Citizen Science Tools Reveal Changes in Estuarine Water Quality Following Demolition of Buildings

Round 1

Reviewer 1 Report

Fig. 2 resolution should be improved.

Table 1. caption:  T column negative t-values should be discussed indicating a reversal in the directionality of the effect, with no bearing on the significance of the difference between groups.

Author Response

1. Fig. 2 resolution should be improved.

Thank you, the suggestion is well taken.  The resolution of figure 2 has been improved and the modified figure has replaced the old one in the revised MS.

In addition to this, for a global readership to know the geographical region of study, figure 1 has been modified to show the location map of the study area with maps of India and Kerala with a shape file of Vembanad Lake as inset (updated figure provided in the revised version).

2. T column negative t-values should be discussed indicating a reversal in the directionality of the effect, with no bearing on the significance of the difference between groups.

We agree. The t-statistic is calculated as (M -M2) / SE. The negative t-values obtained in the analysis does indicate a reversal in the directionality of the effect. The values of Secchi depth, pH and DO have increased with time when compared with week 2.  This has led to the negative sign in the t-value.  We agree that it has no influence on the significance. This clarification has been added to the caption to Table 1.

Author Response File: Author Response.docx

Reviewer 2 Report

In order to help create public awareness on water quality and sanitation, this study highlighted the utility of a low cost device – 3D printed, hand-held Mini Secchi disk (3DMSD) – in combination with a mobile phone application ('TurbAqua') for assessing the water quality of a shallow lake region after a demolition incident. Generally, the topic is interesting for the journal's readership. However, in my opinion some issues should be addressed and improved before the manuscript can be published.

Major comments:

1. Please rewrite the Abstract, which should mainly contain your work and findings, but here you generally neglected the obtained results.
2. In the Section of “Introduction”, some sentences need more references to fully express why you want to do this study.
3. Lines 176: Where >>where.
4. Figure 4: should put in one page. Also, a>>A, b>>B.
5. Discussion: maybe you can give more sentences to explain the uncertainty.

Author Response

1. Please rewrite the Abstract, which should mainly contain your work and findings, but here you generally neglected the obtained results.

Response 1:

Thank you for the critical comments and pointing out a major flaw in the paper.  The abstract has been modified as follows:

Turbidity and water colour are two easily-measurable properties used to monitor pollution. Here, we highlight the utility of a low-cost device – 3D printed, hand-held Mini Secchi disk (3DMSD) with Forel-Ule (FU) colour scale sticker on its outer casing – in combination with a mobile phone application ('TurbAqua') that was provided to laymen for assessing the water quality of a shallow lake region after demolition of four high-rise buildings on the shores of the lake. The demolition of the buildings in January 2020 on the banks of a tropical estuary – Vembanad Lake (a Ramsar site) in southern India– for violation of Indian Coastal Regulation Zone norms created public uproar, owing to the consequences of subsequent air and water pollution. Measurements of Secchi depth and water colour using the 3DMSD along with measurements of other important water quality variables such as temperature, salinity, pH and dissolved oxygen (DO) using portable instruments were taken for a duration of five weeks after the demolition to assess the changes in water quality. Paired t-test analyses of variations of water quality variables between the second week of demolition and consecutive weeks up to the fifth week showed that there were significant increases in pH, dissolved oxygen and Secchi depth over time, i.e., the impact of demolition waste on the Vembanad Lake water quality was found to be relatively short-lived, with water clarity, colour and DO returning to levels typical of that period of year within 4-5 weeks. With increasing duration after demolition, there was a general decrease in FU colour index to 17 at most stations, but it did not drop to 15 or below, i.e., towards green or blue colour indicating clearer waters, during the sampling period.  There was no significant change in salinity from second week to fifth week after demolition, suggesting little influence of other factors (e.g. precipitation or changes in tidal currents) on the inferred impact of demolition waste. Comparison with pre-demolition conditions in the previous year (2019) showed that the relative changes in DO, Secchi depth and pH were very high in 2020, clearly depicting the impact of demolition waste on the water quality of the lake. Match-ups of the turbidity of the water column immediately before and after the demolition using Sentinel 2 data were well in agreement with the in-situ data collected.  Our study highlights the power of citizen science tools in monitoring lakes and managing water resources and articulates how these activities provide support to Sustainable Development Goal (SDG) targets on Health (Goal 3), Water quality (Goal 6), and Life under the water (Goal 14).

