buttman writes:

Theres a little "eye" that looks up at the sky at various points. It
detects if there is cloud, or sky. The sensor is prone to errors,
though. Sometimes if it's broken, it'll say its overcast because some
of the sensors are pointing diagonally and are sensing the side of the
cloud. I had an ATC college-style text book that had a very
informative diagram of how they worked, but that book is long gone.

http://www.cyanogen.com/products/cloud_main.htm

here is an example of one.

Cool--I wouldn't mind having one (useful for photography as well as astronomy
and aviation). However, it's still a ground-based sensor, so it would only
see the first level of clouds. Conversely, a satellite would only see the top
layer of clouds. If there are three or four layers of clouds, how do weather
services discover them?

The same is true for temperature, humidity, and pressure. Pressure you can
probably infer from surface pressure, and temperature you can guess at in a
similar way. Humidity is more vague. I'm curious as to how all of these get
measured aloft.

And what about winds? Weather services seem to have awareness of winds aloft,
but where are they getting the measurements? Winds aloft may have no
correlation with surface winds and can change a lot over short distances. You
could set up probes, but that's a lot of probes to launch and recover even to
cover small areas. You could rely on PIREPs, but that seems kind of hit and
miss. So how is it really done?

How are the paths and speeds of jet streams determined?