Seismic data are frequently acquired as long, continuous time series, but some of the benefits of this approach, such as improved time-frequency resolution and the ability to accurately quantify the noise, are oftentimes overlooked. Moreover, continuous time series facilitates access to low frequencies of the observed wavefields. Well-defined low frequencies are the key for impedance and full waveform inversions, while characterizing the noise background allows one to develop an effective denoising strategy, which in turn reduces the uncertainty of all seismic-derived products. We demonstrate that the multiscale approach utilizing wavelet transforms is an efficient tool to quantify spatio-temporal characteristics of seismic data in different frequency bands. Our analysis uses a portion of finely-sampled continuous time data to characterize signal and noise present in near offsets of data acquired with a Vibroseis source. We find that the data generated by the baseplate coupling with the ground have better low frequency content than the sweep time data. We also identify and attenuate different types of noise, including acoustic and mechanical noise generated by the Vibroseis, harmonic distortion and powerline interference. We address the noise in the 2D complex wavelet domain, linking the bands and scales of the transform to the physical dimensions of seismic data and zeroing the coefficients corresponding to noise. The strategy presented in this paper is efficient for suppressing sourcegenerated noise, surface waves and powerline interference, as demonstrated by comparing raw and denoised correlated data.
Presentation Date: Wednesday, October 14, 2020
Session Start Time: 9:20 AM
Presentation Time: 9:20 AM
Location: Poster Station 13
Presentation Type: Poster
