Address possibility of multiple main peaks in KDE

In some cases, there may be two main peaks in the KDE used to identify the ridges (e.g., when quenched spaxels are included in analysis of the rSFMS). The code should be able to address this case and choose the correct peak. For (r)SFMS, this would be the peak at high SFR, but I'm not sure how generalizable that is to other relationships.

Start by locating the KDE used to identify ridges and review how the current peak is selected. Define how the correct peak should be chosen for rSFMS and how that rule should generalize to other relationships; done means cases with multiple main peaks select the intended ridge without breaking existing analyses.

Written by the indexing model from the issue text.