Implement exact solution to two-box energy balance model

[hdrake]

Currently, we implement an approximate solution to the two-box energy balance model, which is really more of a 1.5 box solution which assumes the upper-box equilibrates instantly with the forcing. I did this because I did not know how to solve the full two-box equations but thankfully Geoffroy et al. 2013 did this for us and the solution is just a sum of exponential green's functions.

[hdrake]

I've implemented this in branch https://github.com/ClimateMARGO/ClimateMARGO.jl/tree/exact-two-layer-solution, but ran into problems getting the optimization algorithm to converge.

[hdrake]

@danielkoll may have had a breakthrough on this