CNN-BiLSTM: A Novel Deep Learning Model for Near-Real-Time Daily Wildfire Spread Prediction

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Please see the attachment.

Comments for author File: Comments.pdf

Author Response

Dear Reviewer1,

First, we would like to express our gratitude for your invaluable comments on our paper. We would also like to extend our thanks for taking the time and effort to provide such helpful feedback. In the following, we explain how we have revised our manuscript in response to your comments and suggestions. We hope that these revisions improve the quality of our manuscript so that it is deemed acceptable for publication in the GIScience & Remote Sensing journal. Please find our detailed response in the following table:

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

Review of “CNN-BiLSTM A Novel Deep Learning Model for Near Real-Time Daily Wildfire Spread Prediction”

 

The manuscript entitled “CNN-BiLSTM A Novel Deep Learning Model for Near Real-Time Daily Wildfire Spread Prediction” introduce a novel deep learning framework for estimating the fire progression on a daily basis. The prediction of fire progression importance in many aspects such as wildfire management and fire weather forecast. This manuscript provides one path way for the wildfire progression in near real-time. I would recommend a minor revision before publication.

 

NDVI as Fuel Loading Proxy: Utilizing NDVI to approximate fuel loading is an interesting approach, albeit with some reservations. Could you specify which NDVI dataset is employed? Is it based on climatology, or does it utilize near real-time NDVI derived from satellite observations? The latter might be compromised by the smoke from fires obscuring the NDVI signal, affecting the model's accuracy.

 

Wind Speed Estimation: Your method derives wind speed information from two ground stations to estimate wind conditions across the entire fire event domain. While this might be suitable for experimental purposes, its applicability to near-real-time forecasting, especially in areas lacking ground station coverage, is questionable. Have you explored incorporating wind data from forecasting systems like GEOS-FP, which could provide comprehensive and timely wind information across various terrains?

 

Author Response

Dear Reviewer 2,

First, we would like to express our gratitude for your invaluable comments for our paper. We would also like to extend our thanks for taking the time and effort to provide such helpful feedback. In the following, we explain how we have revised our manuscript in response to your comments and suggestions. Moreover, we presented the modifications we made in our paper using the “Track Changes” option in Microsoft Word. We hope that these revisions improve the quality of our manuscript so that it is deemed acceptable for publication in the Remote Sensing journal. Please find our detailed response in the following table:

Author Response File: Author Response.docx