Ground-Based Hyperspectral Retrieval of Soil Arsenic Concentration in Pingtan Island, China
, Guangsheng Liu 1, Jinming Sha 4, Zheng Duan 3 and Lanhui Wang 3
Round 1
Reviewer 1 Report (Previous Reviewer 2)
For the hyperspectral retrieval of soil heavy metals concentration, the optimal characteristic bands’ selection, as well as the optimal retrieval models’ selection, remains a challenge. Additionally, satellite-based hyperspectral retrieval still encounters several issues, including atmospheric effects, limitations in temporal and radiometric resolution, and data acquisition, etc. Given this, taking the Pingtan Island, the largest island in Fujian Province and the fifth largest in China, as study area, the optimal characteristic bands were selected from the ground-based full spectrum of soil samples, considering both statistical and physical perspectives. And three linear models (Multiple Linear Regression Model (MLR), Partial least Squares Regression Model (PLSR), and Geographically Weighted Regression Model (GWR)) and three nonlinear machine-learning models (Back Propagation Neural Network Model (BP), Support Vector Machine Regression Model (SVR), and Random Forest Regression Model (RFR)) were tested. Subsequently, the hyperspectral retrieval of soil arsenic content was carried out in Pingtan Island.
In summary, this study is systematic, rigorous, and innovative. In particular, the proposed optimal selection of characteristic bands based on the two criteria of high Pearson’s correlation coefficient and high sensitivity to soil active materials, successfully overcomes the issues of uncertainty and low quality in characteristic bands selection based on Pearson's correlation coefficients. The study provides theoretical and technical support for the monitoring and contamination evaluation of soil arsenic concentration using satellite-based spectroscopy in relatively independent island towns with dense populations, both in China and worldwide. Therefore, the study is of great significance.
In addition, the language throughout the entire paper is generally clear and precise. However, further embellishment is still recommended to improve the quality of a few expressions.
Figure 2: The red box is out of the plot.
The English expression need to be checked for the full manuscript. In addition, some references are not shown correctly on Page 6. It is highly recommended that the authors check through the work.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report (New Reviewer)
Few comments are given in the paper. My main doubt concerns the description of the spectral bands used to estimate the arsenic content. Which and how many bands did you use in the original characteristic bands or the optimal characteristic bands and so on. This should be clarified to make the paper useful to other researchers.
Comments for author File: Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report (New Reviewer)
See attached file
Comments for author File: Comments.pdf
See attached file
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report (New Reviewer)
No more comments
This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.
Round 1
Reviewer 1 Report
Manuscript, entitled (Ground-Based Hyperspectral Retrieval of Soil As Concentra
tion in Pingtan Island, China). The authors reported Overall, the research in Pingtan Island provides theoretical and technical support for the monitoring and contamination evaluation of soil As concentration using satellite-based spectroscopy in relatively independent island towns with dense populations, both in China and worldwide.
Abstract
· Please do not use the pronouns in scientific writing? Such as we tested in line 18. · The full name of the abbreviation such as MLR, PLSR, GWR and ….. should be written in the first time. · Please take care during the writing the most of verbs must be in the past. For example in line 21, the experimental results showed that. · The abstract should be shorted and remove the unnecessary sentencesKeywords: should be arranged alphabetic.
Introduction
· Please highlight in introduction, what is the novelty (originality) of the work? And what is new in your work that makes a difference in the body of knowledge?
· Please add the hint about the basic of Hyperspectra for estimating soil arsenic content?
Materials and methods· In line 183. Please add the area of view for the optic.
Results · Table 1 should be added in M&M. Conclusions · Please write about the limitations of this work in details in conclusion section.Moderate editing of English language required
Reviewer 2 Report
In this paper, the authors are trying to analyze the spatial distribution of arsenic concentration in Pingtan Island using geochemical investigation data and ground-based soil spectra retrieved through hyperspectral analysis. Six machine-learning models, including both linear (MLR, PLSR, and GWR) and nonlinear (BP, SVR, and RFR) models, were evaluated to identify the optimal models. The results indicate that some areas of Pingtan Island are at risk of arsenic pollution, and the RFR and GWR models demonstrate good retrieval performance due to their specific characteristics.
In summary, this study is comprehensive, rigorous, and innovative, providing valuable support for monitoring and evaluating soil arsenic concentration using satellite-based spectroscopy in densely populated towns. But I think some issues remain unresolved, and recommend the work to be revised before acceptance. The comments are as below:
1. In Section 3.2, the interpolation result was masked by the land area extracted from the Landsat 8 OLI image of Pingtan Island in Aug. 2013. Therefore, it is necessary to include relevant information about the OLI image such as its introduction and preprocessing in Section 2.
2. In Section 3.2, to enhance the practical significance of this study, it is recommended to conduct a more in-depth analysis of the impact of soil arsenic pollution on human life, including through food and water sources.
3. In Section 4.1, to enhance our understanding of the sources and export of soil arsenic in Pingtan Island, a relatively independent island town with a dense population, it is recommended to conduct a more thorough analysis of the natural and human factors involved in these processes to provide valuable insights for other island towns.
4. In the discussion of Section 4.2, it is recommended to conduct a further analysis of the accuracy differences between the different modeling results in Section 3.3. This can be done from the perspectives of the model's own characteristics, the characteristics of the source and sink of soil arsenic in Pingtan Island, and its spatial distribution characteristics.
5. Figure 1: The full map of China is not necessary; the map of Fujian province might be more helpful. In addition, the coordinate is too small to read for Pingtan Island.
6. Figure 2: The legend of (a) and (b) are hard to read; The Google Earth image (c) is not a vertical view, but tilted from the south. It should be revised to keep consistent as the above two images.
7. Figure 4: The image seems not clear enough (< 300 dpi), please make it clear.
Additionally, it is recommended to review and refine the language. For example, I prefer arsenic rather than As, the abbreviation. The English can be polished for a more clear and accurate expression.
