Thank you very much for your revised submission to GMD. Following the subsequent response from the reviewers, I am now happy to accept this paper for publication subject to the minor revisions pointed out by Anonymous Referee #1. The comments have been attached below, for your reference.

=====

The authors have addressed the majority of my concerns regarding this manuscript. While the results contained are not exactly novel (e.g., increasing resolution and using multi-model ensembles improve TC forecasts), the paper should now be a sufficient reference for future studies using TYMIP-G7 data to address the many (interesting) questions that are raised in the last section of this manuscript. In particular, I look forward to seeing further studies that study the impact of model configuration on forecasted TCs and the relationship between maximum point-wise intensity and 3-D structure (there are many features worth exploring in Fig. 7 alone). I have relatively minor comments (primarily points of clarification and typos) that should be addressed, but otherwise I feel the manuscript is suitable for publication in GMD.

Page 1, Line 33; "furnish" is awkward. Replace with "produce."

Page 2, Line 33; "Errors in TC track prediction by the JMA operational global atmospheric model at a given lead time have..."

Page 2, Line 39; "... Fenshen predicted recurvature far from the..."

Page 4, Line 8, "calculated" should be "forecasted"

Page 6, Line 11, "MSSG-A is primarily used..." this seems exclusively used (I don't see any coupling for MSSG-O in this paper). It should be noted if any simulations do indeed couple to MSSG-O.

Page 8, Lines 5-6, How is dissipation determined? When a local SLP no longer exists?

Page 10, Lines 1-2. It is interesting that NICAM tends to have the largest AR50s (indicating a broad wind field) but DFSM has a much larger intensity bias (towards stronger storms). This implies that the internal dynamics of the modeled TCs are very different between the two simulations (DFSM has a very tight inner core versus NICAM which has a weaker, broader core, but also a broader wind field). Interestingly, this actually allows for prediction of what Fig. 7 will look like before it is introduced. The authors could note this specific point as an area that needs to be rigorously explored to understand why the two models produce vastly different TCs (I assume due to both dynamical core and physical parameterizations).

Page 24, Line 24, should be "atmosphere-only"

Page 24, Line 26, another relevant reference may be Zarzycki (2016, Tropical cyclone intensity errors associated with lack of two-way ocean coupling in high-resolution global simulations) who showed that a simple slab ocean was not sufficient to properly represent TC cold wakes and that a turbulent mixing parameterization is required.

Page 24, Line 27, "aa" typo.