs in shallow waters (even below the BMLD in HCL shapes), and that DCM lies around the centre of the pycnocline for Gaussian (symmetric) Chla profiles. The discussion of Chla shapes is discussed regarding bibliography but not clearly related to the density profiles. In the end I am not sure if the authors aim to infer subsurface Chla values from BMLD in case there no Chla profiles are available. The relationship of the developed tool and/or shelf seas primary production with man-made structures, as well as possible influence of climate change, is too indirect.
The main goals stated along the paper, which are (i) providing a new analytical tool to systematically tag density profiles, (ii) helping to understand basic processes relating Chla and vertical density, and (iii) providing predictive capability for subsurface Chla at fine scales, are in my view not clearly addressed in the ms in its current status. My recommendation is that the article should be returned for major revision. I encourage the authors to focus on highlighting the improvements provided by the developed tool over other methods and describing how their results address the aforementioned main goals.
I provide some specific comments below, mainly regarding sections 1-3, I hope will help to improve further versions of the manuscript.
------
Specific Comments:
l.14 Abstract and general. The definition/selection of 8 ‘density layers’ instead of other number is not sufficiently justified. These are levels (discrete depths) instead of layers.
l.36. (also l.57). Specific for shelf seas with strong tides. The authors should notice that many shelf seas have small o no tides.
l.41. Bryden et. al 2005 paper here is not adequate. Scale is too broad and main outcomes are superseded by further results of the rapid array and others.
l.89-90. There are no standard methods to MLD identification neither in shelf nor oceanic waters.
l.91. BMLD as an “indicator” of the vertical... Indicator or proxy?
l.96 and others. BMLD is indistinctly referred to as ‘bottom mixed layer depth’ and ‘below mld”. Should address this mismatch.
l.101 ‘this new level of understanding’ sounds a bit presumptuous, maybe just this new algorithm.
l.111. ‘Fig.2’. It is normally requested to cite figures in order, please check.
l.123. ‘standard MSS editing procedure’ requires a reference.
l.132 and others. Not necessary to specify used functions of TEOS-10, this is too much detail.
l.135 et.seq. (section 2.1.1). General, I guess the authors are using Chlorophyll-fluorescence profiles (from a fluorometer) which is not the same as Chlorophyll-a. Should clarify.
l.138. Understand that smoothing/resampling refers only to undulator.
l.149 ‘The analyses were run in R v3.6.3...’ too much detailed. Again in l.204 etc.
l.161-162. I do not understand sentence 'and three equal sections were used to divide the difference between the minimum and maximum Chl-a values into three equal sections'
l.175. Fig.2. why HCL (e) is above HCU (f)? I find this confusing.
l.191 One of the first comprehensive classifications of MLD objective methods available is provided by Thomson and Fine, JAOT, 2003, including curve segmentation aforementioned methods.
l.235 et.seq. why these ad hoc parameters? 2-delta and 90% of the entire profile.
l.240-244. I find confusing that computing the tangent of the angle phi causes issues but computing the angle does not.
l.299 again density layers vs density levels
l.320, Table.2. I miss an explanation for exploring linear regression and ‘one-to-one’ regression. Should intercept of regression be forced to cross zero for any reason?
Section 3.3.1. I find too many numeric details and data in the text, should be embodied in tables or figs. Same issue in 4.2.
Citation: https://doi.org/10.5194/egusphere-2022-140-RC2 -
AC2: 'Reply on RC2', Arianna Zampollo, 08 Jul 2022
Dear reviewer,
We are very thankful for your time and your comments on the paper. According to all the reviewers, we identified some common issues that came across, and we have planned to improve the manuscript following all your advice.
The main points we want to work on are: i) better defining the scope of the paper by deleting the Chl-a shapes from the analyses, ii) simplifying the methods, and iii) providing the code to let users trying with the proposed algorithm.
Below, we describe the main changes we are going to introduce into the paper to address the above points.
