Geoscience Problems Research Articles

Locating terrane boundaries in South China with big geochemical data mining

The basis for understanding continental tectonics to a large degree relies on terrane recognition. However, discrepant identifications and interpretations of the same terrane as well as its boundary are still commonly seen in South China. For years, the central debate has been revolving around the Yangtze-Cathaysia boundary and West-East Cathaysia boundary. Here we characterized the geochemical patterns (or differences) of the proposed terranes and located their boundaries by using a big geochemical dataset of South China, which consists of compositions of 75 elements in 5231 sediment samples. To extract valuable information from this big geochemical dataset, we applied multiple data mining techniques, including principal component analysis, sequential binary partition, Sobel operator, and support vector machine. Our results show that the geochemical patterns vary distinctly at regional scale across both the subblocks (i.e., Yangtze and Cathaysia) and terranes (i.e., Jiuling, Huaiyu, and Youjiang). Accordingly, we conclude that 1) the Northeast Jiangxi fault marks not only the Jiuling-Huaiyu boundary but also the north-eastern Yangtze-Cathaysia boundary; 2) the Yangtze-Cathaysia boundary extends westward along the Jiangshan-Pingxiang-Chaling fault; and 3) a fault zone comprising the Northwest Fujian fault and Zhenghe-Dapu fault possibly demarcates the West and East Cathaysia. The boundaries defined by the contrasting geochemical patterns in this study are also supported by geophysical, Nd isotopic, and geological evidence. Our observations will allow better explanation of where the proposed terranes originated and how South China was assembled. Our work demonstrates that data mining techniques are effective in deriving and integrating information for solving geoscience problems.

Neural network attribution methods for problems in geoscience: A novel synthetic benchmark dataset

AbstractDespite the increasingly successful application of neural networks to many problems in the geosciences, their complex and nonlinear structure makes the interpretation of their predictions difficult, which limits model trust and does not allow scientists to gain physical insights about the problem at hand. Many different methods have been introduced in the emerging field of eXplainable Artificial Intelligence (XAI), which aims at attributing the network’s prediction to specific features in the input domain. XAI methods are usually assessed by using benchmark datasets (such as MNIST or ImageNet for image classification). However, an objective, theoretically derived ground truth for the attribution is lacking for most of these datasets, making the assessment of XAI in many cases subjective. Also, benchmark datasets specifically designed for problems in geosciences are rare. Here, we provide a framework, based on the use of additively separable functions, to generate attribution benchmark datasets for regression problems for which the ground truth of the attribution is known a priori. We generate a large benchmark dataset and train a fully connected network to learn the underlying function that was used for simulation. We then compare estimated heatmaps from different XAI methods to the ground truth in order to identify examples where specific XAI methods perform well or poorly. We believe that attribution benchmarks as the ones introduced herein are of great importance for further application of neural networks in the geosciences, and for more objective assessment and accurate implementation of XAI methods, which will increase model trust and assist in discovering new science.

Modeling of Bandpass GPR Problem by HIE Procedure With Enhanced Absorption

For geoscience problem whose structures are complex and fine along a single direction, hybrid implicit-explicit (HIE) procedure is incorporated with the complex envelope (CE) method in bandpass circumstance. In ground-penetrating radar (GPR) simulation, large amount of low-frequency waves results in unacceptable performance especially in low frequency. To alleviate such phenomenon, higher order concept is introduced into perfectly matched layer (PML) formulation. The efficiency, accuracy, and absorption are illustrated through numerical examples. It can be concluded that the proposed procedure can receive better efficiency, considerable absorption, and remarkable accuracy. Such condition indicates the proposed scheme takes advantages over the afore-mentioned schemes which show its potential in the bandpass GPR problems. Most importantly, it can solve two pivotal issues: 1) it can alleviate time increment problem of the implicit algorithms with small time step and 2) it can keep the unconditional stability with the increment of time steps.

Igneous rock area and age in continental crust

Abstract Rock quantity and age are fundamental features of Earth's crust that pertain to many problems in geoscience. Here we combine new estimates of igneous rock area in continental crust from the Macrostrat database (https://macrostrat.org/) with a compilation of detrital zircon ages in order to investigate rock cycling and crustal growth. We find that there is little or no decrease in igneous rock area with increasing rock age. Instead, igneous rock area in North America exhibits four distinct Precambrian peaks, remains low through the Neoproterozoic, and then increases only modestly toward the recent. Peaks in Precambrian detrital zircon age frequency distributions align broadly with peaks in igneous rock area, regardless of grain depositional age. However, detrital zircon ages do underrepresent a Neoarchean peak in igneous rock area; young grains and ca. 1.1 Ga grains are also overrepresented relative to igneous area. Together, these results suggest that detrital zircon age distributions contain signatures of continental denudation and sedimentary cycling that are decoupled from the cycling of igneous source rocks. Models of continental crustal evolution that incorporate significant early increase in volume and increased sedimentation in the Phanerozoic are well supported by these data.

Development of artificial intelligence models for well groundwater quality simulation: Different modeling scenarios.

Groundwater is one of the most important freshwater resources, especially in arid and semi-arid regions where the annual amounts of precipitation are small with frequent drought durations. Information on qualitative parameters of these valuable resources is very crucial as it might affect its applicability from agricultural, drinking, and industrial aspects. Although geo-statistics methods can provide insight about spatial distribution of quality factors, applications of advanced artificial intelligence (AI) models can contribute to produce more accurate results as robust alternative for such a complex geo-science problem. The present research investigates the capacity of several types of AI models for modeling four key water quality variables namely electrical conductivity (EC), sodium adsorption ratio (SAR), total dissolved solid (TDS) and Sulfate (SO4) using dataset obtained from 90 wells in Tabriz Plain, Iran; assessed by k-fold testing. Two different modeling scenarios were established to make simulations using other quality parameters and the geographical information. The obtained results confirmed the capabilities of the AI models for modeling the well groundwater quality variables. Among all the applied AI models, the developed hybrid support vector machine-firefly algorithm (SVM-FFA) model achieved the best predictability performance for both investigated scenarios. The introduced computer aid methodology provided a reliable technology for groundwater monitoring and assessment.

An enhanced dual IDW method for high-quality geospatial interpolation

Many geoscience problems involve predicting attributes of interest at un-sampled locations. Inverse distance weighting (IDW) is a standard solution to such problems. However, IDW is generally not able to produce favorable results in the presence of clustered data, which is commonly used in the geospatial data process. To address this concern, this paper presents a novel interpolation approach (DIDW) that integrates data-to-data correlation with the conventional IDW and reformulates it within the geostatistical framework considering locally varying exponents. Traditional IDW, DIDW, and ordinary kriging are employed to evaluate the interpolation performance of the proposed method. This evaluation is based on a case study using the public Walker Lake dataset, and the associated interpolations are performed in various contexts, such as different sample data sizes and variogram parameters. The results demonstrate that DIDW with locally varying exponents stably produces more accurate and reliable estimates than the conventional IDW and DIDW. Besides, it yields more robust estimates than ordinary kriging in the face of varying variogram parameters. Thus, the proposed method can be applied as a preferred spatial interpolation method for most applications regarding its stability and accuracy.

Joint interpretation of geophysical data: Applying machine learning to the modeling of an evaporitic sequence in Villar de Cañas (Spain)

An optimal strategy for building realistic geological models must combine different geophysical techniques, each with its advantages and limitations. However, dealing with multiple geophysical datasets to derive a geological interpretation is not straightforward since geophysical parameters are not always functionally related. In this work, we propose an innovative approach consisting of using machine learning techniques to jointly interpret three geophysical datasets (a pseudo-3D resistivity model, a 3D velocity model, and 4 well-logs). These datasets, among others, were acquired to characterize the suitability of an evaporitic sequence for hosting a temporary storage facility of hazardous radioactive waste, which was planned in Villar de Cañas (Spain). Our strategy consisted of integrating both models in a single 3D bi-parametric grid that nested the velocity and resistivity values in each node. We then used a supervised learning algorithm to lithologically classify each node according to a training set based on the borehole data, which acts as ground truth. The training set is composed of classifiers that lithologically label resistivity-velocity pairs. However, the very shallow nodes lack classifiers due to the poor well-log coverage at the top part of the evaporitic sequence. To fill this gap, we computed an unsupervised cluster analysis that provided new classes to complete the training set. Finally, the supervised classification was applied, providing a new 3D lithology model that is far more consistent with the geology than the models derived from each parameter independently. The 3D model also revealed geological features previously unknown, notably the existence of an inactive fault. The proposed method can be applied to integrate and jointly interpret any kind of multidisciplinary datasets in a wide range of geoscientific problems, including natural resource exploration, geological storage, environmental monitoring, civil engineering practice, and hazard assessment. • The joint interpretation of geophysical data improves subsurface characterization. • Machine learning makes the interpreter's job easier while reducing bias. • Combined supervised and unsupervised learning makes the most of data. • Machine learning to enhance geological interpretation of historic datasets. • Geophysical surveys should be planned anticipating their subsequent integration.

Cave and Speleothem Science: From Local to Planetary Scales

Caves occur everywhere on our planet, from the tropics to the high latitudes and from below sea level to alpine settings. Cave morphologies provide clues to their formation mechanisms, and their iconic mineralogical features—stalagmites and stalactites—carry a wealth of paleoenvironmental information encoded in their geochemistry and mineralogy. Recent work demonstrates a striking improvement in our ability to decode these paleoenvironmental proxies, and dramatic geochronological advances enable higher resolution records that extend further back in geologic time. Cave research addresses an ever-increasing range of geoscience problems, from establishing the timing and mechanisms of climate change to uncovering detailed records of geomagnetic field behavior.

P-Kernel Stein Variational Gradient Descent for Data Assimilation and History Matching

A Bayesian method of inference known as “Stein variational gradient descent” was recently implemented for data assimilation problems, under the heading of “mapping particle filter”. In this manuscript, the algorithm is applied to another type of geoscientific inversion problems, namely history matching of petroleum reservoirs. In order to combat the curse of dimensionality, the commonly used Gaussian kernel, which defines the solution space, is replaced by a p-kernel. In addition, the ensemble gradient approximation used in the mapping particle filter is rectified, and the data assimilation experiments are re-run with more relevant settings and comparisons. Our experimental results in data assimilation are rather disappointing. However, the results from the subsurface inverse problem show more promise, especially as regards the use of p-kernels.

Influence of water on the physical properties of olivine, wadsleyite, and ringwoodite

Abstract. The incorporation of water in nominally anhydrous minerals plays a crucial role in many geodynamic processes and evolution of the Earth and affects the physical and chemical properties of the main constituents of the Earth's mantle. Technological advances now allow the transport properties of minerals to be precisely measured under extreme conditions of pressure and temperature (P and T) that closely mimic the P–T conditions throughout much of the Earth's interior. This contribution provides an overview of the recent progress in the experimental studies on the influence of water on physical properties (i.e., diffusivity, electrical conductivity, thermal conductivity, sound velocity, and rheology) of olivine, wadsleyite, and ringwoodite together with their applications. In particular, consistency among various experimental data is investigated, discrepancies are evaluated, and confusions are clarified. With such progress in the experimental determination of transport properties of major mantle minerals, we can expect new insights into a broad range of geoscience problems. Many unresolved issues around water inside Earth require an integrated approach and concerted efforts from multiple disciplines.

A department terminated.

I logged onto a video call last month, not knowing what to expect. Faculty members in my department had just 1 day's notice of the meeting with our dean, which made me wonder, “Is it another budget cut—or worse?” After a quick apology, the dean pivoted to the news: Given the crunch on university finances amid the COVID-19 pandemic, he'd made a difficult decision to terminate the geology department. If we wanted to keep our jobs, we'd have to find another department willing to take us in. His words hit me like a ton of bricks. For 20 years—ever since I received tenure—I assumed I'd always have a job if I kept up my teaching and research. But the pandemic had brought long-standing problems to a head. > “The quality of our research wasn't enough to save us.” Three years ago, the same dean had asked us for help. His budget was in the red because enrollments were falling and a new budget system had moved money out of his college. The geology department had a few large, popular courses, but many of our upper-level classes were highly specialized, attracting only a handful of students. Fewer than 10 students a year majored in geology. We needed larger class sizes and more students to enroll as majors, the dean told us. We convened faculty meetings to discuss how to move forward. Our curriculum hadn't changed much in decades. Our students learned how to identify rocks and stare down microscopes, but they weren't much exposed to many of the more pressing problems in geosciences, such as climate change and groundwater pollution. Some of us, including me, wanted to overhaul the curriculum. But others argued against abandoning our focus on traditional skills and concepts. We ended up making changes only around the edges. We added new elective courses on climate, medical geology, and extraterrestrial life, which attracted hundreds of students—many from other departments. But we didn't change the course requirements for geology majors, and many professors continued to teach the same material in the same way. The number of majors didn't budge. Then came COVID-19. We stayed open and taught students on campus, but enrollments continued to fall. We expected administrators to make cuts, but we did not think entire departments would be axed—especially not ours. Our seven professors collectively bring in hundreds of thousands of federal research dollars every year and publish dozens of papers. But the overhead from our grants did not make up for the lack of tuition dollars. The quality of our research wasn't enough to save us. After the meeting with our dean, my mind raced. What would become of our labs, our grad students, and all the projects in full swing? I had a secondary faculty appointment in a different college on campus, so I might be able to work there and avoid leaving the town where I'd lived for 27 years. But I didn't want to leave my geology colleagues. I wanted a solution that kept us together as a team. Two days after the dean's call, I spoke with a university administrator who prodded me to not give up and asked, “What's your vision?” I spent the weekend speaking with my colleagues and brainstorming a plan. Together, we imagined a wholly new earth science curriculum, one that would prepare students for the challenges of today. We'd rethink all of our courses from scratch, focusing on problems that students—and their potential employers—care about. For instance, instead of simply teaching how rocks crack and weather away, we'd explore how those cracks affect the movement of polluted groundwater. Our new vision looks to the future and leaves the past behind, something we could never bring ourselves to do before the prospect of termination forced us to spring into action. We're not sure how the university will react to our plan. Hopefully, our efforts aren't coming too late. I hope academics elsewhere can learn from our mistakes and take a hard look at their own offerings to make sure they're serving the current generation of students. It's hard to overhaul a department and develop entirely new courses. But it's much harder if you wait until a crisis to act.

Cretaceous to Recent net continental uplift from paleobiological data: Insights into sub-plate support

AbstractThere are many geoscience problems for which constraining histories of uplift or subsidence of Earth’s surface is of direct or indirect importance, for example reconstructing tectonics, mantle convection, geomorphology, sedimentary and chemical flux, biodiversity, glacio-eustasy, and climate change. The least equivocal constraints on timing and amplitude of vertical motions on geological timescales come from the distribution of rock formed in shallow marine environments. However, obtaining enough observations at sufficiently large spatial and temporal scales (∼100–10,000 km, ca. 1–100 Ma) to constrain histories of regional topographic evolution remains challenging. To address this issue, we adapted modern inventories of paleobiological and paleoenvironmental data to generate a new compilation of >24,000 spot measurements of uplift on all continents and numerous oceanic islands. Uncertainties associated with paleobathymetry, post-deposition compaction, and glacio-eustasy are assessed. The compilation provides self-consistent and, in places, high-resolution (<100-km-length scale, <1 Ma) measurements of Cretaceous to Recent (post-deposition) net uplift across significant tracts of most continents. To illustrate how the database can be used, records from western North America and eastern South America are combined with geophysical observations (e.g., free-air gravity, shear, and Pn-wave tomography) and simple isostatic calculations to determine the origins of topography. We explore how lithospheric thinning and mantle thermal anomalies may generate uplift of the observed wavelengths and amplitudes. The results emphasize the importance of large inventories of paleobiological data for understanding histories of tectonic and mantle convective processes and consequently landscapes, climate, and the environment.

Tackling 21st Century Geoscience Problems with Machine Learning

A new cross-journal special collection invites contributions on how machine learning can be used for solid Earth observation, modeling and understanding.

Recent Advances in Geomathematics in Croatia: Examples from Subsurface Geological Mapping and Biostatistics

Geomathematics is extremely important in geosciences, particularly in the geology. The key for any geomathematical analysis is the definition of a typical model to be applied for further prognosis, either through deterministic or stochastic approaches. The selection of the appropriate procedure is presented in this paper. Two different geomathematical subfield datasets were used in subsurface geological mapping and palaeontology and different biostatistics applications, representing important geomathematical subfields in the Croatian geology. The different subsurface interpolation methods tested, validated and recommended for application were used to obtain the best possible outcome in reservoir modelling, in the cases with small datasets. Cross-validation may be chosen as the main selection criteria, applied to the Croatian part of the Pannonian Basin System (CPBS). Recent advances in biostatistics applied in palaeontology and case studies from Croatia are also presented, where biometric studies are of significant importance in fossil biota. Data, methods and problems in geosciences are vast subjects, and address a wide spectrum of fundamental science. Because geology includes subsurface and surface geology, and very different datasets regarding variable and number of data, we have chosen here two representative case study groups with original samples from Northern Croatia. Subsurface mapping has been presented on limited petrophysical datasets from the Northern Croatian, Miocene, hydrocarbon reservoirs. Biostatistics have been presented on very different samples, allowing us to achieve paleoenvironmental reconstructions of the size of relevant fossils, such as dinosaurs or other species and their paleoenvironments. All examples highlight examples of the valuable application of geomathematical tools in geology. The results, cautiously validated and correlated with other, non-numerical (indicator, categorical) geological knowledge, are of enormous assistance in creating better geological models.

Optically Stimulated Luminescence Sensitivity of Quartz for Provenance Analysis.

Finding the source or provenance of quartz grains occurring in a specific location allows us to constrain their transport pathway, which is crucial information to solve diverse problems in geosciences and related fields. The optically stimulated luminescence (OSL) sensitivity (light intensity per unit mass per unit radiation dose) has a high capacity for discrimination of quartz sediment grains and represents a promising technique for provenance analysis. In this study, we tested the use of quartz OSL sensitivity (ultraviolet emission) measured under different preheating temperatures and with blue light stimulation at room temperature (~20 °C) for sediment provenance analysis. Quartz OSL sensitivity measured at 20 °C is positively correlated with the sensitivity of an OSL signal measured using procedures (preheat at 190 °C for 10 s, blue stimulation at 125 °C and initial 1 s of light emission) to increase the contribution of the fast OSL component, which has been successfully applied for sediment provenance analysis. The higher OSL signal intensity measured without preheating and with light stimulation at room temperature allows the use of lower given doses, thus reducing measurement time. Additionally, the OSL sensitivity measured at 20 °C in polymineral silt samples of a marine sediment core is also suitable for provenance analysis, as demonstrated by comparison with other independent proxies. OSL signals obtained through light stimulation at room temperature have thus the potential to considerably expand measurement possibilities, including in situ measurements using portable OSL readers.

Introduction to the Special Collection: Magnetism in the Geosciences—Advances and Perspectives

AbstractMagnetic property research in the geosciences has seen an enormous development during the most recent decade. The underlying special collection published in two AGU journals: Journal of Geophysical Research—Solid Earth and Geochemistry, Geophysics, Geosystems is an outcome of the 11th International Conference on Rock Magnetism, held in Utrecht (The Netherlands) in July 2017 and reflects this full breadth. We recognize four groups in which individual contributions are gathered. (1) Recent advances in fundamental rock and mineral magnetic methods, including discussions on how to assess magnetic grain size, which is centrally important to magnetic stability over geologic timescales and high‐fidelity paleomagnetic recording. (2) Geomagnetism with an emphasis on late Pleistocene paleosecular variation and relative paleointensity records, relevant to geomagnetic field models. (3) Reassessments of classic paleomagnetic field tests and examples of tectonic and magnetostratigraphic applications. (4) The merit of magnetic property analysis in solving geoscience problems, both in soft and hard rocks.

