In complex sedimentary environments, the identification of thin sandbodies and the accurate prediction of their thickness remain challenging, particularly when relying on a single analytical approach. Taking the lower sub-member of the fourth member of the Shahejie Formation (Es4L) in the Chenguanzhuang area of the Dongying Depression as a case study, this study proposes a quantitative prediction method that integrates sedimentary facies constraints with machine learning-based seismic multi-attribute fusion. Based on core observations, well log data, and 3D seismic datasets, the study area is subdivided into two zones: Zone I (shallow-water delta front) and Zone II (shore–shallow lake). Sensitive attributes for each zone are optimized using Pearson correlation analysis and hierarchical clustering, and five machine learning models—SVR, Random Forest, MLP, Ridge Regression, and Lasso Regression—are systematically evaluated. The MLP model is selected for Zone I, achieving R2 values of 0.856 and 0.936 for the training and test sets, respectively, whereas Ridge Regression combined with leave-one-out cross-validation (LOOCV) is adopted for Zone II to mitigate overfitting caused by limited well data, yielding R2 values of 0.864 and 0.779. Compared with conventional linear regression (R2 = 0.45), the proposed approach significantly improves the accuracy of quantitative sandbody prediction, providing a reliable geological basis for hydrocarbon exploration and an effective technical framework for similar complex sedimentary environments.

Despite renewed interest in the energy resources of the Perth Basin, rift propagation, the evolution of fault activity and depocentre development along the Western Australian margin during rifting of East Gondwana are not well established. Rifting resulted in a series of predominantly NNW-SSE trending graben and half-graben extending from the Southern Carnarvon Basin in the north, through to the Perth Basin in the south. Gravity and magnetic data indicate two regions of greater structural complexity along this margin, with the northern of these regions data-poor. We investigate the more southerly region, commonly referred to as the northern Perth Basin, a Paleozoic to Mesozoic depocentre that developed through multiple phases of extension with variable extension directions, recording intracontinental rifting and eventual break-up of Australia and Greater India as part of the East Gondwana interior rift. Recently acquired seismic surveys that better image the pre-Mesozoic stratigraphy, together with legacy seismic and well data, have been utilised to create composite sections perpendicular and parallel to the rift axis. These composite sections reveal a progressive southward shift in fault activity and depocentres from the late Carboniferous through to the Early Cretaceous and that towards the south, many of the N-S trending fault segments previously thought to be active during the Cisuralian (early Permian) initiate much later in the Jurassic or Cretaceous. By contrast, E-W trending faults in the north of the basin, which formed perpendicular or at high angles to the rift axis, become throughgoing faults during the Late Triassic to Early Jurassic. Ultimately, these new interpretations provide evidence for episodic propagation of the rift system toward the south.

