The problem with the radiosonde data is that it's very inaccurate from most site, exhibiting a significan dry bias at higher altitudes, and often not measuring humidity at all at temperatures below -40F. So the data from those soundings is often pretty useless.
As for the claim "
air at altitude doesn't deviate much statewide at any given time", that is false. Humidity deviates as much as clouds deviate. A natural cloud is simply the air made visible by humidity being above 100%. So the variation in clouds indicates the variation in humidity.
When there are no clouds it does not mean there is no variation in humidity, it just means it's all under 100%. The variation is still there, like invisible clouds.
A better way of looking at it is the water vapor satellite products, which use wavelengths sensitive to water vapor, and mostly measure upper atmosphere vapor:
http://www.weather.gov/satellite#wv
You can see it varies quite a bit statewide.
But more interestingly, click on the "View loop: 24 hours" there to see just how much it varies over 24 hours, and how the variance can go in multiple directions.