Sparse isocon analysis: A data-driven approach for material transfer estimation

Isocon analysis has been widely applied to various geoscientific problems as a simple standard tool for quantitative estimation of material transfer. Despite its usefulness, similar to all material transfer calculations, this method generally requires the presumptive specification of immobile elements or the assumption of conservation of mass or volume. However, the validity of such assumptions is particularly controversial. Here we propose a novel data-driven method that automatically estimates the mass gain or loss of elements based on compositional data of multiple samples that have been altered from the original rock without assuming immobile elements. The proposed method uses a mathematical framework, called sparse modeling, that can extract essential information from high-dimensional datasets based on the sparsity of the system. In this case, it is assumed that some elements show higher immobility than others (i.e., the material transfer of such elements is near zero). By optimizing the evaluation function, the immobile elements are automatically selected. By inputting only the bulk compositional datasets, the user can obtain the material gain or loss with total mass change ratio for each sample relative to the reference (original) rock. The effectiveness of the method is validated and discussed using synthetic and natural sample data. Software packages are available from the authors in MATLAB function and Excel workbook forms.

Permeability prediction and diagenesis in tight carbonates using machine learning techniques

Abstract Machine learning techniques have found their way into many problems in geoscience but have not been used significantly in the analysis of tight rocks. We present a case study testing the effectiveness of artificial neural networks and genetic algorithms for the prediction of permeability in tight carbonate rocks. The dataset consists of 130 core plugs from the Portland Formation in southern England, all of which have measurements of Klinkenberg-corrected permeability, helium porosity, characteristic pore throat diameter, and formation resistivity. Permeability has been predicted using genetic algorithms and artificial neural networks, as well as seven conventional ‘benchmark’ models with which the machine learning techniques have been compared. The genetic algorithm technique has provided a new empirical equation that fits the measured permeability better than any of the seven conventional benchmark models. However, the artificial neural network technique provided the best overall prediction method, quantified by the lowest root-mean-square error (RMSE) and highest coefficient of determination value (R2). The lowest RMSE from the conventional permeability equations was from the RGPZ equation, which predicted the test dataset with an RMSE of 0.458, while the highest RMSE came from the Berg equation, with an RMSE of 2.368. By comparison, the RMSE for the genetic algorithm and artificial neural network methods were 0.433 and 0.38, respectively. We attribute the better performance of machine learning techniques over conventional approaches to their enhanced capability to model the connectivity of pore microstructures caused by codependent and competing diagenetic processes. We also provide a qualitative model for the poroperm characteristics of tight carbonate rocks modified by each of eight diagenetic processes. We conclude that, for tight carbonate reservoirs, both machine learning techniques predict permeability more reliably and more accurately than conventional models and may be capable of distinguishing quantitatively between pore microstructures caused by different diagenetic processes.

Machine Learning for the Geosciences: Challenges and Opportunities

Geosciences is a field of great societal relevance that requires solutions to several urgent problems facing our humanity and the planet. As geosciences enters the era of big data, machine learning (ML)—that has been widely successful in commercial domains—offers immense potential to contribute to problems in geosciences. However, geoscience applications introduce novel challenges for ML due to combinations of geoscience properties encountered in every problem, requiring novel research in machine learning. This article introduces researchers in the machine learning (ML) community to these challenges offered by geoscience problems and the opportunities that exist for advancing both machine learning and geosciences. We first highlight typical sources of geoscience data and describe their common properties. We then describe some of the common categories of geoscience problems where machine learning can play a role, discussing the challenges faced by existing ML methods and opportunities for novel ML research. We conclude by discussing some of the cross-cutting research themes in machine learning that are applicable across several geoscience problems, and the importance of a deep collaboration between machine learning and geosciences for synergistic advancements in both disciplines.

Convolutional neural networks as aid in core lithofacies classification

Artificial intelligence methods have a very wide range of applications. From speech recognition to self-driving cars, the development of modern deep-learning architectures is helping researchers to achieve new levels of accuracy in different fields. Although deep convolutional neural networks (CNNs) (a kind of deep-learning technique) have reached or surpassed human-level performance in image recognition tasks, little has been done to transport this new image classification technology to geoscientific problems. We have developed what we believe to be the first use of CNNs to identify lithofacies in cores. We use highly accurate models (trained with millions of images) and transfer learning to classify images of cored carbonate rocks. We found that different modern CNN architectures can achieve high levels of lithologic image classification accuracy (approximately 90%) and can aid in the core description task. This core image classification technique has the potential to greatly standardize and accelerate the description process. We also provide the community with a new set of labeled data that can be used for further geologic/data science studies.