Dear Authors and Editor,
I have read the article entitled "Application of novel hybrid model for Land subsidence susceptibility mapping" line by line carefully and it is worth mentioning that the article has many ambiguities and questions in my opinion, which I have mentioned below. Therefore, considering the title, aims and analyzes done in the article, as well as referring to the subject and the less studied aspects of geology and geomorphology, I think that this article has not met enough standard and hence, it is not acceptable for Geological Journal and I reject it.
• As can be seen from Figure 2, this is not a geological and geomorphological phenomenon called "land subsidence" but gully erosion. The shape and morphology of land subsidence are not mainly such linear and it can be surrounded an extent environment. According to this figure, most of the geological formations belong to the Quaternary (low level piedmont fan and valley terrace deposits) which are young and erodible alluviums formations and containing mainly marl, gypsum and salt that are more susceptible to gully erosion instead of land subsidence. Therefore, from the morphological point of view, the phenomenon of the mentioned hazard of the article is ambiguous and questionable.
• The article does not provide information on the effect of groundwater wells and their uncontrolled extraction, which as a result leads to subsidence, and the authors only emphasize on the modeling procedures. It is only mentioned in the discussion that there are deep wells in the study area and how does this justify this important issue?
• Unfortunately, the geological and geomorphological aspects of the subject have been less addressed and the importance of modeling results has been emphasized, and this is not appropriate for an international journal, like Geological Journal, where the geological aspects of the subject is very important.
• To what extent identifying areas prone to subsidence, that I believe are not subsidence, has helped managers to control this geological phenomenon so far? Unfortunately, identification maps of these phenomena alone are not a good solution for managing and preventing these phenomena in arid and semi-arid regions of the world.
• The authors have finally concluded that the GBM-EG hybrid model is recognized as an appropriate model. Other authors may reach different conclusions with this dataset and other models in the same area, indicating uncertainty in the discussion of modeling with mathematical algorithms. However, in the last decade, all the efforts of natural hazard researchers have only led to the preparation of a susceptibility map, and this trend is worrying for the prevention and management of natural hazard in the future.
• Most importantly, the discussion of this article has not poorly written but unfortunately does not address any of the geological issues. In other words, no geophysical studies have been conducted on the status of groundwater aquifers. In the areas with high groundwater declining, the least susceptibility to land subsidence has been seen in this work, and this result is not acceptable.
• Other contradictions can be also observed in the article: for example, in the areas with high drainage density, a high susceptibility to subsidence can be observed, while the result should be the opposite. Also, how is it possible the areas that are less distance from the faults acquire the least susceptibility to the occurrence of this phenomenon? In addition, the slope angle had the least effect on the occurrence of land subsidence, and this is scientifically unacceptable. This area is mostly plain and has a low slope. How can the least impact are found comparing with slope aspect, which in the plains it cannot be related to land subsidence. I believe that the declining in groundwater resources should be the most important factor for land subsidence occurring.
• In the article, the effect of factors affecting the occurrence of this phenomenon has not been discussed, in other words, how these factors have affected the land subsidence? The authors concluded that the hybrid model (GBM-EG) has the greatest ability to predict the occurrence of land subsidence; while the effect of conditioning factors has been investigated by only the random forest (RF) model and the obtained results have many inconsistencies. Regarding Fig.5, land use and elevation in the two combinations of training datasets (50:50 and 60:40) and other combinations (70:30, 80:20 and 90:10) were, respectively, known as the most important conditioning factors. How has elevation been the most effects on the occurrence of land subsidence? I could not find reference/s in this regard to refer by the authors?
Overall, instead of using different combinations in the training dataset, the authors could discuss and analyze most of these phenomena by referring to the field surveys and observing these phenomena more closely and studying the geology of the region. Therefore, the results of these analyzes are very ambiguous and the authors have not been able to discuss geology and geophysics well and there are many contradictions about the main reasons for this phenomenon.

Dear Authors,
While thanking the respected authors of the manuscript, it is noted that unfortunately, many doubts and questions from the previous stage were not well responded by the authors, for example: “Other contradictions can be also observed in the article: for example, in the areas with high drainage density, high susceptibility to subsidence can be observed, while the result should be the opposite. Also, how is it possible the areas that are less distant from the faults acquire the least susceptibility to the occurrence of this phenomenon? In addition, the slope angle had the least effect on the occurrence of land subsidence, and this is scientifically unacceptable. This area is mostly plain and has a low slope. How can the least impact are found compared with the slope aspect, which in the plains it cannot be related to land subsidence. I believe that the decline in groundwater resources should be the most important factor for land subsidence occurring”.
As I mentioned above, there are some contradictions between the modeling results and reality. Authors just responded, “Sometimes the effect of one parameter causes the effect of another parameter to disappea”. I believe that the authors could not well justify some contradictions as the critical comments. In other words, the authors have paid much attention to the modeling aspects of the case, while the conceptual framework of this work brings many ambiguities and questions. Therefore, this version of the manuscript cannot be published.