Enhancing the Performance of Machine Learning and Deep Learning-Based Flood Susceptibility Models by Integrating Grey Wolf Optimizer (GWO) Algorithm

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The article addresses an important and increasingly relevant topic, namely flood risk, using cutting-edge techniques such as machine learning and deep learning.

The article is interesting and well-written. The literature review on the subject is comprehensive, and the objectives are clearly stated. I just have a few comments, particularly regarding the presentation of the "Materials and Methods" section.

- Section 2.3. Flood conditioning factor: The paragraph on Flood conditioning factors, although comprehensive, should provide additional details on where the same input data can be obtained. For instance, in reference to the Landsat-8 images, it would be beneficial to specify the season and exact dates considered. Additionally, an explanation of the rationale behind these specific choices should be included.

-  Please, add additional details also un the 2.4. Multicollinearity assessment.

- The sequence of metrics, techniques, and tests listed in paragraph 2.6 "Validation Techniques" should more appropriately be referred to as "performance measures" of the models rather than validation techniques.

-  The paragraph accompanying section 3.2 "Frequency Ratio (FR) Result" is currently difficult to read and not very informative. Given that Table 2 already provides the numerical results, it is recommended to enhance readability and comprehension by directly highlighting the factors with the highest contribution potential to flooding in the investigated basin. Presenting these key factors more directly and perhaps emphasizing them within the table itself would significantly improve the clarity and effectiveness of the presentation.

- Could additional information be provided about the mentioned "natural breaks classification method" (line 369)?

Overall, while the paper is thorough and well-structured, addressing these points would improve the presentation and clarity, making the study even more robust, accessible, and reproducible by readers.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This study applied two ML (SVR and XGBoost) models and one DL (RNN) model to create the susceptibility map, and later, an optimization algorithm (GWO) was integrated into these three models to enhance the performance. It has been confirmed that all six models effectively predict susceptibility and perform well. The modeling ascertained that the RNN (AUC: 0.956) model shows the highest performance among the three models, underlining that the DL model is better than the ML models. Furthermore, the integration of GWO enhanced the performance of all three models, with the RNN-GWO (AUC: 0.968) model having higher performance. The RNN-GWO-based map depicts 8.05% of the MRB as highly susceptible to flooding. The most influential conditioning factors are SPI, geomorphic units, LULC, stream density, TWI, and soil types. The paper is interesting and well-written but has drawbacks:

1. The authors contribution and scientific novelty of the paper are not clear. the authors use well-known models and techniques. 

2. The authors should present a scheme or workflow of the proposed methodology.

3. The authors didn't present clearly what set of metrics is specific for this task.

 4. Subpart 4.1 should be expanded with statistical analisys which should be performed.

5. The results of the paper are not clearly presented. What did the authors try to say? Did they choose the best method or only evaluate SOTA methods or what?

Comments on the Quality of English Language

The quality of English is acceptable.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf