This paper is much improved, the method is described well and the authors now carefully interpret their results in terms of spatial climate patters.
I have only two criticisms.
1. It remains unclear to me how the the results from the bias-corrected ensemble were used in the PCA.
Does the PCA use the ensemble median (ie the results shown in Fig 1 & 2)?
2. Interpretation of chagnes in indices from PC1.
>> All indices have higher values in future climate. This can be interpreted as lower difference between seasons (increase of PC1),
Even with the much clearer description of the pca method and terminology, I am still unable to see how this conclusion can be justified. Surely, equation 3 implies that if the temperature data for all months were to increase by one standardized unit, then the future PC1 indices would increase by 1 x sum(T1:T12) (where T1:T12 are values for PC1 from Table 4), ie a change of ~1.1. That is, I do not see how a change in the PC1 indices can be taken to imply a change in seasonality when it appears a non-seasonal (constant) change would also affect the PC1 future indices? Surely you have to subtract the spatial mean of the future data from the future data, rather than use the means from the present-day data, in the standardization? I feel the authors must address this question. |