b'AEGC 2023Short abstractsmachine learning and artificial intelligence and it offers a secureThe area under investigation - the Kambala mining district REST API to connect directly to other software. - lies within the south-central section of the Archaean Norseman-Wiluna greenstone belt, 60 to 100 km SSE of Nonlinear LS-RTM based on seismic scale separationKalgoorlie in Western Australia. This study focuses on the and full wavefield multi-parameter FWI. Bluebush Line cluster of deposits located 10 to 15 km east of the Widgiemooltha Dome. The deposits are Kambalda type ystein Korsmo, Yang Yang, Nizar Chemingui, Andrey Pankov,komatiite channelised nickel ore bodies. Major ore bodies Antonio Castiello and Andrew Long include Cameron - Kambalda, Lawry and Republican nickel deposits. Principal geological units of the area distinguishable PGS in geophysical data are Kambala Komatiite, Lunnon Basalt and Least-squares migration (LSM) is a linear inversion processBlack Flag Group of felsic volcaniclastic and siliciclastic rocks. with a convex objective function. It updates the reflectivity byNickel deposits are located along the contacts of volcanic rocks matching the high wavenumber perturbations in the observedand sedimentary beds.and synthetic data. The linear property of the inversionElectromagnetic data were used to outline conductive relates to the fixed background model, meaning that theaccumulation of Fe - Ni sulphides. Magnetic data were used to velocity field is assumed known and kept unchanged duringunderstand the tectonic pattern which could act as conduit the inversion, making it feasible to estimate with a linearfor hydrothermal fluids. Gravity data aided in identification solver. The synthetic data are generated based on the Bornof intrusive bodies and their contacts with sedimentary approximation, where only the first order scattering term issequences. As for AEM, we re-processed VTEM data from utilised in the forward modeling step. In practice, this means2009, 630 linear km, with spacing of 200 m. The application of that only the near- to mid-reflection angles can be utilised inInduced Polarisation modelling allowed us taking into account the inversion process. also the IP distortions. The outcome are both a more reliable On the other hand, Full Waveform Inversion (FWI) is a highlyconductivity model and the distribution of the chargeability nonlinear inversion with a complex objective functionin the subsurface. The conductive nickel deposits are well consisting of many local minima. The aim of FWI is to minimiseresolved, often close to sharp conductivity contrast, likely the misfit between observed and modeled data. Refractionsfavored by faulting.and diving waves have proven to be robust and straightforward to utilise in FWI. Reflections can also allow velocity estimationRepurposing geotechnical data for energy transition.beyond the penetration depth of diving waves. For this to be feasible, the modeling engine must initiate the completeJess Kozmanwavefield; diving waves, refractions and reflections (beyond Born approximations). Katalyst Data ManagementWe will demonstrate how a new nonlinear data-domain LS-RTMPetroleum exploration geotechnical disciplines have a can invert for the reflectivity while simultaneously refining thedocumented history of utilising well managed and curated FWI velocity model. The implementation takes advantage ofdata to shorten project cycle times, reduce operational risk, the similarities between RTM and FWI and a unique imagingand deliver higher quality data driven decisions that optimise condition to carefully update the velocity and reflectivityasset productivity. Recent initiatives to provide open-source without any leakage between the two models. In particular,platforms for data visualisation and interpretation on the reflectivity changes caused by density variations are notcloud have also made those data sets available to new groups erroneously mapped as velocity updates. The inversion utilisesin exploration, as geotechnical experts are re-tasked with the full acoustic wavefield through an alternative formulationexploring for new opportunities in sustainable resources.of the wave-equation parametrised in terms of velocity andRecent case studies show that existing optimum industry best vector reflectivity. The nonlinear LS-RTM results show significantpractices for cloud ingestion, enrichment and consumption structural improvements, more focusing, and better faultof digital datasets can be applied to support strategic choices imaging compared to RTM. in asset selection and development for sustainable energy resources. The lessons learned from applying principles Regional mineral potential evaluation and fast trackfor making cloud data findable, accessible, interoperable and re-usable are enabling embedded data workflow targeting using existing EM, magnetics and gravityprocesses and procedures that deliver value in the area of case study from the Eastern Goldfields, Westernhydrogen exploration and storage, geothermal, and carbon Australia. sequestration projects. The value of digital data in reducing decision latency is easier to measure as monitoring of cloud Peter Kovac 1 Antonio Menghini2 and Andrea Viezzoli2 data storage and delivery metrics becomes more visible. 1 Peter KovacLarge digital transformation projects have also led to more 2 Aarhus Geophysics executive level representation for digital data managers and their support of data driven decision making in the resource In this paper we present a mineral potential evaluationsector. Experience with using digital datasets on the cloud case study from the Eastern Goldfields Superterrane, infor predictive modelling under uncertainty is being used to the Kalgoorlie Terrane. We would like to demonstrate theevaluate the data necessary to execute strategic selection effectiveness of EM, magnetic and gravity data geologicalof sustainable assets, and geoengineering data sets from interpretation in the early stages of the mineral explorationinstrumented facilities are available to enable real-time cycle, using a sample legacy dataset downloaded from theoptimisation of existing processing workflows to make them archives of publicly available datasets. more sustainable.FEBRUARY 2023 PREVIEW 112'