The evolution of the major rivers originating from the SE Tibetan Plateau has been a research hotspot due to a close connection between tectonic events, geomorphological shifts and river formation. This study reviews and compiles a large group of provenance analyses including zircon UPb dating and whole-rock geochemistry, in order to provide a systematic interpretation of the drainage evolution of the extensive rivers, mainly represented by the modern Red River, Mekong River, Pearl River, Yangtze River, the coastal South China rivers, etc. There are numerous debates over the potential existence of a paleo-Red River, in other words, whether the Yangtze upper and middle reaches, Mekong and other SE Asian rivers partly or collectively formed a single drainage system during Miocene (or even earlier), and eventually generated thick sedimentary sequences in the South China Sea (SCS). The fragmentation of this continental-scale river is speculated to result from several river captures and reversals during the Cenozoic before evolving into the present-day drainage framework. However, this hypothesis has been increasingly doubted, and our review shows a lack of robust evidence supporting the presence of a unidirectional N-S drainage, since zircon UPb signatures of SE Tibetan River sands, relict Cenozoic terrestrial deposits and contemporaneous offshore sediments fail to match in consistence. Instead, the Mekong River possibly had not achieved its current form until the Middle Miocene, possibly triggered by the Tibetan Uplift to enhance the SE Asian summer monsoon precipitation. Large uncertainties remain over the timing of the Yangtze River formation, with a wide range of age estimates extending from the Earliest Miocene to Holocene. During the early Cenozoic, it was the combination of axial topographic pattern and the prolonged extensional setting of the South China margin that collectively controlled sediment supply and distribution, and induced a dominant eastward transport pathway from eastern Indochina into the southern depression of the northern SCS. It wasn't until the Early Oligocene (or even later) that the topographic inversion of SE Asia by westward to eastward tilting accelerated the headwater erosion and drainage basin enlargement. The Pearl River thereby experienced a significant inland expansion during the Late Oligocene, and reached its near-modern delineation since the Early Miocene. At the same time, some rivers across coastal South China, such as the Min and Jiulong Rivers might also have extended farther to the west. In any case, regional tectonic activity between the Tibetan Plateau and the marginal sea basins shaped the geomorphological and topographic characteristics in a fairly complicated way, and controlled the overall source-to-sink patterns and fluvial system evolution.The SE Asian continental margin was dominated by a long-lived extension from the Late Mesozoic to Early Cenozoic. The marginal sea basins, including the SCS and the East China Sea (ECS) basins, have been preserved with exquisite sedimentary records of this geological history. Nevertheless, due to the limited drilling penetration, low-resolution seismic profiles, and the lack of a reliable stratigraphic framework, the sedimentary evolution of the offshore basins is still controversial from different perspectives. In this review, we also illustrate the Early Cretaceous–Early Miocene sedimentary evolution of the South and East China Sea regions by synthesizing our newly obtained results and dataset from previous literature, and by displaying the lithofacies patterns in sequential palinspastic restorations. Our preferred paleogeographic scenarios incorporate the conjugate relationship between continental margins and the birth-and-demise of the presumed proto-SCS. The lithofacies patterns, depositional environments as well as the provenance dataset, commonly demonstrate the sedimentary responses to the significant Meso–Cenozoic shift of magmatism and geodynamics along the SE Asian margin, and together have implications for the mechanisms of continental rifting and oceanic crust accretion during the Cenozoic.

Sediment routing systems are commonly investigated by applying the well-established U-Pb isotopic system to date the crystallization of detrital zircon grains in order to reveal the ultimate source of the dated component in a sediment. However, detrital zircon U-Pb geochronology provenance analysis cannot distinguish between sediment supplied from first-cycle erosion of crystalline basement and multicycle reworking of older basin components. Instead, analysis of Pb isotopes in detrital K-feldspar tends to track first-cycle sediment, as this mineral is more labile than zircon. Here we present new in situ Pb isotope ratios of detrital K-feldspar from the Mungaroo and Brigadier Formations (Northern Carnarvon Basin, North West Shelf, Australia) and K-feldspar crystals of potential (meta)igneous crystalline source rocks. These new K-feldspar Pb isotope ratios, in combination with detrital zircon U-Pb geochronology, allow for more holistic investigation of sediment provenance and refinement of sediment recycling histories. The Albany Fraser Orogen/Wilkes Land, the Musgrave Province, the Pilbara Craton, and other northern terranes (Lhasa Block, southwest Borneo) across a broad region of former northeast Gondwana are identified as the main source regions of the sampled intervals of the Mungaroo and Brigadier Formations. Additionally, differences between the relative volumetric significance of source regions interpreted from K-feldspar and zircon datasets reveal the extent of sediment reworking along a major sediment pathway. We propose a model of relatively more recycled material from the distal Albany Fraser Orogen/Wilkes Land, entraining additional recycled material from the central Australian Musgrave Province and finally mixing with directly sourced sediment from the proximal Pilbara Craton.

The post-rift succession of the Northern Carnarvon Basin (north-western Australia) has commonly been interpreted as a relatively simple and uniform sequence that records a transition from siliciclastic to carbonate sedimentation deposited on a passive margin. However, given the significant vertical and lateral variation in seismic facies visible on seismic data, this interpretation likely oversimplifies the depositional history of the margin. Regional composite seismic lines that cross most of the basin, integrated with lithological and biostratigraphic information from exploration wells, provide the context for a better understanding of the depositional processes and environments that characterize the post-rift continental margin succession. We show that the sedimentary sequences deposited above the Valanginian breakup unconformity contain a wide variety of seismic facies that can be linked to a number of different marine depositional environments, with the greatest lateral variation occurring in the Turonian – Rupelian and in the Tortonian – Present. The former interval consists of three dominant seismic facies, namely polygonally faulted, incised, and parallel bedded, which we interpret to indicate a lateral transition from an environment primarily dominated by fine-grained pelagic/hemipelagic deposition to one dominated by energetic bottom currents that created both depositional and erosional features, such as contourite drifts and associated moats. The latter interval is expressed by sigmoidal and continuous reflections which pass laterally into more chaotic reflection packages, which we interpret as clinoforms and mass-transport complexes (MTCs). The development of bottom currents may be connected to changes in circulation associated with the opening of oceans adjacent to the northwest margin of Australia, while the MTCs may indicate increased regional seismic activity and slope instability resulting from the development of collisional plate boundaries. Definition of these sequences highlights the significant changes that have occurred in the sedimentary processes that operated on the margin, and their potential link to its tectonic evolution.

In the northern Perth Basin (Western Australia), the Early Triassic Kockatea Shale is the primary petroleum source rock. Possible source rocks in the Northern Carnarvon Basin are more varied and include the Upper Jurassic Dingo Claystone as well as the Early Triassic Locker Shale. Biomarker analyses were conducted on petroleum samples from these basins to understand the nature of the petroleum systems. Many of the analysed petroleum samples contain carotenoids (okenane, chlorobactane and isorenieratane) derived from photosynthetic sulfur bacteria, suggesting that their source rocks were deposited under conditions of photic zone euxinia (PZE) and/or derived from microbialites. In the northern Perth Basin, the major lithofacies contributing to the source rock are dark coloured mudstones deposited under PZE conditions and/or derived from microbialites. In the southern Perth Basin, the potential source rock is either Permian, Jurassic or Cretaceous in age as indicated by the low concentrations or absence of carotenoids and the Triassic biomarker n -C 33 alkylcyclohexane. There is also a possibility that the Lower Triassic Locker Shale is the source rock of petroleum in the Tubridgi field on the Peedamullah Shelf of the Northern Carnarvon Basin, based on the similarity of biomarkers to Perth Basin petroleum sourced from the Kockatea Shale. However, the possibility of charge from the Upper Jurassic Dingo Claystone cannot be entirely excluded. Supplementary material: biomarker dataset is available at https://doi.org/10.6084/m9.figshare.c.6452153

Geological and geophysical interpretation is characterised by large and localised datasets that are extremely expensive to acquire. There are clear advantages in applying deep learning techniques to such datasets, but this requires a large amount of suitable data for effective training. Creation of training data can be time-consuming, but novel countermeasures such as Generative Adversarial Networks (GANs) are an enticing alternative with the potential to alleviate this issue by providing synthetic data that is representative of the real data. This paper details an investigation of the potential of GANs for the augmentation of labelled seismic facies from a mass transport complex in training a convolutional neural network (CNN) for facies classification. The study adopts a specific GAN approach, known as the Conditional Style-Based Logo GAN (LoGANv2), because of its capability to generate conditioned output with improved training stability. By using LoGANv2 to synthesise examples that mimic the behaviour of the real data, based on a 3D seismic dataset from the North Carnarvon Basin in Australia, the accuracy of the traditional CNN for facies classification improved from 38% (the benchmark result where no augmentation was applied) to 52%, 61%, 94% and 74% for four trials with different degrees of augmentation. The results show that increasing training size, either through manual annotation (trial 2) or standard manipulations such as image rotation (trial 3 and trial 4), improves the performance of LoGANv2 and hence the CNN classification. The experiments indicate that although standard manipulation can increase the diversity of the training dataset, the balance between diversity and consistency within the datasets is also important, and this should be optimized for LoGANv2 to achieve better performance. In addition, we explored the adaptability of the proposed LoGANv2-CNN approach to unlabelled samples from another survey to demonstrate the robustness and flexibility of the model.

