Conference Agenda

Mine wastes and tailings, long viewed as environmental issues, are increasingly recognized as anthropogenic resources whose reprocessing offers a dual contribution: the recovery of secondary raw materials and the remediation of contaminated sites that put local communities at risk. The Structured Anthropogenic Resource Assessment methodology for UNFC (SARA4UNFC), developed within the HE-project FutuRaM, operationalizes the UN Framework Classification for Resources (UNFC) for anthropogenic resources. It organizes a systematic assessment along the UNFC axes, environmental and socio-economic viability (E), technical feasibility (F), and confidence in quantities (G). As a first stage, it was designed to support approval processes by capturing the regulatory minimum requirements that a project must meet to be implemented. SARA4UNFC has already been applied in various case studies involving different types of waste streams, including three case studies on mining waste, and has demonstrated that it delivers consistent, transparent, and comparable results.

For mining-related projects, including tailings, compliance with the statutory minimum requirements is necessary but not sufficient. Social acceptance, which is decisive for whether a project can actually be realized, depends on voluntary sustainability commitments that go beyond the minimum legal requirements: community health near active and historical sites, regional employment, and responsible environmental management, to name a few. This contribution presents initial findings from ongoing work at LMU Munich to extend SARA4UNFC with a structured sustainability dimension. The extended framework enables voluntary commitments to be made transparently and comparably assessable, supporting tailings reprocessing as a practice that reconciles resource recovery, environmental remediation, and social acceptance.

Accurate estimation of groundwater occurrence probability requires predictive frameworks that preserve hydrogeological consistency while remaining robust to spatial generalisation and environmental heterogeneity. Conventional machine learning approaches often treat groundwater observations as independent samples, limiting their ability to represent spatially structured subsurface processes and reducing reliability when applied to unseen geographic regions. To address these limitations, we introduce GeoHydroTwinNet (Geospatial–Hydrogeological Twin Network), a physics-informed graph learning framework that integrates spatial relational learning with the incorporation of hydrogeological constraints within a unified dual-stream architecture. The framework jointly combines spatial dependency modelling and hydrogeological priors within a single training objective, enabling physically consistent estimation of groundwater occurrence probability under spatially unseen conditions. GeoHydroTwinNet was evaluated using a harmonised 30 m spatial-resolution dataset integrating topographic, geological, hydrographic, land-cover, vegetation, climatic, and groundwater monitoring data across Khuzestan Province, Iran. Model performance was assessed using leakage-controlled 16-fold spatial cross-validation designed to prevent spatial dependence between training and validation regions. Results show that GeoHydroTwinNet produces spatially coherent groundwater occurrence probability maps with improved predictive performance compared with benchmark machine learning models, including XGBoost and RF. Beyond accuracy, the proposed framework maintains physically consistent spatial patterns under geographic generalisation while providing interpretable outputs for groundwater resource assessment and decision support. These results demonstrate the potential of physics-informed graph learning as a robust framework for probabilistic groundwater modelling in environmentally heterogeneous and data-limited regions.

Understanding how deformation bands form in porous rocks requires realistic representation of grain-scale cohesion and damage evolution. We present a new intergranular bonding model for polyhedral particles in the open-source Discrete-Element-Method (DEM) framework Yade, and investigate its ability to reproduce deformation-band development under compactional-shear conditions. The model incorporates a mixed-mode tensile–shear bond failure formulation, allowing grain-scale cohesion loss to develop during deformation.

Cohesion between neighboring polyhedral grains is introduced through bonded contacts that apply equal and opposite forces between interacting particles. The bonds can fail in tension, shear, or a combination of both according to an elliptical failure criterion. Polyhedral particles were chosen to better represent the geometry of natural grains. Each grain is represented by a single numerical body. This avoids the need to construct grains from multiple bonded spheres and eliminates the additional internal contacts and bonds required by such approaches.

Simulations are performed in a 2D compactional shear setup. Samples with different initial porosities and sorting characteristics are considered. During deformation, bond failure localizes into narrow low-porosity zones that resemble deformation bands observed in natural porous rocks. Strain and cohesion loss become concentrated within these localized regions, while the surrounding material experiences less deformation.

The results suggest that intergranular bonding is an important component of DEM models aimed at reproducing strain localization in porous rocks. The polyhedral model also allows future implementation of grain fracturing without discretizing grains into smaller elements. The approach provides a useful basis for future studies of deformation-band evolution and fault-zone initiation.

For rock engineering design the rock strength is one of the key parameters. However, it has been shown in numerous studies, that rock strength is dependent on several experimental conditions such as loading rate, confining stress, or moisture content; common standards and recommendations account for these. End friction effects are known to have an influence on rock strength as determined in the laboratory, however, this effect has not been considered in detail, which may lead to the under- or overestimation of rock strength.

In this study, Ruhr sandstone and Gildehaus sandstone, were investigated with respect to the influence of end face friction on their uniaxial compressive strength, deformation modulus and Poisson’s ratio at varying deformation rates. Tests were carried out at strain rates of 10^-3 s^-1, 10^-5 s^-1 and 10^-7 s^-1 using either graphite powder, Teflon sheets or no lubrication to vary frictional conditions at the specimens end faces.

The results indicate that both, the uniaxial compressive strength as well as the deformation modulus are affected by the frictional conditions and are significantly reduced when Teflon is used as a lubricant. Furthermore, the effect of the lubricant on the uniaxial compressive strength and deformation modulus is independent of strain rate, whereas the effect of end face friction was larger in the rock that exhibited higher strength. Poisson's ratio seemed to be influenced by end face friction although further investigation is required. Dominant failure mode changes from shear to tensile as end face friction is reduced.

Carbonatitic magmas represent a volumetrically minor but petrologically and economically important component of Earth’s magmatic budget. Their origin is mostly related to direct partial melting of carbonate-bearing mantle sources, extreme differentiation of CO₂-rich, SiO₂-undersaturated silicate magmas, or enrichment in a carbonate component through limestone-dolostone assimilation by basic-ultrabasic silicate magma at shallow-mid crustal depths. Carbonate-silicate melt interaction was investigated at 1GPa in piston-cylinder experiments performed on three SiO₂-undersaturated natural rocks from the Bohemian Massif: a polzenite (BM1), a monticellite-bearing melilitite/vesecite (BM2), and an olivine basanite (BM3). Each composition was mixed with 10, 30 and 50mass% synthetic CaCO₃ and equilibrated at 1200 and 1300°C to evaluate carbonate assimilation capacity and compare resulting liquids with those obtained at 0.2GPa using the same starting materials [Lustrino et al., 2022; https://doi.org/10.1130/G49621.1].

Increasing CaCO₃ addition drives experimental glasses toward lower SiO₂ and higher CaO contents, with BM2+30 and 50%CaCO₃ yielding microcrystalline groundmasses with carbonatitic–kimberlitic compositions. Interstitial calcite occurs only in experiments with high crystallization degrees and CaO-rich, SiO₂-poor residual melts, highlighting the combined role of crystallization pathways and bulk composition in controlling calcite saturation. No physical separation between liquids of different composition was observed, suggesting that SiO₂-poor and CaO- and CO₂-rich residual melts from basic to ultrabasic rocks may remain a single homogeneous phase at depths comparable to the continental Moho. These preliminary results indicate that strongly SiO₂-undersaturated mildly alkaline magmas can record crustal carbonate assimilation, a process that is not incompatible with the coexistence of a carbonate component in upper mantle sources (as proposed for Pleistocene Intra‑Apennine Province).