Improve radiative model logic (Mavrin) at low temperatures

[theo-brown]

Currently WIP.

Idea is to:

• Use the high-temperature (coronal) Mavrin model where possible
• When the temperature drops to low, switch to using the low-temperature (edge, non-coronal) Mavrin model

Potential later work:

• Do something smart when we drop below the threshold of the non-coronal Mavrin model.
• Move things around so these models are kept in a generic location (e.g. physics?)

[theo-brown]

Example cooling rate curves.