Adding significance to LMR plots of correlation statistic

[CommonClimate]

The standard plot for field verification (e.g. Fig 7 of Hakim et al 2016) gives no indication of significance. Yet, because of persistence and the relatively low # of degrees of freedom, r values well above 0 might be statistical artifacts.

I suggest re-using the correlation function from Pyleoclim's Stats.py to test for this, and covering all the areas of the graph that are not significant at the 5% level with black dots.

The code is written in Python 3.5, so it may make sense to wait until LMR moves to Python 3 world.