Analyses of the sedimentary record, in particular the origin, routing and timing of deposition, can provide insight into a vast array of different geological processes over Earths history. However, understanding the provenance of very fine-grained sediments, particularly clay minerals, is typically challenging given their difficulty in being precisely dated. Yet the provenance of clay material may potentially provide greater detail on more distal detritus relative to other more conventional sand-sized provenance indicators (e.g. zircon). Triple quadrupole laser ablation inductively coupled plasma mass spectrometry, which allows in-situ Rb-Sr dating, offers several potentially significant advances in the geochronology of fine-grained Rb bearing mineral phases. Nonetheless, dating of very fine-grained sediments, especially those formed in the Phanerozoic after the terrestrial vegetative evolution, is challenged by the inevitable mixing of potentially different generations of Rb bearing minerals within the volume ablated by the laser (~90,000 µm3). Rb-rich minerals will dominate the isotopic budget of any mixture and may help to constrain age components within the sample. In this study, we present in-situ Rb-Sr data of six variably biostratigraphically constrained Paleozoic shale samples from the Barnicardy-1 drill core (Canning Basin, Western Australia). We develop a new Monte-Carlo modelling approach to improve the constraint on Rb-Sr ages by confining the range of initial Sr-ratios from analysed mixtures via a priori knowledge of the samples stratigraphic level. Rb model age components (~620 Ma) imply clay detritus was sourced from the Paterson Orogen and helps refine the stratigraphy of the Canning Basin.

Detrital zircon U-Pb geochronology has enabled advances in the understanding of sediment provenance, transportation pathways, and the depositional age of sedimentary packages. However, sample selection and processing can result in biasing of detrital zircon age spectra. This paper presents a novel approach using in situ detrital zircon U-Pb measurements on thin-sections to provide greater confidence in maximum depositional ages and provenance interpretations. New U-Pb age data of 310 detrital zircon grains from 16 thin-sections of the Triassic Mungaroo Formation from two wells in the Northern Carnarvon Basin, Australia, are presented. Whilst detrital zircon age modes are consistent with previous work, there are some differences in the relative proportions of age modes, which are partly attributed to a lack of small grains in hand-picked grain mounts. The relative sample bias is quantified via grain size comparison of dated zircon (in thin-sections or hand-picked mounts) relative to all zircons identified in bulk-mounts and thin-sections. The youngest age mode (c. 320-195 Ma) is consistent with an active margin to the north. The dated zircons reveal a single grain defined maximum depositional age of 197 Ma for the upper part of the Mungaroo Formation, suggesting deposition may have continued into the Early Jurassic.

With the rapid development of deep learning technologies, data-driven methods have become one of the main research focuses in geophysical inversion. Applications of various neural network architectures to the inversion of seismic, electromagnetic, gravity and other types of data confirm the potential of these methods in real-time parameter estimation without dependence on the starting subsurface model. At the same time, deep learning methods require large training datasets which are often difficult to acquire. In this paper, we present a generator of 2D subsurface models based on deep generative adversarial networks. Several networks are trained separately on realistic density and stratigraphy models to reach a sufficient degree of accuracy in generation of new highly detailed and varied models in real-time. This allows for creation of large synthetic training datasets in a cost-effective manner, thus facilitating the development of better deep learning algorithms for real-time inversion and interpretation.

