Spectral Line Removal in the LIGO Data Analysis System (LDAS)

TL;DR

This paper discusses the implementation and application of spectral line removal techniques in the LIGO Data Analysis System to improve gravitational wave data analysis by reducing narrow frequency band noise.

Contribution

It introduces a linear model-based line removal method implemented in LDAS and demonstrates its effectiveness on LIGO S1 data.

Findings

Effective removal of spectral lines from LIGO data.

Reduction of noise in gravitational wave signals.

Improved sensitivity for stochastic background searches.

Abstract

High power in narrow frequency bands, spectral lines, are a feature of an interferometric gravitational wave detector's output. Some lines are coherent between interferometers, in particular, the 2 km and 4 km LIGO Hanford instruments. This is of concern to data analysis techniques, such as the stochastic background search, that use correlations between instruments to detect gravitational radiation. Several techniques of `line removal' have been proposed. Where a line is attributable to a measurable environmental disturbance, a simple linear model may be fitted to predict, and subsequently subtract away, that line. This technique has been implemented (as the command oelslr) in the LIGO Data Analysis System (LDAS). We demonstrate its application to LIGO S1 data.