Geological Survey of Finland,Ministry of Economic Affairs and Employment

Inversion of gravity data is an important method for investigating subsurface density variations relevant to mineral exploration, geothermal assessment, carbon storage, natural hydrogen, groundwater resources, and tectonic evolution. Here we present a scientific machine-learning approach for three-dimensional gravity inversion that represents subsurface density as a continuous field using an implicit neural representation (INR). The method trains a deep neural network directly through a physics-based forward-model loss, mapping spatial coordinates to a continuous density field without predefined meshes or discretisation. Spatial encoding enhances the network's capacity to capture sharp contrasts and short-wavelength features that conventional coordinate-based networks tend to oversmooth due to spectral bias. We demonstrate the approach on synthetic examples including smooth models, representing realistic geological complexity, and a dipping block model to assess recovery of structures at different depths. The INR framework reconstructs detailed structure and geologically plausible boundaries without explicit regularisation or depth weighting, while reducing the number of inversion parameters as the problem size grows bigger. These results highlight the potential of implicit representations to enable scalable, flexible, and interpretable large-scale geophysical inversion. This framework could generalise to other geophysical methods and for joint/multiphysics inversion.

Ecological restoration has emerged as a conservation approach for peatlands, which often experience increased leaching of dissolved organic carbon (DOC) and soluble nitrogen (N-tot) and phosphorus (P-tot) following drainage. Restoration, however, introduces new disturbances that temporarily elevate leaching, while monitoring peatland runoff remains challenging, highlighting the need for new methods to assess restoration success. We studied changes and connections of porewater and runoff quality across three pristine and five drained and restored boreal peatland sites, quantified runoff loads, and developed regression models to estimate loads using porewater data. After restoration, median runoff DOC, N-tot, and porewater and runoff P-tot concentrations showed respective increases of 35, 34, 67, and 224% compared to the drained state, peaking within 1-4 years before declining. Porewater DOC and N-tot levels started decreasing immediately. Porewater concentrations approached near-pristine levels within roughly 4-10 years depending on the parameter and site, whereas runoff levels stayed elevated through tenth year with only site-specific exceptions. Porewater and runoff DOC and N-tot were strongly correlated, while P-tot correlations varied among drained, restored, and pristine states. DOC, N-tot, and P-tot loads initially rose after restoration but declined over the following >3 years post-restoration to 61.3, 1.52, and 0.038 kg/ha/yr, typically falling below drained state levels. Regression models overestimated produced annual runoff loads from porewater concentrations by an average of 50.9%. Models predicted DOC most accurately, while performance for N-tot and P-tot was weaker and varied across sites.

Scandio-fluoro-eckermannite (IMA 2024-002), a new Sc-dominant amphibole-supergroup mineral, has been discovered in the Bayan Obo REE-Nb-Fe polymetallic deposit, China. The new mineral was collected from banded Fe-REE ores that have formed due to the fenitization caused by carbonatite intrusion, in the Main and East open pits at Bayan Obo. Associated minerals include monazite, bastnasite, magnetite, biotite, fluorite, bazzite, thortveitite, and magnesio-fluoro-arfvedsonite. The new mineral occurs as euhedral to subhedral crystals and aggregates, appearing both as inner zones of a crystallization sequence from scandio-fluoro-eckermannite to magnesio-fluoro-arfvedsonite as well as homogeneous fine-grained particles, reaching up to 350 mu m in size and similar to 7 wt% in Sc2O3 content. Scandio-fluoro-eckermannite displays a light yellow to light blue color under plane-polarized transmitted light, with perfect cleavage on {110}, non-magnetic, and no fluorescence. The hardness is 5-6 by analogy to eckermannite, and the calculated density is 3.097 g/cm(3). Electron microprobe analyses determined the main components (average value in wt%): Sc2O3 6.39; SiO2 54.30; MgO 13.42; Na2O 8.38; Al2O3 1.29; MnO 1.47; CaO 1.21; K2O 0.47; FeOcalc 6.43; Fe2O3calc 3.80; F 3.01; H2Ocalc+ 0.67; F equivalent to O-1.27; total 99.74. The composition normalized on the basis of 24 anions (O, OH, F, Cl), with the assumption of (OH, F, Cl) = 2 apfu, corresponds to the empirical formula (A)(Na0.52K0.09 square(0.39))(Sigma 1.00) (B)(Na1.81Ca0.19)(Sigma 2.00) (c)(Mg2.87Fe2+ Mn-0.77(3+) 0.18Sc0.80Fe3+ (0.41))(Sigma 5.03) (T)(Si7.78Al0.22)(Sigma 8.00) O-22(W)[F-1.36(OH)(0.64)](Sigma 2.00). It leads to the simplified formula (Na,square)(Na,Ca)(2)[(Mg,Fe2+)(4)(Sc,Fe3+,Mn3+)][(Si,Al)(8)O-22)](F,OH)(2) and the ideal formula NaNa2(Mg4Sc)Si8O22F2. The crystal structure was refined in the monoclinic system, space group C2/m (#12). Its unit-cell parameters are: a = 9.8212(3) angstrom, b = 18.0866(5) angstrom, c = 5.3091(2) angstrom, beta = 103.767(4)degrees, and Z = 2, with the a:b:c ratio of 0.543:1:0.294. The crystal-structure refinement indicates that Na is the dominant cation at the A(m) and M(4) sites, Mg is the dominant cation at the M(1) and M(3) sites, Sc is the dominant trivalent cation at the M(2) site, and F is the dominant cation at the O(3) site. Therefore, this is the Sc-dominant variety of fluoro-eckermannite. This discovery highlights the importance of amphibole in controlling Sc in this type of ore-forming system. Scandio-fluoro-eckermannite might also be used as a potential recorder to investigate the enrichment process of Sc in the Bayan Obo deposit.

This study investigated the combined effects of ground granulated blast furnace slag (GGBFS) content (0 - 46%) and carbonation conditions on salt freeze-thaw scaling resistance of concrete. Standard 7 days curing resulted in minimal scaling (< 1.0 kg/m(2)), while accelerated carbonation for similar duration amplified scaling values to 1 kg/m(2), 3.3 kg/m(2), and 4.5 kg/m(2) in 0%, 15%, and 46% GGBFS concretes, respectively. Extended carbonation (30 days) further increased scaling to 1.2, 5.7, and 7.4 kg/m(2) for these mixtures, with 70% of total scaling occurring within the first 14 freeze-thaw cycles. The strong correlation (R-2 = 0.873) between carbonation depth and scaling depth confirmed that approximately 50% of the carbonated layer was removed during the slab test. Specimens with 0% and 46% GGBFS showed 56% and 8% reduction in large capillary pores at paste level (2 & micro;m/voxel), respectively. While concrete specimens with 46% GGBFS exhibited 22% increase in macropores (>= 15 & micro;m), indicating carbonation-induced microcracking. Despite these porosity changes, capillary sorptivity remained unchanged across all concretes, while total water absorption decreased between 12 - 16%, suggesting that carbonation fills pore bodies without blocking connecting throats. Cement paste with 46% GGBFS contained 47% less portlandite than paste with 0% GGBFS. Thermogravimetric analysis showed the formation of calcium carbonate (27% in paste with 46% GGBFS after 30 days carbonation) greater than the amount attributable to portlandite alone, confirming decomposition of calcium silicate hydrate. This decalcification combined with carbonation induced microcracking and deterioration of air-void system, might weakened the binding matrix and explained the severe salt freeze-thaw scaling in carbonated slag concretes.

Understanding the heterogeneity of deformation in exhumed lower-crustal terranes contributes to elucidating mechanisms of lower-crustal deformation and paleotectonic processes. Here, we focus on the Kynsikangas shear zone (KSZ), a prominent structural discontinuity that is part of a network of crustal shear zones within the southern Svecofennian orogen. The shear zone overprints metagranitic rocks, gneisses and migmatites. A detailed field-based structural examination of the shear zone considered the patterns, intensity, kinematics and rheology of deformation. A pronounced scale-dependent heterogeneity in the geometry and intensity of mineral shape fabrics as well as in the deformation kinematics are structural hallmarks of the KSZ. The analysis of deformation kinematics and mineral shape fabrics revealed that the KSZ formed by strike-slip dominated, left-lateral transpression under NW–SE shortening. Mesoscopic shear bands and S-C fabrics delimit phacoids, which are reminiscent of scaly fabrics, known from active crustal shear zones formed at low metamorphic grade. We, therefore, interpret these structures as high-temperature equivalents of scaly fabrics, which may constitute slip transients in a frictional-viscous material. Our analysis also confirms that small-scale kinematic indicators may not portray the regional sense-of-shear and, thus, questions their use for identifying shear zone kinematics in general. Due to its distinct kinematics and orientation, the KSZ most likely served as a northwest-trending tectonic transfer zone among an anastomosing pattern of mostly easterly striking regional shear zones during later stages of the Svecofennian orogeny.