Generating a Long-Term Spatiotemporally Continuous Melt Pond Fraction Dataset for Arctic Sea Ice Using an Artificial Neural Network and a Statistical-Based Temporal Filter
Round 1
Reviewer 1 Report
Dear authors,
you have conducted research at a very high level. High-latitude regions are attracting more and more attention due to their undeniable role in shaping the climate of our planet. The relevance of your research does not raise questions against the backdrop of ongoing climate change.
I would like to wish you success in your further research.
Best wishes
Comments for author File: 
Comments.pdf
Author Response
Responses to Reviewers
We would like to thank the anonymous reviewers for their valuable comments that helped us to revise and improve the presentation and technical context of our paper. In the following sections, we addressed all the issues according to the comments and suggestions of reviewers. The corrections made in this revision were highlighted in blue. All numbered items (pages, equations, figures, and references) are consistent with those in the revised manuscript, excepted if otherwise stated. For convenience, the comments of reviewers are repeated below in italics.
Responses to Reviewer#1
Comments and Suggestions for Authors
Dear authors,
you have conducted research at a very high level. High-latitude regions are attracting more and more attention due to their undeniable role in shaping the climate of our planet. The relevance of your research does not raise questions against the backdrop of ongoing climate change.
I would like to wish you success in your further research.
Author’s reply: We would like to thank you for your positive evaluation of our work. To improve the presentation of this manuscript, we added discussions about the shortcomings, prospects, implications, and potential applications of this study in the revised manuscript. Please refer to the revised manuscript for more details (the corrections and revisions were highlighted in blue). Thank you very much.
Author Response File: Author Response.pdf
Reviewer 2 Report
This is an excellent manuscript. I only have the following small problems that the authors need to improve.
At the end of the introduction, please indicate what is the objective of this manuscript?
L.121-122, The input data of the neural network includes the solar azimuth, solar zenith, apparent azimuth and apparent zenith of the MOD09GA product. The above parameters were not used as input data in previous studies, please explain the necessity of adding these parameters.
L.257, It is recommended to add the seven bands and four azimuth parameters of the MODIS product to the MPF simulation.T o prove the necessity of adding the azimuth parameter.
L.257, It is recommended to join the model jointly established by all the data from May to September, and then evaluate its accuracy.
L.269, Figure 7 The evaluation parameters of model accuracy should use R2, and the standard used in the whole text should be unified.
L.269, The R2 of the temporal filter is 0.43, please explain the reason and ideas for improvement in the discussion.
The Discussion section adds some comparisons between this study and other related studies.
In addition, I think it should also explain the shortcomings and prospects of this manuscript.
Author Response
Responses to Reviewers
We would like to thank the anonymous reviewers for their valuable comments that helped us to revise and improve the presentation and technical context of our paper. In the following sections, we addressed all the issues according to the comments and suggestions of reviewers. The corrections made in this revision were highlighted in blue. All numbered items (pages, equations, figures, and references) are consistent with those in the revised manuscript, excepted if otherwise stated. For convenience, the comments of reviewers are repeated below in italics.
Responses to Reviewer#2
Comments and Suggestions for Authors
This is an excellent manuscript. I only have the following small problems that the authors need to improve.
At the end of the introduction, please indicate what is the objective of this manuscript?
Author’s reply: Thanks for your valuable suggestion. We have reorganized the last paragraph of introduction section, and added descriptions about the motivation and objective of this manuscript. Please see Page 3, Lines 97-113 for more details.
L.121-122, The input data of the neural network includes the solar azimuth, solar zenith, apparent azimuth and apparent zenith of the MOD09GA product. The above parameters were not used as input data in previous studies, please explain the necessity of adding these parameters.
L.257, It is recommended to add the seven bands and four azimuth parameters of the MODIS product to the MPF simulation. To prove the necessity of adding the azimuth parameter.
Author’s reply: Thank you for your suggestions. We have revised the manuscript to explain the necessity of adding these parameters (incident/viewing geometries).
