Explore Feasibility of Merging Stations

[TrentonBush]

Many stations appear to be the same facility but split into multiple pieces (eg Denver International Airport is broken into one period 1975-1994 and another 1994-present). We want to explore how consistent those records are before (hopefully) merging them.

Idea:
take a seasonally-balanced few years of data on either side of the merge point and compare differences to differences at simulated split points within the consistent data periods

If the gap between the periods is very small, we could look for jump discontinuities as well