I probably shouldn't have said "substantial" since what I really meant was "not cancelled out in the way mwengler describes".

I don't think I can actually do the calculation without an estimate of the typical gradient of coastal land in the US (i.e., the conversion factor from sea level rise to shrinkage) but let's make a crude guess and see what happens. So, North America has an area of about 25M km^2 so our square is about 5000km on a side. Expected sea level rise by 2100 is about 0.5m (I've seen wildly inconsistent figures for this, though). Let's suppose that sea-level land has a typical gradient of 1 in 50, so that a 0.5m rise means a 25m shrinkage in the usable land. Then the total amount of land lost would be about 20000km x 25m = 20km x 25km = 500 km^2, roughly comparable to the area of San Francisco.

This is probably an underestimate: North America is wigglier than our square model (so more coast relative to its area) and I suspect that actually coastal land is flatter than 1 in 50.

So it's a small fraction of the total area (as of course was obvious from the outset) but personally I'd consider it a substantial loss if an area the size of San Francisco fell into the sea.