b'AEGC 2021Short abstracts135: Optimising 3D coal seismic imaging with pre-stackby others, which may or may not meet their target depth or depth migration processing requirements. There is a growing demand for time series datasets to be more accessible to facilitate alternative Mr Xiaodong Lu1, Mr Alan Meulenbroek 1 and Mr Karel Driml1 processing methods, particularly on HPC infrastructures, which enable processing of time series datasets at full resolution and 1 Velseis Pty Ltd running of larger models with more ensemble members and PSDM is used routinely in the oil and gas industry. However,uncertainty quantification.seismic processing in the coal industry does not routinelyTo address these issues, the GeoDeVL project experimented exploit this more advanced imaging technique. This paperwith a rapid open, transparent field-to-desktop-to-publication discusses the application of PSDM to a 3D coal volume acquiredworkflow to process and publish MT time series datasets using in the Bowen Basin in Qld in 2018. the new 15 Petaflop Gadi supercomputer at NCI. To do this, The target coal seams in the case-study area vary in depth from 5parallelised codes were developed to automate the generation 0m to ~300 m. The data quality varies significantly over the surveyof Level 0 to 1 time series data. Creating time series data area. The poor data quality area is caused by a combination oflevels for 95 Earth Data Logger stations now takes minutes, higher surface elevation and thick tertiary cover. Time processingversus days and weeks previously taken using more traditional provided a high-quality image where data quality was good butprocessing methods.not where data quality was poor. Poor statics meant that NMOThe process developed under the GeoDeVL project showed velocity picking was problematic in this area. how geophysicists can now work with less processed data and PSDM was utilised to derive a velocity model which aimedtransparently develop their own derivative products that are to improve the imaging of the target seams in the poor-datamore tuned to the specific parameters of their use case. Further, area. Two different initial velocity models were tested. The firstas new processing methodologies and/or higher capacity used the PSTM velocity. The second aimed to include velocitycomputers become available, the rawer forms of earlier surveys information of the near-surface. It was derived by performingare still available for reprocessing. Comparable trials in HPC PSDM on a set of shallow constant velocity models with the aimprocessing decades ago led to widespread use of HPC in the of imaging the base-of-weathering reflector. The associatedpetroleum exploration industry: will these results lead to similar velocities which maximized the stack response were picked anduptake of HPC in the minerals exploration industry?interpolated to create the initial model.Final PSDM stacks derived from the different initial models141: Seismic velocity analysis in the presence of AVO produced images superior to the PSTM stack. Reflectors whichpolarity reversals by fuzzy c-mean clusteringwere uninterpretable on the PSTM stack were imaged well onDr Javad Khoshnavaz 1, Dr Duy Thong Kieu2, Prof Hamidreza both PSDM stacks. Additionally, imaging of reflectors was subtlySiahkoohi3 and Mr Andrej Bona4improved in the good-data area. The PSDM stack derived using the near-surface model produced the best image. Anisotropy1 Formerly at Curtin Universityparameters calculated from this model were also more realistic. 2 Hanoi University of Mining and Geology3 Institute of Geophysics at the University of TehranWhile the workflow used to image the base-of-weathering4 Curtin Universityreflector has been used to derive an initial velocity model, the workflow can also be used as an alternative method for derivingRoutine seismic imaging algorithms often require velocity models. a statics solution. Therefore, velocity analysis plays a crucial role in the accuracy of velocity models and imaging. Seismic velocity analysis in CMP 136: Using the NCI Gadi supercomputer todomain has been dominated by the use of a powerful coherency revolutionise processing of MT time series data: resultsmeasurement tool, called semblance. Although this works quite from the GeoDeVL experiment reasonable for most of practical cases, it is incapable of dealing with polarity variations across moveout curves caused by faulting Dr Nigel Rees1, Mr Sheng Wang2, Dr Ben Evans1, Dr Bruceor AVO anomalies of class II. In this research, we proposed an Goleby3, Prof Lesley Wyborn 4, Dr Tim Rawling5, Dr Kelseyinversion-based velocity analysis algorithm that is based on Druken1 and Dr Rui Yang1 fuzzy c-means clustering, which has been recently considered in geophysical concerns. We apply the proposed algorithm of a 1 National Computational Infrastructure synthetic data example and a field CMP gather and compare the 2 Australian National University corresponding outcomes with the results obtained by semblance 3 OPM Consulting Australia analysis. The results suggest the effectiveness of the proposed 4 Research School of Earth Sciences/Australian Nationalalgorithm in the case of polarity variations.University5 AuScope 142: Application of seismic imaging to target the MagnetoTelluric (MT) time series datasets are expensive toPaleozoic basement underneath Tertiary basalts in the acquire, can be high volume (100s of terabytes), and thenorthwest Tasmania for exploration and miningtime taken to publish (measured from collection to release) often takes more than two years. Time series datasets haveMr Chuang Wang 1, Dr Gerrit Olivier1 and Dr Martin Jutzeler1been notoriously hard to access: most data providers only1 University of Tasmaniamake derivative MT transfer functions (EDI files) and model outputs accessible online. Hence, MT practitioners can beIn Tasmania, high-resolution reconstruction of lithospheric reliant on the data processing from raw data to be conductedstructure is of great significance. The application of seismic AUGUST 2021 PREVIEW 86'