2. In the Section of “Introduction”, some sentences need more references to fully express why you want to do this study.

Response 2:

More references have been added in the introduction.

Line 75 – Nitin et al., 2014 (reference 2)

Line 87 – Sakshi and Malik, 2018 (reference 5)

3. Lines 176: Where >>where.

Response 3:

Sorry for the careless mistake committed. Modified in the revised MS (line 216 of revised MS).

4. Figure 4: should put in one page. Also, a>>A, b>>B.

Response 4:

All suggested changes have been incorporated in the revised MS.

5. Discussion: maybe you can give more sentences to explain the uncertainty.

Response 5:

As desired by the reviewer, the following additional information regarding uncertainties in the observations have been added towards the end of discussion in the revised MS.

Citizen science activities, necessarily, rely on simple tools to overcome gaps in observations. The utility of such observations will depend ultimately on the quality and reliability of the observations. Uncertainties associated with Secchi Disk measurements and Forel-Ule colour index measurements have been discussed in detail by Brewin et al. [10]. Furthermore, we recognise the importance of providing adequate training to the citizen scientists, to ensure data quality. The ways by which errors can be reduced have been presented in the training manual produced as part of the project work and is an open-access publication [29], which was part of the background material provided to the citizen scientists at the training course provided to them. When satellite observations are included, there are additional uncertainties that arise from mis-matches between satellite and in situ observations. Because of the limited number of match-up observations available from this study,we could not carry out a detailed comparison of satellite and in situ observations. But this remains a goal for the future.

Author Response File: Author Response.docx

Reviewer 3 Report

Menon et al. used the two basic ocean color parameters, Secchi Disk depth (SDD) for water clarity and Furel Ule index (FU) for the color of water, collected by citizen scientists using a portable tool (3D mini Secchi Disk) in a lake near four demolished buildings, to investigate how the building demolition influenced the aquatic system nearby in terms of water quality. This study shows potential in obtaining useful field data for scientific use not only from the academic field work /expeditions but also from a wider community in a more convenient way, e.g., data can be collected by common citizens who do not have to be very experienced in corresponding disciplines. As this manuscript was submitted as a Communication but not a research article, I assume that this is aiming to deliver rapid and important preliminary research findings, and provide an efficient method which can be used for water quality monitoring for inland water systems. The whole manuscript is easy to understand because a straightforward method was used and analysis/verification performed is also simple. Though this manuscript fulfills the requirements of a communication paper, in my point of view, there are still several aspects that need to be addressed with more clarity.

The authors used “estuarine water quality” in the title while in the manuscript authors referred to the Vembanad lake. Without any prior knowledge of the geographical locations of the lake, one may get confused whether the study area is a lake or an estuary or in between.

The authors mentioned a few times very briefly that citizen scientists were involved to measure the SDD and FU index data, however readers might be much more interested in how these citizens were selected, how they were trained, what their general backgrounds are and how many of them participated in this investigation, as this information can explain how the citizen science could be operated and whether there might also be influence on the quality of the data measured.

I understand that as a communication paper it should not be too long, however it seems that some information was not well described or justified in the manuscript, especially in the data and method section. See detailed comments below.

L23 MSD to 3DMSD to be consistent?

L31 full name of SDG

L59 on the banks of water bodies – too general, banks of a lake or river? I also felt “water bodies” were used quite often in the manuscript but might not fit some context.

