Fig 14 from
Rehfeld, K., Marwan, N., Heitzig, J., and Kurths, J.: Comparison of correlation analysis techniques for irregularly sampled time series, Nonlin. Processes Geophys., 18, 389-404, doi:10.5194/npg-18-389-2011, 2011.
http://www.nonlin-processes-geophys.net/18/389/2011/npg-18-389-2011.html full paper in PDF available.
This recent paper looks at the difficult problem of correlating data where each data has a different X axis, that is, each Y point is at a different X, something is irregular about the timescale.
Conventional math won’t work. The choice is either make the data comparable or use lesser known correlation algorithms. None of this is entirely satisfactory. The blunt message is you can’t.
All that is left doing your best with what can be done, is into approximations and heuristics.
The obvious field where this problem is common is paleoclimatology.
At the end of the paper the authors show a very interesting result from tests.
This suggests climatic modal difference between the Medaeval Warm Period and the Little Ice Age, and suggest this might be to do with change in the prevailing monsoon wind. Cause or effect cannot of course be known.
Tallbloke's Talkshop 
Very Harry Potter stuff here. how can you show anything worthwhile with so much uncertainty?
It is no better than the climate models touted by the powers that be.
Verily they be pottering with ‘airy data.
I’m inclined to agree except I haven’t looked at the whole problem they are using as some kind of example. Conceptually I find the idea of changes in linkage attractive.
Came across the paper during trawling looking for clues on practices within the climatic community.