Dear authors,

Thank-you for the much improved manuscript and response to reviewers. I don't think it needs to go back to the reviewers as you have largely addressed their comments, however there are a few remaining minor issues.

Fig 1b - TG locations (plural).
Regarding the use of place names, Aberdaron is marked on the map but not necessary to the text, and the Llyn Peninsula isn't on the map (Fig 1). I suggest you remove one and add the other to the map.

line 106 These shallow water effects are enhanced around the island because of curvature on the directions of current flow. I don't understand this. Rephrase?

Caption of Table 1 refers to tidal phase but there's not phase info in that table.
Table 1 - Better not use dd/mm/yy for dates at all, it's confusing for Americans (though at least it's in the caption). Apparently the house standard is 25 July 2007 (dd month yyyy), 15:17:02 (hh:mm:ss) . yyyy-mm-dd HH:MM is probably also OK if you're short of space.

Throughout: I suggest getting rid of "Phase" to refer to the measurement campaigns and replace by "Deployment" or similar. Otherwise there's phrases like "phase 1 residuals" which is confusing for no good reason.

The data is available to download as required. Thank-you. You might consider also depositing it in a suitable respository of similar data.

line 184: I really dislike "obviously", it's either patronising or a wallpaper-word. I guess it's there to placate those reviewers to whom it's obvious.

line 204-208: Fig 3 has red & black curves? & level differences are not plotted, so line 208 is not right.

Fig 4 a and b : It's not ideal that the coloured cells are aligned with the arrows on their corners rather than centred, though I know it's a bit awkward to correct in matlab. It means the green cells are misalinged with the longest arrows. And it makes me wonder if it flags a possible problem as matlab will default to plotting interpolated data rather than the actual values of each cell. Please check that the plot is as intended. It is important as it affects the argument about TPXO9 success in replicating the tide. Also please edit the colour scale so it doesn't saturate at 2.5, so we can see how high it gets. (c) has a of maximum only about 1.6 - which cell is represented? Perhaps draw a box round those cell(s)? Why is it less than 2.5 from the colour scale?

It does look like the TPXO9 model sees the island in some fashion, through averaged bathymetry, as there is is an increase in current speed through the channel. It would be interesting, though not essential, to see the model bathymetry.

lines 284-306: You've used one method to calculate energy from obs, and another for the energy from TPXO9? It makes it hard to compare... can you use the same for both? Also there's some repetition here.

line 319 : astronomic speed . Speed during GA tide?

line 350: marked with arrows? These have now gone? If you do add more, can I suggest magenta, for better contrast with clouds.

Table 2 would be less cluttered and easier to read if you got rid of most of the lines. "Horizontal lines should normally only appear above and below the table, and as a separator between the head and the main body of the table." Also some of the vertical lines - I suggest you group in pairs.

You have bypassed reviewer 2's question on uncertainty on tidal constituents. I know why, it's not calculated in TASK, and is not trivial, but I think you do need to comment a bit further on this.
The natural place for the non-tidal standard deviation would be in Table 2 with the results rather than Table 1. What is the non-tidal residual standard deviation of the observations if TPXO9 tides are assumed correct? What if only M2+S2+M4 (from each) is used (ie exactly how much is omitted in all the ignored constituents?)

And finally, there's a lot of typesetting problems. I think the typesetter will help with these at the next step but please help by setting the equations and variables in italics where appropriate, and ensuring tidal constituents are correctly set. m s-1. Etc. Please check the proofs carefully.

Regards,

Dr Joanne Williams