L75 In the above text the authors mentioned the influence of the building demolition on waters from coast and estuary, why was a lake then finally chosen here? More background information should be helpful. 

L98 “from damage” – what damages for example?

L101-102 “Daily sampling was conducted during the second week, and the sampling interval was increased from daily to once every three days in the subsequent weeks” - not clear. increased from daily to once every 3 days? reads like a decrease of sampling interval.

L109-110  reads a bit repeated with previous sentences in this paragraph

L115 water colour – better replace it with FU index as water colour is not specific enough and could mean more.

L116-118 “A total of 12 samplings in duplicate were done during the sampling period for salinity, DO and pH whereas, 96 observations for Secchi depth and water colour were available for the analysis” -  not clear. A total of 12 sampling of each parameter (salinity, DO and PH) at each station or this means something else? How can one link this 12 samplings to Figure 2? How many SDD and FU measurements were there for each station (96/7)? Though data were uploaded as supplement it is uneasy to understand without further explanations.

L171 “Salinity and temperature parameters were adjusted to the average values for the study” - what average, on which basis? How does this relate to the satellite S2A/S2B data?

L173-L177 The algorithm the authors used here to retrieve SDD from Sentinel 2 data was developed by a study carried out in Portuguese waters which might violate when applying it to Indian waters. How was this algorithm justified to be suitable in this study area?

Figure 2: Fig 2A SDD unit is (m) instead of (m-1); dissolved oxygen has no data for week 5 and also on y-axis label week 3 (2020) is not shown properly.

L208-209: in week 2 already high FU (>18) was already observed, not two weeks after (which means it started from week 3)

L213 why using return to 15 here? before it was 15 or below?

L226 overestimation of SDD - why is that, might be due to violated algorithm applied in this area? any statistics between satellite derived and in situ data?

L234 “Secchi depth at the demolition sites different only by 0.18 meters” – this refers to satellite derived SDD?

L247 Less green algae -  how do you know there were less green algae in the water? No any information or data about phytoplankton types or chlorophyll content were introduced in the study.

Conclusion could be more brief and better organized- the first half paragraph reads a bit tedious.

Author Response

Response to Reviewer 3 comments

1. The authors used “estuarine water quality” in the title while in the manuscript authors referred to the Vembanad lake. Without any prior knowledge of the geographical locations of the lake, one may get confused whether the study area is a lake or an estuary or in between.

Response 1:

Vembanad Lake is the largest lake along the south west coast of India.  Northern arm of the lake opens into the Arabian Sea at two places imparting estuarine features to the lake.  The southern arm receiving bulk of fresh water input from rivers has more features of a lake.  Due to the continuity, both arms are together referred to as the Vembanad Lake (Ramamritham et al. 1986, Unnithan et al., 2001). As the study area falls in the northern arm of the lake with predominant estuarine features, the term ‘estuarine water quality’ was used in the title.

As the reviewer suggested, there are chances of confusing the reader with the terminology.  Hence a brief description about Vembanad Lake along the lines of the paragraph above has been added in the modified MS (Lines 118 - 123 onwards in section 2.1).

2. The authors mentioned a few times very briefly that citizen scientists were involved to measure the SDD and FU index data, however readers might be much more interested in how these citizens were selected, how they were trained, what their general backgrounds are and how many of them participated in this investigation, as this information can explain how the citizen science could be operated and whether there might also be influence on the quality of the data measured.

Response 2:

Citizens participating in the study were local inhabitants who volunteered to undertake the measurements, and were trained by our team. Details regarding the establishment of citizen science network and their training are described in the following paper: Grinson George, Nandini N. Menon, Anas Abdulaziz, Robert J. Brewin, Pranav P, Gopalakrishnan A, Mini K. G, Somy Kuriakose, Shubha Sathyendranath, Trevor Platt (2021).  Citizen scientists contribute to real-time monitoring of lake water quality using 3Dprinted mini Secchi disks – Frontiers in water (accepted). This additional information has been added to the revised MS.

I understand that as a communication paper it should not be too long, however it seems that some information was not well described or justified in the manuscript, especially in the data and method section. See detailed comments below.

3. L23 MSD to 3DMSD to be consistent?

Response 3:

Sorry for the oversight.  Changes have been made to make the word 3DMSD consistent throughout the revised MS.

4. L31 full name of SDG

Response 4:

Expanded form Sustainable Development Goals given in abstract with the acronym in parenthesis.

5. L59 on the banks of water bodies – too general, banks of a lake or river? I also felt “water bodies” were used quite often in the manuscript but might not fit some context.

Response 5:

The term water bodies in line 59 was used in a general context.  In other places, water body refers to Vembanad Lake which is mentioned immediately before using the term water body.

6. L75 In the above text the authors mentioned the influence of the building demolition on waters from coast and estuary, why was a lake then finally chosen here? More background information should be helpful. 

Response 6:

The study area description has been modified to improve clarity for the readers and a synoptic scale image is added for indicating how the lake is connected to the coastal waters and to the tributary where demolition happened. Modifications have been made in the description of study area as follows:

Two apartment complexes, twin towers of “Alfa Serene” and “Holy FaithH₂O” were located on the shores of a narrow tributary opening into Vembanad Lake, the largest lake along the south west coast of India.  The northern arm of the lake opens into the Arabian Sea at two places, imparting estuarine features to the lake.  The southern arm receiving bulk of fresh water input from rivers has more features of a lake.  Both arms are interconnected and are referred to as the Vembanad Lake [11,12].

7. L98 “from damage” – what damages for example?

Response 7:

Physical and structural damages to the nearby buildings and other associated causalities. Text modified accordingly.

8. L101-102 “Daily sampling was conducted during the second week, and the sampling interval was increased from daily to once every three days in the subsequent weeks” - not clear. increased from daily to once every 3 days? reads like a decrease of sampling interval.

Response 8:

Sorry if the statement was misleading.  The sentence has been changed as the sampling frequency was reduced from daily to once every three days in the subsequent weeks (Line 137 of the revised MS).

9. L109-110  reads a bit repeated with previous sentences in this paragraph

Response 9:

Agree with the reviewer.  The line has been deleted in the revised MS.

10. L115 water colour – better replace it with FU index as water colour is not specific enough and could mean more.

Response 10:

Comment well taken and throughout the revised MS, water colour has been replaced with FU index.

11. L116-118 “A total of 12 samplings in duplicate were done during the sampling period for salinity, DO and pH whereas, 96 observations for Secchi depth and water colour were available for the analysis” -  not clear. A total of 12 sampling of each parameter (salinity, DO and PH) at each station or this means something else? How can one link this 12 samplings to Figure 2? How many SDD and FU measurements were there for each station (96/7)? Though data were uploaded as supplement it is uneasy to understand without further explanations.

Response 11:

12 samplings indicate the number of visits made the scientific team in the four weeks for taking measurements.  3DMSD measurements were more in number as there were additional measurements by citizens and multiple readings were available from each station, thus leading to 96 weekly-averaged observations of Secchi depth and water colour.

12. L171 “Salinity and temperature parameters were adjusted to the average values for the study” - what average, on which basis? How does this relate to the satellite S2A/S2B data?

Response 12:

Apologies for confusion caused to the reader.  The sentence in the methodology refers to the use of salinity and temperature as ancillary parameters for the C2RCC processor.  To derive Secchi depth from S2A/S2B, L1C subsetted products were computed using C2RCC processor.  The weekly averaged salinity and temperature values from our in situ observations were used as input into the C2RCC model (Soomets et al., 2020).

To avoid confusion, the relevant part in methodology has been modified with the sentence above (Line 209-210 of section 2.2 of revised MS).

