I found many works that demonstrate the non-stationarity of NAO in relation between the various types of data (precipitation, temperature etc ...). But I haven't found a statistical study of NAO data. Can anyone help me? Are there such studies? In order to be able to correct correlations, the time interval must be long enough (but not too long) so the two phenomena related are known to be considered stable. I did statistical stationarity tests, now I need a bibliographic confirmation.