Applied Geochemistry最新文献

Deep learning algorithms have become a cutting-edge technology for mining geochemical survey data to identify geochemical patterns related to mineralization. Similarities in the origins of the same types of mineral deposits may result in similarities in the observed geochemical anomalies to some extent. However, image-based models have limited ability to model the relationships between samples because fixed-size pixel patches are not connected. Graphs have enormous potential to capture spatial information, which can model the complex nonlinear spatial relationship between vertices and edges and effectively measure the spatial relationships between metallogenic information and geochemical survey sites. The graph-based model considers both labeled and unlabeled data as inputs, allowing the relationships between individual samples to be incorporated into the network. In this study, we used a graph-based model, the graph attention network (GAT), to recognize geochemical anomalies associated with gold mineralization in the Xiaoqinling–Xiong'ershan region of the western Henan Province, China. The GNNExplainer, a method for explaining the predictions of graph-based deep learning tasks, was used to determine the importance of the input features of the trained GAT model. The results indicated that the model regards Bi, Cd, Mn, Zn, and Mo as important geochemical elements for mineralization. Meanwhile, a comparative study of GAT with a convolutional neural network suggests that the graph-based model can be applied to effectively identify mineralization-related geochemical patterns, and geochemical anomalies recognized by GAT are more representative of gold mineralization.

Machine learning methodologies can provide insight into Brønsted-Guggenheim-Scatchard specific ion interaction theory (SIT) parameter values where experimental data availability may be limited. This study develops and executes machine learning frameworks to model the SIT interaction coefficient, ε. Key findings include successful estimations of ε via artificial neural networks using clustering and value prediction approaches. Applicability to other chemical parameters is also assessed briefly. Models developed here provide support for a use-case of machine learning in geologic nuclear waste disposal research applications, namely in predictions of chemical behaviors of high ionic strength solutions (i.e., subsurface brines).

Understanding the difference in groundwater flow between glacial and interglacial periods is crucial for predicting the impact of future climate changes on groundwater movement. This study assesses the difference in groundwater flow between the last glacial period (LGP) and the postglacial period (PGP) in fractured mudstones of the Horonobe area, Japan, by combining the data for stable isotopes (δD and δ¹⁸O) and Cl⁻ concentration of fracture and pore waters with radiocarbon (¹⁴C) age. The isotopic compositions of fractures and pore waters indicate that groundwater at 28–250 m deep in a borehole closest to the recharge area comprises meteoric water, recharged under the same climates as the present. The fracture water has isotopic compositions more similar to meteoric water than the matrix pore water near the fracture. The ¹⁴C age of fracture water suggests meteoric water recharge during the PGP. At greater depths in the borehole and sampling points in other boreholes, the isotopic compositions indicate the mixing of glacial meteoric and altered connate water, with the fracture water having comparable isotopic compositions with the matrix pore water. The recharge timing of meteoric water is inferred to be the LGP or before based on ¹⁴C dating. These results suggest that the meteoric water recharged during the PGP flows at a shallow depth, whereas the meteoric water recharged during the LGP intruded to greater depths. This result is consistent with previous inferences from surface geophysical and geological surveys that the depths of local valleys during the LGP were greater by < 50 m than the present ones and enhanced the downward hydraulic gradient. Combining the chemical and isotopic compositions of groundwater with ¹⁴C age helps assess the groundwater flow during the LGP and PGP in fractured rocks.

Available equilibrium constant data for reactions of germanium with inorganic ligands in aqueous solution have been critically evaluated. Even though the relevant literature is sparse and mostly rather old, we have established a working thermodynamic description of germanium in aqueous, multicomponent media for its most important interactions with inorganic ligands. These thermodynamic parameters will be useful in environmental and (eco)toxicology studies. However, within the limitations of the presently available literature, significant uncertainties are inescapable. The implications for thermodynamic modelling in general are far-reaching.

