Paper Title: Evaluation of Environmental Factors in Landslide Prone Areas of Central Taiwan using Spatial Analysis of Landslide Inventory Maps

This study select satellite imagery sources that its acquisition date should be close to disaster events for reflecting real geomorphological change. So, all of the satellite imagery sources could not be identical or provided by single satellite data acquisition system. Of course, SPOT series and FORMOSAT-2 satellites had different spatial resolutions. To deal with the problem, this study collects and uses aerial photo close to the event-based landslide inventory to which systematically improves the quality of data content and data structure so that it can definitely promote accuracy and reduce inconsistency between classified results from different satellites


Original
Thanks for the comments.
To avoid excessive figures and reduce unnecessary information, the authors will try the best to combine those similar GIS-data layers, or statistical results of identical regions into single figures within reason as to not diminish the quality of the study and then polish the revised paper to make it more readable, the data more clear.Under your direct indications, the Introduction section of the manuscripts will be rewritten substantially and introductory information found in other sections will be moved to this section.After revision, the full paper will be expected to be well-structured, more concise and highlighted the research results and data value.Thanks for the comments.
The two sections described above "Methods and spatial data" and "Results" will be reconstructed and rephrased to reflect data and methods and results of this study.To avoid misunderstandings, the authors will remove the words above "primary contributors" and "secondary contributors".
The original paragraphs will be amended in detail and also add some reference to support the above description.Thanks for the comments.
The original sentence has not described NDVI method or value precisely so this causes misunderstandings.
In a single, identical area, satellite imagery acquirement may be affected by seasonal changes, the angle of incidence of the Sun, atmospheric radiation and capture angles, which results in inconsistence of NDVI range for each difference period images.
Therefore, this paper chose some of samples plot on the single Item No.

Original Paper Comments Author's Response Page
Line Text image to define NDVI threshold to separate vegetation and nonvegetation areas with reference GIS-layers such as roads and land use maps to identify whether it belongs to landslide area.
The revised manuscript will, for example, take given satellite images and utilize NDVI results to illustrate vegetation change and the original sentence will also be amended according to the above mentioned.Additionally to increase clarity, some references to NDVI method of land cover detection will be added.Thanks for the comments.

Item
The authors will follow your suggestion to reduce some information to highlight the standpoint of Section 3.3.
Thanks for the comments.

Similarly, the new landslide ratio as the ratio of new landslide area to the watershed area.
13 Page 15, lines from 2 to 10: these introduction is irrelevant to the Section, and should be removed.
In In some parts of the paper Uchiogi is mispelled as Uchiughi, these errors will be corrected in the revised paper.
According to Uchiogi formula, one can assume that for a given watershed, triggered landslide emperica equation for a rainfall event under 200mm of critical rainfall can be stated as follows: One can find that if the accumulated rainfall is up to 200mm and is inputted into this formula, the estimate of the new landslide ratio by this event can be calculated as follows: The calculated result shows that the new landslide ratio is zero which means no landslides or slope failures occurred during Item No.

Original Paper Comments Author's Response Page
Line Text available in Japanese, which is not acceptable), in which the same formula is used and the value zero is obtained for cumulated rainfall smaller than critical rainfall, as it should be.
this event which does not match with physical phenomena of the exceedance of critical rainfall for slope failure.
To modify the imperfection of the original formula, the authors suggest the addition of one constant(C) to the original formula.C aims to represent that when the accumulated rainfall equals or exceeds the critical rainfall a certain amount of new landslide occurs in a watershed.The correction will make the original formula more reasonable and enhances its applicability.The modified formula is as follows: In terms of obtaining the value of C parameter, the data for each rainfall-induced landlide event and its corresponding accumulated rainfall will be compiled for statistical regression based analysis on the suggesting governing formula (see the figure).The figure shows that the C parameter is a constant and seems like a truncated value of the y-axis(new landslide ratio).
From the perspective of physical significance, it implies when a rainfall event reaches its critical rainfall, this parameter(C number of figures could be reduced, for example, summarizing figures from 1 to 3 into a single figure; the same applies to other cases. the study area.This Section is too long and not useful to make the point of the manuscript (see specific comments).Thanks for the comments.Shenmu area has been affected by serious sediment-related disasters and the hazard history information is used to point out and classify the landslide areas of differing proneness.The section aims to deliver hazard history of over 20 years and the effects of heavy rainfall and typhoon events highlighting and strengthening the legitimacy of the research and the context which it represents.Without the hazard history, the validity of the report becomes more questionable and is seemingly incomplete.The inclusion of hazard history information in landslide research is seen as industry-standard as can be seen by many different researches.Nonetheless, the authors will should present strictly the results of this study(see specific comments) Figure.Photos of dark grey argillite and grey slate the following, the Authors use the Uchihugi formula to calculate the new landslide ratio from the magnitude of the rainfall events.They modify the original formula adding a parameter, C. How do they obtain the value of C parameter quoted in figure 17? What do they mean by "initial increment landslide ratio"?Which is the physical meaning of the constant they introduce?Page 15, line 15: "However, when the rainfall parameters of Uchihugi empirical model reach the critical rainfall, the new landslide in the watershed becomes zero."This sentence is not clear to me!It seems that when the value of cumulated rainfall is larger than the value of critical rainfall the new landslide becomes zero.I checked in the article "S.-J.Chiou, et al.: Evaluating Landslides and Sediment Yields Induced by the Chi-Chi Earthquake and Heavy Rainfalls" (the suggested reference is actually only Thanks for the comments.
follow your suggestion to remove relevant information and reconstruct the Discussion and Conclusions sections based on main obtained results.In addition, landslide potential map of the three temporal periods pre-1999 Chi-Chi earthquake, from 1999 Chi-Chi earthquake to pre-and posttyphoon Morakot would be also discussed and validated in the revised manuscript.