March 04, 2005

Tectonics 2

Ok, take a look at this.

The histogram on the left has two distinct peaks. This is because there are two different types of rock making up the bulk of the surface of the Earth, one of which is less dense and "floats" on top of the other. It is for this reason that we have continents separated by deep oceans rather than a world entirely covered in shallow seas.

The cumulative frequency curve on the right could be used as a sanity check for plate tectonics simulations. Any simulation should produce a similar curve, if it's attempting to mimic an Earth-like world.

There are three types of fault, spreading, transform, and subducting. In a simulation we can pretty much ignore transform faults (unless we're interested in earthquakes), because they don't do much. Spreading faults are found at the borders of plates that are moving away from each other. The denser of the two types of rock is produced there, creating new ocean floor. If this were the only type of fault on Earth, we would have nothing but a world-wide shallow sea.

Subduction faults are found where two plates are moving towards each other. One plate inevitably wins the collision and forces the other to bend downwards into the mantle. This produces a deep ocean trench along one side of the fault, which explains the sharp downward spike at the right end of the cumulative frequency curve. The interesting thing is what happens on the other side of the fault. Through some bit of thermo-chemical magic I don't understand yet, the subducting plate triggers volcanic activity about 100km away from the trench. The volcanos produce the second, less dense type of rock, which forms island chains. The Indonesian islands of Java and Sumatra were formed like this, as were Japan and the Aleutian islands. If the activity goes on long enough, entire continents can be formed. South America is the best example of this, with a deep trench off the west coast, followed by a volcanic mountain chain (the Andes) just inland.

These mountains partly account for the spike on the left side of the cumulative frequency curve (they then quickly erode down to relatively flat continents). The rest of it is accounted for by continent-continent collision, where the subducting continental rock is buoyant enough that it doesn't so much subduct as get shoved underneath the other plate. The Himalayas were produced in this way.

So, from a simulation point of view, you not only have to keep track of the thicknesses of the two types of rock at every point on your world, but also the locations of spreading faults, and the locations and polarity of subduction faults.


Post a Comment

<< Home