Short-Term Response of Chlorophyll a Concentration Due to Intense Wind and Freshwater Peak Episodes in Estuaries: The Case of Fangar Bay (Ebro Delta)

Round 1

Reviewer 1 Report

This manuscript addresses the Chla-concentration variability at short time scales in a shallow and microtidal estuarine system (Fangar Bay). The role of wind and freshwater discharges is analyzed from satellite imagery and in-situ moorings. Intense wind events cause Chla to increase, whereas weak winds (breezes) do not trigger high Chla values. It is hypothesized that nutrient resuspension from the bottom, due to the enhanced turbulent mixing by wind, favors the increase of Chla. No correlations are found between freshwater discharges and Chla values. 

Overall, this Reviewer's impression is positive. However, the manuscript needs to be improved before I can recommend it for publication. 

The manuscript is compelling and fits nicely the scope of the Special Issue (and, thus, of Water journal). The study of the short-term variability of Chla due to wind, tide and/or river discharges is a topic that (a) goes further the well-known seasonal variations of Chla, and (b) is gaining interest in the coastal physical oceanography community. The methodology, which is described meticulously, is coherent with the objectives the authors pursued. The manuscript imaginatively combines the analysis of data from comprehensive field campaigns and satellite imagery. I think this particular issue will be appreciated by potential readers of Water. Although there is room for improvement, the paper is overall well-written and easy to follow. English usage could be improved, though. The structure of the manuscript could be improved at some points, particularly in the results description and its link with figures (see specific suggestions on how to improve the manuscript below). 

My major concern is regarding the conclusion that Chla concentration increases are due to resuspension of nutrients. I have (reasonable) doubts about that. 

Certainly, resuspension of nutrients is a plausible hypothesis. Furthermore, it is clear that wind-generated turbulence is the factor that controls resuspension (directly or indirectly through wind-waves). However, resuspension of Chla-containing biomass, such as algal microorganisms that grows at the bottom and margins (microphytobenthos), could be another very plausible mechanism that explains Chla concentration increases. This is a very common mechanism that drives Chla variability in estuaries and bays. In fact, a substantial part of the biomass that contribute to the Chla measurements may have its source at the bottom. Although with the available data is difficult to be conclusive, there are clues in authors' results that indicate that this could be indeed the case. (a) There are strong correlations between mixing, Chla and turbidity (Figs. 2 and 4). (b) Several snapshots in Fig.5. that show the highest concentrations of Chla near the coast, where the bay shoals and light penetrate deeper in the water column, thereby favoring the growth of algal microorganisms at the bottom. (c) The lack of correlation with freshwater inputs, which are usually loaded with nutrients. Also, Delgado (1989) in ECSS, https://doi.org/10.1016/0272-7714(89)90007-3 already discussed the presence of microphytobenthos in the Fangar Bay. There are evidences in this and other estuaries and bays that mud and microphytobenthos could be resuspended simultaneously from bottom and shoals by erosion due to wind, waves, and tides (de Jonge & van Beusekom (1995) in LO, https://doi.org/10.4319/lo.1995.40.4.0776 ; Díez-Minguito & de Swart (2020) in JGR:Oceans, https://doi.org/10.1029/2019JC015188). I suggest that the authors discuss this in section 4.

Specific comments:

———

L38. This reference is out of place here. Neither related to habitats nor ecosystem characteristics. Move it where you mention wind-induced estuarine circulation in the bay for the first time. 

L40. More broad citations would be appreciated here. 

L50. …due to an increase in turbulent mixing (by wind?). 

L50. Episodic. …event?

L51. open.

L60. It would be nice to identify those in a map of the study area. 

L64-67. Could the authors cite a few of them? 

L68. Please be more precise here. ‘Shallow, microtidal, wind-dominated environment’?

L96. Rain events / rainfalls…. 

L96. Local mean sea level at the coast. 

L96. NE winds, which …, are responsible

L111. …is (highly?) stratified

L114-115. Why is that? Please explain briefly. 

L125. Please indicate here sign criteria for currents. 

L189. I suggest to change event labels to avoid confusion (M2 is a tidal constituent!). 

Figure 3. No turbidity data here? 

L195- In my view, it is difficult sometimes to follow the authors’ story if they are continuously going from one figure to another… I suggest to warm the readers that the results are presented ‘by events’ not ‘by figures’.  

L206. field campaign/survey

L206-208. I suggest to move this sentence elsewhere (introduction?). Here simply state the satellite data is confronted with insitu data. 

Figure 5. In my view, Fig.5 should be presented earlier than Fig.4. Makes more sense to me. After all, Fig.2 and 3 show insitu data and Fig.4 compares insitu and satellite data. 

