OR/14/014 Part 3: Data processing
Beamish D and White J. 2014. TellusSW: airborne geophysical data and processing report . British Geological Survey Internal Report, OR/14/014. |
Processing is currently performed in CGG ATLAS.
Magnetic Data Processing
The magnetic data is first de-spiked and decimated to 20 Hz.
Removal of system Parallax and Heading
The instruments recording magnetic intensity, magnetic gradient and various altitude parameters are at physically different locations to the instrument recording the aircraft position. In addition, each instrument requires a slightly different amount of time to process its raw inputs (frequency, voltage, etc.) and compute the raw data recorded by the acquisition system. The cumulative result of these effects is that data readings are recorded at a different time position to that at which they were actually acquired. As a result, the position with which they are recorded is offset in time from the position at which they were acquired. In order to correct for this discrepancy, a time offset is applied to the data readings to restore them to their correct time and thus location. The required offsets for each aircraft and each instrument are determined in field tests as described in Tests and calibration.
Parallax corrections were applied to the data as per the tests carried out and detailed in Parallax Control. Heading corrections were applied to the data as per the tests carried out and detailed in Heading Test.
Correction for temporal magnetic ('Diurnal') variation
Diurnal base station values were corrected for secular variation by removing the International Geomagnetic Reference Field (IGRF) field value at the base station location, computed at the time of measurement. This method is more effective than subtracting a simple constant since the regional field at any location is varying with time. The corrected (residual) diurnal field values were then subtracted from the interpolated and compensated magnetic data.
Correction for regional magnetic variation (IGRF)
Regional effects of the earth’s magnetic field were removed by subtracting from each reading the value computed from the IGRF 2010 model. IGRF was computed at each survey data point using the actual time that that point was surveyed.
GPS height is also used in IGRF computation as the large variations in height above sea level encountered in the project area yield significant variation in computed IGRF values. Failure to use actual aircraft height (above the spheroid) in such cases has been seen to cause additional levelling problems in regions of complex terrain.
Since IGRF removal is a regional field subtraction, no base value is applied to the result. Consequently, corrected values are true residuals in the sense that they range from negative to positive values. The computed IGRF values subtracted are provided in the final data.
Levelling
After correction for diurnal magnetic variations, residual levelling errors remain in the data. These result from the difference between diurnal activity measured at the base station and diurnal activity at the aircraft position. Errors also result from residual aircraft heading errors and imprecision in measuring aircraft position (vertical and horizontal).
A three stage procedure is then applied to the data to produce the final levelled magnetic data.
Gridding
Levelled data were then gridded to produce 2D maps of the data. Refer to Table A5.1 in Appendix 4 for gridding parameters.
Residual magnetic intensity
The residual magnetic intensity (RMI) is calculated from the total magnetic intensity (TMI), the diurnal, and the regional magnetic field. The TMI is measured in the aircraft, the diurnal is measured from the ground station and the regional magnetic field is calculated from the International Geomagnetic Reference Field (IGRF). The first step is to remove the low frequency component of the diurnal from the TMI which is extracted from the filtered ground station data. The average of the diurnal is then added back in to obtain the resultant TMI. Next, we tie line level and micro-level the TMI data. The regional magnetic field, calculated for the specific survey location and time using the IGRF model, is removed from the resultant TMI to obtain the RMI.
Magnetic first virtical derivative
The first vertical derivative (1VD) was calculated in the frequency domain from the final grid values to enhance subtleties related to geological structures.
Reduction to the pole
The residual magnetic intensity was reduced to the pole using a 2D frequency domain operator, working from the gridded values of the levelled magnetic data. The calculation, assuming all induced magnetization, was based on the following magnetic field parameters:
- Magnetic declination: 2.5°W.
- Magnetic field inclination 65.5°N.
Analytic signal of magnetics
To improve on the definition of the magnetic structures, the analytic signal was calculated from the residual magnetic values. The analytic signal calculation is the Pythagorean sum (square root of the sum of the squares) of the three derivatives x, y and z. This was calculated in the frequency domain by using a 2D operator working from the gridded values of the levelled magnetic data.
Calculated horizontal gradient
The calculated horizontal gradient was derived from the residual magnetic intensity grid to enhance the high frequency content of the data and attenuate the low frequency background. It was normalized to nanoteslas per metre based on the cell size.
Horizontal gradient gridding
The levelled anomalous magnetic field grid are generated from measured lateral and calculated longitudinal gradients as follows:
The lateral gradient was levelled to correct for DC offsets from line to line, that are caused by hysteresis, by computing the average lateral gradient for each line.
A horizontal gradient grid was then computed from the measured lateral and computed longitudinal gradients and the forward transform computed.
The frequency domain grid for the total field was computed by use of the transfer function below:
where kx and ky were the wave numbers in the x and y directions and k2 was:
This transform was computed for every point in the frequency domain except kx=ky=0. This point represented the DC value of the total field, which was computed by calculating the average total field value for the survey area.
The reverse FFT was then computed for the above frequency domain grid and the DC value above added.