7.3 Seismic noise and signal processing techniques

Seismic noise, from natural and human sources, can muddy seismic recordings. Understanding its origins and impacts is crucial for accurate data interpretation. Signal processing techniques help clean up this noise, improving the quality of seismic data.

Filtering methods, like bandpass and notch filters, remove unwanted frequencies. Advanced techniques, such as Fourier transforms and deconvolution, further enhance signal quality. These tools are essential for extracting meaningful information from seismic recordings.

Seismic Noise

Sources and Characteristics of Seismic Noise

Top images from around the web for Sources and Characteristics of Seismic Noise

• 

  SE - Improving the quality of empirical Green's functions, obtained by cross-correlation of high ... View original

  Is this image relevant?

• 

  SE - Introducing noisi: a Python tool for ambient noise cross-correlation modeling and noise ... View original

  Is this image relevant?

• 

  SE - Introducing noisi: a Python tool for ambient noise cross-correlation modeling and noise ... View original

  Is this image relevant?

• 

  SE - Improving the quality of empirical Green's functions, obtained by cross-correlation of high ... View original

  Is this image relevant?

• 

  SE - Introducing noisi: a Python tool for ambient noise cross-correlation modeling and noise ... View original

  Is this image relevant?

1 of 3

Top images from around the web for Sources and Characteristics of Seismic Noise

• 

  SE - Improving the quality of empirical Green's functions, obtained by cross-correlation of high ... View original

  Is this image relevant?

• 

  SE - Introducing noisi: a Python tool for ambient noise cross-correlation modeling and noise ... View original

  Is this image relevant?

• 

  SE - Introducing noisi: a Python tool for ambient noise cross-correlation modeling and noise ... View original

  Is this image relevant?

• 

  SE - Improving the quality of empirical Green's functions, obtained by cross-correlation of high ... View original

  Is this image relevant?

• 

  SE - Introducing noisi: a Python tool for ambient noise cross-correlation modeling and noise ... View original

  Is this image relevant?

1 of 3

• Ambient noise originates from natural and anthropogenic sources continuously present in seismic recordings
• Microseisms result from ocean wave interactions with the seafloor, generating low-frequency seismic waves
  • Primary microseisms (0.05-0.1 Hz) produced by direct ocean wave pressure on the seafloor
  • Secondary microseisms (0.1-0.5 Hz) caused by wave-wave interactions creating pressure fluctuations
• Cultural noise stems from human activities (traffic, machinery, construction)
• Wind and atmospheric pressure changes contribute to low-frequency seismic noise
• Thermal noise affects high-frequency seismic recordings, especially in borehole instruments

Impacts and Measurement of Seismic Noise

• Signal-to-noise ratio (SNR) quantifies the relationship between desired seismic signals and background noise
• SNR calculation involves dividing signal amplitude by noise amplitude
• High SNR indicates clearer seismic signals, while low SNR suggests noise dominance
• Noise levels vary with location, time of day, and seasonal factors
• Long-term noise studies help characterize site-specific seismic backgrounds
• Seismic stations often include multiple instruments to differentiate between noise sources
• Noise reduction techniques include site selection, instrument design, and data processing methods

Signal Processing Techniques

Filtering Methods for Noise Reduction

• Filtering removes unwanted frequency components from seismic data to enhance signal quality
• Bandpass filter allows a specific range of frequencies to pass while attenuating others
  • Low-cut filter removes frequencies below a specified threshold
  • High-cut filter eliminates frequencies above a certain limit
• Notch filters target and remove specific narrow frequency bands (power line interference)
• Wiener filtering optimizes signal extraction based on known noise characteristics
• Adaptive filtering adjusts filter parameters in real-time to respond to changing noise conditions
• Spatial filtering utilizes multiple seismic stations to reduce noise based on wave propagation properties

Advanced Signal Processing Techniques

• Fourier transform converts time-domain seismic signals into frequency-domain representations
  • Fast Fourier Transform (FFT) efficiently computes discrete Fourier transform
  • Enables analysis of signal frequency content and design of frequency-based filters
• Deconvolution removes the effects of the recording system and Earth's response from seismic data
  • Improves temporal resolution and removes unwanted reverberations
  • Spiking deconvolution aims to compress the source wavelet
  • Predictive deconvolution removes periodic multiples from seismic records
• Stacking combines multiple seismic traces to enhance signal and suppress random noise
  • Common Midpoint (CMP) stacking improves signal-to-noise ratio in reflection seismology
  • Diversity stacking weights traces based on their signal quality before summation
  • Beam forming utilizes array geometry to enhance signals from specific directions

Frequency Analysis

Spectral Analysis Techniques

• Spectral analysis examines the frequency content of seismic signals
• Power spectral density (PSD) quantifies signal power distribution across frequencies
• Spectrogram displays how frequency content of a signal changes over time
  • Useful for identifying time-varying noise sources and seismic phases
• Coherence analysis measures correlation between different frequency components
• Multitaper method provides robust spectral estimates for non-stationary signals
• Wavelet analysis offers time-frequency decomposition with variable resolution

Frequency-Based Signal Enhancement

• Fourier transform facilitates frequency-domain analysis and filtering of seismic data
  • Enables identification of dominant frequencies in seismic signals and noise
  • Allows design of optimal filters based on spectral characteristics
• Bandpass filtering selectively preserves desired frequency ranges in seismic data
  • Butterworth filters provide smooth frequency response with minimal ripple
  • Chebyshev filters offer steeper roll-off but introduce passband ripple
  • Elliptic filters combine steep roll-off with controlled ripple in both passband and stopband
• Filtering techniques address specific noise issues in seismic recordings
  • Low-pass filtering removes high-frequency noise (cultural noise, instrument artifacts)
  • High-pass filtering eliminates long-period noise (tilt, temperature variations)
  • Band-reject filtering targets narrow-band noise sources (power line interference)
• Time-frequency filtering adapts filter parameters based on signal characteristics over time
• Empirical mode decomposition separates signals into intrinsic mode functions for targeted filtering