It has been widely reported that the snow, melt ponds, and ocean water are non-Lambertian, and the directional surface reflectance of Arctic sea ice varied significantly with the incident/viewing geometries. To capture the reflectance anisotropic properties, the incident/viewing geometries were used as input variables for establishing the relationship between surface reflectance and shortwave surface albedo of Arctic sea ice using an ensemble back-propagation neural network. In this study, a similar approach was carried out for predicting the MPF of Arctic sea ice, and the incident/viewing geometries were also used as input variables.
Please see Page 5, Lines 188-196 for more details.
L.257, It is recommended to join the model jointly established by all the data from May to September, and then evaluate its accuracy.
Author’s reply: We are grateful for your suggestion. We have added a figure for representing the performance of GA-BPNN model for all the data (from May to September). Please see Page 9, Lines 280-282 for more details.
L.269, Figure 7 The evaluation parameters of model accuracy should use R2, and the standard used in the whole text should be unified.
Author’s reply: Thank you for your suggestion. We have revised the manuscript as your recommendation. Please see Page 9, Line 283 for more details.
L.269, The R2 of the temporal filter is 0.43, please explain the reason and ideas for improvement in the discussion.
Author’s reply: Thanks for your suggestion. We have added a discussion to explain the reason and ideas for improvement in the discussion section.
Although data gaps can be efficiently filled using the statistical-based temporal filter, the gap filling methods with higher accuracy are still required for generating spatiotemporally continuous MPF datasets. In this study, only the relationship between the neighboring days was used in the gap filling of the MPF dataset, and this assumption may be invalid due to snow fall/melt, and sea ice drifts, which would result in a relatively poor gap filling accuracy under certain circumstances. In the future, the temporal correlation, spatial autocorrelation, and the relationship with the passive remote sensing data can be incorporate to generate a seamless MPF dataset. In addition, the deep learning and inpainting methods can also be used for developing new gap filling methods, which have shown promising results in recent studies.
Please see Pages 18-19, Lines 466-475 for more details.
The Discussion section adds some comparisons between this study and other related studies.
In addition, I think it should also explain the shortcomings and prospects of this manuscript.
Author’s reply: Thank you for your suggestion. We have added the findings and implications of the comparisons between this study and other related studies. Please see Page 18, Lines 422-445 for more details. In addition, we have reorganized the discussion section and added more details about the shortcomings and prospects of this manuscript.
The shortcomings of this study and the issues that need to be addressed in the future are listed as follows:
(1) The spatial resolution and validation accuracy of the proposed NENU-MPF dataset need to be improved in the future. Owing to the inconsistency of the footprints of the MPF dataset and the validation data, our results are only roughly consistent with the validation dataset. To address this issue, a labeled dataset for melt ponds in the Arctic derived from high spatial resolution satellite imageries is needed. There is a potential to generate a more accurate MPF dataset using deep learning and a training dataset, and the validation of the MPF datasets would also benefit from a labeled dataset of melt ponds. In addition, developing new validation and uncertainty qualification methods over Arctic sea ice region are also required in future studies.
(2) Although data gaps can be efficiently filled using the statistical-based temporal filter, the gap filling methods with higher accuracy are still required for generating spatiotemporally continuous MPF datasets. In this study, only the relationship between the neighboring days was used in the gap filling of the MPF dataset, and this assumption may be invalid due to snow fall/melt, and sea ice drifts, which would result in a relatively poor gap filling accuracy under certain circumstances. In the future, the temporal correlation, spatial autocorrelation, and the relationship with the passive remote sensing data can be incorporate to generate a seamless MPF dataset. In addition, the deep learning and inpainting methods can also be used for developing new gap filling methods, which have shown promising results in recent studies.
(3) Satellite observations acquired by multiple platforms and sensors can be jointly used to extend the temporal span of the MPF dataset. The method proposed in this study can be adapted to other remote sensing data, e.g., Advanced Very High Resolution Radiometer (AVHRR) and Visible Infrared Imaging Radiometer Suite (VIIRS) data. The combined use of AVHRR, MODIS, and VIIRS data has the potential to extend the temporal span to more than 40 years (from 1981 to the present).
Please see Pages 18-19, Lines 455-481 for more details.
