The paper presents a novel method to identify the location and development stages of offshore wind turbines, at a global scale, for the period 2016-2021, using radar imagery of the Santinel-1mission. The methods proposed employs deep learning based object detection, linking the SyntEO approach for automatic synthesis of large annotated training data and a cascade of two convolutional neural network models. This allowed the identification of, first, the offshore wind farm areas, and second, the offshore wind farm turbines. The detection of OWT development stages was realized by applying two consecutive peak finder algorithms to the swap profiles of the time series imagery.

 The data generated and method used are new, the workflow is thoroughly documented and described with sufficient detail. The resulted DeepOWT data set is openly available in a readily usable format compatible with geo-spatial software like QGIS. One strong aspect is the comparison of the resulted data set with the OSM database and the GOWT v1.3 data set, which validates and emphasizes the advantages of the resulted Deep OWT data set. The identified locations of OWTs are also consistent with the data provided by other open-source data repositories such as EmodNET.

The research is highly relevant in the context of the accelerated energy transition, leading to the high scale deployment of offshore wind farms. The results can represent valuable inputs in large-scale environmental impact assessments of renewable infrastructure, at a global scale. This can contribute to a more coherent, strategic and knowledge-based decision-making process related to future deployments of OWFs, when considering potential impacts on the marine environment. Furthermore, an updated data-base of OWFs and their development status can refine the outputs of regional and global energy models.

The work should be completed at various points. From this perspective, the authors should work to improve the manuscript and prepare it for a journal publication.

Some (Minor) of the changes required are:

1. In the abstract, the main challenges should be described briefly and then the main proposed solutions should be mentioned. It is not necessary to discuss the details of the gaps.
2. An additional English grammar and spelling check should be performed. Moreover, some sentences have to be reformulated in order to present in a clearer way the ideas and the findings of the proposed work.
3. Each acronym should be explained the first time it appears in the text, even if it appeared in the abstract. Checks all abbreviations in text: each word should start with capital to explain an abbreviation. Add a table with the title to the article. Note: The list of abbreviations should be in alphabetical order
4. The “2 Related Research” background section is very weak. Authors should develop this section properly by reading new articles and references about “Sentinel 1” and “CNN”. In the introduction, please list the main contributions of this paper and the research question, and the motivations. What are the main advantages to select CNN?
5. Line 25> “also affects stakeholders”, Please give some examples of such as stakeholders.
6. Line 63> “offshorewind”, Please check. Can be “offshore wind”
7. Figure 7> Please check the horizontal axis of figure 7 “ quarterly time series”.
8. Line 365> “by” can be deleted.
9. All formulas must be transferred to the methodology section. Are these formulas part of the results of your research?
10. Specify the Materials and Methods section and the results section of your research. "It can be developed based on the following: Introduction, Case Study, Materials and Methods, Results, and Conclusion. Display the article with a better structure."