Photic zone euxinia (PZE) has previously been identified in the Early Triassic Kockatea Shale of the northern Perth Basin, based on the presence of biomarkers such as isorenieratane, which is derived from isorenieratene produced by green sulfur bacteria. However, green and purple sulfur bacteria can also occur in microbial mats. In this study we present a basin-scale assessment of biomarkers associated with open water column PZE and/or microbialites. The lithofacies from the Early Triassic of the northern part of the northern Perth Basin consist of dark coloured mudstones (black to dark grey) with microbialites, which were deposited away from basin margins. These samples are found to contain okenane, chlorobactane and isorenieratane derived from carotenoid pigments of purple, green-green and green–brown sulfur bacteria, respectively. These biomarkers are not observed in the light coloured mudstones (medium grey) formed under oxic conditions in a tidal environment with higher clastic input close to the basin margins in the southern part of the basin where shallow marine sandstones were also deposited. Okenane and chlorobactane were abundant in facies containing microbialites which developed in a shallow water setting on intra-basinal structural highs. The development of oxic conditions near the basin margins in the Perth Basin provided refuges for organisms during the end Permian mass extinction event. Okenane was more abundant in the microbialite facies compared with the dark coloured mudstones deposited under PZE. C33 n-alkylcyclohexane (n-C33 ACH) has previously been described as a biomarker associated with ecosystem collapse during the lower Triassic, and its ratio relative to the C34 n-alkane was elevated in facies that contain abundant microbialites. Mercury (Hg) to total reduced inorganic sulfur (TRIS) and Hg to total organic carbon (TOC) ratios are positively correlated, supporting the development of euxinia as sulfide sequesters Hg. The high Hg/TRIS values in microbialite facies support mat development with high Hg concentrations. The values of δ13COM and δ34S were isotopically lighter in microbialite facies when compared with mudstones formed under PZE, although samples deposited under oxic conditions showed the isotopically lightest δ13COM and isotopically heaviest δ34S. The variation in δ13COM probably reflects the different carbon fixation pathway of various sulfur bacteria, while it appears that the difference of δ34S values between PZE and microbialites is a result of differences in the microbial community structure and the higher relative abundance of purple sulfur bacteria. In addition, fluctuations between PZE and oxic conditions were identified throughout the sampled intervals attributed to fluctuations in the depth of chemocline. Under the shallower chemocline, PZE developed widely in the basin, even in shallower waters. Under the deeper chemocline PZE was absent or limited to deeper water, and oxic conditions developed in shallower water. The fluctuations of PZE and oxic conditions in the northern Perth Basin suggest the development of multiple episodes of harsh environmental conditions after the end-Permian mass extinction, similar to those identified in other regions (e.g., Meishan in China and Peace River in Canada).

This paper presents a semi-quantitative analysis of gravity-driven deformation along the Namibian margin using extensive 2D depth converted seismic data. The geometries, internal characters and distribution of gravity-driven systems were investigated through regional and detailed seismic studies. The research shows that surficial slumps are typically ca. 50 m thick and are characterised by contorted seismic facies commonly occurring along the slopes of the margin. They commonly funnel and cluster within high relief areas such as canyons and pre-existing landslide scars. These contrast with coherent slides that are up to 2 km thick which extend laterally along the margin for tens to hundreds of kilometres. Slides preferentially occur in the proximal part of the margin and are constrained within the main margin depocenters. Here, high sedimentation rates and loading promote the generation of distinct, weak, overpressured layers that favour initiation of sliding of relatively coherent sediment masses. This research also shows that one-third of volume of the post-rift sediments on the Namibian margin were affected by slides and slumps. This demonstrates that gravity-driven deformation is a key geological process that can strongly modify the evolution of rifted passive margins.

With the increased size and complexity of seismic surveys, manual labeling of seismic facies has become a significant challenge. Application of automatic methods for seismic facies interpretation could significantly reduce the manual labor and subjectivity of a particular interpreter present in conventional methods. A recently emerged group of methods is based on deep neural networks. These approaches are data-driven and require large labeled datasets for network training. We apply a deep convolutional autoencoder for unsupervised seismic facies classification, which does not require manually labeled examples. The facies maps are generated by clustering the deep-feature vectors obtained from the input data. Our method yields accurate results on real data and provides them instantaneously. The proposed approach opens up possibilities to analyze geological patterns in real time without human intervention.

Abstract Variably oriented dolerite intrusions outcrop in the Albany–Fraser Orogen along the south coast of Western Australia with previously unknown ages but where previous studies interpreted Mesoproterozoic to Cretaceous emplacement. Here, we place temporal constraints on seven mafic intrusions across ∼150 km of coast using zircon U–Pb, apatite U–Pb, and plagioclase 40Ar/39Ar geochronology, coupled with whole-rock major and trace-element geochemistry, that reveal late Mesoproterozoic to potentially Early Cretaceous crystallisation ages. Three intrusions metamorphosed to greenschist facies are likely associated with either the emplacement of the ca 1210 Ma Marnda Moorn large igneous province or Stage II Albany–Fraser Orogeny, both of which were associated with the assembly of Rodinia. Three unmetamorphosed dykes have (probable) Neoproterozoic to lower Cambrian emplacement ages, likely associated with the ca 550–500 Ma Kuunga Orogeny during Gondwana assembly. The final sill, also unmetamorphosed, strikes perpendicular to the other six intrusions, shows unusual Pb anomalies and contains inherited zircon that has been reset by a Permian or younger event, pointing towards magmatism in southwestern Australia during the breakup of Gondwana. The new results provide hitherto unrecognised mafic intrusive evidence for modification of Proterozoic crust, potentially associated with Rodinia assembly, Gondwana assembly and Gondwana breakup in southwestern Australia. KEY POINTS Variably oriented mafic dykes in southwest Australia are dated by zircon U–Pb, apatite U–Pb and plagioclase 40Ar/39Ar methods. The dykes are related to Rodinia assembly (ca 1200 Ma), Gondwana assembly (ca 550 Ma) and, probably, Gondwana breakup (ca 135 Ma). These new ages provide evidence for mafic activity clearly linked to the supercontinent cycle.

