Obv I am not tip.
If NAO is static, determining factors on sensible weather will be highly correlated to the more unstable PNA index, since when you lower degrees of freedom, you can more accurately correlate outcomes to the degrees of freedom that remain (as the entire domain of outcomes is reduced when one degree of freedom is static).
In essence, if NAO is taken care of, then indices that are oscillating such as PNA are better for differentiating sensible weather outcomes