L212-213. Overestimate, underestimate

L235. You mean ‘inversion of the classical (or positive) estuarine circulation’, I guess. There is estuarine circulation in both cases. Thus, please, indicate explicitly in the NW case that a classical estuarine circulation is observed, i.e., outflow near the surface and inflow near the bottom. 

L243-244. ‘Turbidity *also* increases during this event’ . Ok, but note that nothing is mentioned above regarding turbidity in other events (boxes M1, M2, and M3). Moreover, it would be nice to plot in Fig.4 turbidity, too! Biomass also contributes to turbidity and it could provide further insights about the origin of the Chla peaks. 

L256. Could the authors provide estimates for mixing or mixing rates? Potential energy anomaly or a Richardson-like number as a function of time? I think they have the data. This would allow to quantify the stratification-destratification by wind. 

L269-270. This lack of correlation between freshwater discharges and Chla concentrations is interesting. Typically, freshwater discharges from crops carry a lot of nutrients. 

L285-318. As authors mention, nutrient resuspension by enhanced levels of turbulent mixing is just one possible hypothesis that may explain the increase in Chla concentration. However, there are others. Resuspension of Chla-containing biomass such as algal microorganisms (microphytobenthos) is another very plausible one (which is very common in estuaries and bays!) With the available data, authors cannot be certain. Therefore, please discuss another possible mechanisms. 

L319-320. Do you mean in the deepest parts? If so, Figure 5 seems to show the opposite behavior, i.e., higher Chla concentrations near the coast. Am I right? 

Figure 7. I think this is a great figure. It would be far more clear if the authors could schematize that Chla increases are somehow related to resuspension of matter (nutrients or Chla-containing biomass) from the bottom. 

L342-345. I guess this supports the idea of resuspension of biomass from the bottom. 

L352-355. Are Sentinel-3 data products better in this sense? Discriminate turbidity from fluorescence Chla measurements is a problem with insitu measurements, too… 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

The study sought to link physical forcing of Fangar Bay by winds to chlorophyll-a dynamics, with important implications for water quality and aquaculture. The relevance of the study is quite high, as providing direct responses between water quality and physical forcing is difficult, and the observations provided here, if analyzed more extensively, could provide an ideal framework for other studies to consider.

However, as the research is currently presented, there are many issues with presentation and how the results are analyzed, limiting the authors' conclusions. The authors also focus almost exclusively on work in Fangar Bay in the Introduction and Discussion, rather than extending the topic to other systems. Beyond limiting the scope of the work, another consequence of this approach is that the fundamental drivers of change that the authors reference are not well cited. For example, what work has been done in other systems that could be leveraged here?

To draw conclusions between meteorology and satellite observations, very clear, statistically-supported links are required. Here, the authors rely on overlapping observations, rather than drawing clear conclusions through statistical analysis. An example of how this could be made more clear would include a simple correlation between wind velocity and chlorophyll-a, ideally with a temporal lag considered (e.g., a cross-correlation analysis).

This submission is very relevant with significant potential, but further work should be considered to improve the manuscript for future submission.

1) The author's thoroughly describe the Sentinel-2 sensors, but never mention what chlorophyll-a algorithm is used. The knowledge about the sensors stated here is widely available and/or known. What isn't known is what the author's did beyond the level 1c product - what atmospheric corrections were applied? What chlorophyll a algorithm was used? Did the author's consider a total suspended matter algorithm to consider the effect of winds on resuspending sediments, which may bias chlorophyll a estimates? Presumably Fangar Bay is a very optically complex body of water, particularly after wind events, with additional issues of adjacency to land that can impact chlorophyll a estimates. It isn't unreasonable to use a standard chlorophyll a product and proceed with analysis, particularly with relatively good agreement between in situ and satellite estimates (note, these are not matchups, but do show similar temporal trends). However, the process of how chlorophyll a was derived from Sentinel-2 reflectance products needs to be thoroughly described to better understand potential biases in the product and build confidence in the authors' findings.

2) Please consider a stick plot for wind time series plots (Fig. 2a and 3a). This will make it much easier to draw comparisons between data. Additionally, turbidity in Fig. 2g would likely benefit from a log scale, to better view change over time. I did not find Fig. 6 relevant. Fig. 7 could be useful to the reader, but requires more detailed analysis prior, and ideally a more developed visualization. In its current format, it illustrates a direct response between winds and flow, but also appears to convey that any winds will induce a response in chlorophyll a. I don't think this is what the authors' want to convey, so further developing this figure to clearly summarize the key findings is needed and would be quite helpful.

