Adjustments for baseline shifts in far-fault strong-motion data: An alternative scheme to high-pass filtering

Traditional processing methods of accelerometric strong-motion records rely on band-pass filtering to remove contaminating noise. While filtering of low-frequency noise is often desirable to reduce the distortion of displacement and velocity waveforms, application of crude filtering can lead to the loss of long-period components of earthquake ground motion. This paper focuses on a baseline adjustment algorithm, which reduces noise-induced distortion of ground-motion accelerometric signals, and unlike high-pass filtering, retains potentially useful information at long periods. A brief description of the state of the art techniques in strong-motion data processing is presented, followed by a detailed formulation of the proposed baseline adjustment algorithm and some examples of its applications to recently recorded accelerograms. This method is found effective for high-quality far-fault data where the amplitude of the recorded signal is of structural engineering significance.