It guess I thought it would be obvious, but this is intended to be (a) a simple model and (b) a bit of social-media based education (based on the first day of a quarter-long undergraduate seismology class). Of course there is oversimplification.

This one is on a motor, so stress is continuous and smooth. Unfortunately it means we can't easily stop the system after a quake, which is what I usually do to give students time to measure the slip. Hence the video--allows them to measure after the fact.

Oh, there's a lot that I skimmed over here, so yes to the need for more detail. Appreciate this, though!

Great point and important correction. Thank you!

Of course! If you need a higher resolution video let me know.

Anyhow, I just love my little multi-brick model, and I love how similar it is to these finite fault models of REAL QUAKES. The brick system was built for me by one of my students years ago and I wanted to share, particularly with those of you who teach about earthquakes.

We don't know how much each spring is stretched. We don't know how much friction is holding things in place. These are not measurable parameters. We just know that small quakes are common and large quakes are rare.

Ultimately, we have to just wait to see what the Earth does.

But slowly, over time, those springs will stretch again. Sometimes--actually most of the time--we'll get small slip events. Small earthquakes. But they can also fail in a big way.

You can imagine little spring between those rectangles. Some are really stretched out after that quake, while others were stretched before and have now released a lot of that elastic energy.

Note that the size of those "bricks" is a function of the model. Some models break the fault into much, much smaller blocks. Ultimately it's a continuous slab of rock, not discrete blocks. But the more chunks you use, the more complex and computationally intensive the model.

Each chunk of the model (the rectangles on the fault plane) are like bricks in the video. Some, shown in red, moved a LOT (more than 15 m, or 50 ft!). Other "bricks" moved a much smaller amount--only a few ft. The whole thing took over two minutes to finish slipping.

This figure shows such a model for the 2010 M8.8 Maule (Chile) earthquake (fig. from USGS). The gridded area is the portion of the Nazca plate that moved in that event. You can see that it is broken up into chunks for the model.

Here's the cool part. We see something very much like this in what are called "finite fault models".

In short, we can't predict when the earthquake will happen. And we can't predict how big it will be.

You might notice that the small quakes are pretty common. And they're not always the bricks closest to the front--sometimes bricks near the back slip. It's hard, in fact it's impossible, to predict which bricks will fail, when they will fail, and whether they'll push others to slip too.

You can see that the springs stretch to different degrees: some part of the plate store more elastic energy (mostly near the source of stress). Sometimes only a single brick slips: that's a small quake. Other times, the whole thing moves. That's a much bigger event.

Back to my video. The plate is represented by all of those bricks. Elastic energy can be stored between them, so sometimes part of the plate moves, and sometimes lots of it does (in practice, you never get an entire plate moving, but my model is too small to show that).