3) The manuscript relies heavily on analysis with eyes. The work would be significantly improved with statistical relationships between parameters, and more robust analysis between parameters. For example, a cross-correlation analysis between wind and chlorophyll-a within an observation window suitable for the measurements would better describe how Fangar Bay responds to winds. Ideally, this would be done in situ, and then see if it holds for satellite observations. In the event in situ data isn't available for comparisons, there are modeled winds (e.g., NCEP NCAR reanalysis data products) available that could provide a suitable stand in product. The authors could then consider relationships between wind velocity and chlorophyll response based on a finding of a lagged response that would build confidence in a clear relationship between the two parameters. Examples to consider are a cross-correlation analysis, which would find the degree of lag between winds and chlorophyll a response (e.g., a 1 day lag between chlorophyll a and wind velocity). From here, a correlation plot may reveal a relationship, and whether a threshold wind velocity is required for a response in chlorophyll a.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

This paper has been well discussed how environmental factors changed from external factors have ecologically affected the estuary. This is generally logical but requires minor revisions. Please refer to the points below.

Abstract: In the abstract, I hope that the core content will be written more clearly and the research goals will be clearer.

Line 71-77: Please move to another paragraph or delete it.

Line 100: How much fresh water introduced the estuary? (seasonally?)

Fig.1 Add scale to the map. It is also difficult to determine the location of the river.

Fig. 2,3 Capital and small letter. M6wind. Letter size.

Fig. 4 It is necessary to explain the difference between the field and satellite results in M6.

Fig. 5 it seems to be changed more clearly. (ex. Place of m1 m4~)

Line 271: Where is it written about correlation?

Line 346-364: Probably, benthic algae re-float and contribute to total chl.a because of low depth. Please discuss.

Fig. 7 It is still difficult to understand why chlorophyll is increased in both cases. Please provide additional explanation.

Line 388 It seems to be an exaggerated interpretation. The breeze is difficult to destroy the stratification, and it is unlikely that the stratification strength will be strong at low depths (just 4 meter).

Author Response

Please see the attachment

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Authors did a good job addressing Reviewers comments. In my opinion, the manuscript can go to production after a minor revision of text and format edition.

L25 (here and elsewhere) 'Chla-contanining biomass' or 'Chla concentration' 
L70. In addition, remote sensing of... 
L173-192. Normal (not bold) fonts
L184-185. Please, improve format of inline equations.
L189. Capitalize (or not) Chlorophyll-a. Please be consistent throughout the whole manuscript. 
L197. Reference to Figure not found. Please add. (Here and elsewhere)
L219. sun glitter instead of sunlight reflection? 
L232-233. \rho normal font (not bold)
L246. I think it would be helpful for the reader if the authors add the expresion of the Wedderburn number here. 
L254-266. Multiple references to figures are missing. 
L268. Breeze episode: Multiple references to figures are missing (check the whole manuscript out here and elsewhere). Please use not italic fonts here. 
L270-274. Italic font for captions, I think.
L307. '...). Nevertheless...' (missing blank)
L341. Unhopefully? I guess it should read 'Unfortunatelly, the lack of nutrients...'
L374. Missing reference. 
L388. Another, missing reference. 
L425. Skip 1 line. 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

Overall Comments

The revisions overall have improved the manuscript. The authors would benefit from ensuring some of the responses to reviewers, incorporated into the manuscript, better fit into the overall flow of the paper.

My primary concern is the use of the standard Chl a algorithm available in SNAP. Soriano-González et al. (2019) provided an approach that works well in Fangar Bay for estimating Chl a, using ACOLITE which is specifically designed for high spatial resolution sensors (i.e., Sentinel-2). Those authors recommended the NDCI algorithm to estimate Chl a in Fangar Bay. While the authors show generally good agreement in trends (Fig. 5), it should also be noted that some of the more direct matchups that may be suitable for typical cal/val analysis of satellite data still show ~300% differences (e.g., July 22, 2019 in Fig. 5b). Considering that validated products for this system exist, I would highly recommend the authors use these approaches. At the very least, Rrs within SNAP should be applied to the NDCI algorithm if the authors find ACOLITE not accessible for this paper. This will significantly improve how this work is received.

The maps provided in Fig. 4 are also a bit concerning. The authors note that bottom reflectance may be an issue in Fangar Bay in the Discussion. Based on the turbidity of the system, I suspect satellite products are fine. However, the fact that some of the pixels over Punta del Fangar are reported as valid water pixels makes me question how well the land is being masked, what pixels are being flagged and the overall approach. I understand satellite data needs to be available in an operational sense. However, the quality of the data being used still needs to be assured. I recommend the authors re-visit the algorithm they are using, ensure the appropriate masks are applied, and disregard analysis of land-adjacent pixels (e.g., observations of high chlorophyll along the shore). These steps will ensure this work is well received by the diverse, interdisciplinary community it is intended for.

