60 research outputs found
Evidenze geofisiche di emissioni di fluidi nel Golfo di Trieste (Nord Adriatico)
No full textThe presence of fluids in the sediments of the Gulf of Trieste (Northern Adriatic) produces surface occurrences that have been identified by high resolution seismic data. In particular, south-west of Grado, six sites have been pinpointed with evidence of accumulation of fluids within the sediment, dispersion of fluids in the water column, and presence on the sea bottom of rocky outcrops whose genesis is assumed to be linked to the precipitation of calcium carbonate methane-derived
Artificial intelligence applications for accurate geothermal temperature prediction in the lower Friulian Plain (north-eastern Italy)
No full textGeothermal energy as a sustainable and clean energy source depends on the accurate estimation of reservoir temperatures. Understanding aquifer temperatures is crucial for optimizing low-enthalpy geothermal system exploitation. Advances in predictive algorithms can improve geothermal efficiency, while conventional methods of indirect temperature measurement and assumptions in geochemical analysis lead to uncertainties. As a solution, this study presents a comprehensive evaluation of six machine learning algorithms including eXtreme gradient boosting (XGBoost), decision tree, generalized regression neural network, extreme randomized trees, radial basis function, and elastic net. We employed essential performance metrics including coefficient of determination (R2) score, root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and variance accounted for (VAF) to elucidate their predictive accuracy and generalization potential in the lower Friulian Plain (north-eastern Italy) where a geothermal reservoir is present. Among the algorithms scrutinized, XGBoost emerges as a predictive exemplar, achieving a remarkable R2 score of 0.9930 on the test dataset, with consistently low RMSE of 0.788, MAE of 0.587, MAPE of 1.909, and high VAF of 99.30, reaffirming its exceptional precision and robustness. It is worth noting that the other four models show slightly weaker performance than XGBoost, while elastic net shows moderate predictive power, which illustrates the complexity of the database. The Wilcoxon signed-rank test confirmed the superior performance of XGBoost in estimating geothermal temperatures compared to other algorithms, with statistical evidence supporting its precision and reliability. A Monte Carlo simulation for uncertainty analysis underlined the importance of model selection, accuracy and uncertainty management in the planning of geothermal projects in the lower Friulian Plain. A sensitivity analysis was performed to identify the main factors influencing the temperature prediction. Among the parameters considered, bicarbonate the highest significance at 0.51, which is essential for accurate temperature prediction because of its buffering capacity which directly influences water’s thermal properties. Magnesium and electrical conductivity each contribute with 0.11, also play significant roles due to their impact on the water’s heat retention and distribution capabilities. Water depth, with a value of 0.08, also has a significant influence on the temperature profiles in prediction models. In summary, the accurate prediction of XGBoost for the temperature of aquifer in carbonate reservoirs in the lower Friulian Plain, underline its value for optimizing geothermal resources and highlight most important influences on temperature
The Ross Sea formation: enquiring the sensitivity of basin architecture to prior conditions, with numerical models and a parameter search
No full textThe basins composing the 1000-km wide West Antarctica Rift System (WARS), derived from extensional dynamics lasting from the Cretaceous to the Middle Neogene, bear evidence of a peculiar evolution through time: a transition from a diffuse to a localized thinning style and a migration of the focus of deformation, which likely progressed towards the cratonic domains of West Antarctica. Using the current observations, we aim at identifying which inherited starting conditions [1] result in outcomes compatible with the present-time structures and which do not allow so, unless other factors are accounted for.
To this aim, we turn to an extensive grid search in the parameter space, running a large number of forward numerical models to cover the possible permutations of parameters under test. We use the open source Underworld2 code [2] with a simplified scheme of starting conditions and kinematics boundaries, for lithospheric-scale 2-D thermomechanical models. We analyse the results obtained by changing a great number of parameters, including initial geometries of the crust and lithosphere, different rheologies, inherited structures, such as strain-weakening scars and thermal remnants of slabs.
We identify that a high crustal thickness (more than 45 km) is required to accommodate the first rifting phase (170 km ca. of cumulated extension, [3]) without producing crustal necking and eventual ocean formation. Parameters that favour a weaker strength profile, chiefly temperature (due to a thicker crust and/or a shallow lithosphere-asthenosphere boundary), are also required to avoid an early transition to localized deformation, in agreement with previous studies [4]. Smaller scale features, such as partition in multiple sub-basins, require additional factors, such as inherited weak-zone seeds (“scars”) in the crust and mantle, which are likely remnants of previous compressive phases [5].
