Evaluation of Historical CMIP5 GCM Simulation Results Based on Detected Atmospheric Teleconnections

Round 1

Reviewer 1 Report

The study discusses atmospheric oscillations, which are found using the correlation method of Wallace and Gutzler. This method finds remote patterns of co-variability, which are not always oscillations. More analysis must prove the oscillatory nature of these patterns. Moreover, the oscillation is described a dipole. Oscillations have phases not poles. I suggest the authors replace atmospheric oscillations by teleconnection patterns.  

Grid point to grid point comparisons can be an objective analysis when all grids have the same number of points that correspond to the same geographical locations. CMIP 5 models do not have the same grids or resolutions. Interpolations introduce small errors. Reanalyses used for validation also have different resolutions and grids. This point needs a discussion in the manuscript.  

L34: synoptic scale usually denotes low- and high- pressure areas on weather maps. The synoptic-scale systems are not modes of climate variability.  

L36-38: An atmospheric oscillation is not synonymous with a teleconnection. For example, mechanisms explaining the North Atlantic Oscillation include the extra-tropical variability and teleconnections from the tropics. 

L45-46: Please explain how teleconnections can be detected based on decadal periods. Harding et al. 2011 uses monthly anomalies, which only capture the interannual variability. Does the authors refer to inter-decadal differences noticed for NAO (e.g., Weisheimer et a. 2016)?  

L152-153: “grid cells with correlations that are in one‑to‑one correspondence with each other” one-to-one refers to a correlation value of 1? Please be more specific.  

L278-279: Please clarify whether the threshold value of correlation corresponding to the 25th percentile is the same for reanalyses and models.   

Figure 8: Please clarify the meaning of grey squares, blue dots, and red dots.  

Minor comment: 

AT500HP is an unusual acronym for the 500-hPa geopotential height. Z500 is commonly used.  

L159: “permutation test” -> a permutation test 

L160: “each grid points which consists the PotACs” -> each grid point in a PotAC  

L466-467: “not only for the 30‑year‑long periods but also for the 10‑year‑long periods.” -> for both the 30-year and 10-year periods.  

References 

Weisheimer, A., Schaller, N., O'Reilly, C., MacLeod, D.A. and Palmer, T. (2017), Atmospheric seasonal forecasts of the twentieth century: multi‐decadal variability in predictive skill of the winter North Atlantic Oscillation (NAO) and their potential value for extreme event attribution. Q.J.R. Meteorol. Soc, 143: 917-926. doi:10.1002/qj.2976 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Dear authors,

I am not comfortable with your using “Atmospheric oscillations” as synonym for “teleconnections”. It may be common in the field, but it is simply not oscillations but patterns of variability you are talking about.Therefore, I suggest that you reformulate the first sentence and accordingly the rest of Introduction.

This is a long paper with many details, and their order of appearance is mixed up. While it is clear that lots of work has been done, I am not sure what new knowledge we can gain from this paper. 

You find that the GCMs typically reproduce the teleconnections in similar geographical areas as the reanalyses, but there are signifcant differences concerning their intensities and the position of the most intense regions.  Based on the comparison of the results of the stability patterns and the loess regression curves, you found that the CMCC‐CMS model is the only GCM which performs well with respect to the stability patterns and loess regression curves. Why is this? What aspect in this particular model do you think has contributed to this outcome? What the developers of other models can learn from this result?

In your conclusions, you claim that your method provides temporal evolution of the quality of the GCMs. This is unclear and should be clarified.

If the paper is to be resubmitted, it should be shortened and the value of the method clearly demonstrated. Please consider reducing jumpiness in the text

Author Response

Please see the attachment.

Author Response File: Author Response.docx