The scope of the paper will be clarified by focusing on the BMLD (base of the pycnocline) and its use as a proxy for the depth of maximum Chl-a (DMC) in shelf waters. To date, the paper is packed with many details regarding the co-occurrence at the same depth of any density layer (that we will rename as “level”) (e.g. AMLD, BMLD, DHP and Max N2) and DMC. The current structure of the paper reports first the comparison for all the profiles together (section 3.2) and then the comparison for each Chl-a shape (section 3.3). However, the length of the paper and the amount of information has increased the confusion among all the reviewers, who struggled to identify the main scope of the paper and often focused mainly on issues referred to Chl-a shapes. On the contrary, we have written this paper to promote a different point of view in investigating subsurface Chl-a by using density profiles. Hence, the main aim of the paper is to highlight the BMLD as a useful tool to predict and investigate DMCs in shelf waters. The vertical distribution of DMCs nearby BMLDs suggests that this variable has an ecological relevance when we investigate the vertical distribution of Chl-a subsurface patches, and we suggest its use in further research (enlarging these applications in the Discussion). However, this point does not come across easily, and we decided to delete all the analyses related to Chl-a shapes to focus mainly on the use of the BMLD and its potential. The following paragraphs will be deleted: 2.2 in the methods will not include Chl-a shape identification, 3.3 in the results, 4.1 and 4.2 in the discussion. However, understanding the physical processes underpinning the vertical distribution of each Chl-a shape is an open question, and the presented results showed how each shape exhibits a different association of DMCs with the pycnocline. Hence, we are interested in detailing this question in another paper, to avoid hiding the main scopes of this paper, which are i) proposing a method to extrapolate the base of the pycnocline from density profiles and ii) evaluating its association with the vertical distribution of Chl-a (regardless the Chl-a shape).
The second and third points (“simplifying the methods” and “providing the code to let users trying with the proposed algorithm”) are ensuring that the reader fully understands the method and its potentialities. For this reason, we will reduce the number of details regarding the algorithm in paragraph 2.4 and we will focus on the requirements, limitations, and circumstances in which the method can be used. Since paragraph 3.1 describes what is considered a correct or wrong identification, and is a repetition of the methods, we decided to integrate it into the methods together with figure A1. Moreover, we will upload the code of the function on GitHub, where an example will be also provided. The details regarding the structure of the function will be reported in the supplementary material to allow people to replicate, improve and use the code. Therefore, Figure 3 and part of the methods will be moved to supplementary materials.
The removal of Chl-a shapes from the paper will change the discussion section, which will be reduced and will focus on describing the relationship between density and Chl-a profiles. We will review the physical variables that are playing a role in the definition of BMLD and AMLD, and their association with the vertical distribution of maximum Chl-a in the water column. Figure A2 will be moved to the main text to better understand the vertical distribution of the depth-integrated Chl-a with regard to each density layer (AMLD, BMLD, DHP and Max N2).
Here we respond to your main specific comments:
“To be clear in this point, I think there is likely a detailed analysis of the large profiles dataset that support the development and use of their algorithm instead of others, but at this stage this is not easily assimilated by the reader.”
Thank you for the comment. We agree that the number of the information reported in the paper is too large and this leads all the reviewers to focus on different outcomes instead of focusing on BMLD and its potential. Your comments indicated to clarify the scope of the paper and we intend to do it by following the points we described in the first section of the response.
“If I understand properly, Chla profiles have not been classified or clustered following any systematic objective method but manually. Therefore, one of the main strengths of the work (i.e. providing automatic algorithms to process large amounts of profiles) weakens.”
Unfortunately, we could not classify the Chl-a shapes using an automated algorithm although we tried a cluster classification of them. We followed a few papers that tried to classify the shapes of Chl-a, but their methods were not successful with our dataset. However, the main point of this paper is not to find a method to classify Chl-a shapes (which is still an open issue), but proposing a different way of investigating the relationship between density and Chl-a. The identification of BMLD in the water columns was not described before in the literature, and its use is a valuable tool to investigate subsurface processes (underlying the pycnocline structure). Therefore, this paper intends to focus on the BMLD and Chl-a, leaving aside the classification of Chl-a in shapes. In future analyses, we are interested in pursuing the investigation of different processes underpinning each Chl-a shape which will be more suitable for a different publication (different research questions from what we want to present in this paper).
“The discussion of Chla shapes is discussed regarding bibliography but not clearly related to the density profiles.”
Considering the large number of changes that we are going to apply to this paper, the discussion will be reviewed focusing on the density and Chl-a profiles, and the physical variables underpinning the surface and deep mixing layers.
“In the end I am not sure if the authors aim to infer subsurface Chla values from BMLD in case there no Chla profiles are available.”
Since the relationship between BMLD and DMC can be described by a linear regression, the DMC can also be inferred by looking at the BMLD or DHP. This will be more clarified in the paper.
