All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
The last reviewer seems satisfied with the recent changes made by the authors, and therefore I can recommend this article for acceptance.
[# PeerJ Staff Note - this decision was reviewed and approved by Mehmet Cunkas, a PeerJ Section Editor covering this Section #]
The author has addressed my concerns, and I recommend publishing this version.
-
-
**PeerJ Staff Note:** Please ensure that all review, editorial, and staff comments are addressed in a response letter and that any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate.
**PeerJ Staff Note:** It is PeerJ policy that additional references suggested during the peer-review process should only be included if the authors agree that they are relevant and useful.
**Language Note:** When you prepare your next revision, please either (i) have a colleague who is proficient in English and familiar with the subject matter review your manuscript, or (ii) contact a professional editing service to review your manuscript. PeerJ can provide language editing services - you can contact us at [email protected] for pricing (be sure to provide your manuscript number and title). – PeerJ Staff
The article focuses on the downscaling of land surface temperature (LST), which is an important research direction in current remote sensing science. Land surface temperature (LST) is a key parameter in the study of climate, hydrology, ecology, and urban heat island effect, but existing satellite sensors (e.g., MODIS) provide LST data with coarse resolution, which makes it difficult to meet the needs of fine-scale studies. Therefore, downscaling LST data from coarser resolution (e.g., 1 km) to higher resolution (e.g., 90 m) is of great scientific significance and application value. The article realizes this goal through the random forest regression model and verifies the transferability of the model, which provides a new technical path for research in related fields.
I think this research has some innovative and realistic values. But it still needs modification and improvement before publication.
The technical route adopted in the article is clear and reasonable. The process of constructing the random forest regression model is also fully described, including steps such as parameter optimization and cross-validation, which provide the scientific validity and reliability of the study. But can we consider combining more environmental parameters (e.g., soil moisture, atmospheric parameters, etc.) for model training in the future to further improve the prediction accuracy and generalization ability of the model?
The results of the study show that the random forest regression model performs well on the test data in the main regions, with low root mean square error (RMSE) and high coefficient of determination (R²), which suggests that the model is capable of accurately predicting fine-resolution LST data.
However, the following topics need to be addressed and clarified:
Although the model shows some adaptability in different seasons and regions, the prediction accuracy decreases in some months (e.g., May and June), which suggests that the model may need to be further optimized when dealing with complex climatic and terrain conditions.
1. Revise the descriptions of the topography, climate, and land use types of the two study areas. The current descriptions are overly brief and lack sufficient contextual detail. Include information on major vegetation types, soil classifications, and other relevant environmental characteristics to provide a clearer understanding of the study setting.
2. Update the introduction to reflect recent advancements in the field. While the importance of urban land surface temperature (LST) and the thermal environment is appropriately introduced, several cited references are outdated. Replace or supplement these with literature from the past five years, particularly studies that incorporate recent applications of machine learning in urban thermal analysis, to enhance the relevance and currency of the research.
For example:
- Exploring the cooling intensity of green cover on urban heat island: A case study of nine main urban districts in Chongqing.
- Machine Learning for Urban Heat Island (UHI) Analysis: Predicting Land Surface Temperature (LST) in Urban Environments
**PeerJ Staff Note:** It is PeerJ policy that additional references suggested during the peer-review process should only be included if the authors are in agreement that they are relevant and useful.
3. Analyze the observed differences in model performance across different months in greater depth. Specifically, explain why prediction accuracy is lower in May and June. Investigate whether these variations are influenced by seasonal climatic changes, land use dynamics, or other environmental factors, and provide supporting evidence or hypotheses.
4. Further examine the model’s reduced performance in secondary regions compared to primary regions. Explore potential contributing factors such as topographic complexity, land use heterogeneity, or climatic variability. A more detailed analysis will help clarify the limitations and applicability of the model across different spatial contexts.
5. Improve the quality of all visual materials. Many of the figures currently lack sufficient resolution, and several legends are unclear or difficult to interpret. Enhance image clarity and ensure that all graphical elements are legible and effectively support the text.
-
-
-
What is the methodological contribution of the study? The study only uses the widely used random forest model for the downscaling. Random forest, while effective, is a well-established algorithm in geospatial modeling and remote sensing, and its use alone does not reflect spatial information and characteristics. The lack of methodological contributions, especially for using spatial characteristics, has critical impacts on the accuracy of the study for the downscaling of MODIS LST to 90 meters. There is no clear justification for the choice of random forest among so many machine learning approaches, and no comparison with alternative downscaling techniques, including both spatial and learning approaches.
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.