b'Education matters models under different hydrogeologicalmachineSVM), digital data (i.e. gamma-conditions using a synthetic approach.ray (g-ray) spectrometry and apparent A method is presented to couple TEMconductivity (EC a )) used in combination and surface NMR to derive hydraulicor independent of each other, suitable conductivity using the Markov-Chaintransect spacing of the digital data (i.e. Monte Carlo method. Thirdly, a modelling5, 10, 20, 30, 40, 60, 80 m) and number framework is presented to coupleof soil samples (i.e. 120, 110,, 10) for multiple geophysical techniques of TEM,calibration. I test these using a validation borehole NMR and audio frequency(i.e. 40) data set. The comparisons were magnetotellurics (MT) with stochasticevaluated considering the agreement groundwater modelling through thebetween measured and predicted CEC ensemble-smoother method. Thisusing Lins concordance correlation approach allows the uncertainty incoefficient (LCCC) and accuracy using geophysical data to be expressed asroot mean square error (RMSE).a prior probability distribution that can be incorporated and accountedThe results indicate that for the DSMUnderstanding deep crustal architecture for in groundwater model inversion.of topsoil CEC, the Cubist with anis an ongoing challenge in earth This method was applied to a potentialintermediate number of calibrationsciences, largely because in most cases in-situ recovery site in Kapunda, Southsamples (i.e. 80) using in combinationthese deep structures cant be sampled. Australia to estimate the regional-scaleboth g-ray and EC awas optimal inThe Durham Ox High is a geophysical impact of a hypothetical operation trialterms of agreement (LCCC = 0.79). Foranomaly interpreted in the deep crust in a probabilistic manner. Together,subsoil, a smaller number (i.e. 30) of soilthat extends across the Stawell and these three novel studies represent asamples for calibration was required toBendigo Zones in northern Victoria, contribution to the coupling betweenachieve good agreement (LCCC = 0.89).and into southern NSW. The Durham geophysics and hydrogeologicalIn terms of accuracy, the accuracyOx High is characterised by high modelling. (RMSE = 5.42 cmol(+)/kg) of subsoil CECamplitude, long wavelength magnetic was satisfactory, as it was less than halfand gravitational responses, and MSc/MPhil theses standard deviation (SD) (7.55 cmol(+)/kg)although several hypotheses have been of measured CEC. While not the same forproposed to explain these anomalies, it Tom Zhao, University of NSW: Digital soiltopsoil CEC, the accuracy (RMSE = 1.93remains poorly understood. However, mapping of soil cation exchange capacitycmol(+)/kg) was not as satisfactory as itrecent interpretations from deep (CEC) using different models, sample sizeswas over half measured topsoil CEC SDcrustal reflection seismic surveying, and transect spacing in a sugarcane field in(1.68 cmol(+)/kg). south of this anomaly, provided Proserpine, QLD direct evidence for improved controls The results also showed that while g-rayon the deep crustal architecture in alone was superior to EC adata forthe Stawell and Bendigo Zones. The prediction, better results were achievedseismic survey suggested that the mid when both digital data were used into deep crust consists of dense, thick combination. In terms of a suitableand imbricated block of Cambrian transect spacing for collection ofmafic volcanics. These mafic volcanics digital data to predict topsoil CEC, theare overlain by Cambrian-Ordovician small transects spacing (i.e. 5 m) wasturbidites where Devonian (post-recommended. For subsoil prediction,Tabberabberan) S-type granites have larger transect spacing may still beintruded through to the surface (e.g. appropriate (i.e. 560 m). The DSMLake Boga Granite and Pyramid Hill approach overall enabled topsoilGranite). This study aimed to resolve and subsoil prediction of CEC withthe geometry and depth for the deep The soil cation exchange capacitygood accuracy and small residuals,crustal architecture associated with the (CEC) is an important property becauseparticularly at large calibration data setsDurham Ox High using potential field it influences soil structural stability,(i.e. 80). The final DSM of topsoil andmodelling and granite geochemistry. nutrient availability, pH and reaction tosubsoil CEC therefore do allow for theNew petrophysical and high-resolution fertilisers. To assist Australian sugarcaneimplementation of the six-easy-stepsland-based gravity data were acquired farmers balance sugarcane-yield andnutrient management guidelines forfrom the overlying granites to constrain minimise fertiliser run-off, the six-easy- lime. As the soil pH in most area here istheir geometries. Furthermore, this steps nutrient management guidelineslower than 5.5, the larger northern halfstudy provided insight into the source were developed. In this research, a digital(22.74 ha) requires a small ( 2.5 t/ha)rocks causing the Durham Ox High, soil mapping (DSM) approach is used toapplication rate with the southern halfusing a laser-ablation split-stream understand the spatial variation in topsoilrequiring intermediate (35 t/ha) totechnique to simultaneously acquire (00.3 m) and subsoil (0.60.9 m) CEC.large (5 t/ha) rates. U-Pb ages and Hf isotopic compositions With regard to DSM, various aspects ofon zircons extracted from the Lake the current materials and methods areMatthew Shrimpton, The University ofBoga and Pyramid Hill Granite. These compared and contrasted, includingMelbourne: Deep crustal architecture ingranites provided as compositional choice of model (i.e. linear mixed modelsouth-eastern Australia: Insights from split- constraints because crustal melting toLMM, regression krigingRK, Cubist,stream zircon analysis and potential fieldproduce S-type granitic melts has been random forestRF and support vectorgeophysics suggested to occur deeper in the crust. DECEMBER 2020 PREVIEW 32'