Permeability Characteristics of Improved Loess and Prediction Method for Permeability Coefficient

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The authors present permeability tests carried out on loess as they are or mixed with cement or lime. The authors linked permeability coefficient to cement or lime content, dry density, particle size (e.g. d50, d60), coefficient of uniformity and curvature. They studied the factors that most influence the coefficient of permeability and they applied the SVM in order to predict the permeability coefficient. The paper need a major revision because contains inaccurate technical terms and the experimental procedures need to be explored further as indicated below.

Page 1 line13: specify particle size characteristics (e.g. d50, d60)

Page 1 line13: replace “particle gradation” with “grain size distribution”

Page 1 lines 14-15: “Variable pressure head infiltration test” is not clear. “Infiltration test” is a term usually used for field tests, “variable pressure head” in geotechnics is usually defined “falling head permeability test”

Page 1 line 18: replace “inhomogeneity coefficient” with “coefficient of uniformity”

Page 1 line 36: “pore ratio”: do you mean “void ratio”? (see also page 2 line 46)

Page 2 line 83: Replace always in the manuscript “particle analysis” with “grain size distribution” or “particle size distribution” or “grain size analysis”

Page 2 line 85: please define in brackets the formula of Cu and Cc

Table 1: with “Static gravity” I think you are referring to “Specific weight (Gs)”

Table 1: Please specify how the maximum dry density and optimum water content were determined (was a Proctor test performed?).

Table 1: Please specify in the text that Atterberg limits were performed.

Page 4 paragraph 2.2: just from the title of the paragraph it is clear that the test was performed as falling head permeability test. Instead it is necessary to specify in the text that a falling head permeability test was performed, specify also the direction of flow, the test time (after how long was the permeability coefficient determined also in reference to the curing time of the cement?)

Page 7 lines 145-146: it is preferable to use “Content” rather than “concentrations”

Page 13 line 302: Move the definition of the acronym SVM to line 296

Page 14 equation 2: define “a”

Page 14 lines 335-336: there is a repeated sentence

Page 15 line 349: Define the acronym RBF

Page 17 paragraph 4: Authors should carefully evaluate the real need to include the figures contained in this paragraph. The discussion should be related to the data obtained in this paper in reference to the literature data. If the graphics have been taken from other papers, the necessary copyright permissions are required (see instructions of the Journal).

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The submitted article entittle “Permeability Characteristics of Improved Loess and Prediction  Method for Permeability Coefficient", investigates the influence of  dry densities, particle size characteristics, particle gradation, and admixture contents and types on  the permeability of improved loess, focusing on the Malan and Lishi loess.  Furthermore,  developes a predictive model for permeability using a support vector machine (SVM), The article addresses an interesting topic.

However, to further enhance the clarity and impact of the title, we suggest the following refinements:

In order to make the article more useful to the scientific community, you should indicate how you have arrived at the data indicated in Table 1: If you have made laboratory measurements, indicate the method (even if there is same standard in this respect, indicate which one) that you have used and the measuring instruments. If you have taken them from other scientific studies, please describe them.

As in Table 1, Table 2 lacks the standards and instrumentation used to perform the tests.

You talk about comparing several prediction models, the SVM, a regression line and a neural network, but in the article you only indicate the characteristics of SVM. For the article to have value from a scientific point of view, the results must be supported by studies that are reflected in the article. For example, it talks about the Mean Squared Error, but at no point does it indicate which are the values of this error for each of the models it compares.

Therefore, strongly recommend a major revision, because in its present form it cannot be published.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors revised the paper as suggested by the reviewer improving the quality of the manuscript

Reviewer 2 Report

Comments and Suggestions for Authors

With the changes made, the article has improved considerably, so in my opinion, it can be published.