The ground above the fault changed its direction of motion.

If this is like the Cascadia episodes, the reversal in the direction of ground motion indicates there is a slow earthquake on the megathrust rather than the usual state of that part of the megathrust being locked. When the slow earthquake ends, the ground will go back to its usual sense of motion of moving inland as the fault again is storing stress rather than releasing it by slipping.

Its annoying that the reporters took this slant of backwards motion - there have been many slow earthquakes now documented, although perhaps this is a larger-than-normal one.

I took the quiz and did poorly - and I think the quiz is the kind we try to avoid giving students because of ambiguous questions, not to mention "correct" answers that are wrong.

1. For example, why do they say 50% rather than 75% of American earthquakes are in Alaska? The 1964 earthquake alone probably dominates the moment and energy release of historical events in the US. And the probably next 9 largest earthquakes were also in Alaska. Current thinking has the New Madrid earthquakes as relatively minor M7.5s. If they are counting earthquakes, the reason 90% of earthquakes are not in Alaska is because no one cares to count all those millions of small quakes, such as are counted in California.

2. The numbers I've heard for the 1976 Chinese quake are 750,000 - 1M, closer to 1M.

3. The 1960 deaths were low both because the South American west coast is sparsely populated AND because of foreshocks.

4. Chile 1960 caused $1M in damage to the California coast. We're supposed to remember whether that earthquake caused $1M or $5M, when damage estimates vary by a factor of 2 or more depending whether it is insured costs or (very) rough estimates are cited? $4.5 isn't right for the 1964 Alaska earthquake, either, which had $8M in damage to Crescent City.

I did better the second time I took the test.