Common Interval calculation for chronology-level (non-running) statistics

[kanchukaitis]

In datasets with uneven distribution in time, calculation of chronology-level statistics (especially rbar, but also e.g. EPS and SSS) requires selecting a common interval (or, alternatively, electing to deal with the challenge of non-overlapping series). Currently in ARSTAN, chronology-level statistics are calculated on all pairwise comparisons with at least 20 years of overlapping data. In dplR, the user has several choices for common interval selection, maximizing the number years, number of cores, or a combination of both. Finally, the older way of doing this in ARSTAN finds an optimal period maximizing sample depth and number of years.

In dplPy, ideally the user would be able to elect to use any of these 4 methods, or to specify a common interval of their choosing.