[water] bad calibrations pass through qa/qc checks

[rfiorella]

Bad calibration values are passed through qa/qc checks currently, and produce implausible (or impossible) isotope ratios.

[bumbanian]

Many of these would be identifiable w/ a screening criterion for slope and/or R^2. Should explore, but I expect ~0.9 - 1.1 is a reasonable range for the regression slope parameter. Maybe start there?

[rfiorella]

Currently no r2 threshold in calibrate_ambient_water_isotopes - seems an obvious place to start to improve this, and would make code structure more consistent w/ carbon calibration.