Implementing the COARE v3.6 algorithm

[andrew-s28]

The COARE v3.6 algorithm was a bit harder to implement than the v3.5 algorithm, in part because it's a bit harder to define what has changed in this version (at least insofar as the use cases of this code is concerned, i.e., not gas transfer).

The most important changes that I can identify from the COARE v3.6 source code are:

• changes of the sea-state/wave-age parameters
• the implementation of a parameterized sea state if based on wind speed and wave age
• use of a function to calculate albedo for shortwave radiation flux rather than the constant 0.945 parameter used in v3.5
• use of variable salinity in calculated saturation specific humidity (this was supposed to be added here but was missed, so is actually found in Adds salinity dependence to the saturation specific humidity calculation #39)

However, the albedo function used in the v3.6 source code utilized a massive lookup table and required user entry for longitude and fractional date, which complicated input data (it would be great if this was consistent between versions). Instead, this implements a middle ground, by providing albedo as a function of latitude but assuming the hour is noon and the day is the equinox, so it's not an exact replica of v3.6 (unlike v3.5, which more or less is an exact replica).

Since the only changes are minor parameter changes and the new sea-state parameterization, so for the most part, running v3.6 will give you the same results as v3.5 to several decimal places for most variables. The only place it's really relevant is in these niche use cases. Therefore, I'm going to add documentation that recommends the user to use v3.5 unless they have a specific reason to prefer v3.6.