[1] Perron, P., Le Pourhiet, L., Guiraud, M., Vennin, E., Moretti, I., Portier, É., & Konaté, M. (2021). Control of inherited accreted lithospheric heterogeneity on the architecture and the low, long-term subsidence rate of intracratonic basins. BSGF - Earth Sciences Bulletin, 192. https://doi.org/10.1051/bsgf/2020038
[2] Mansour, J., Giordani, J., Moresi, L., Beucher, R., Kaluza, O., Velic, M., Farrington, R., Quenette, S., & Beall, A. (2020). Underworld2: Python Geodynamics Modelling for Desktop, HPC and Cloud. Journal of Open Source Software, 5(47), 1797. https://doi.org/10.21105/joss.01797
[3] Brancolini, G., Busetti, M., Coren, F., De Cillia, C., Marchetti, M., De Santis, L., Zanolla, C., Cooper, A.K., Cochrane, G.R., Zayatz, I., Belyaev, V., Knyazev, M., Vinnikovskaya, O., Davey, F.J., Hinz, K., 1995. ANTOSTRAT Project, seismic stratigraphic atlas of the Ross Sea, Antarctica. In: Cooper, A.K., Barker, P.F., Brancolini, G., (Eds.), Geology and Seismic Stratigraphy of the Antarctic Margin. Antarctic Research Series, vol. 68, https://doi.org/10.1029/AR068
[4] Huerta, A. D., & Harry, D. L. (2007). The transition from diffuse to focused extension: Modeled evolution of the West Antarctic Rift system. Earth and Planetary Science Letters, 255(1–2), 133–147. https://doi.org/10.1016/j.epsl.2006.12.011
[5] Talarico, F., Ghezzo, C., & Kleinschmidt, G. (2022). The Antarctic Continent in Gondwana: a perspective from the Ross Embayment and Potential Research Targets for Future Investigations. In Antarctic Climate Evolution (pp. 219–296). Elsevier. https://doi.org/10.1016/B978-0-12-819109-5.00004-
Morphoneotectonics and lithology of the eastern sector of the Gulf of Trieste (NE Italy)
No full textThe paper aims to describe and map the geomorphological and lithological features of the Gulf of Trieste and its eastern coasts and to define its neotectonic behaviour by means of the analysis of the morphoneotectonic evidences.
The final map, produced at a scale of 1:30,000, shows the outcome of field investigations carried out along the coast and the sea bottom and a detailed geomorphological classification of the coastline. Published and new data coming from the analysis of archeological remains, geomorphological and sedimentological sea-level indicators and geophysical researches are discussed in order to provide a complete overview of the study area
The role of inheritance and kinematics in shaping the West Antarctica Rift System (WARS): A parametric study
No full textExtensional tectonics in continental settings usually results in lithospheric stretching of narrow (e.g. Upper Rhine Graben), as well as wide regions (e.g. Basin and Range). Some rift systems bear evidence of a transition between the two styles: the West Antarctica Rift System (WARS) have likely progressed from diffuse to focused rifting (Cretaceous - Middle Neogene). The system currently covers a length of 1000 km, at the boundary between East and West Antarctica and is composed of four main basins (Victoria Land Basin, Central Trough, Northern Basin and Eastern Basin). The deformation pattern and available geological reconstructions suggest that, at least for some part of the rifting, the extension occurred concurrently in multiple sections.Inheritance of prior structural, thermal, and rheological heterogeneities is likely a controlling factor in this evolution [1]. Consequently, with the goal of identifying the most likely initial conditions, we designed a series of 2-D numerical models, analysing the effect of variations in the temperature field, rheology, accumulated strain, distribution of extensional pulses on the basins’ geometry. To this aim, we used the open source Underworld2 [2] and BGR-02 and ACRUP2 profiles [3], as 2-D analogue of the WARS structures in the Southern Ross Sea.The results show that the models most consistent with the observations are those that include inherited weakness zones at the cratonic boundary. Early onset of focused extension often occurs, with high sensitivity to the pattern of inherited weakening.[1] Perron et al. (2021). DOI:10.1051/bsgf/2020038[2] Mansour et al. (2020). DOI:10.21105/joss.01797[3] Trey et al. (1999). DOI:10.1016/S0040-1951(98)00155-
Diffuse Cretaceous-Cenozoic rifting in the Southern Ross Sea: the influence of inheritance and kinematics
No full textContinental Rift systems often involve narrow regions, which accommodate all the stretching. In some cases, the initial extension occurs with a diffuse style and may successively produce a narrow rift. An example is the West Antarctica Rift System, bearing evidence of the concurrent formation of multiple basins normal to the rift axis. This rift system has undergone extension between the Cretaceous and the middle Neogene age (105 to 11 Ma [1, 2]), due to the sea floor spreading in the northwestern Ross Sea. It is composed of three main basins (Victoria Land Basin, Central Trough, and Eastern Basin), which cover a present-day length of 900-1000 km, encompassing the lateral contact between the cratonic domains of East Antarctica and West Antarctica Phanerozoic lithosphere. The different basins, bounded by structural highs, exhibit significant variations in the thickness and thinning of the underlying crust and lithosphere. This multiple-basin pattern suggests that, at least for some part of the rifting, the deformation occurred in a diffuse way, instead of being localized in a small portion of the rift system [3].
