I think your idea of coupling emergency signals (those that a community can put together in a hurry), yet visible from above may be the missing link. The way I wrote the initial idea, a major drawback is that it's not something people could do in a crisis. At best, it's an early distress signal before a crisis occurs.

Good point on folk not having time to setup a large distress sign during a crisis. Perhaps the important thing is to define a global ground-to-air distress sign. So that regardless of the size of the sign anyone seeing it would immediately recognize it as a distress signal.

I like your idea on detecting unusual changes. What I was worried about was how to filter out false signals. Also, like the idea of auto detecting changes using algorithms. All these together, a unique distress pattern, algorithmic and automated processing may make this a workable concept!