Dear Allen,

It is a real pity that the Copernicus system has not allowed me to post my editor comment before you submitted the new version of your communication. I therefore enclose it below, instead of posting it. I noted that most of this comment is still relevant, although some bits of your replies are now integrated in the new version. But I feel that you can much more take advantage from the questions raised by the referees.

Best regards, Daniel


Editor comment on “Explanation of the values of Hack’s drainage basin, river length scaling exponent” by A.G. Hunt.

I believe that both referees provided insightful reviews with important and stimulating questions that are furthermore often convergent. Unfortunately, the author’s replies seem often abrupt, which is a problem for a discussion journal, and rather unconvincing. This is particularly the case with the obvious question that river network domains and Earth surface cannot be fully reduced to 2D vector spaces, because they are 2D manifolds that have dynamical interactions. This means that there is indeed place for discussion on the potentials and limitations of insights from a current distribution through a random (flat) 2D distribution of conductivity to the dynamics of river networks. This does not mean that the latter may not give insights in the former, but this is not immediately granted.

It is also difficult to understand that when the author brings some new arguments in his replies, including after a new literature search, he surprisingly states that he does not want to accordingly revise his communication. This is the case for the uncertainty estimates of the Hack’s exponent and the question of the precise meaning of “the strong disorder limit”. In the latter case, he discusses in his reply a little bit precise statements by Sheppard et al. 1999, but does not yet provide “hints on how much non-binary and heterogeneous” is this limit in comparison with other heterogeneous processes, as suggested by referee 2. Let me emphasize that the referred paper indeed distinguishes various percolation models (with a plural) rather than to rely on a given theory. By the way, it seems rather obvious that “numerical values yielded by given percolation models with empirical estimates” should have been read “numerical values yielded by given percolation models with respect to empirical estimates” to avoid any unnecessary polemic.

Overall both referees considered that this short communication points out an interesting coincidence between a given range of empirical estimates of the Hack’s exponent and that of numerical exponents of percolation models, but made understand that it is presently a huge overstatement to claim that the latter gives an explanation of the former. In particular, they pointed out that there are possible alternative explanations. Consequently, the referee 2 suggested that “a more neutral title would be more adequate”, a request in agreement with several questions from referee 1. I agree with this suggestion that should help to accordingly revise this communication to become more informative and less declarative. In this direction, the detailed comments from both referees have to be taken into account.

Editor comment on “Explanation of the values of Hack’s drainage basin, river length scaling exponent” by A.G. Hunt.

I greatly appreciate that, after a few email exchanges, the author re-examined his replies to the referees and took into account their suggestions to somewhat moderate his claim with respect to the complexity of the problem and the uncertainty on the range of the empirical exponent estimate. I also appreciate there is a given attempt to give some insights on the non-binary percolation models and on their heterogeneity type. I am somewhat disappointed that other suggestions remain rather ignored, in particular the difference between a convoluted 2D manifold and a 2D (flat) vector space.
Overall the paper is now publishable by NPG after a last check (e.g., see “…but one of them. In particular,…” in the introduction).

Daniel Schertzer