Historically, surface complexation model (SCM) constants and distribution coefficients (K_d) have been employed to quantify mineral-based retardation effects controlling the fate of metals in subsurface geologic systems. Our recent SCM development workflow, based on the Lawrence Livermore National Laboratory Surface Complexation/Ion Exchange (L-SCIE) database, illustrated a community FAIR data approach to SCM development by predicting uranium(VI)-quartz adsorption for a large number of literature-mined data. Here, we present an alternative hybrid machine learning (ML) approach that shows promise in achieving equivalent high-quality predictions compared to traditional surface complexation models. At its core, the hybrid random forest (RF) ML approach is motivated by the proliferation of incongruent SCMs in the literature that limit their applicability in reactive transport models. Our hybrid ML approach implements PHREEQC-based aqueous speciation calculations; values from these simulations are automatically used as input features for a random forest (RF) algorithm to quantify adsorption and avoid SCM modeling constraints entirely. Named the LLNL Speciation Updated Random Forest (L-SURF) model, this hybrid approach is shown to have applicability to U(VI) sorption cases driven by both ion-exchange and surface complexation, as is shown for quartz and montmorillonite cases. The approach can be applied to reactive transport modeling and may provide an alternative to the costly development of self-consistent SCM reaction databases.

The Mediterranean Sea is a unique archive of past climatology, as this region is sensitive to both global climate fluctuations and manmade disturbances. Our understanding of these processes is however, limited, due to the absence of high-resolution multi-element proxies in this region. Here, I present the first high-resolution record of trace-elements (TE) in Mediterranean aragonite veremtid reefs, spanning the last millennium. Modern vermetid TE contents are mostly in agreement with other marine biogenic aragonites, reinforcing the potential use of veremtid reefs as paleo-environmental proxies. The down-core U/Ca and Sr/Ca fluctuations resemble the main climatic events of the last millennium, potentially linked to changes in sea surface temperature (SST). Pb and Cd records are associated with anthropogenic pollution and demonstrate trends related to growing industrial activity in the Anthropocene. Carbonate records of Al, Fe and Rb are attested as potential proxies to infer terrestrial inputs. Fe/Ca is used to decouple dust from riverine sources, while Rb/Ca is proposed for differentiation between sources of dust. The sub-millennial multi-element study enables the observation of independent environmental patterns within a single archive, and the decoupling of anthropogenic imprint from natural variability.

The fluctuations of the Rare Earth Elements and Yttrium (REE-Y) concentrations on exhumed carbonate normal fault scarps may reveal the number and size of paleoearthquakes that exposed the scarp subaerially. This is because, prior to each large-magnitude earthquake, narrow (<50 cm) sections of the fault plane which are in direct contact with the soil become enriched in REE-Y before they are exhumed co-seismically, together with deeper, non-enriched, scarp sections. Following exhumation, depletion in REE-Y commences on both the enriched (i.e. ‘soil rupture zone’) and non-enriched (i.e. ‘rock rupture zone’) scarp sections. Although these processes are commonly described to occur on carbonate scarps, the mechanisms through which they operate remains poorly understood. Here, we present a series of laboratory tests that mimic the natural process of REE-Y enrichment/depletion to elucidate the mechanism of REE-Y impregnation. Our results indicate a fast uptake of REE-Y by the carbonate plane, when in contact with soil, either as (REE, Y)₂(CO₃)₃ precipitate or by adsorption on calcite surfaces. The source of REE-Y in soil solution is released in a “pulses” due to alternations of dry and wet periods, characteristic of Mediterranean climatic conditions. Organic matter oxidation during the first rain events, triggers the Mn reductive dissolution and the release of REE-Y into the soil solution. The pH decrease due to organic matter dissolution is buffered by calcite, especially in the vicinity of the scarp, where calcite dissolution and re-precipitation occurs with a marked pH oscillation between 9.3 and 7.7. Further, comparison of these results with empirical data from three co-seismically exhumed fault scarps in Greece and Italy places quantitative constraints on the timing of these processes: the REE-Y enrichment within the ‘soil rupture zone’ may reach a maximum of ∼50% in about 500 years (+0.53 μg/kg/year), while the REE-Y depletion from the scarp is slow (−0.021 μg/kg/year), with a maximum recorded retention time of ∼16 ka. These enrichment and depletion characteristics work together to preserve paleoearthquake signal on carbonate scarps. Thus, this methodology is a valuable tool for quantifying the number of past earthquakes on carbonate fault scarps and allows more targeted use of expensive dating techniques (i.e. with cosmogenic nuclides) in order to derive the precise timing of these paleoearthquakes.

