9 hours ago, Janov said:

That is one of the last unsolved problems in aviation: Where does the wind come from?

There is literally no way to find out, people have suggested putting up red/white "tubes" of fabric to show the direction, use instruments to read it (like they used to have on church steeples), or even use some sort of method to record these readings and then transmit them with the use of invisible waves...

 

There are a basic exercises like rectangular pattern and S-turns. They pretty much give good idea where wind is coming from after one or two iterations.

Also old school navigators used to visually  observe drift over the ground and based on that guesstimated wind direction. In fact some old school airplanes used to have windows specificity design for navigator to observe ground and stars. 

When I was flying in the RW I always got the winds report from ATIS and then the tower.  Since I spent a lot of time flying out of and into uncontrolled fields, I would hit Unicom and ask for "winds & active".  Finally, I would overfly the field (if no one responded) and look at the windsock.

For simming, I check the latest metar report for the field (I have several ways to access this) and play the ATIS report.  I have found that the winds reported are actually pretty accurate and plan my approach (and crosswind compensation) accordingly.  What is NOT accurate in the sim are the ATC runway assignments - it seems they are wrong about 90% of the time, even when they are reporting winds properly...

When in the air (and not using AP), I simply watch my flight line - whether flying visually, following a VOR radial, or following a GPS indicator if the needle or line moves, then I obviously have a crosswind in play - I compensate until the aircraft maintains the proper path.