[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).