1) You can't cover the whole planet (ecological disaster). You can't even cover large swaths of it for the same reason.
2) Most of the solar energy hits the oceans. That energy is crucial to keeping the Earth a relatively stable temperature. Interrupting it is very unwise.
3) Human population and energy needs are expected to continue to rise. Humans need three things: space, food, and energy. Solar energy sacrifices the first two in the name of the third. This is a lot like ethanol to me. Corn for energy or corn for food? Most sensible people should choose food for numerous reasons; solar radiation is a huge component of food production. There's only specific locations where solar energy makes sense. Anywhere solar energy competes with food, we need to choose food. You can't build them in forests. Mountain tops usually have huge construction/maintenance impediments on top of lack of demand. I could go on. I'm not going to do the math but what is absolutely certain is that there is a point, even if it were completely free, the cons will exceed the pros. It may not be at 2030, it may not be at 2100, but it is there and it is sooner than anyone would like it to be.
30% is the global average. In other words, that's average.
The only solid albedo number I saw so far is from that Oregon link which said 30%. Those panels should be fine in a desert but the desert surface should be checked for albedo first. If it has particularly high albedo (e.g. your link's 45% number), it really shouldn't be built there. If you leave the albedo of 45% alone and build where it is 15% increasing it to 30% through the reflectivity of the panels, it effectively cools the planet a tiny bit which is a win in my book. This is especially true in deserts where there's little to no humidity.
See thread:
http://www.techpowerup.com/forums/threads/solar-roadways.202415/
It fails the basic math test in virtually every department (monetary cost, electrical cost, electrical generation, and so on).