Seismic Noise — Types and Mitigation
Seismic Noise — Types and Mitigation
Introduction
Seismic data is never recorded in perfect conditions. Alongside the reflections we want, the sensors also capture unwanted energy known as seismic noise. Noise can obscure geological features, distort amplitudes, and reduce the reliability of interpretation. Understanding seismic noise — and how to mitigate it — is essential for producing high‑quality seismic images.
This article explains the major types of seismic noise, where they come from, and how modern processing workflows suppress them.
1. What Is Seismic Noise?
Seismic noise is any unwanted energy recorded in seismic data that does not correspond to true subsurface reflections. Noise can originate from:
The environment
Acquisition equipment
Human activity
Geological complexity
Processing artifacts
Noise is broadly classified into coherent and incoherent types.
2. Types of Seismic Noise
A. Coherent Noise
Coherent noise has a recognizable pattern and can mimic real reflections.
Examples:
Ground roll
Multiples
Guided waves
Swell noise
Airwave energy
Why it matters: Coherent noise is dangerous because it can be mistaken for geology.
B. Incoherent Noise
Random, uncorrelated energy that reduces signal clarity.
Examples:
Wind noise
Instrument noise
Random environmental noise
Why it matters: Incoherent noise reduces signal‑to‑noise ratio and masks subtle features.
C. Acquisition‑Related Noise
Noise introduced by the acquisition system.
Examples:
Cable strum
Crossfeed
Sensor coupling issues
Why it matters: Poor acquisition can permanently degrade data quality.
D. Processing‑Related Noise
Artifacts introduced during processing.
Examples:
Migration smiles
Aliasing
Over‑filtering artifacts
Why it matters: Processing noise can mislead interpreters if not properly controlled.
3. Noise Mitigation Techniques
A. Filtering
Bandpass filters
FK filtering
Tau‑P filtering
Goal: Remove unwanted frequencies or dips.
B. Ground Roll Suppression
FK fan filters
Radon filtering
Adaptive subtraction
Ground roll is one of the most common noise types in land data.
C. Multiple Attenuation
SRME
Radon demultiple
Wave‑equation demultiple
Multiples are coherent and often require advanced techniques.
D. De‑Noising Algorithms
Median filters
FX de‑noising
Curvelet transforms
Machine‑learning de‑noising
Modern AI‑based de‑noising is becoming increasingly effective.
E. Acquisition Improvements
Better sensor coupling
Optimized source‑receiver geometry
Noise‑resistant equipment
Good acquisition reduces the need for aggressive processing.
4. Challenges in Noise Mitigation
Avoiding signal loss
Preserving amplitude fidelity
Handling complex geology
Balancing noise suppression with resolution
Preventing over‑processing
Conclusion
Seismic noise is unavoidable, but with the right techniques, it can be effectively managed. Understanding noise types and mitigation strategies is essential for producing clean, interpretable seismic data that supports accurate geological decisions.