The payback on solar is long term. Unlike thermal, where it is relatively cheap to create the plant and the primary cost of generating electricity is the fuel, there is a very high installation cost per kwh for solar. It is the upfront cost that still makes solar relatively expensive compared to most traditional power generation.
Thus, if the solar installations get destroyed at a high rate, solar costs per kwh will be very high indeed in Japan.
I think, but I cannot prove or cite anything to prove my theory, that roof-mounted lighter panels could be framed up to limit the damage because they are lighter and could be mounted on frames pinned to structural support with shake compensators to prevent the panels from being destroyed by anything but very severe local quakes. A solar collector is a much heavier piece of equipment that would be harder to secure. However in any area exposed to tsunami or severe storms, solar panels might not do well. Siting is always an issue in Japan.
The relatively high rate of destruction of all Japanese buildings and infrastructure is one reason why costs of power generation in their country are so high. It's a problem with any type of power generation, but it would have relatively far greater impact for types of power that require very high installation costs. In the past, Japanese companies and individuals have coped by building relatively cheap buildings in high-destruction areas, and by careful siting of larger, more expensive buildings on rock beds, plus shock absorbing systems.
I think designing and installing quake-resistant roof-mounting systems for solar panels could be done without too much additional cost. They would still be very vulnerable to storm damage anywhere near the coast, and frankly, most of Japan's population is on the coast.
You only have to look at their roofs to see that they are designed to be very resistant to fire and storm damage. They do have a governmentally supported system of quake insurance, but it is costly and it only pays 50% of the damage.
Still, the problem for Japan is that solar power couldn't keep the country running, so if they decide to deep-six nuclear plants, they would have to build a lot more thermal plants to compensate. This would have very bad effects on their balance of trade, because they have to import most their fuel.
Tidal turbines, which I realize are not an established technology, might be far more suitable for Japan. They already have a ton of hydro plants and most of those plants are pumped hydro.
http://www.fepc.or.jp/english/energy_electricity/location/hydroelectric/index.htmlFEPC page on power plants:
http://www.fepc.or.jp/english/energy_electricity/location/index.htmlSolar power isn't new in Japan - I know in the mid 2000s they were second in the world in solar power generation - but it is still a very small percent of total generation. Japan is also a very urbanized society, and a high manufacturing society. Its government also has the highest level of public debt/GDP in the world - over 200% at this point. They are in such a severe crisis now that it is hard to see the way out.
Map of principal power plants in Japan
Example of how they handle loading factors: