Review Reports

Round 1

Reviewer 1 Report

This paper is well written. I have no major criticisms. Only very minor ones. dealing with a few language issues - minor text editing (see below). This is an interesting approach to show ways to use Sentinel 2 data, along with ground truth, to estimate chlorophyll. It shows utility, and also cautions other researchers about potential limitations. 

1907484

Chlorophyll not capitalized unless at start of sentence. Please check entire paper (see line 57; it is capitalized there, and should not be). 

 

Units, such as nm, should have a space between units and value; line 63, should have “438 nm” and “676 nm”. Check entire paper. 

 

Line 97, superscript the 2 for km².

 

Line 107, replace the “N – “ with “N, "

 

Line 144: "First" instead of "Firstly"

Line 152: there's an accent or quote after "et al."; remove the accent or quote

Line 166: should be "to be retrieved with" not "to retrieved"

Line 185: not "These tunes were " but instead "This tuning procedure was "

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

I have reviewed the manuscript "Assessment of human-induce effects on sea/brackish water chlorophyll-a concentration in Ha Long Bay of Vietnam processing in Google Earth Engine". I added the following comments and suggestions:

Line 57: it is unnecessary to use capital letter for chlorophyll.

Line 97: please use a superscript for km²

Figure 1: I understand that Vietnam and Viet Nam are both correct. However, I would suggest using just one form.

 Line 256: please add further references.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

 

In this manuscript, Quang et al studied the chl-a concentration around the Ha Long bay and Cua Luc bay in Vietnam, a region with rapid urban population growth, active tourist activities, and coal mining. The imagery data from Sentinel-2/MSI data from 2019-2021 is used to derive the chl-a concentration based on a conventional OC-2 algorithm and analyzed using Google Earth Engine. The results are calibrated through a NOAA chlorophyll-a dataset with improved performance. The authors then explored the seasonal changes of the chl-a against various zones involving coal mining,  tourism, Aquaculture, lake and urban activies, and found high correlations except with coal mining.

 

This work is very interesting for the study and monitor water quality and of value to provide quantitative inputs for environmental protection and policy making. The data and methodology are generally solid. However, I feel a few technical details are not clear. A few suggestive comments are provided below.

 

1.     Line 99-101, the authors mentioned that wastewater and solid waste from nearly regions are coming to the bay. Meanwhile the bay is also a popular tourist site. Could you comment on whether there are regulation or facilities to control the wastewater drainage to the bay? I am also interested how this study can help to form or refine policy to protect the bay.

2.     Line 126, the Sentinel-2 L2A data from GEE data are used. Could you add more information on who uploaded and maintained the data?

3.     Table 1, comparing the bandwidth with the central wavelength, the bandwidth is really very large which will make many bands overlaps. I am not very familiar with the MSI data, but is this the reason that the MSI data is not optimal for the ocean color studies which usually require high spectral resolution and SNR?

4.     Line 195, why R^2 is selected for this study as the cost function to minimize? Have you considered other functions, such as RMSE? Another issue is that the chl-a values various significantly. Would it be useful to consider it in log scale?

5.     Line 204, for the model calibration, different results using calibrated and uncalibrated model are compared. Could you explain more how the OC-2 model is calibrated?

6.     Line 216, just to confirm whether the R^2 values of 0.85 and 0.81 refer to the results in Fig 3B and 3C? It seems the R^2 values don’t change much. My worry is that the smaller value of chl-a may be not represented well by the cost function which results in the poor prediction skill for small chl-a values as shown in both Fig 3B and 3C.

7.     Fig 4, would it be useful to plot chl-a in log scale here too? Many of the smaller chl-a values are overlapped?

8.     Line 249, the chla values varies between 0 and 100mg/m^3 (also in Fig 5), would the calibration method still valid? Since the maximum chl-a value in Fig 3 seems less than 20mg/m^3.

9.     Line 288, since GEE is mentioned in the title, it would be helpful to provide more information about GEE. Could you comment on the CPU resources used for this study?

10.  Fig 5 and 6, are these plots made by GEE?

 

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx