Examining Thresholding and Factors Impacting Snow Cover Detection Using Nighttime Images
Round 1
Reviewer 1 Report
Please thoroughly review the manuscript for its grammar.
Introduction: The author should focus on adding reviews and descriptions of related works focusing on snow cover detection using nighttime images. What is state-of-the-art? Their findings and challenges, etc.
When citing two or more researchers, put them in chronological order…. see page 2 line 53 & 54 (Foster, 1983; Wiesnet et al., 1981) edit as (Wiesnet et al., 1981; Foster,1983). Check the entire manuscript for the same.
Methodology: Make the section more concise. The flow chart must be placed at the beginning of the methodology section. Describe this in the manuscript accordingly.
Discussion: add some last references in this section.
Conclusion: The section needs to be reconstructed. Add some key recommendations for decisionmakers
Author Response
Reviewer 1:
“Please thoroughly review the manuscript for its grammar.”
The MDPI English service has review the manuscript grammar.
“Introduction: The author should focus on adding reviews and descriptions of related works focusing on snow cover detection using nighttime images. What is state-of-the-art? Their findings and challenges, etc.”
We agree and we have added reviews and descriptions of three other works in the Introduction (Huang et al., 2018; Liu et al., 2021; Miller et al., 2013) (Lines 60-79)
“When citing two or more researchers, put them in chronological order…. see page 2 line 53 & 54 (Foster, 1983; Wiesnet et al., 1981) edit as (Wiesnet et al., 1981; Foster,1983). Check the entire manuscript for the same.”
We have changed the citations to numerical order in accordance to the journal guidelines.
“Methodology: Make the section more concise. The flow chart must be placed at the beginning of the methodology section. Describe this in the manuscript accordingly.”
Agree, we have substantially reduced the material and methods section, especially the data preparation has been made more succinct (also in response to another reviewer). The flowchart is included as the Graphical Abstract. According to the journal guidelines the image used in the Graphical Abstract cannot be used again in the text.
“Discussion: add some last references in this section.”
We agree and reference now how our results compare to previous results in this section. We added two additional references in Lines 390-395
“Conclusion: The section needs to be reconstructed. Add some key recommendations for decisionmakers”
We have added additional recommendations to help in the decision on hoe use the results in the conclusion (Lines 512-516)/
Reviewer 2 Report
Please see the attached file.
Comments for author File: 
Comments.pdf
Author Response
Reviewer 2:
“’(a land cover product… NDVI values)’. What is the purpose of this sentence? It is recommended to be removed. (Lines 143-145)”
Agree. We have followed the recommendations and removed the line.
“Since the preprocessing method is simple, its description should be condensed and the figures can be removed. (2.2 Data Preparation, or 2.3. Data Preparation)”
We have removed the figures (Figure 2 and 3 in the first version of the manuscript), and the text has been also condensed accordingly.
“’The thresholding algorithms were implemented using the Python Programming language and the scikit-image processing library (Walt et al., 2014)’. Why this information is important? It is recommended to be removed (Lines 199-200)”
We agree with the recommendation and have removed the lines.
“The basic principles and procedures of the seven thresholding algorithms should be briefly described. (2.4. Methodology)”
We added the basic principles of the algorithms in Lines 235-240. In addition, we guide the readers to a paper by (Sekertekin, 2021) where the algorithms are explained in more detail in the context of remote sensing.
“Overall error is not used in the manuscript so the sentence is suggested to be deleted (Line 247). More evaluation metrics should be added. For more usage of the confusion matrix, it is recommended to refer to the relevant papers:
Ground-based evaluation of MODIS snow cover product V6 across China: Implications for the selection of NDSI threshold. Sci Total Environ, 2019,651:2712-2726
STAR NDSI collection: a cloud-free MODIS NDSI dataset (2001-2020) for China. Earth Syst Sci Data, 2022, 14:3137-3156”
We have deleted the suggested lines.
Furthermore, we have added Cohen’s Kappa to all confusion matrices in the manuscript as an additional evaluation metric. We have also cited the relevant papers in our manuscript.
“Section 3 and Section 4 should be integrated. The analysis should follow the data, otherwise it is difficult to understand them.”
We were uncertain about the expectations of this comment. We asked our editor contact for explanation on this topic. Still waiting for a response, but for the time being, we assume Section 3 is Result, and Section 4 is Discussion. The journal layout requirements requires us to have separate Result and Discussion sections.
“The evaluation based on MODIS NDSI dataset is not very reliable, especially in forested areas. It is necessary to evaluate snow cover detection by using in-situ snow depth data.”
We agree that in-situ measurements are more reliable. Because our case study areas covered multiple countries (therefore measurements accross countries may not be consistent), and there were limited number of in-situ measurements for our areas (we had access to measurements from the US, but small numbers in our areas), there would be limited statistical validity. for these reasons, we chose MODIS NDSI, which was also used previously as a reference dataset in Huang et al., (2022) .
“With only one collection date selected for each case study area, the total sample is inadequate. It is recommended that all eligible images be used for statistical analysis throughout the study period”
We definitely agree, that more images would lead to more statistically significant results. At the moment searching for eligible images (cloud free, full moon, etc.) is a manual process and it showed limited potential candidates according to our eligibility criteria. Most studies in the field ((Liu et al., 2021), (Miller et al., 2013)) use only one or a small number of images in their research. (Huang et al., 2022) did use multiple images for one case study area, however in that study they were specifically looking at the impact of different lunar phases (so also one image per lunar phase). Since our approach focuses on comparing the relative performance of different algorithms we trust it already provides valuable insights for selecting an algorithm. A study on temporal effects on the performance is an additional idea worth exploring in a subsequent paper. And we would like to address that in the future.
