The sandstone-type uranium deposit of the Kelulun Depression is the first industrially valuable uranium deposit discovered in the Hailar Basin. This study performed a systematic examination of 17 sandstone samples from the Yimin Formation in the Kelulun Depression based on various analytical techniques. The findings of the current study were synthesized with previous research to investigate the impact of the redox conditions and the tectonic background of the source area, as well as the paleoclimatic evolution of the Yimin Formation on uranium mineralization. The elemental Mo, U/Th, V/Cr, Ni/Co, and V/(V + Ni) ratios indicate that the paleowater was in an oxygen-rich environment during the deposition of the Yimin Formation. Additionally, the C-value, Sr/Cu, Al₂O₃/MgO, and Rb/Sr ratios indicate that the Yimin Formation was formed in a paleoclimate characterized by arid-to-semi-arid conditions. The geochemical characteristics of the observed elements indicated that the sediment source of the Yimin Formation was mainly felsic rocks from the upper continental crust, the weathering of the rock was weak, and the tectonic background was a passive continental margin. Coffinite is distributed in the form of cementation and stellates within or around pyrite crystals, and uranium-titanium oxide is mostly distributed in an irregular granular distribution in the biotite cleavage fractures of the study area. In summary, the findings of this study reveal that the tectonic settings, provenance, uranium source, paleoclimate, and oxygen-rich paleowater of the Yimin Formation have important geological significance for the large-scale uranium mineralization of the Kelulun Depression.

The Ordos Basin is a key district for sandstone-hosted uranium, yet host-rock controls and uranium sources remain debated. We integrate measured sections, whole-rock geochemistry, and detrital zircon U-Pb-Lu-Hf data from the Cretaceous Huanhe Formation (Yihewusu, northern Ordos) to resolve provenance, transport, and enrichment pathways. Uranium enrichment is concentrated in feldspathic-lithic sandstones deposited in proximal fluvial-lacustrine settings. Detrital zircon ages define three clusters—Phanerozoic (500–200 Ma), Paleoproterozoic (2000–1700 Ma), and Neoarchean (2600–2300 Ma)—with Proterozoic grains >60%, indicating derivation from Archean–Paleoproterozoic TTG gneisses, granulites, and khondalites of the Yinshan Block and the northern Central Orogenic Belt. Zircon εHf(t) values (−10.84 to +7.76) and crustal model ages (3.2–2.1 Ga) record substantial Meso- to Neoarchean crustal growth in the source terranes. Critically, Permian-Cretaceous intermediate-felsic igneous rocks along the northern margin of the Western North China Block—marked by elevated U, Th/U > 5 (indicative of U loss), pervasive feldspar micro-fractures, and proximity to basin-margin uranium belts—are identified as the principal uranium reservoirs. We propose a dual uranium supply: soluble uranium mobilized from leached igneous rocks during weathering and fluid-rock interaction, and U-enriched detritus delivered to the basin. Uranium concentrated in redox-sensitive, feldspathic-lithic sandstones of the Huanhe Formation, which effectively trapped advected uranium at proximal facies transitions. These findings establish a direct genetic link between basin-margin uranium sources and in-basin mineralization, providing a predictive framework for regional uranium exploration in North China.

Calcrete-type uranium prospectivity prediction is challenged by the strong heterogeneity of multi-source geoscientific raster datasets, weak anomaly responses, and the lack of explicit heterogeneous information organization in conventional deep learning models. In this study, the Yilgarn Craton of Western Australia was selected as the study area, and a geology-guided fixed-group fusion ResUNet model (GGF-ResUNet) was developed based on 12-channel multi-source geoscientific raster datasets. At the input stage, the evidence layers were divided into four fixed geoscientific proxy groups according to their data modality and geological interpretation, namely gravity, aeromagnetic, radiometric, and geochemical groups, and intra-group channel weighting together with inter-group gating was introduced to enhance the hierarchical representation and adaptive fusion of heterogeneous information. Ablation results showed that GGF-ResUNet achieved better performance than the baseline ResUNet, with AUC increasing from 0.9340 to 0.9740 and F1-score improving from 0.7264 to 0.8356. Further comparative experiments with Attention U-Net, U-Net, SegNet, and FCN showed that GGF-ResUNet achieved comparatively better quantitative performance and more spatially coherent prediction results under the current experimental setting. Without substantially increasing model complexity, the proposed method improves the representation and integration of heterogeneous geoscientific information and provides a feasible technical pathway for calcrete-type uranium prospectivity prediction under weak-anomaly conditions.