b'FeatureEstimating Nth-order vertical gradientsExamples The fourth vertical gradientIn this section we test the new method on airborne-derivedFigure 2 shows the same data as Figure 1 but now for the total magnetic intensity (TMI) data from the Ebagoola datafourth vertical gradient. There is excellent correlation between set from a survey flown in Queensland in 1990 with a nominala) the fourth horizontal gradient calculated using the operator flying height of 120 m and flown in an east-west direction (datain equation 12, b) the the fourth vertical gradient calculated available from Geoscience Australia). We apply the method tofrom the horizontal gradient and c) the fourth vertical gradient two example calculations, the first is calculating the first verticalcalculated using the Fourier transform method. The correlation gradient and the second is the fourth vertical gradient. between the HG and FFT VG methods is 0.949.The first vertical gradient Discussion and conclusionsThe upper plot in Figure 1 shows the TMI data. The lower plotThe main purpose of this paper was to demonstrate the in Figure 1 shows the computed first horizontal gradient (usingfeasibilty of calculating the appropriate gradients using a the operator in equation 9); the first vertical gradient from thederivative operator for the horizontal gradient followed by horizontal gradient in the Fourier domain using equation7computing the vertical gradients by simply multiplying by iN (from HG) and the first vertical gradient using the standard[sgn()] and not by N in the frequency domain. This method Fourier transform method (via FFT). We corrected for the sign ofreduces the amount of amplification of the noise in the higher the horizontal gradient as it is a function of the flight direction.frequencies. It is not essential to filter the potential-field data The lower plot in Figure 1 shows that there is less noise onbefore implementing this method.the vertical gradient from the horizontal gradient than on the vertical gradient from the FFT. The correlation coefficient of theThe amplitudes of the real and imaginary components are not first vertical gradient with the horizontal gradient with thosechanged by the method, only the phase or shape is changed. calculated from the FFT is 0.983. This means that there must be some significant signal in the Figure 1. The data are from a section of line 2781 from the Ebagoola Survey.Figure 2. The displays are the same as in Figure 1, except now the data The upper plot displays the TMI data. The lower plot shows the computedare thefourth-order horizontal gradient and the two fourth-order vertical first-order horizontal gradient, the first-order vertical gradient computed fromgradientscalculated from the horizontal gradient (HG) and via the Fourier the horizontal gradient (HG) and the standard first-order vertical gradienttransform (FFT). computed by the Fourier transform method (FFT).46 PREVIEWJUNE 2024'