Specific Comments

Line 16: I suggest changing “is basically” to “are primarily”

Line 18: “campaigns” should be “campaign”

Line 19: I suggest changing “Chl a at surface” to “surface Chl a”

Line 28: “deserved” should be “deserve”

Line 38: Replace “used to be” with “are”, and remove “a” before “a complex”

Line 45: Remove “a” before “a observed”

Line 52: I suggest changing “how” to “what”

Line 70: “episodic” should be “episodes”

Lines 70-72: “In addition, Remote sensing of small and complex coastal areas still faces several challenges so inter-comparison exercises with in situ data are needed [11].” I recommend the authors remove this sentence. This references cal/val efforts for complex coastal areas, and the results published here are not such an effort. Certainly, the trends in Fig. 5 are compelling, but the lack of direct matchups makes it difficult to discern how well the SNAP Chl a products are truly working in this system (see comments below for lines 183-186).

Line 78: “discuss the presumably” should be “discusses the presumable”

Line 98: I do not think the added “which” should be here. I recommend removing and maintaining as it was in the previous version.

Line 130: “whit” should be “with”

Line 138: “station” should be “stations”

Line 157: I suggest changing “during” to “for”

Lines 183-186: It’s my understanding that the IOP model in SNAP built for Case 2 waters was designed with data from the North Sea. Do the authors consider Fangar Bay optically similar to the North Sea? I understand that the trends in Figure 5 seem reasonable; however, Soriano-González et al. (2019) recommended the NDCI algorithm for Chl a estimates in Fangar Bay. I would highly recommend using this approach, as it has been established to work in Fangar Bay. I do not think the “off-the-shelf”, pigment-based chlorophyll-a algorithm provided within SNAP is suitable for use in Fangar Bay without validating it within this system. My concern is accepting this in its current format for publication will set a poor precedent for other researchers.

Line 197: It may be an issue with my version, but here and elsewhere I am viewing an “Error! Reference source not found” in place of a figure.

Lines 205-209: The sentence starting with “In addition, […]” covers 4 lines worth of text. I recommend breaking up this sentence for easier reading and clarity.

Line 216: Add an “n” to “Sentiel-2”

Line 220: “also show also increases” should be “also show increases”

Figure 4: Based on these results, it looks like the Punta del Fangar is tidally inundated. Regardless, I would expect a complete land mask for this portion, but the masking is irregular between scenes. Have the authors considered using ACOLITE, as in Soriano-González et al. (2019)? I am not entirely familiar with how land is masked within SNAP, but it appears there may be issues with the current processing. I encourage a strong look at this so readers have full confidence in the satellite chlorophyll-a results.

Line 226: “increases” should be “increase”

Line 230: “follows” should be “follow”

Lines 231-234: I don’t agree with the added sentences here. The comparison of in situ and satellite Chl a is really just a general evaluation. In terms of a true intercomparison, you would need many more direct comparisons. For example, ESA and NASA generally hold validation matchups as within +/- 3-4 hours. Developing relationships for your dataset doesn’t build trust in the algorithm performance in a true validation sense.

Line 236: “wind” should be “winds”, or “calm” should be “calms”

Line 244: “peaks” should be “peak” and “shows” should be “show”

Lines 246-250: This paragraph feels a bit out of place. Can it be better incorporated into an existing paragraph?

Lines 280-281: I suggest changing “These peaks are originated by the increased pumping of water.” to “These peaks originated from increased pumping of water.”

Line 301: I suggest changing “face” to “consider”

Lines 333-335: The Chl a estimates will be most uncertain near the shoreline due to land adjacency effects. I do not recommend using this data to justify this process.

Line 341: “Unhopefully” should be “unfortunately” and “fully” should be “full”

Lines 355-357: Please re-word this sentence: “Also, strong wind intensity duration during sea-breeze used to last few hours in comparison to E-NE and NW events that are associated at duration of few days being thesea-breeze less effective.”

Line 359: I suggest changing “from bottom at the water column” to “from the bottom of the water column”

Line 409: Remove “the” before “the co-responsible”

Line 423: I suggest changing “a good skill of Chl a evolution at wind and freshwater peaks episodes in a forecasting context” to “an improved understanding of Chl a evolution from episodic wind and freshwater discharge events for forecasting purposes.”

Line 442: “process” should be “processes”

Line 443: “camping” should be “sampling” and “in improve” should be “on improving”

Lines 444-446: “New Sentinel products (such as Sentinel-3) with the ability to measure turbidity and Chl a jointly with an increasing of the temporal and spatial resolution may benefit the Chl a evolution analysis at short-term.” Sentinel-3 does not have the spatial resolution required to accurately characterize Fangar Bay. Full resolution OLCI data is 300m; thus, if any pixels were reported, the coverage of Fangar Bay would still be incomplete. I recommend the authors remove this sentence.

Author Response

Please see the attachment

Author Response File: Author Response.docx