Wind's Sources and Drains (see live map!!) I was pointed out by a friend to this website that shows live map of wind in US. It sometimes show interesting places where all the wind seems to converge and vanish. 
What's the origin of such "wind drains"? Do pure geological phenomena lead to simple cancellation of wind vectors or more complex effects? Is that (can that be) somehow explored in terms of wind energy?
 A: Air is more or less incompressible for the purposes of this discussion, and there is no way that the average wind speed can converge onto one spot, or else air would pile up there. But this suggests the resolution, since the piled up air will simply move the other way at the top of the atmosphere. There is a third dimension suppressed, and the opposing divergence at higher altitude makes the total divergence zero.
I cannot think of another reasonable reason why this structure would persist. But this would mean that at some places near the convergence line in the central USA, you will observe winds flowing in opposite directions at the bottom and top of the atmosphere.
A: That map is showing surface winds (typically measured 10 metres above the ground).
It is quite possible for the surface wind that's been graphed here to be two-dimensional: without poking around in the original metadata, and in the algorithm used to graph it, it's rather hard to tell. When many winds seem to converge at a point, and there looks to be a wide-area rotation going on, then either surface winds seems to converge and vanish at a point, because the energy in surface winds is getting diverted from lateral to vertical movement; or surface winds appear to emerge from a point, because the reverse is happening.

More broadly, when winds vanish, any combination of several things may be happening:


*

*when fronts collide, they can push the moving air upwards, thus increasing the kinetic energy of the air higher up, but making the wind at the surface still

*all winds blow themselves out: surface friction turns kinetic wind energy into low-grade heat. The higher the roughness of the surface, the faster this happens.

*particular configurations of landscape can cause wind to change direction and height, meaning it might vanish from the surface-wind charts

*it's not unusual for the source data to come in the form of zonal and meridional components ($u$ and $v$ vectors), averaged over an hour. To take an extreme example of how that can mislead: if you have a wind that's 10 m/s south for 30 minutes, that then changes direction to be 10 m/s north for 30 minutes, then the average wind in that hour would be shown as zero, in $u$-$v$ wind vectors.

A: Wind is, in general, powered by convection.  The ground heats up because the sun shines on it.  The top of the atmosphere is generally cooler because heat escapes from it into space in the form of infra-red radiation. (The ground can't radiate into space in this way because of greenhouse gases in the atmosphere, which provide it with a kind of thermal blanket.)  Since hot air is more buoyant than cold air, it tends to want to rise, and this results in the formation of convection cells, where hot air rises in some places and cool air comes down in other places.  
As the other answers have pointed out, the site you linked to shows only the wind speeds at the surface.  There are plumes of hot air rising to the top of the atmosphere, and what you see on the map is the air being sucked into the bottom of the plume.  If you could see the whole thing in 3D you would see that air rising and then spreading out again at the top of the atmosphere, and then eventually coming back down again in the high-pressure regions where the surface-level winds are diverging.
Because the Earth rotates, there is also a Coriolis force which makes the air tend to rotate as it gets drawn inward toward the plume. You can see this rotation on the map as well.  When the same thing happens with a much greater intensity over the ocean, we call it a hurricane.
