Thank you for your detailed responses to the comments of the three referees. There is just one matter raised by RC2 that I would like you to address:

In the response to RC2, Point 6:
6- Figure 3: In a correlation plot, error bars of measurements are needed. Did the regression procedure take into account errors of the measurements?

Response:
Errors bars were added as shown in Fig. 3, page 18 of the revised version.

Thank you for doing this.

Errors of the measurements were not taken into account in the regression procedure.

Can you please do this. It should be straightforward to calculate the gradient which takes into account errors on both X and Y data, for example using the orthogonal distance regression technique, or perhaps others.

Many thanks.