incorperate parameterization into model structure

[sdtaylor]

the current setup works well for parametrizing the complex (uniforc, unichill, gdd) models using scipy.optimize. But scipy.optimize is overkill for the new simpler models (naive, linear_temp). So a solution is to switch the optimizer from the current location in scipy_bootstrapped_estimates.py to models.py, allowing for a simple solution for the simple models. This would also allow different scipy.optimize.differential_evolution arguments for the more complex models.

[sdtaylor]

this is in the new pyPhenology package, but I won't be doing it for this study.