“The relationship of the developed tool and/or shelf seas primary production with man-made structures, as well as possible influence of climate change, is too indirect.”
We agree that passing from the surface and deep mixing layers to man-made structures and climate change is a considerable step, however the mixed layer depth is influenced by physical variables that are likely to be affected by both climate change and man-made structures, especially those from the offshore renewable energy. Hence, we consider it essential to mention that understanding the vertical distribution of density and Chl-a is important to guide the investigation of disturbances' effects (climate change and offshore renewable energy) on the right physical variables. Identifying a tight overlap between BMLD and DMC suggests that physical variables close to the seabed (e.g. bottom temperature) are indeed key variables to address the effects of disturbances on primary production. On the contrary, the exclusive investigation of the surface processes (up to AMLD) may lead to partial conclusions about the effects due to climate change or man-made structures. Hence, describing the implications of BMLD in characterising the effects of climate change or man-made structures e.g. wind turbine foundations (which are likely to impact the mixing of the water column) on the ecosystem is useful to give a context of the potential uses of these variables in further investigations.
“The main goals stated along the paper, which are (i) providing a new analytical tool to systematically tag density profiles, (ii) helping to understand basic processes relating Chla and vertical density, and (iii) providing predictive capability for subsurface Chla at fine scales, are in my view not clearly addressed in the ms in its current status.”
We hope that the points described at the beginning of this response will clarify the aims of the paper.
Citation: https://doi.org/10.5194/egusphere-2022-140-AC2
-
AC2: 'Reply on RC2', Arianna Zampollo, 08 Jul 2022
-
RC3: 'Comment on egusphere-2022-140', Anonymous Referee #3, 20 Jun 2022
General comments
The manuscript (MS) presents results on relationships between chlorophyll and density profiles in shelf sea conditions. The work is based on in situ data and focuses on the need to investigate if and how the main features of deep chlorophyll maximum conditions can be related (or even inferred) to observations of density profiles.
The topic is relevant for the scientific community due to the growing need to understand biogeochemical processes and impacts on them by climate modifications or human interventions in coastal and shelf areas. A great effort has been spent by the Authors to address the topic with a set of novel and relatively high-complexity tools. However, even if on one side these novel methods are an added value of the MS, on the other hand their description is probably excessively detailed and not easily readable resulting in a quite long MS that probably lacks of clarity also in the results, discussion and conclusions.
I suggest to deeply revise the MS in order to better highlight its scientific value.
Specific comments and Technical corrections
The authors should consider to address the following aspects to improve the MS quality.
- Improve the methods section readability, reducing the description of the methods and avoiding non-necessary details and repetitions in the methods.
- Generally improve the quality of the figures; moreover tables and figures could be used to resume results in a more comprehensive way. As a consequence, the results section could be focused on the most relevant results and significantly shortened.
- Discussion section is quite long and one of the focus of the MS (wind farm impacts) risks to be lost by the reader. The Authors should consider to focus the Discussion on few relevant themes.
In the following some minor points are listed. Maybe additional minor issues could be revised, however I think that these aspects can be better addressed in a next review step after a deep revision of the manuscript.
In the abstract, it should be clearly stated that the work is based on experimental measurements. Moreover, the sentence about the eight density layers and the three different portions of pycnocline could be modified in order to better describe the methods applied. If I correctly understood: i) pycnocline has been identified on a set of density profiles; ii) three portions of the density have been identified with respect to the pycnocline (above, centre, below); iii) density profiles have been classified in eight types; iv) in the density profiles, possible proxies the position of deep chlorophyll maximum have been investigated. If corrected, these steps should more clearly result in the abstract.
L. 19 in the abstract: instead of “<=120 m”, consider “depth <=120 m”.
L. 42: Maybe “processes” instead of “effects”.
Some parts that could be removed or significantly reduced in the introduction, since they seem repetitions or they are not very informative with respect of the MS objectives:
L. 46-47: “The vertical […] in the marine environment”.
L. 66-68: “The exclusive […] needs to be investigate further”.
L. 82-90.
L. 95: Are you meaning “the distance” instead of “the depth”?
L. 101: “the performance of these two proposed density layers” can be misleading, since it is not evident what a density layer performance mean. Maybe it could be rephrased with “we compared results with other relationships between density layers and Chl-a proposed in literature”.
L. 107-114: The first paragraph of the Methods section seems more suitable to the introduction. Consider if it can be feasible in the new version of the manuscript.
