Graph-Represented Broad Learning System for Landslide Susceptibility Mapping in Alpine-Canyon Region
, C. L. Philip Chen 2,*, Feng Qing 3, Xingmin Meng 4, Yan Zhao 4, Tianjun Qi 5 and Tianyao Miao 6
Round 1
Reviewer 1 Report
Your research looks very interesting and the results potentially are significant. But you absolutely have to write a solid and much more critical discussion. Without a deep discussion of the results conclusion's significance drops very much.
181 – 187 – visual interpretation for landslides is proper, but landslide susceptibility mapping on the base of post-landslide maps could lead to wrong results. Parameters (slope/aspect/NDVI/land use etc.) were calculated in the case of pre-landslide or post-landslide remote sensing data?
197 – “specify 21 features from…” reader does not know what these features are. There is more information later, but they are not clear enough. An additional table of these features should be added with information about the type of data also. As I suppose, rainfall data are not on 50x50m resolution... Now, we do not know what environmental factors have been taken into account.
206 – curvature, Roughness, and Topographic Position Index factors are standardly used at high-resolution DEMs. Are these factors show anything on 50x50m resolution?
206 – 208 – “Plane curvature (…) to understand the process of erosion and runoff formation.” Yes, of course, but erosion and runoff are rather not a factors for landslide susceptibility…
219 – did 1-year rainfall (R1y) differs in scale of one valley?
440 – “…almost all the factors…” – what does it mean? Which one not? R1y for example?
449-454 – did you think about the “rain/wind aspect”? In my region, wind-aspect plays a crucial role in local changes in precipitation. Probably in your region is similar – wind directions cause local higher rainfall and this is more important than eg. solar radiation
461-469 – this should be in the discussion section
470 – discussion section is very poor. Authors should discuss: 1) probable mistakes of their research, 2) show limitations of their methods, 3) show which analyzed factors resulted as ‘not influenced on model’, but for nowadays knowledge plays important role in landslide 4) influence of input data (for ex. resolution) on results, 5) etc. etc. etc.
508-515 – I agree with the significant importance of these findings. But their importance is much lower until it won't be discussed much deeper and with bigger self-criticism, with the role of input data.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The paper "Graph-represented Broad Learning System for Landslide
Susceptibility Mapping in Alpine-canyon Region" is an interesting paper. But there are some points that need to be addressed.
1- What is the novelty of your study?
2- Please move figure 1 of your study to the study area section.
3- The literature needs improvement. There are several new techniques used in this domain you can refer to them to enrich your literature. for instance
https://doi.org/10.3390/rs13194011
https://doi.org/10.1016/j.catena.2020.104833
https://doi.org/10.1007/s11069-021-04844-0https://doi.org/10.5194/gmd-2022-119
4- The legend in Figure 3seems not correct. for example, the lithology varies from 0 to 5. What does it mean?
5- You need to use a suitable format for the equations based on journal standards.
6- Figure 4 is not clear. It needs further imprivements.
7- The discussion is poor. need to expand and compare your finding with a minimum of 3 similar studies. Also, mention the limitation of your proposed work.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Please see attached.
Comments for author File: Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
no more comments
Reviewer 2 Report
A graphical abstract is needed /.
