Add geo_strf_steric_height function

[DocOtak]

First crack at #205.

What should the tests look like? I tried throwing in the values in the documentation and got reasonably close answers, but it looks like the interpolation method really matters.

The geo_strf_dyn_height test values were calculated in write_geo_npyfiles.py

[efiring]

For consistency with the current version, I suggest adding a "test_steric_height_mrst" that is a slight modification of test_dyn_height_mrst in gsw/tests/test_geostrophy.py.

In addition, I think the docstring needs to include a condensed version of what is in the Matlab help text for this function, or a clear and usable pointer to that information. As far as I can see, the only value of the function is its docstring--without that, there is really no point in making a whole function that just returns a constant times the output of another function. And in fact, its output, one particular (and arbitrary, as far as I can see) definition of "steric height" based on a "global average" parameter, is itself pointless, in my view.

[efiring]

Good condensation. Slight tweaks:
"pressure divided by the constant value..." -> "pressure, divided by the global mean surface value..."
and append "(see page 46 of Griffies, 2004)" to the end of the sentence, or maybe the end of the docstring. If numpydoc has a
"References" section, add the Matlab doc ref for Griffies to it; otherwise, stuff into the parenthetical addition above, or tack it on the end of the docstring without its own section.