The factors controlling these deformation styles have been identified in the inheritance of structures and thermal/rheological heterogeneities [4], which acted concurrently with the extensional kinematics in shaping the present-day rift architectures. Therefore, an improved knowledge on how different thermo-structural initial conditions (e.g. lateral contacts, thermal transients, accumulated strain softening) influence the outcome of rifting may help identify the most likely state at the onset of rifting. To this purpose, we implement a series of numerical models, testing several starting structural conditions (rheology, temperature, prior damage) and distribution of extensional velocity (a single phase or multiple pulses, for the same total extension) that could trigger this peculiar diffuse deformation pattern.
To build a 2-D simplified geometry of the structures of the rift system, we took as a reference the seismic profiles BGR-02 and ACRUP2, normal to the rift axis, along the 77° S parallel [5]. We assumed an initial crustal thickness of about 50 km and a kinematic pattern consisting of two main distinct extension phases, covering the Cretaceous-Cenozoic interval [1, 6].
Modelling was carried out using the open source Underworld2 code [7], which relies on Lagrangian integration point finite element approach and provides a Python API to construct, run, and visualize the output of geodynamic models. The results show that the models that are more consistent with the observations require the existence of peculiar a-priori inherited features. In addition to the role of inheritance, diffuse patterns are favoured, for the same extension amount, by slow and long-lasting rifting phases, with respect to fast and short time pulses.
This work was carried out in the context of PNRA project "Onset of Antarctic Ice Sheet Vulnerability to Oceanic conditions (ANTIPODE)".
[1] Behrendt et al. (1991) https://doi.org/10.1029/91TC00868
[2] Granot & Dyment (2018) https://doi.org/10.1038/s41467-018-05270-w
[3] Huerta & Harry (2007) https://doi.org/10.1016/j.epsl.2006.12.011
[4] Perron et al. (2021) https://doi.org/10.1051/bsgf/2020038
[5] Trey et al. (1999) https://doi.org/10.1016/S0040-1951(98)00155-3
[6] Davey & De Santis (2006) https://doi.org/10.1007/3-540-32934-X_38
[7] Mansour et al. (2020) https://doi.org/10.21105/joss.0179
Artificial Intelligence in Deep Geothermal Energy: Trends, Insights, and Future Perspectives
No full textDeep geothermal energy, known for its stable base load power and resilience to environmental fluctuations, is increasingly recognized as an important renewable energy source. Yet, its development is constrained by subsurface variability, high exploration costs, and operational inefficiencies. Artificial intelligence (AI) can analyze complex data, reveal patterns, and support predictive modeling to lower costs, shorten timelines, and improve efficiency. This review aims to evaluate how AI can address these barriers by systematically synthesizing its applications in deep geothermal research. A structured Web of Science search and multi-stage screening yielded 183 peer-reviewed journal papers, classified across eight research areas: reservoir characterization, exploration and resource identification, system optimization, seismic monitoring and risk assessment, drilling optimization, hybrid energy systems, environmental impact and sustainability, and techno-economic analysis. Our analysis shows that since 2020, AI applications in geothermal energy have expanded exponentially, surpassing overall AI growth rates. China and the United States dominate research output, followed by Germany, Turkey, Canada, and India. Advanced algorithms are increasingly preferred: convolutional neural networks for spatial modeling and image interpretation, recurrent neural networks for time-series forecasting, physics-informed AI, Bayesian frameworks, and autoencoders advance uncertainty quantification and data reconstruction. The novelty of this review lies in its comprehensive cross-domain synthesis of AI applications in deep geothermal energy, using a unified algorithm–input–output–performance lens. This structured mapping enables comparisons not possible in earlier overviews, reveals methodological strengths, identifies effective approaches for different geothermal tasks, and uncovers underexplored areas such as environmental assessment and techno-economic analysis
Processing and interpretation of multichannel seismic lines acquired in the Drygalski Basin of the western Ross Sea, Antarctica
