b'Seismic window Seismic windowHiSeis, wanted to provide clients with theAs I understand it, mineral deposits big picture - an interpreted product toare precipitated from mineralised help the client understand the geology.fluids (magmatic or hydrothermal) that It is no surprise that hard rock miningtransport the extracted minerals from geophysicists are not familiar withdepth (Figure 2). The fault distribution interpreting seismic data (I must admit Iis important because it may focus the would be lost if asked to do somethingtransporting fluid or act as a barrier with EM data) so HiSeis aim to use theirto flow. The search for minerals starts expertise to provide a model that alsowith a structural model which includes incorporates all the available gravitya 3D fault framework. Then, based on and magnetic data. Australian mining ismodelling seismic amplitude variations, faced with declining grades and resourcean understanding of the signature of Michael Micenkodepletion and as the search moves deeperhydrothermal alteration (or some other Associate Editor for Petroleumtraditional methods struggle. Seismic ischaracteristic) is used to determine the mick@freogeos.com.au the only method that can provide goodprobability of key rock types to guide results at depths greater than 2 km. Theexploration to areas of high potential Hard rock seismic application of the seismic method is not(Figure 3). With this volume of data limited to surface seismic, and somerequiring analysis, Machine Learning Following my article on seismic forsurveys have recorded downhole methods(ML) is being used to interpret the data mineral exploration (Preview Feb 2021)such as vertical seismic profiles (VSP) andand determine probabilities. The seismic I received a note from Reg Nelsoncross hole imaging. A robust workflow hassignature is often not an obvious solid describing briefly his experiments inbeen developed for describing possibleanomaly with clear edges, but a fuzzy recording seismic over mineral bodiesoutcomes and includes sonic logging inzone of higher probability caused by when he worked with the SA Mineswellssomething that has generally notsubtle changes in density and perhaps Department some time ago. Regbeen acquired in minerals exploration.Pwave velocity.recorded data over Olympic Dam andBut good sonic logs are a prerequisite forHistorically seismic was used at the various areas on the Stuart shelf, andmodelling the contrasts associated withend of mine life, if at all, to check if even ventured as far as Mt Isa. expected rock types.The results were underwhelming, but one survey at Lake Torrens recorded reflections that correlated with some volcanics when, in desperation, Reg applied a 250 Hz low cut filter (Figure1). In a paper published in 1984 Nelson noted that targets were becoming deeper and more elusive such that the depths were at the limit of investigation of conventional electrical methods. In this environment only seismic reflection surveys could image the geology - if there were sufficient contrasts in acoustic impedance. His modelling showed ore bodies had a characteristic signature with density variations having more contrast than velocity. He concluded seismic could be useful if the signal bandwidth was at least 2 octaves and the upper frequency was greater than 200 Hz. These early efforts were hampered by the recording instrument that had only 24 channels, which is primitive compared to todays instruments that record thousands of channels.Regs letter prompted me to have another look at seismic for mineral exploration. I contacted Antony Brockmann and he took me through technology that hasFigure 1.Shot records from Stuart Shelf trials varying low cut filter60 Hz (left), 250 Hz (centre) and progressed significantly in the last ten120 Hz (right)with 500 Hz high cut (modified from Nelson 1984). Correlatable reflections are only apparent years. He explained that his company,on the high frequency record (centre).AUGUST 2021 PREVIEW 38'