References:
Liu, D.; Zhang, Q.; Wang, J.; Wang, Y.; Shen, Y.; Shuai, Y. The Potential of Moonlight Remote Sensing: A Systematic Assessment with Multi-Source Nightlight Remote Sensing Data. Remote Sens. 2021, 13, 4639, doi:10.3390/rs13224639.
Huang, Y.; Song, Z.; Yang, H.; Yu, B.; Liu, H.; Che, T.; Chen, J.; Wu, J.; Shu, S.; Peng, X.; et al. Snow Cover Detection in Mid-Latitude Mountainous and Polar Regions Using Nighttime Light Data. Remote Sens. Environ. 2022, 268, 112766, doi:10.1016/j.rse.2021.112766.
Miller, S.D.; Straka, W.; Mills, S.P.; Elvidge, C.D.; Lee, T.F.; Solbrig, J.; Walther, A.; Heidinger, A.K.; Weiss, S.C. Illuminating the Capabilities of the Suomi National Polar-Orbiting Partnership (NPP) Visible Infrared Imaging Radiometer Suite (VIIRS) Day/Night Band. Remote Sens. 2013, 5, 6717–6766, doi:10.3390/rs5126717.
Sekertekin, A. A Survey on Global Thresholding Methods for Mapping Open Water Body Using Sentinel-2 Satellite Imagery and Normalized Difference Water Index. Arch. Comput. Methods Eng. 2021, 28, 1335–1347, doi:10.1007/s11831-020-09416-2.
Reviewer 3 Report
The manuscript entitled "Examining Thresholding and Factors Impacting Snow Cover Detection Using Nighttime Images” aimed to map snow cover using nighttime satellite data in different regions and then compare the results with snow cover that derived from MODIS".
The manuscript is interesting and it is appropriate to be published in Remote Sensing, after addressing the following points.
1) The manuscript needs revision for language and grammar.
2) The introduction must be improved. For example, (i) select three or four papers that match your research's aim and then criticise (if possible) and summarize each one in three or two lines and (ii) try to follow the introduction structure done by Huang et al., (2022). You are nearly there.
3) Page 4, Table 1. Third column, MODIS not Modis. The time of VIIRS is 01:30 (fourth column).
4) Page 4, Data Collection. Provide a table and put (i) the website link, (ii) dates and (iii) spatial resolution for each dataset.
5) Page 4, Data Preparation. How did you deal with the different spatial resolution between land cover data (10 m), VIIRS DNB (500 m), NDVI (500m or 1 km).
6) Figures 2 and 3. I suggest adding a true or false colour satellite image that manifests the snow cover.
7) Page 17, line 435. The classification accuracy of the land cover is also a source of uncertainty that may affect the results.
8) I do like your work. However, all figures are unprofessionally produced, find someone who specializes in cartography to help you, OR just look to Huang's paper.
Author Response
Reviewer 3:
“The manuscript needs revision for language and grammar.”
We have requested the MDPI English service to review the grammar.
“The introduction must be improved. For example, (i) select three or four papers that match your research's aim and then criticise (if possible) and summarize each one in three or two lines and (ii) try to follow the introduction structure done by Huang et al., (2022). You are nearly there.”
We have added the summaries of three other papers in the introduction. (Lines 60-79). Thank you.
“Page 4, Table 1. Third column, MODIS not Modis. The time of VIIRS is 01:30 (fourth column).”
We have changed the column names accordingly, and have changed the time of capture from UTC to the local time.
“Page 4, Data Collection. Provide a table and put (i) the website link, (ii) dates and (iii) spatial resolution for each dataset.”
We have added Table 2 to the manuscript which includes, the name of the dataset, the dates, and the spatial resolution.
“Page 4, Data Preparation. How did you deal with the different spatial resolution between land cover data (10 m), VIIRS DNB (500 m), NDVI (500m or 1 km).”
We agree that it was unclear in the previous version. We added that information to section 2.3 in the lines 223-224
“Figures 2 and 3. I suggest adding a true or false colour satellite image that manifests the snow cover.”
We agree that adding color composites would have made the figures more informative, nevertheless, due to a requirement from another reviewer, these images were dropped from the manuscript as that section was condensed.
“Page 17, line 435. The classification accuracy of the land cover is also a source of uncertainty that may affect the results.”
We agree the accuracy of the land cover data could impact the results. Therefore, we have added a reflection on that in lines 463-465.
“I do like your work. However, all figures are unprofessionally produced, find someone who specializes in cartography to help you, OR just look to Huang's paper.”
We acknowledge that the figures did require more work. We have updated all the remaining figures in the manuscript. We took inspiration from Huang’s paper and roughly based the design of our new figures on their figures. Thank you for this suggestion.
We clearly see our manuscript improved following this reviews and suggestions, we look forward to your additional feedback.
Kind regards,
The Authors.
References:
Huang, Y.; Song, Z.; Yang, H.; Yu, B.; Liu, H.; Che, T.; Chen, J.; Wu, J.; Shu, S.; Peng, X.; et al. Snow Cover Detection in Mid-Latitude Mountainous and Polar Regions Using Nighttime Light Data. Remote Sens. Environ. 2022, 268, 112766, doi:10.1016/j.rse.2021.112766.
Round 2
Reviewer 2 Report
The authors have answered all of my concerns.
