Physical instrumental vetoes for gravitational-wave burst triggers

We present a robust strategy to veto certain classes of instrumental glitches that appear at the output of interferometric gravitational-wave detectors. This veto method is "physical" in the sense that, in order to veto a burst trigger, we make use of our knowledge of the coupling of different detector subsystems to the main detector output. The main idea behind this method is that the noise in an instrumental channel X can be transferred to the detector output (channel H) using the transfer function from X to H, provided the noise coupling is linear and the transfer function is unique. If a nonstationarity in channel H is causally related to one in channel X, the two have to be consistent with the transfer function. We formulate two methods for testing the consistency between the burst triggers in channel X and channel H. One method makes use of the null stream constructed from channel H and the transferred channel X, and the second involves cross correlating the two. We demonstrate the efficiency of the veto by "injecting" instrumental glitches in the hardware of the GEO 600 detector. The veto safety is demonstrated by performing gravitational-wave-like hardware injections. We also show an example application of this method using 5 days of data from the fifth science run of GEO 600. The method is found to have very high veto efficiency with a very low accidental veto rate.