The short answer is that a tall hill or mountain represents a broad elevation change that's contiguous with the surrounding landscape. Difference in air quality from sea level is not a simple function of elevation, but of air flow and direction. You actually have to treat the air near the surface as a flow.
Check out this extremely kickass old film: http://youtu.be/7SkWxEUXIoM?t=29s The drag near the solid surface greatly slows down, and introduces turbulence to, an otherwise laminar flow. This means that the air near the surface of the earth is doing a lot more interacting with the earth than the other air, and is getting and keeping a lot more heat and dissolved gases. Above that turbulent layer, you just have cold dry air. It loses heat, water condenses, gases fall out.
What can you use as an indicator for the height of the wet air layer? The height of the local trees! Their tops are about at the top of the survivable boundary layer in the air. Desert plants are short, rainforest plants are tall. Interestingly, this might lead you to wonder where the tallest trees in the world are.
Probably in some place where there's an ocean wind that blows into a blind valley, right? Because that thick boundary layer would just pile up and up, right?
Here's an elevation map of California: http://www.netstate.com/states/geography/mapcom/images/ca_h....
And here's where the Giant Sequoia redwoods are: http://www.yosemite.ca.us/library/sequoias_of_yosemite/distr...
So being on top of a tall building is like being on a spike up above the livable atmosphere for trees. Unless your city is built in a place that already had giant trees, it won't have much success growing them at heights above their native height.
Unless you choose very specific trees.