Add test dataset for two-station

[aappling-usgs]

CDI Proposal Tier 1: Add a small test dataset with two stations of dissolved oxygen observations on which the new two-station estimation capabilities can be demonstrated and routinely tested.

[mdodrill-usgs]

The dataset proposed in #469 needs to be revisited:

• Ian's data release file 2_VFTS_and_One-station_model_input.csv has the lag already applied to upstream DO columns. I didn't catch this, thinking that the lag still needed to be applied. Therefore, the current implementation isn't correct.
• The file 2_VFTS_and_One-station_model_input.csv could possibly serve as an example of preformatted data (after helper functions apply lag).
• docs: ?two_station_example needs updated, correct source (link to data release, if still appropriate).
• Look at additional files in data release (+ what Ian sent me) for example starting point for "raw" data -- this is the form that input formatting functions should act on.