No full textThe present Master degree thesis work focused on the re-processing and interpretation of IT90AR60 and IT90AR61A-B multichannel seismic lines acquired in the western Ross Sea (Antarctica) in order to improve the above-mentioned seismic lines and realize a seismo-stratigraphic interpretation to enhance the comprehension of the Drygalski Basin evolution located in the western Ross Sea, Antarctica. Two kinds of multiple attenuation techniques were used: SRME and predictive deconvolution. The re-processed seismic lines were interpreted by tracing different horizons according to seismo-stratigraphic criteria. In particular, it was possible to recognise and date two erosional unconformities: the Unconformity B represents a huge Miocene hiatus (14.7-4 Ma) and the Unconformity T represents a more recent erosional event of Late Pleistocene (3 Ma). Four different units were recognised: Unit 4, Unit 3, Unit 2 and Unit 1. The Unit 4 could represent glacio-marine sediment sub-horizontally stratified except for some zones where it was intensively deformed and disrupted by faults; Unit 3 is characterized by discontinuous, irregular and chaotic reflectors, which show a downlap geometry from coast seaward above the basal unconformity B providing progradational features interpreted as ice-proximal sediment; Unit 2 shows seismic facies with chaotic and massive geometry and with blank zones, probably due to till sediments lenses occurring on the side of the basin; Unit 1 is characterized by locally distributed moraines occurring in the depocenter of the basin. Moreover, N-S oriented faults were traced and, some of these, were correlated between the two seismic profiles. Once the interpretation was completed, isobate and isopach maps were created in order to understand the shallow stratigraphy in depth of the Drygalski Basin and the thicknesses of the deposited units. In addition to this, morphobathymetric data were analysed by providing information about the evolution of the Antarctic Ice Sheet. In fact, it was possible to observe that the ice sheet was grounded on the seafloor and, as a result of climate changes, it started to retreat, as testified by the presence of grounding zone wedges and morainal banks
Predictive modeling of shallow tunnel behavior: Leveraging machine learning for maximum convergence displacement estimation
No full textAccurate prediction of maximum convergence in unsupported, shallow tunnel construction is crucial for optimizing the lining and ensuring tunnel safety. Machine learning (ML) algorithms, especially through boosting techniques, enable effective solution of complex engineering problems and demonstrate their capabilities in problem solving and optimization. In this study, the FLAC 3D package was used to create a robust and validated database of 954 datasets. Five tree-based ML algorithms, including extreme gradient boosting (XGBoost), adaptive boosting (AdaBoost), gradient boosting machine (GBM), histogram-based gradient boosting (HGB) and categorical boosting (CatBoost), were used to predict the maximum convergence displacement for unsupported shallow tunnels. For the test dataset, XGBoost outperformed the other models with an excellent coefficient of determination of 0.9633, a minimum mean absolute error of 0.0021 and a low root mean squared error of 0.00725. HGB followed closely behind, and GBM and CatBoost showed strong performances, while Adaboost was less effective. The superior performance of XGBoost highlights its effectiveness in predicting maximum convergence in shallow tunnels. An in-depth sensitivity analysis within the XGBoost model showed the significant influence of soil elastic modulus on the maximum convergence displacement in unsupported tunnels. The remarkable results achieved by the XGBoost algorithm on our complex tunnel convergence predictions illustrate the profound ability of ML to tackle complicated geotechnical challenges. This interdisciplinary collaboration demonstrates the potential of advanced algorithms to improve safety and efficiency in construction, underlining the crucial role of technology in tackling complex problems and establishing a new paradigm for innovation in the field
- Eocene initiation of Ross Sea dextral seismogenic faulting and implications for East Antarctic neotectonics
No full textThe Ross Sea region of the East Antarctic plate provides evidence for intraplate tectonic activity in
Cenozoic times. Still unresolved are the cause, timing and kinematics of this intraplate tectonism. By
integrating and discussing the different (kinematic and temporal) signals of Cenozoic tectonism, intraplate
dextral shearing is recognized as the main tectonic regime controlling the structural architecture of the Ross
Sea region from the Mid-Eocene (c. 40–50 Ma) onward. We speculate that propagation and persistence of this
tectonic regime through time constitutes a feasible seismogenetic framework to explain past and current
tectonism in the Ross Sea region
- …