The Campanian source strata of the Tano Basin, Ghana, have garnered exploration interest based on the recent discoveries of hydrocarbon shows in some offshore wells. This study reports on the bulk geochemical composition (Total organic carbon content and Rock-Eval pyrolysis parameters), molecular markers, and stable carbon isotope (δ¹³C) composition of eleven Campanian sediments from TP–1 and Ankobra–1 well, a shallow and deep-water well in the Tano Basin respectively. Based on the analysis of total organic carbon content and Rock-Eval pyrolysis parameters (S1, S2, and Tmax), it was found that the TP-1 Campanian sediments contain indigenous hydrocarbons. Contrariwise, the Campanian samples from the Ankobra-1 well are mainly composed of non-indigenous hydrocarbons, possibly originating from offshore drilling fluid contaminants. Distribution of saturated and aromatic hydrocarbons, including normal alkanes, acyclic isoprenoids, steranes, naphthalenes, phenanthrenes, dibenzothiophene, dibenzofuran, fluorenes, and their alkyl derivatives, as well as stable carbon isotope (δ¹³C) values, were analyzed to gain insight into the biological source input, depositional environment, and thermal maturity of the organic matter (OM) in the sediments. The molecular and isotope results indicated that the OM in the indigenous hydrocarbons is from mixed organic sources, with a predominance of terrestrial plants, whereas lower marine organisms dominate the non-indigenous hydrocarbons. Moreover, the molecular ratios obtained from our analysis indicated that the OM present in the indigenous hydrocarbons was deposited under suboxic conditions in a marginal marine environment. Contrarily, the non-indigenous hydrocarbons exhibit a strong association with marine carbonate facies. Additionally, the maturity assessment based on biomarker ratios, indicated that the OM within the Campanian sediments from TP-1 well is in the early phase of oil generation whereas the OM in the non-indigenous hydrocarbons is in the peak-late mature phase. In conclusion, our geochemical findings have revealed the hydrocarbon potential and heterogeneity of hydrocarbons within the Tano Basin's Campanian source strata induced by non-indigenous hydrocarbons. The geochemical data obtained from our study provide evidence of the potential influence of non-indigenous hydrocarbons on the misrepresentation of source input, depositional environment, and thermal maturity of organic matter (OM) in source rock analysis.

Early detection of CO₂ leakage through monitoring is important to ensure long-term safety for geologic carbon storage (GCS). A geochemically informed leak detection (GILD) model has been developed for groundwater chemistry monitoring at CO₂ injection sites. The GILD model integrates a geochemical model that simulates fluid chemistry changes in CO₂ leakage events and a Bayesian belief network (BBN) model that evaluates monitoring observations to identify leakages. The geochemical model is implemented using Geochemists’ Workbench to assess fluid chemistry changes as a result of small CO₂ leakage in an above-zone monitoring interval (AZMI) formation with varying mineral assemblages and background fluids. Response functions are fitted to the output of the geochemical model and are translated to conditional probabilities in the BBN model. The BBN model gives operational prediction of the leak probability given a set of groundwater monitoring measurements and the probability of detecting a leak at a given magnitude. The detection capabilities of multiple monitoring parameters are compared. For aquifers that contain calcite, it is valuable to incorporate other monitoring parameters with pH to increase the sensitivity of detection. For aquifers with no calcite, pH alone is a sensitive parameter. This research illustrates a method of identifying CO₂ leakage into aquifers with both geochemical and statistical tools.

This work presents a three-year continuous observation of ozone (O₃) and fine particle (PM_2.5), as well as their precursors in Qingdao, China, from September 2018 to August 2021. The annual concentrations of O₃ and PM_2.5 were measured as 35.9–44.5 ppbv and 31.6–34.2 μg m⁻³, respectively. Analysis of the interannual variations of O₃ and PM_2.5 concentration indicated the effective control measures for PM_2.5 and O₃ pollution in Qingdao in recent years. Nevertheless, we still observed 85 O₃ episodes and 80 PM_2.5 episodes during the whole observation. And it was found that the fraction of secondary inorganic aerosols (SIA) in PM_2.5 mass significantly increased during PM_2.5 episodes. By employing the Lagrangian photochemical trajectory model (LPTM), we investigated the roles of local production and regional transport (i.e., short-distance transport and long-distance transport) that play in the O₃ and SIA formation in these episodes. The contribution of local production, short-distance transport, and long-distance transport to O₃ concentration was calculated as 36%, 25%, and 39%, respectively. Long-distance transport seemed to play a more significant role in the SIA formation, accounting for 56% of SIA concentration. Moreover, the calculation results of the relative incremental reactivity (RIR) showed that both O₃ and SIA formation were mainly affected by NO_x and VOCs emissions during the process of long-distance transport. The observed unexpectedly high contribution of long-distance transport to O₃ and SIA formation suggests that more extensive regional joint prevention and control policies on NO_x and VOCs emissions are warranted to mitigate secondary air pollution.