Quality control of gridded aeromagnetic data

Abstract

Many gridded images of aeromagnetic data include artefacts and 'unlikely' correlations of anomalies across flight lines. This distorts the anomaly shapes and patterns used to interpret the data, as well as introducing (or removing) power at certain frequencies which may be important for further processing. Some of the features may be due to poor data quality, inappropriate choice of survey parameters, poor gridding algorithms or inappropriate choice of the parameters used by the gridding algorithm. We develop a method for quality control of aeromagnetic grids that uses a new gap-filling algorithm, the Kaiser-Bessel window function and the fast Fourier transform (FFT) algorithm to calculate the grid's power spectrum. Visual inspection of the power spectrum enables many gridding and data problems to be diagnosed including over-smoothing, notch filtering, ineffective levelling, undersampling and directional bias caused by the sampling pattern and/or gridding algorithm. Existing methods for extrapolating and filling data gaps in grids prior to applying the FFT are shown to distort the power spectrum and can make quality control difficult. An alternative based on thin-plate spline interpolation to the data gaps was developed and used successfully. The method uses two columns and two rows around the edge of the grid and enforces continuity in grid values and derivatives (up to 2nd order) by fitting a thin-plate surface to these edge pixels. The quality control tool was used to determine sensible input parameters for bidirectional splines and minimum curvature, two of the main gridding methods used for aeromagnetic data, and to suggest the better method for a given situation. We found that bi­directional splines can exploit the fast sampling along lines but distort anomaly shapes when the line spacing is too wide. Minimum curvature is unable to exploit the fast sampling along lines but produces less distortion of anomalies at wide line spacing.