Dear Authors,

although you have replied appropriately to the referee's comments, I find that the result of your revision is not yet ready for publication. Although I am not a native speaker of English I have the strong feeling that the language (words, grammar, logic) needs significant improvement. As there is one author from Bristol, it should be possible for you to find a colleage there who can help in this respect.

A couple of minor issues follow:

1) Eq. 1: As there are (2X+1)*(2X+1) grid points in your smoothing square shouldn't the normalising factor be (2X+1)**2 ?

2) line 297: replace "logarithmic" by "exponential".

3) Caption of Figure 7 mentions a panel 7D, but neither does the main text mention it nor is a figure 7d present.

4) line 347: What is a correlation score? Do you simply mean the correlation between two time series? I.e.
Cor=mean[(x-mean(x))(y-mean(y))]/(sigma_x*sigma_y) ?

5) line 372: small deviations.

6) lines 395 and 426: Referee 2 strongly suggested to replace "calibration". Please do it.

Dear Authors,

I accept your manuscript for publication. It seems that the language has been improved quite much. Congratulations!

Please write equation 1 as follows:

RV_a,b = ..... RV_i,j,

That is, write it as an equation (with an = sign), use the index pair (a,b) of your point of interest on the left hand side, and the summation index pair (i,j) on the right hand side.

Please correct also the following typos:

line 18: see level.

line 317: data (not datas).