Thank you for your responses to the reviewers. I can now encourage you to produce a revised version along the lines you have described. I will send the paper back to the reviewers to confirm that you have addressed the main points.

I am personally not quite convinced by your response to Rev 1 point 3 (stationarity) . While this is an interesting answer it does not quite answer the question the reviewer asked, which is what the model says about the correlation between LD accumulation and SWWA rainfall. While I understand your point about the mechanism, I don't quite see why you can't extract the exact diagnostic from the model, as the comment implies. Please see if you can address this in a better way when you prepare your new version.

Dear authors
When I looked at your track changed version, it seemed that the scale of changes you had made was rather small and not really commensurate with what the two reviewers requested (both asked for major revisions). Reviewer 1 has looked at the paper again and confirms exactly this opinion. I noted in my previous editorial comment that your response did not answer the referee's request (point 3) about stationarity, and asked you to address this - but you did not. In view of this please prepare a further version of the manuscript, paying particular attention to the three points that the reviewer has emphasised. Once these are addressed, I will consider it worthwhile to ask the reviewers to spend time on the paper again. Thank you for acting on this.

Dear authors
Thank you for the revisions you have made to your paper. As you will see on most issues the two reviewers are happy with the new version and believe it should be published. However I concur with rev 1's concern about the stationarity of the relationship between LD accumulation and precipitation in SWWA (as shown in Fig S6). In the supplement you dismiss this as hardly relevant, but it is precisely the relationship you assume in extending your reconstruction, and it deserves a fair comment in both the main text and the supplement.

Specifically rev 1 and I both understand that the instrumental data you have showed a good relationship over the last century between LD accumulation and precipitation in SWWA. In extending the reconstruction you assume that this remains the case, and maybe it does. However your climate model output in Fig S6 quote clearly shows that, at least in the model, the relationship can disappear or even show an opposite sign for century long periods. I appreciate that this doesn't affect your broader discussion about influences on SWWA rainfall, and that it could be an artefact in the model. But Fig S6 cannot just be dismissed: I am quite sure if the result had been the opposite that you would have made a big deal out of it.

I am therefore minded to accept the next version of your paper subject to:
a) you deal with the minor points raised by both reviewers and
b) you add a statement of caution in the main text that the model does not show a stationary relationship (Fig S6) between LD accumulation and precipitation in SWWA, and that if this is really true it undermines the case for using LD accumulation alone as a proxy for SWWA rainfall. This should be borne in mind when assessing the reconstruction shown.
I would find it hard to accept a version that does not include such a statement in the main text, although of course you are welcome to argue against it if you feel you have a case.

Thank you for your further revisions. While I remain concerned that the assumption of stationarity is crucial to using your measurements as a reconstruction of SWWA rainfall, and that the only evidence we have is against it, I accept that you have gone some way to warning readers, and that the underlying work is worth publishing. I am now accepting the manuscript.