L. 112: I suggest to consider to replace “identify” with “to identify”.
L. 113 “is evaluated by comparing the vertical distribution of subsurface Chl-a”: to clarify the comparison cited in the sentence, I suggest to consider the following rephrasing “is evaluated thanks to comparison of BMLD with the vertical distribution of subsurface Chl-a.
L. 121: The indication of the years (from 2000 to 2014) can be moved at L. 118, where the time length of measurements is cited firstly.
Some details about instruments could be probably removed:
L. 122-123: “Temperature and conductivity […] editing procedure”.
L. 130-133: From “In situ” to the paragraph end.
L. 141: “predict” is a word that is usually relate to forecast, in this sentence maybe “interpolate” is more appropriate.
L. 144 From “The pre-processing” to the paragraph end: my impression is that this sentence can be shortened removing non-necessary details, or delated.
L. 152: Maybe DCM is a more usual way to identify the subsurface (or deep) Chl-a maximum. However, I understand that the authors are aiming at defining an abbreviation for the depth of the Chl-a mximum (that is not strictly DCM, indeed); I suggest to consider CMd (Chl-a maximum depth) to avoid confusion with DCM.
L. 154: Here Eq. 1 is cited, but It appears three page later. Usually equations are cited more closely to their appearance in a MS. Consider to move the equation and the first time it is cited closer.
L. 164-165: Consider to rephrase as follows: “Only 2% of the profiles were excluded from the dataset due to unclear subdivision or very different shapes.
Fig. 2: It would be more consistent to indicate with a letter (a, b , etc.) each sub-plot of the figure (the left plot is not labelled with a letter). In the right plots, the “Depth” arrow should point toward the bottom.
L. 182: I think that “rectangles” is more suitable than “squares”.
L. 186: “Among” (capital A) instead of “among”.
L. 200: Maybe “maximum squared buoyancy frequency” instead of “maximum buoyancy frequency squared”?
L. 210-216 illustrate characteristics of AMLD and BMLD and methods to identify them, however AMLD definition and identification methods are discussed also at lines 189-195. Consider to condensate in a unique paragraph.
L. 224-227 seem a repetition of the strategy adopted in the MS.
L. 228-292: Please, consider to move detail of this method to an Appendix.
L. 346-360: these lines contain some repetitions of details provided in Methods section. They can be significantly shortened or removed.
L. 392: A bracket is missing after Fig. 4c.
L. 440: “amount of phytoplankton” is perhaps misleading, since chlorophyll is evaluated here (and not phytoplankton biomass).
L. 514-516: “demonstrates” seems quite strong in this context. Please, consider “suggest” or “indicate”.
L. 649-651 and L. 655-660 provide valuable discussion points.
Citation: https://doi.org/10.5194/egusphere-2022-140-RC3 -
AC3: 'Reply on RC3', Arianna Zampollo, 08 Jul 2022
Dear reviewer,
We are very thankful for your time and your comments on the paper. According to all the reviewers, we identified some common issues that came across, and we have planned to improve the manuscript following all your advice.
The main points we want to work on are: i) better defining the scope of the paper by deleting the Chl-a shapes from the analyses, ii) simplifying the methods, and iii) providing the code to let users trying with the proposed algorithm.
Below, we describe the main changes we are going to introduce into the paper to address the above points.
The scope of the paper will be clarified by focusing on the BMLD (base of the pycnocline) and its use as a proxy for the depth of maximum Chl-a (DMC) in shelf waters. To date, the paper is packed with many details regarding the co-occurrence at the same depth of any density layer (that we will rename as “level”) (e.g. AMLD, BMLD, DHP and Max N2) and DMC. The current structure of the paper reports first the comparison for all the profiles together (section 3.2) and then the comparison for each Chl-a shape (section 3.3). However, the length of the paper and the amount of information has increased the confusion among all the reviewers, who struggled to identify the main scope of the paper and often focused mainly on issues referred to Chl-a shapes. On the contrary, we have written this paper to promote a different point of view in investigating subsurface Chl-a by using density profiles. Hence, the main aim of the paper is to highlight the BMLD as a useful tool to predict and investigate DMCs in shelf waters. The vertical distribution of DMCs nearby BMLDs suggests that this variable has an ecological relevance when we investigate the vertical distribution of Chl-a subsurface patches, and we suggest its use in further research (enlarging these applications in the Discussion). However, this point does not come across easily, and we decided to delete all the analyses related to Chl-a shapes to focus mainly on the use of the BMLD and its potential. The following paragraphs will be deleted: 2.2 in the methods will not include Chl-a shape identification, 3.3 in the results, 4.1 and 4.2 in the discussion. However, understanding the physical processes underpinning the vertical distribution of each Chl-a shape is an open question, and the presented results showed how each shape exhibits a different association of DMCs with the pycnocline. Hence, we are interested in detailing this question in another paper, to avoid hiding the main scopes of this paper, which are i) proposing a method to extrapolate the base of the pycnocline from density profiles and ii) evaluating its association with the vertical distribution of Chl-a (regardless the Chl-a shape).
