b'Discovery of the Havieron Au-Cu deposit, WAFeatureFigure 5.2.Residual of the terrain corrected gravity.Figure 3.Digital terrain model.Even with the density-optimised terrain-corrections, the gravity image is disappointing. Many short wavelength highs and lows of up to half a milligal, in some cases single point readings, disrupt a simple ovoid that might be expected from the magnetic image. We believe these variations to be caused by topographic undulations at the base of the unconsolidated cover. The model discussed below simulates most of these variations using irregularly shaped polygons a few metres to several tens of metres thick and having density contrasts of around 1 gm/cc (to simulate a bedrock ridge) or -1 gm/cc (to simulate a trough). These high frequency variations are almost an order of magnitude stronger than what might be expected of measurement noise in what are judged to have been well executed surveys.Figure 4.Bouguer gravity image created using the industry standardModel and methodcorrection density of 2.67 gm/cc.In an era when geophysical interpretation is dominated by sand dunes disrupt the Bouguer gravity image when over- inversion algorithms, we maintain that value can still be corrected using the standard density of 2.67 gm/cc, but thegained from the old fashioned method of forward modelling disruption is minimised using a density of 1.50 gm/cc which isused here. The importance for a geophysical model to have consistent the notion that the dunes consist of loose dry sandgeological credibility cannot be overstated, and, unlike most (Telford, etal., p25). Terrain corrections, using an algorithminversion codes, forward modelling leaves the user in control based on the method of Hammer (1939), were applied but onlyof the direction a model takes through the many iterations that three points in the far south west yielded corrections greaterare invariably needed. Furthermore, forward modelling helps to than normal measurement noise (0.02 mGal). The effect ofdevelop a keen understanding of just how ambiguous potential the sand dunes is not apparent in the revised gravity imagefield data is, and it provokes critical thought throughout (Figure 5.1) which shows a clear bulge in the contours of thethe arduous process required to arrive at some final model gravity gradient. After removing the regional trend, a residualthat is offered as an interpretation of the data. The writers gravity image exhibits a weak irregular ovoid (Figure 5.2), thedo not recommend against the use of inversion algorithms approximate centre of which is located slightly more thanwhen the user feels they add value to a project; however, we 200 m south of the magnetic peak. prefer forward modelling because of the control maintained on the final form of the model and because of the ease with which numerical experiments can be performed to test the significance of key model bodies, as well as alternate models. In addition, forward modelling is easily linked to the phase/scatter diagram method (Hanneson, 2003), also called the MagGravJ method, which provides a quantitative link between the physical properties of the model bodies and geological aspects of the rocks that the bodies represent.The geophysical data was simulated using a forward modelling algorithm based on the theory of Talwani (1960, 1961). The tops and bottoms of model bodies are horizontal, and the bodies and have polygonal outlines with sides that are defined by vertical lines through the vertices which can be altered collectively (to simulate a plunging body with the same shape) or individually (to simulate a body shape that Figure 5.1.Terrain corrected gravity; density 1.50 gm/cc. changes with depth).43 PREVIEW AUGUST 2022'