Magnetic depths to basalts: extension of spectral depths method

By assuming that anomalies in a TMI grid arise from thin magnetic layers, spectral depth estimates to one or more equivalent magnetic layers may be obtained repeatedly across large areas. A general formula is derived for the purpose and demonstrated on the TMI grid of the Northern Territory.

Although spectral depth determination has played a role in magnetic interpretation for over four decades, automating the procedure has been inhibited by the need for manual intervention. This paper introduces the concept of a slope spectrum of an equivalent layer, to be used in an automated depth interpretation algorithm suitable for application to very large datasets such as the complete Northern Territory aeromagnetic grid.

In order to trace the extensive basalts across the Northern Territory, profiles of spectral depths have been obtained at 5 km intervals across the NT stitched grid of total magnetic intensity (TMI). Each profile is a graph from 0 to 1000 m of the probability of a magnetic layer occurring at each depth. Automating the collection of the 50 000 profiles required the development of a formula that relates slopes along the power spectrum to depths to an equivalent magnetic layer. Model slabs were populated with a large number of randomly located dipoles and their power spectra correlated with modelled depth to provide the formula. Depth profiles are too noisy to be used singly, but when a series of depth profiles are lined up side-by-side as a transect, significant magnetic layers can be traced for large distances. Transects frequently show a second layer. The formula is quite general in its derivation and would apply to any mid-latitude area where significant magnetic bodies can be modelled as extensive layers.

Because the method requires a radial power spectrum, it fails to provide signal at depths much shallower than the flight line spacing. The method is convenient for a fast first pass at depth estimation, but its horizontal resolution is rather coarse and errors can be quite large.