The extensive discussion indicates the need to considerably revise the manuscript in response to the reviewer comments. Hence the authors are invited to submit a manuscript that is revised according to their suggestions given in the Discussion.

The authors have responded comprehensively and satisfactorily to the reviewer comments. I note that some reviewer comments criticized the overall approach of socio-hydrological modelling; the authors were naturally not in a position to consider all suggested improvements, but they well defended their current approach (and better explained the study’s rationale) in response to the reviewers’ concerns. Moreover, the paper now better describes the model assumptions and caveats, also involving a sensitivity analysis. I agree that the study shall be restricted to the Murrumbidgee, i.e. not including a comparison with Republican.
I’d like to add a few (technical) issues to be addressed, as follows.

* Fig. 7 (and others): the labels a), b) etc. are not shown in some figures, please add this information, otherwise readers are confused.
* Please consider whether Figs. 6 & 7 can be merged so that readers can directly see to what extent the observed(?) and simulated curves are in line with each other – if the agreement / simulation quality is something you’d like to highlight with these figures. In some cases (irrigation area and use after ca. 1990) they don’t match well, which should be explained - I may have overlooked such an explanation though.
* Fig. 11: Explain abbreviations “W” and “πd” in the lower right plot (d).
* Two reviewers raised general concerns on the type of modelling used here, which are not directly addressed in the revision. I leave it to the authors whether they’d like to refer to the open discussion of the former manuscript version (in HESSD) for once, where they provide a balanced discussion of this issue.

Please have a final check whether the labeling of Fig. 6 a-h is optimal: for example, c is total irrigated area while h is irrigated water extraction - shouldn't it be the same labels? The same holds for a and e, and b and f, respectively.