If you want to use more high resolution data over larger areas you'll have to learn quite a lot about "view dependent continuous level of detail surfaces", streaming of terrain data and out of core visualization.
For an example of how this may be have a look at: http://globe.sintef.no/
You'll find a few useful references in the documentation section http://globe.sintef.no/documentation/datasets.html
. For another useful approach have a look at the "grand canyon demo" on the jogl pages.
I have also enyoyed the references found in the Vterrain pages: http://www.vterrain.org/