The second and third points (“simplifying the methods” and “providing the code to let users trying with the proposed algorithm”) are ensuring that the reader fully understands the method and its potentialities. For this reason, we will reduce the number of details regarding the algorithm in paragraph 2.4 and we will focus on the requirements, limitations, and circumstances in which the method can be used. Since paragraph 3.1 describes what is considered a correct or wrong identification, and is a repetition of the methods, we decided to integrate it into the methods together with figure A1. Moreover, we will upload the code of the function on GitHub, where an example will be also provided. The details regarding the structure of the function will be reported in the supplementary material to allow people to replicate, improve and use the code. Therefore, Figure 3 and part of the methods will be moved to supplementary materials.
The removal of Chl-a shapes from the paper will change the discussion section, which will be reduced and will focus on describing the relationship between density and Chl-a profiles. We will review the physical variables that are playing a role in the definition of BMLD and AMLD, and their association with the vertical distribution of maximum Chl-a in the water column. Figure A2 will be moved to the main text to better understand the vertical distribution of the depth-integrated Chl-a with regard to each density layer (AMLD, BMLD, DHP and Max N2).
Here we respond to your main specific comments:
“However, even if on one side these novel methods are an added value of the MS, on the other hand their description is probably excessively detailed and not easily readable resulting in a quite long MS that probably lacks of clarity also in the results, discussion and conclusions.”
Thank you for your comment. We agreed the paper is long and many details in the methods (section 2.4) can be moved to supplementary materials. As we mentioned in the first section of this response, we intend to shorten the manuscript and delete the sections with Chl-a shapes, which can be part of a further paper focused on understanding the physical variables underpinning each Chl-a shape.
“Improve the methods section readability, reducing the description of the methods and avoiding non-necessary details and repetitions in the methods.”
We hope that reducing the number of analyses and details will improve the readability. The methods will be eased, and repetitions between methods and results will be solved into a unique section that describes the algorithm's use. Moreover, the details of the algorithm will be moved to supplementary materials, and the code will be provided in a GitHub repository.
“the results section could be focused on the most relevant results and significantly shortened.”
Deleting the sections referring to Chl-a shapes will strongly decrease the amount of information in the paper, and clarify the BMLD’s uses (main aim).
“Discussion section is quite long and one of the focus of the MS (wind farm impacts) risks to be lost by the reader. The Authors should consider to focus the Discussion on few relevant themes.”
As described before, we intend to reduce the discussion and focus on the density and Chl-a profiles. The discussion will focus on describing the relationship between density and Chl-a profiles, reviewing the physical variables that are playing a role in the definition of BMLD and AMLD, and their association with the vertical distribution of maximum Chl-a in the water column. Moreover, we consider it essential to mention that understanding the vertical distribution of density and Chl-a is important to guide the investigation of disturbances' effects (climate change and offshore renewable energy) on the right physical variables. Identifying a tight overlap between BMLD and DMC suggests that physical variables close to the seabed (e.g. bottom temperature) are indeed key variables to address the effects of disturbances on primary production. On the contrary, the exclusive investigation of the surface processes (up to AMLD) may lead to partial conclusions about the effects due to climate change or man-made structures. Hence, describing the implications of BMLD in characterising the effects of climate change or man-made structures e.g. wind turbine foundations (which are likely to impact the mixing of the water column) on the ecosystem is useful to give a context of the potential uses of these variables in further investigations.
Citation: https://doi.org/10.5194/egusphere-2022-140-AC3
Arianna Zampollo et al.
Arianna Zampollo et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 504 | 143 | 11 | 658 | 10 | 47 |
- HTML: 504
- PDF: 143
- XML: 11
- Total: 658
- BibTeX: 10
- EndNote: 47
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1