13. L173-L177 The algorithm the authors used here to retrieve SDD from Sentinel 2 data was developed by a study carried out in Portuguese waters which might violate when applying it to Indian waters. How was this algorithm justified to be suitable in this study area?

Response 13:

We used the results from Rodrigues and colleagues (2020) because of the quality of their results. However, we acknowledge that it is important to note that the results are from a different locality. This has been added in the revised paper.

14. Figure 2: Fig 2A SDD unit is (m) instead of (m-1); dissolved oxygen has no data for week 5 and also on y-axis label week 3 (2020) is not shown properly.

Response 14:

Sorry for the mistake.  Fig. 2A y-axis has been modified as per suggestion of the reviewer.Changes have been made in the figure in the revised MS. 

DO measurement was not taken during week 5 due to technical reasons and it is already mentioned in the MS.

15. L208-209: in week 2 already high FU (>18) was already observed, not two weeks after (which means it started from week 3)

Response 15:

Mention in the text was to two weeks after demolition, the first set of observations after the demolition, when FU colour index was high. Over time, the index decreased, depicting movement to clearer waters.

16. L213 why using return to 15 here? before it was 15 or below?

Response 16:

Below 15, i.e., from 14-10 FU index is indicative of green colour, changing to bluish green (9-5) and then blue (4-1).  From FU index 15 or above, the colour is a mix of green and brown.  Brown is indicative of murky waters or abundance of suspended material in the water.  In our earlier studies (unpublished data), the lowest FU index noted in Vembanad Lake was11. But we agree that the phrase “return to 15” is misleading, and the text has been modified accordingly.

17. L226 overestimation of SDD - why is that, might be due to violated algorithm applied in this area? any statistics between satellite derived and in situ data?

Response 17:

We chose the Rodrigues et al (2020) algorithm because it gave Secchi depth values that were in the right range, when compared with in situ observations. The match-up observations were too limited (around 3) for rigorous match-up analysis. The overestimation reported is when weekly averages were compared, and this could be from a number of reasons, including the mis-matches between satellite and in situ observations. This has been clarified in the revised text (lines: … and lines …)

18. L234 “Secchi depth at the demolition sites different only by 0.18 meters” – this refers to satellite derived SDD?

Response 18:

Yes, and represents the maximum difference observed in the satellite time series. The sentence in the revised MS has been modified to bring in clarity regarding this (Line 275-276 of revised MS).

19. L247 Less green algae -  how do you know there were less green algae in the water? No any information or data about phytoplankton types or chlorophyll content were introduced in the study.

Response 19:

Reference to green algae is based on the FU colour. Green waters have FU index less than 15. The text has been modified for clarity.

20. Conclusion could be more brief and better organized- the first half paragraph reads a bit tedious.

Response 20:

Conclusion has been redrafted as follows in the revised MS (Lines 394-412):

Anthropogenic disturbances such as civil structure demolitions near a lake can deteriorate the local water quality due to debris falling into the water and the leachates from the shattered civil structures. A citizen science network and 3DMSD linked to a mobile application were used for assessing the water-quality in the water near the demolition sites for a period of 5 weeks after demolition. Being the first case of a court-directed demolition to protect the coastal regulation zone in India, the public were aware of the importance of environmental protection. However, direct observations on the nature of the potential environmental impacts were lacking hitherto. The short-term and long-term impact of such demolitions on the nearby aquatic systems could be scientifically assessed in this study. Citizen science-based data on FU colour scale and Secchi depth, along with in-situ water quality variables such as salinity, temperature, pH, and dissolved oxygen helped to achieve an almost real-time assessment of the water quality pre and post demolition in the Lake area.

