[lnx01]

Actually, I think it's measuring the intensity of light between two points. It has two perpendicular lasers of some wavelength which then interfere with one another. If they interfere perfectly you measure a zero, if the interference is off by some amount you measure a deviation in intensity from the i^ or the j^ direction.

You can eliminate passing trucks, earth tremors, mining, asteroid impacts etc simply by applying a band-pass filter that excludes measurement frequencies outside of the range predicted by the equations.

No expert though, just guessing.

[sleavey]

No filtering is applied to data containing such anthropogenic/geological noise after it is measured to remove such effects. In fact, such data is usually junked and not used for analysis. The sites have thousands of witness channels which listen for things like trucks, ground motion, magnetic storms, etc. that could possibly influence the mirrors in the way a gravitational wave would. If the same signal appears in both the gravitational wave channel and some auxiliary sensor, it's thrown away.

Instead, the mirrors are highly isolated from the ground (suspended from pendulums, motion damped by actuators, that sort of thing) so that such effects do not have significant impact on the motion of the mirrors.

In any case, seismic noise can't be fully isolated and creates a sensitivity wall below around 10Hz. To get sensitivity much below 10Hz, you have to go to space (look up LISA).

[mturmon]

Take a look at the time series from the first detection:

http://www.caltech.edu/news/gravitational-waves-detected-100...

The frequency range varies (it's kind of a chirp) as the black holes spiral in towards each other.

But the frequency at the end is in the auditory range. The instrument is literally measuring a displacement in the same range as many of the things you mention (traffic, mining, ...). A band pass filter will not do.

[iaw]

This is pretty close to the setup I recall. It blows my mind that they have a visual link that long.