We would like to thank the reviewer for the positive evaluation of our work and for his/her valuable comments that helped us to improve the presentation of our work.
Author Response File: Author Response.pdf
Reviewer 3 Report
By using an artificial neural network and a statistical-based temporal filter, the authors tried to generate a long-term spatiotemporally continuous melt pond fraction dataset for Arctic sea ice.
Interesting and well-writing manuscript.
Short comments
Table 1 should be moved to the next section.
In the discussion section, the authors should develop their findings about the comparison of NENU-MPF dataset.
Of course, the results of this study improve your knowledge of the long-term variations in the Arctic sea ice during recent decades. What about the implications? Please add to the conclusion section the implications of this study.
Author Response
Responses to Reviewers
We would like to thank the anonymous reviewers for their valuable comments that helped us to revise and improve the presentation and technical context of our paper. In the following sections, we addressed all the issues according to the comments and suggestions of reviewers. The corrections made in this revision were highlighted in blue. All numbered items (pages, equations, figures, and references) are consistent with those in the revised manuscript, excepted if otherwise stated. For convenience, the comments of reviewers are repeated below in italics.
Responses to Reviewer#3
Comments and Suggestions for Authors
By using an artificial neural network and a statistical-based temporal filter, the authors tried to generate a long-term spatiotemporally continuous melt pond fraction dataset for Arctic sea ice.
Interesting and well-writing manuscript.
Short comments
Table 1 should be moved to the next section.
Author’s reply: Thanks for your suggestion. We have moved Table 1 to Section 2 (Materials and Methods). Please see Page 3, Lines 133-134 for more details.
In the discussion section, the authors should develop their findings about the comparison of NENU-MPF dataset.
Author’s reply: Thanks for your valuable suggestion. We have reorganized the discussion section, and added our findings about the comparison of NENU-MPF dataset and implications of this study.
The comparison results indicate that the newly generated NENU-MPF dataset has a much longer temporal span (from 2000 to 2020) than that of UB-MPF (from 2002 to 2011) and UH-MPF (from 2000 to 2011) datasets, and is more spatiotemporally continuous compared to the other datasets. Thus, it has potential advantages in term of representing the dynamics and evolution of the MPF of Arctic sea ice more realistically. For example, the long-term trend in MPF of the Arctic sea ice derived from the NENU-MPF dataset can be considered more accurate and reliable, and the NENU-MPF dataset can be used as a reference for calibrating and developing the melt ponds evolution models. In addition, there are also many potential applications for the newly generated MPF dataset of Arctic sea ice, such as improving the treatment of sea ice albedo in the global climate models, qualifying the uncertainty of sea ice con-centration dataset, analyzing the interannual variability in melt processes of Arctic sea ice, and estimating the contribution of melt ponds to the sea ice albedo feedback mechanism.
Although differences in the MPF values of the different datasets were identified, a significant decreasing MPF trend can be found in most of these datasets (Figure 15). The large bias and opposite trend of the UH-MPF dataset are mainly due to the assumption in the spectral unmixing method that the spectral reflectance of the sea ice components is invariant, which resulted in large uncertainties in estimating the MPF when the sea ice surface changed significantly during the melting season. In addition, the data gaps due to cloud obscuration also increased the uncertainty of the MPF trends derived from these datasets. The significant decreasing trend of the MPF in the Arctic can be explained by the decline in the sea ice extent, and the fact that the melted area of first-year ice (where the melt ponds mainly appear) has increased dramatically due to global warming, leading to a decrease in the MPF of the Arctic sea ice in recent decades.
Please see Page 18, Lines 422-445 for more details.
Of course, the results of this study improve your knowledge of the long-term variations in the Arctic sea ice during recent decades. What about the implications? Please add to the conclusion section the implications of this study.
Author’s reply: Thank you for your suggestion. We have added the implications and potential applications of this study in the sections of conclusion and discussion. Please see Page 19, Lines 498-502, and Page 18, Lines 426-434 for more details.
We would like to thank the reviewer for the positive evaluation of our work and for his/her valuable comments that helped us to improve the presentation of our work.
Author Response File: Author Response.pdf
