Update e_s to a more accurate formulation and add e_i (based on Murphy and Koop, 2005)

[isilber]

Description Of Changes

In this pull request, I've modified the equilibrium vapor pressure (e_s) calculation to a more accurate formulation based on Murphy and Koop (2005; reference added). I've also added a method to calculate the ice equilibrium vapor pressure (e_i) based on the same manuscript as well as a unit test for this method. Different variations (using different input coefficients) of Teten's formula have been used in the literature (e.g., Bolton, 1980; Alduchov and Eskrige, 1996) but the use of the exponential form of this type of parameterization means that the fits are valid in a relatively limited range of temperatures. The MP2005 formulation on the other hand is valid for a much broader range of temperatures.

I was contemplating whether to add an optional input parameter to saturation_vapor_pressure for the water phase with a default to ice, but then thought that it might be better to add a separate method for ice, thus making code and documentation easier to follow.
Also, I would recommend using the "equilibrium vapor pressure" term rather than "saturation vapor pressure" as that is a more physically accurate term, but that would likely result in plenty of mess after changing the terminology and method terms.

Checklist

• [ V] Fully documented

[CLAassistant]

All committers have signed the CLA.