Petra, GPS data in (southern) California have only been recorded since about 1986 (at least over wide areas), and continuous GPS has only been recorded at a lot of stations for the last few years. Here is the logic and science behind using GPS data to predict areas with a high hazard from earthquakes. Most earthquakes occur on locked faults...faults which are not slipping. But, the blocks of crust on either side of the fault are moving continuously. GPS measure this motion...but only get the full motion at a distance, perhaps 20 km, from the fault (and even at 20 km, is not quite the full motion). The crust is flexing, like if you bend a stick with a crack in it. Eventually, the stick breaks along the weakness...that is the earthquake. Between earthquakes, if you have 2 GPS stations 100 m either side of a fault, you will see very little motion unless the fault creeps at or near the surface. During an earthquake, you will see the motion on the fault...but the motion 20 km on eikther side will be small, because it has been moving all along. So, to the questions...if the blocks are moving at constant rates, then motions measured over 100,000 years or 10,000 years, using geology, might be the same as motions measured over 15 years. But, it has been found that the GPS data are "contaminated" by the post-seismic effects of large earthquakes over the last 50 or 100 years, and by ground water withdrawal, etc. And, now it is starting to look like the block motion is not constant....so that the GPS rates over 15 years are not the same as the geologic rates over 10,000 years. When I say block motion, I do not mean motion of the plates at distances of 500 km from the San Andreas...these motions have been pretty constant over the last 3 million years for the Pacific-North American plate pair. It is blocks with dimensions of tens or a couple 100 km that may change motions. As for having a thrust event...from GPS, you can see that the blocks are not just sliding by each other, but moving towards each other also...so , if the San Andreas has a pure strike-slip earthquake, then there has to still be a thrust event. Usually, the thrust event will occur on a different fault...one that is not as steep. Also, there are bends along the San Andreas fault...the blocks slide past each other on NW-SE segments, but that same motion causes convergence across WNW-ESE segments. Thus, there is a lot more thrusting expected, and seen in the geology, and seen in the earthquakes, at the latitude of the Mojave segment of the San Andreas fault....Chris