Magnetometer Data Processing in Electromagnetic Surveys

TIPS :Accurate magnetometer data processing is the cornerstone of successful electromagnetic survey interpretation. This guide details the critical steps of data reduction, including diurnal correction and IGRF removal, to isolate geological anomalies. We explore advanced techniques for cultural noise removal and micro-levelling, ensuring your magnetic data reflects true subsurface geology rather than external interference.

In the realm of geophysical exploration, acquiring raw magnetic data is only the first step. The true value lies in magnetometer data processing, a rigorous workflow designed to strip away external noise and regional trends to reveal subtle subsurface anomalies. Whether you are conducting an electromagnetic survey for mineral exploration or mapping unexploded ordnance (UXO), the quality of your final interpretation depends entirely on the precision of your data reduction.

This guide outlines the standard operating procedures for processing Total Magnetic Intensity (TMI) data, adhering to international geophysical standards.

Ⅰ. The Physics of Magnetic Data Reduction

Magnetic sensors measure the scalar magnitude of the Earth’s magnetic field vector at a specific location and time. However, the value recorded by the instrument ( $T_{obs}$Tobs​ ) is a composite of several components. To interpret geological features, we must isolate the anomaly caused by local variations in crustal magnetization ( $T_{anomaly}$Tanomaly​ ).

The fundamental equation governing magnetic data reduction is:

$$T_{obs}=T_{IGRF}+T_{ext}+T_{noise}+T_{anomaly}$$Tobs​=TIGRF​+Text​+Tnoise​+Tanomaly​

Where:

• $T_{IGRF}$TIGRF​ is the main geomagnetic field generated by the Earth’s core.
• $T_{ext}$Text​ represents external temporal variations (diurnal and storm effects).
• $T_{noise}$Tnoise​ includes cultural interference and instrument drift.
• $T_{anomaly}$Tanomaly​ is the target signal.

Goal: The objective of processing is to solve for $T_{anomaly}$Tanomaly​ by systematically removing the other components.

Ⅱ. Pre-processing and Quality Control (QC)

Before applying mathematical corrections, raw data must undergo strict Quality Control. This phase determines the “usability” of the survey data.

• Data Merging: Synchronizing the magnetometer readings with high-precision GPS time stamps. Time synchronization errors greater than 0.1 seconds can lead to significant spatial positioning errors, especially in dynamic (moving) surveys.
• Spike Removal: Identifying and removing impulsive noise caused by sensor vibration or sudden electromagnetic interference. We typically use a median filter where any data point deviating more than 3 standard deviations ( $\sigma$σ ) from the local mean is flagged.
• Sensor Heading Correction: For optically pumped magnetometers (e.g., Cesium, Potassium), the sensor output can vary slightly depending on its orientation relative to the geomagnetic field vector. This “heading error” must be calibrated and removed.

Ⅲ. Geomagnetic Corrections

The largest component of the measured field is the Earth’s main field. We must remove this to visualize local anomalies.

1. International Geomagnetic Reference Field (IGRF) Removal

The IGRF is a mathematical model describing the Earth’s main magnetic field. It is updated every five years. To calculate the residual anomaly, we subtract the modeled IGRF value from the observed data.

$$T_{residual}=T_{obs}-T_{IGRF}(lat,lon,elev,time)$$Tresidual​=Tobs​−TIGRF​(lat,lon,elev,time)

Depending on the survey location, the IGRF gradient can range from 20 to 60 nT/km. Failure to remove this accurately will obscure regional geological trends.

2. Diurnal Variation Correction

The Earth’s magnetic field fluctuates throughout the day due to solar wind and ionospheric currents. These variations, known as diurnal variations, can range from 20 nT on a quiet day to over 100 nT during a magnetic storm.

There are two primary methods for correction:

• Base Station Correction: A base station magnetometer records the field at a fixed location at high frequency (e.g., 1 Hz). These readings are subtracted from the rover data based on time. This is the most accurate method for high-resolution surveys.
• Leveling to a Reference: If no base station is available, data is leveled to a quiet-day curve, though this is less precise.

Ⅳ. Noise Attenuation and Cultural Filtering

In modern environments, “cultural noise” is a significant contaminant. This includes power lines, fences, pipelines, and vehicles.

