I was able to view the article for free at
http://www.sciencedirect.com/scienc...d=918210&md5=f2f7d4d2356ab5452fb98ca6fca46333
I don't know if this will work for everyone.
When I tried it from home, it required a $30 download fee, but I could get it from work. It is an interesting article, and embraces the issues presented by the Hurst phenomenon (unlike the unscientific stance of many posters here who dismiss it without understanding it - even making absurd allegations that we should not carry out legitimate scientific research in case it affects policy - a disgraceful position for any scientist).
Based on a quick read of the paper - without confirming any of the detailed analysis inside - the paper embraces LTP as an important part of climate analysis (as did Rybski et al.). A good test is applied within the paper, which appears to be more sensitive than that applied by Koutsoyiannis et al.
I also agree with the paper that there are several problems with the present palaeoclimate reconstructions; Halley notes his results are very sensitive to scaling errors (cf. von Storch and Zorita), plus highlights a sensitivity to spatial sampling of the instrumental record, a problem which can only be worse for the reconstructions. These sensitivities could match or exceed the magnitude of the effect being measured (discussed in the paper), and therefore it is difficult to have full confidence in these results until these issues are properly resolved.
Halley observes the same inconsistencies between the statistics derived from the reconstructions, in particular that the reconstructions yield non-overlapping confidence intervals for many statistics. Again, this suggests the present state of palaeoclimate reconstructions are inadequate for this purpose.
So for many aspects of the study, Halley agrees with and reinforces the position presented by Prof. Koutsoyiannis. The statistical result is an interesting one, and needs the palaeoclimate community to step up and resolve several issues on their doorstep.