All of the modern nuclear plant designs can be described as 'Earthquake-Lite'. That is, they are designed to be situated on firm ground in the eastern US. For this, they have set a nice clean (and somewhat arbitrary) value of 30%g in US-speak. This has been scaled to the US standard response spectrum, and can be generally thought of as 50 cm/sec, wherever people are more intelligent about these things.
This was set at about the probability of 10-4 per year for standard sites in the east, but by coincidence is the minimum level for damage to decent structures in an earthquake. In fact, I was told that the 'Earthquake Hunters' don't even leave their offices for earthquakes less than 30%g. These are the people who write reports on industrial facility damage for earthquake experience. In Taiwan, which has the best seismic monitoring in the world, they have decided that 50 cm/s PGV is the threshold level for damage to housing stock.
That means these paper-plants are not suitable for 'bad' sites. They cannot go on the hanging walls of major thrust faults such as Hamilton or Bruce. Especially not California, Turkey, Japan, or the fore-range of compressive mountain ranges, such as China or Alberta. Those places are the homes of 'super-quakes' which probably throw things at 200 cm/s!
Of all the designs, the AP is probably the one that can best be 'beefed up' to a higher level.
A site evaluation for the AP is, therefore, an exercise in determining that the seismic ground motions cannot exceed 50 cm/s for a suitably low probability, such as 10-4 or 10-5 per year. Unfortunately, this is a bit more difficult than most people think, since this is out of the range of general human experience. The only way to get to these low levels of probability is to look at the entire world and trade space for time. In other words, if you can get a probability for a very large area, such as the world, then you can go to much lower odds if you specifically determine a small area. For example, with a deck of cards, the odds are one in four that you'll pull up a spade, but the odds go down another order of magnitude for pulling up an ace of spades.
Unfortunately, Canada is not really set up for this line of thinking. As with old generals, we are always ready to fight the last war. What we used to do is find a good number for peak acceleration, and design a custom plant around it. We did that for Darlington, with the very low number of 8%g, even though subsequent internal doubling made that much higher.