Seismic wave propagation in a one-dimensional model – part 1
Marcin Polkowski
Translated from Polish. Originally published on morony.pl on 1 Jan 2010.
Part 1 of 2 in the series Seismic waves in a 1D model
- Seismic wave propagation in a one-dimensional model – part 1
- Seismic wave propagation in a one-dimensional model – part 2
Some people are surprised that you can study geophysics at the Faculty of Physics of the University of Warsaw, since it's more geo than physics. It turns out there are two kinds of geophysics: geophysics and geophysics. I decided to turn my attention to the latter. Of all the topics geophysics covers, the one that interests me most is seismology, so that's what I want to pursue.
I'm a firm believer that you can't learn anything from theory alone, so I decided to back up my education with some hands-on work. Unfortunately, doing an actual seismology experiment takes a fair amount of money, so my first steps went into computer modeling. Apparently there's still a lot to be done in this field, so I have at least a few good reasons to give it a go.
Since I've been a fan of the Matlab environment for about a year now, I used it to build my model.
A short theoretical introduction for (almost) non-physicists
The model shows the directions in which seismic waves travel inside the Earth; it doesn't deal with things like wave frequencies.
Seismic waves don't travel in straight lines, because wave velocities inside the Earth aren't constant – in a one-dimensional model they depend only on depth. A wave travels between two points along the path that takes the least time (for those interested: Snell's law, Fermat's principle).
Where wave velocities change smoothly with depth, things are fairly simple – the rays (i.e. the directions of the waves) just bend.
Where there is a sharp boundary (a step change in velocity), wave conversion also takes place. I should mention here that there are different kinds of seismic waves. Inside the Earth we deal with longitudinal waves (labeled P) and transverse waves (labeled S) – for those interested: Seismic wave. P and S waves travel at different velocities. When a wave (P or S) hits a boundary, four waves are created at that point: a reflected P, a reflected S, a refracted P and a refracted S. For various reasons (total internal reflection, zero S-wave velocity) fewer than four waves sometimes come out of the conversion. On top of that, the energy and amplitude of each converted wave are different and depend on the properties of the boundary.
The model
In the first version of the model, which I've just finished, I consider a cross-section through the center of the Earth, i.e. I study ray paths in two dimensions.
I assumed a one-dimensional structure of the Earth (more precisely, of the velocity distribution – meaning that P and S wave velocities and density depend only on depth), following the Preliminary Reference Earth Model (PREM) (Dziewonski & Anderson, 1981).
I assumed 6 boundaries at which wave conversion occurs (6368 km, 6346.6 km, 6151 km, 5701 km, 3480 km, 1221.5 km). For each boundary I calculated the conversion coefficients (i.e. the amplitudes of the resulting waves relative to the amplitude of the incident wave, as a function of the boundary's parameters and the angle of incidence).
One of the results
A model like this can produce a whole lot of interesting results. Here I want to show one of the first ones (the computations take a very long time) – an animation.
By running the model described above – generating 360 P-wave rays at a source 241 km deep and numerically simulating how they evolve (their paths, reflections and refractions) – I managed to create a really nice animation showing how waves travel through the Earth's interior. If the video below doesn't work, you can watch it here. Since the legend didn't turn out great, let me explain: P waves are shown in green and S waves in red.
In my humble opinion it looks really good :).
That's it for the introduction to the problem – when I get more interesting results I'll definitely share them. I also plan to share some observations about Matlab, because I've learned a lot of interesting things over the past couple of weeks.
Next in Seismic waves in a 1D model: Seismic wave propagation in a one-dimensional model – part 2 →