Support for point observations (wind profilers, rawinsondes as point obs)

[rcjackson]

I eventually wish to see a way of constraining against point observations. This may have to be done with a separate cost function that weighs in how far certain points deviate from the point observations. This would be useful for integrating in data points from wind profilers, surface based observations (METARs, etc.), uninterpolated radar gates, and from rawinsonde data whose horizontal displacement varies throughout the launch.

[rcjackson]

#53 is some initial work on this. There is now support for constraining observations from ASOS data downloaded from the Iowa Mesonet Archive.