The facies distribution of the Lower Triassic Kockatea Shale in the northern Perth Basin is an important factor for assessing the Lower Triassic source rock potential in the basin. Seismic interpretation and well analyses demonstrate that Permian-aged faults are responsible for the morphology of the basin in the Early Triassic and created remnant topography that controls the deposition of Lower Triassic sediments. Cores from petroleum wells and their correlated wireline logs show the lithological variation and lateral distribution of facies across the base of the Kockatea Shale. Darker coloured mudstones were deposited under anoxic conditions in the deeper part of the basin, while lighter coloured mudstones and tidally influenced sediments with bioturbation were deposited under oxic conditions in a shallow marine setting. Microbial mats developed in shallower water depths on structural highs (Beagle Ridge and Turtle Dove Ridge) in the anoxic parts of the basin. The distribution of microbial mats is controlled by two factors during the period of ecological recovery from the end-Permian mass extinction: 1) the presence of topographic highs and, 2) distance from the edge of the basin. The source rock potential of both darker coloured mudstone and microbial mat facies is demonstrated by TOC values between 0.3 and 3.4% and HI values between 39 and 579 mg/gTOC. On the other hand, the lighter coloured mudstones, deposited under more aerobic tidal settings, have poor source rock quality and TOC values between 0.2 and 0.5% and HI values between 59 and 127 mg/gTOC.

The North West Shelf of Australia has experienced numerous rift events during its prolonged evolution that most likely started in the Lower Palaeozoic and continued through to the formation of the present day passive margin in the Lower Cretaceous. Carboniferous and Permian is associated with rifting of the Lhasa terrane, a phase extension in the Lower and Middle Jurassic associated with the separation of the Argo terrane Upper Jurassic to Lower Cretaceous extension culminated in the separation of Greater India and Australia. Investigations based on interpretation of extensive, public domain seismic data, combined with numerical mechanical modelling, demonstrate that crustal structure, rheology and structural fabrics inherited from older events exert a significant control on the architecture of younger rifts. Defining the older, more deeply buried rift episodes is challenging, but with seismic data that now images deeper structures more effectively, it is clear that NE-SW oriented Carboniferous to Permian aged rift structures control the overall geometry of the margin. Variations in the timing, distribution and intensity of that rift may account for some of the complexity that governs the Triassic – a failed arm of the rift system might account for the accumulation of thick sequences of fluvio-delatic sediments in an apparent post-rift setting, while active deformation and igneous activity continued elsewhere on the margin. A renewed phase of extension began in the latest Triassic in the western part of the Northern Carnarvon Basin, but became progressively younger to the NE. High-resolution mechanical numerical experiments show that the dual mode of extension that characterises the Northern Carnarvon Basin, where both distributed and localised deformation occurs at the same time, is best explained by necking and boudinage of strong lower crust, inherited form the Permian rift event, proximal to the continental margin, and a subdued extensional strain rate across the distal extended margin. A very clear and consistent pattern of ENE oriented extension, which interacts obliquely with the older NE-SW oriented Permian aged structures, is apparent across the whole of the Northern Carnarvon Basin and extends north east into the Roebuck and Browse Basins. This is at odds with the NW-SE oriented extension predicted by the separation of the Argo terrane which occurs at this time. This may be explained by the detached style of deformation that characterises the Mesozoic interval. Alternatively, the separation of Greater India may have exerted a stronger influence on the evolution of the margin during the Jurassic than hitherto recognised.