Availability of satellite remote sensing data for the region and the scientific analysis of the data helped us to evaluate the weekly progression in water-quality. The water quality during the second week of demolition and consecutive weeks up to the fifth week showed significant increase in pH, dissolved oxygen and Secchi depth over time, i.e., the impact of demolition waste on the Vembanad Lake water quality was found to be relatively short-lived, with water clarity, colour and DO returning to levels typical of that period of year within 4-5 weeks. A key question is whether the time scale of recovery of the water quality is sufficiently rapid to avoid lasting damage to the local ecosystem.

We have demonstrated here the application of simple, hand-held tools for evaluating the time-dependent impact on water quality resulting from the demolition of high-rise buildings on the banks of a lake, illustrated the advantages of engaging local communities in such evaluations through citizen science, and demonstrated how freely-available satellite data can complement in situ observations. The approach is readily adapted to other contexts and localities, and could easily become an essential ingredient of maintaining water quality for sustainable regional development.

Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

Thanks for addressing my comments. Though I still feel that the SDD algorithm developed for other water types is not the best to estimate SDD in this lake, I do understand that during this short period there were not enough in situ data available. However, I wish that the authors perform such studies in the future based on adequate data to generate more convincing results. 

I see some of the replies to the comments are not always matching with the input of the manuscript, e.g. the reasons why SDD was overestimated are listed differently in the replies and the MS. And those reasons sound not so persuasive to me. I also see incomplete replies, e.g., lines... and lines...Just a gentle reminder that please be more careful and responsible in a response to reviewer.

I do acknowledge the idea of citizen science and the preliminary application carried out from it. It is a quite nice start after all. Though holding some different opinions I still think it is worth to be published as a communication paper.

Author Response

Comment 1:

Thanks for addressing my comments. Though I still feel that the SDD algorithm developed for other water types is not the best to estimate SDD in this lake, I do understand that during this short period there were not enough in situ data available. However, I wish that the authors perform such studies in the future based on adequate data to generate more convincing results.

Response:

We gratefully acknowledge the support received from the reviewer in improving the quality of the manuscript. Further, we do agree with the reviewer on the need for developing the SDD algorithm. For developing an algorithm as indicated by the reviewer, we require a better dataset. Definitely, as part of the future studies we will perform this task.  

Comment 2:

I see some of the replies to the comments are not always matching with the input of the manuscript, e.g. the reasons why SDD was overestimated are listed differently in the replies and the MS. And those reasons sound not so persuasive to me.

Response:

Sorry for the oversight.  All the reasons cited in the MS for the over-estimation of SDD by satellites were not given in the reply to reviewers. In the absence of statistical analyses between in situ and satellite data, what we could do was to list out the possibilities for the over-estimation.  This is the main reason why we have resorted to estimation of relative changes in satellite data between weeks.  In the paper, we are only claiming that corroborating the in-situ observation, an improvement in water quality was evident in the satellite derived data also by the 5^th week after demolition. Our effort indicated a qualitative output from these estimations.

As noted in the revised MS, "there was a systematic over-estimation of Secchi depth compared with the in-situ observations when weekly averages were compared. This difference could have been from a number of reasons, including the use of weekly averages for the comparison, and the differences in the scales of observation, apart from the use of an algorithm that had been established for Portuguese waters" (Lines 254-260 of revised MS).

Comment 3:

I also see incomplete replies, e.g., lines... and lines...Just a gentle reminder that please be more careful and responsible in a response to reviewer.

Response:

We apologize for the apparent carelessness in noting line numbers. We meant no disrespect, but our version of the responses shows that all line numbers are duly noted. We suspect that the error occurred when uploading the responses and that there might have been an incompatibility between the ms word s/w that we used, and the libreoffice s/w that the journal uses. We will make double sure that such problems do not arise when we upload our responses this time, by uploading them in pdf version.

Comment 4:

I do acknowledge the idea of citizen science and the preliminary application carried out from it. It is a quite nice start after all. Though holding some different opinions, I still think it is worth to be published as a communication paper.

Response:

We thank the reviewer for supporting the publication of this manuscript as a communication paper.