The Oil Drum: Europe

Wind CAN be predicted in advance, and most wind turbines have a flat output curve over a broad range of wind speeds (such as 12 to 25 m/s). (the exact shape is a matter of economic design, most generators are smaller than max capacity of the blades. Smaller generators allow for lighter towers, it is just economics).

There are many places in the world where one can predict that a wide geographic range will see wind speeds between 12 & 25 m/s for a dozen hours in advance. And if there are, say twelve wind farms in a row behind an advancing front; one can see changes in the leading farm and predict the timing and results for the eleven behind with some confidence once operating history has been developed.

BTW, large fossil fuel plants (such as 500 MW) "trip" fairly frequently, yet the grid adapts and rarely crashes. How can 1.5 MW WTs (with high rotating inertia) slow down a bit in their output and yet the grid just cannot deal with this !?

If a few handfuls of WTs go offline due to excessive winds, NO BIG DEAL ! Your thunderstorm example was "stuff & nonsense" for a widespread grid. Just a triviality.

New Zealand says that they can take 35% wind without further study or grid modification. The only requirement is a minimum of geographic diversification.