There are a few things that bother me about the 'elevator' concept:
1) air friction. Since air is not still, and has an overall nonzero motion over any air column, you have this enormous atmospheric drag on the tether, integrated all the way from ionosphere down to the surface. I don't know if anybody actually tried to calculate the drag force resulting from air currents -- is it not going to be a huge show-stopper, requiring constant boosts to the geostationary 'terminal', and making the surface-end swing around wildly with changing weather patterns (or if it's attached to the surface, wouldn't the tether end up vibrating like a string)?
2) precipitation. Starting a couple of miles above ground, there are likely to be real problems with ice build-up.
3) expensive maintenance. Once the tether begins to wear out, there's no way to repair it; you must replace the entire length of it. And then you face the problem of actually taking the tether down without having it crash on the surrounding structures (after all, the orbital end of it is moving a heck of a lot faster than the Earth's surface!) Of course, you also must have mechanisms that would constantly monitor the tether's health over its entire length, and remove depositing minerals from its surface.
4) Simple Newtonian action-reaction. You start with a slow-moving object on the surface. But, as it rises up the tether, its orbital speed increases. This creates a backward tug on the tether (not to mention a downward tug by the object moving up). Would this not mean that the orbital end of the tether would need constant boosting? If so, just how much less energy would you need to get into orbit? Would the energy savings be worth the expense of the tether?
If you know of a web-site that answers my questions, I'd appreciate a pointer...
------------------
I am; therefore I think.