For a second glance I noticed, that Africa-Megaflyover hi-res strip, what appears only by GE-desktop default view imagery, is totally offset and probably also a bit oversized!
Tried to find proper location on base imagery (timeslider on) still cannot identify.
Post about a dust-devil by member willi1 with location (attached kmz): HERE. / zoom out to get whole hi-res strip in view after download and switch TimeSlider on and off!
Brilliant oGEC thread started by Frank4 available with many more, even "unmarked strips" of this National Geographic Project, where same problem mentioned, but then let.
Checking further locations, the same offset, wrong-sized "overlays" we meet. (let these be both "marked" or "unmarked" sites)
Somewhat low-grade piece of work by Google?, NatGeo...? or by whom? How could it be Q-passed and integrated to GE-Layers with such poor implementation of such a great (and besides all, so interesting) database? Or its because there were no hi-res imagery available at all anywhere that time to fit overlay to?