1. Micro-levelling

Airborne and ground surveys often suffer from “line-to-line” shifts caused by inconsistent diurnal correction or sensor drift. Micro-levelling algorithms (such as the differential median filter) adjust flight lines or traverse lines to a common datum without distorting the geological signal.

2. Decorrugation

This process removes stripe-like noise parallel to the survey lines. It involves separating the data into components parallel and perpendicular to the flight/traverse direction and filtering the perpendicular component.

3. Cultural Noise Masking

Automated algorithms can detect the sharp, high-amplitude gradients typical of fences or power lines. These data points are flagged and interpolated over, effectively “masking” the noise.

Ⅴ. Gridding and Derivative Enhancement

Once the data is cleaned, it must be interpolated onto a regular grid for visualization and derivative calculation.

1. Gridding Algorithms

We typically use Minimum Curvature gridding for potential field data as it honors the data points while producing a smooth surface.

• Cell Size: The grid cell size should generally be 1/4 to 1/5 of the traverse line spacing to satisfy the Nyquist sampling theorem.

2. Reduction to the Pole (RTP)

At mid-to-high latitudes, magnetic anomalies are asymmetric due to induced magnetization. RTP is a mathematical transformation that shifts the anomaly to be directly over its source body, assuming vertical magnetization.

Note: RTP is unstable at low latitudes (near the magnetic equator). In these regions, Reduction to the Equator (RTE) is preferred.

3. Derivatives

Derivatives enhance shallow features and define edges.

• First Vertical Derivative (1VD): Enhances high-frequency (shallow) anomalies.
• Analytic Signal: Useful for locating the edges of magnetic bodies independent of magnetization direction.

Ⅵ. Comparison of Processing Techniques

The following table summarizes when to apply specific processing techniques based on survey objectives.

Processing Step	Primary Function	Best Application Scenario
Base Station Correction	Removes time-varying external fields	High-precision mineral exploration, Archaeology
Micro-levelling	Removes line-to-line shifts	Airborne surveys, Large-scale ground grids
Upward Continuation	Suppresses shallow noise, emphasizes deep sources	Regional geological mapping, Basement depth estimation
Analytic Signal	Edge detection (amplitude only)	Low latitude surveys, Complex remanent magnetization

Ⅶ. Engineering & Instrumentation Considerations

As an engineer, selecting the right processing workflow depends on the hardware used.

Proton Precession vs. Overhauser vs. Optically Pumped

• Proton Precession: Lower sample rate, lower sensitivity. Requires less aggressive drift correction but suffers from lower resolution.
• Overhauser: Excellent for base stations due to low drift and low power consumption.
• Optically Pumped (Cs/K): High sample rates (up to 1000 Hz). Essential for UAV/Drone surveys. Processing must account for high-frequency vibration noise.

GPS Integration
For gradiometer configurations (using two sensors), the baseline distance must be rigid. Processing involves calculating the gradient:

$$\mathrm{\nabla }T=\frac{T_{sensor2}-T_{sensor1}}{d}$$∇T=dTsensor2​−Tsensor1​​

Where $d$d is the separation distance. This gradient data naturally suppresses regional fields ( $T_{IGRF}$TIGRF​ ) and distant cultural noise, simplifying the processing workflow.

Ⅷ. Conclusion

Effective magnetometer data processing transforms raw numbers into a geological map. By rigorously applying diurnal corrections, removing the IGRF, and filtering cultural noise, we ensure that the anomalies we interpret are real.

For engineering firms and exploration companies, investing in advanced processing is not just a technical requirement; it is a risk mitigation strategy. It prevents costly drilling mistakes and ensures that the subtle signatures of economic targets are not lost in the noise.

Reference Sources

Title	Core Content	URL
NOAA National Centers for Environmental Information	Official source for IGRF models and geomagnetic calculators.	https://www.ncei.noaa.gov/
SEG – Society of Exploration Geophysicists	Standards and publications regarding magnetic data interpretation.	https://www.seg.org/
USGS Geomagnetism Program	Real-time monitoring of geomagnetic storms and diurnal variations.	https://www.usgs.gov/programs/geomagnetism
IEEE Xplore – Geoscience and Remote Sensing	Technical papers on advanced filtering algorithms for magnetic data.	https://ieeexplore.ieee.org/
BGS – British Geological Survey	Comprehensive guides on magnetic survey techniques and corrections.	https://www.bgs.ac.uk/

---

FAQ