Estimates for the Effective Permeability of Intact Granite Obtained from the Eastern and Western Flanks of the Canadian Shield
Round 1
Reviewer 1 Report
Review of the manuscript minerals-853911 entitled "Estimates for the Effective Permeability of Intact Granite Obtained from the Eastern and Western Flanks of the Canadian Shield" by Selvadurai et al.
The paper by Selvadurai et al. deals with the use of surface permeability measurements to obtain estimates for the effective permeability of granite cuboids from the eastern and western flanks of the Canadian shield, locations that are considered for a potential deep geological repository. The experimental setup was developed by one of the authors and presented with exhaustive details in a previous publication (Selvadurai & Selvadurai, 2010, Surface permeability tests: experiments and
modelling for estimating effective permeability, Proc. R. Soc. A 466, 2819–2846). It follows that sections 2.2 (the experimental facilities and procedures) and 3 (theoretical aspects) can be drastically reduced to focus on the data obtained on new materials and their analysis.
Looking at the surface permeability data we notice that there is a major difference between both granites : the permeability of the Lac du Bonnet granite is one order of magnitude lower than that of the Stanstead granite. An immediate consequence of this is that the cumulative volume vs. time curve, which should be linear to ensure that we are in the steady state regime, is indeed linear for the granite with the highest permeability (Fig. 7) but shows a pronounced curvature for the granite with the lowest permeability (Fig. 9). This means that the condition of steady state flow is not strictly fulfilled in the second granite during a run. The authors are partly aware of this problem since they run several tests at each location on the surface of the cuboid and present the measured maximum and minimum permeability values. But they use a mean value for the slope of each flow curve, which is not correct since the curve is not linear. If the time to enter the steady state regime is larger than the testing time as is suggested by Fig. 9, the authors should check whether the permeability derived by taking the final slope of the flow curve for each successive test is not decreasing. This would mean that the last value is closest to the 'true' permeability. If the successive values are randomly distributed, the authors should at least use the final slope for their permeability derivation.
In section 5, the authors derive interior permeability estimates by using a kriging procedure. What they obtain is a discretized volume distribution of local permeabilities. These local permeabilities are scalar, meaning that local anisotropies at the elementary volume scale are erased. This may have consequences on the calculation of the effective one-dimensional permeability and the flow pattern presented in Figs. 20 and 21. At this point, the question is: if anisotropies exist at the elementary volume scale, does the combination of surface permeabilities measurements and kriging technique give some insight on them? If yes, why did the authors use scalar values for local permeabilities? If not, how much confidence can we place in Figs. 20 and 21, when applied to repositories where flow pattern are a key issue.
In summary, this paper is really interesting and well-written. Some rearranging is suggested to focus on what is really new. Two issues need to be addressed as suggested above. Once this has been done, a revised version of this manuscript may be ready to be considered for publication.
Author Response
Please see the attachment
Author Response File: 
Author Response.docx
Reviewer 2 Report
Review of
Estimates for the effective permeability of intact granite obtained from the Eastern and Western Flanks of the Canadian Shield
By
Selvadurai et al
This manuscript presents experimentally derived surface permeability measurements and numerical simulations to extrapolate internal permeabilities on centimetre scale cubic blocks of two types of granite from opposite ends of the Canadian shield. The two main objectives, as I understood, were to compare the permeabilites of the two rocks and the suitableness for use as radioactive waste repository storage rocks, and provide a method for estimating internal permeabilities from surface measurements. I found the manuscript topic somewhat underwhelming and flawed. The main result of the manuscript is the acquisition of two new averaged permeability values for the two granites (Stanstead and Luc de Bonnet) which demonstrates that the two rocks have permeabilities that vary by about an order of magnitude. So essentially the ‘new’ contribution is two numbers and this is not really ground-breaking. The main result is also somewhat diminished by the authors acknowledgement that at depth, where there is a lithostatic pressure, the permeabilities of the rocks will be quite different anyway. So I’m left with an overall feeling of what was the purpose of this study really if the results can’t be applied to waste repositories at depths of a few hundred metres? Also, I found the rationale for the modelling section that attempted to estimate the internal permeability from the surface measurements lacking. What is wrong with simply measuring the internal permeability by taking a core sample, which would also, by the way, be much easier to measure the permeability as a function of confining pressure? Why should the surface permeability of each sample be different, that wasn’t really explained well, are the differences in surface permeabilities across the same sample related to weathering? So I think the rationale needs to be rethought or explained better. Given this and several other points, that I list below and give in an annotated PDF, I suggest that the manuscript be reconsidered following major revisions.
As well as the many comments and questions given in the annotated PDF I list some other concerns here:
I must first comment that I have never seen a manuscript with so many references (131!). I also comment that more than 30 of these references are studies from the authors. The reference list needs to be reduced by at least half, if not more. Some of the citation in the introduction is completely inappropriate, for example in line 64 where the reader is told that a company has been considering using plutonic rocks to host radioactive waste and then advised to read 38 papers on this topic with no more specifics given! This needs to be changed.
Another oddity is the use of significant figures when quoting the permeability values which are used with great inconsistency throughout, this needs to be fixed as it is very messy and hard to follow. The main result of the manuscript given in line 572 is that Stanstead granite has a permeability of 5.7 x 10-18 m2 and Lac du Bonnet granite a permeability of 0.9 x 10-19 m2, but that is reported in the manuscript as 57 x 10-19 m2 and 0.97 x 10-19 m2. So why the discrepancy in accuracy between the two data sets, if anything I would have though it possible to report the higher perm Stanstead with more accuracy than the almost 1 order of magnitude less permeable Lac granite, not the other way around. Throughout, the reporting is a bit of a mess.
The authors should give more background on the local geology and the condition of the outcrops from where the rocks were sourced. More information on the anisotropy would also be welcome. I cannot understand the rationale for performing 50 odd spatial measurements of permeability if, as the authors acknowledge, the rocks are isotropic. What is trying to be gained from those measurements, no rationale is given. Why should the permeability be different in any one face and why should it vary spatially? Also, the reader is not given any information about each face of the sample (other than a number), and the spatial distribution of the various measuring points is also not sufficiently described in the methods.
There is also no attempt to describe the measurement errors. How sure are the authors that the distribution of permeabilities they measure across the sample face are not simply within the error of measurement. These are vanishingly low permeabilities (i.e. 10-18 – 10-19) and so can small differences really be resolved?
Comments for author File: 
Comments.pdf
Author Response
Please see attachment
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
I could not see how the authors have commented on or corrected any of the suggestions or comments that i made in the annotated PDF. These seem to have been either completely or largely ignored. I would ask that this is revised.
