Plateau–orogenic belts host a substantial share of global gold resources, yet quantitative prospectivity mapping is challenged by complex mineralization and strongly heterogeneous, multi-scale datasets. Using the Mayoumu area (Tibet) as a representative orogenic gold district, we develop an integrated multi-source workflow that fuses remote-sensing alteration information with regional geochemical and structural constraints within an ensemble-learning framework. Alteration anomalies were mapped from GF-5 hyperspectral imagery using mixture-tuned matched filtering (MTMF) and from Sentinel-2 multispectral imagery using the iCrosta method to extend alteration signals across scales. Geochemical anomalies were extracted from 1:200,000 stream-sediment data through isometric log-ratio (ILR) transformation and robust principal component analysis (RPCA). At the same time, ore-controlling structures were quantified using Euclidean-distance-to-fault layers. Three Boosting-based ensemble models—gradient boosting decision tree (GBDT), extreme gradient boosting (XGBoost), and light gradient boosting machine (LightGBM)—were trained to predict mineral prospectivity. Performance was evaluated using confusion matrix metrics and ROC–AUC, and key predictors were interpreted using SHAP. All three models achieved AUC values > 0.90, with LightGBM performing best (AUC = 0.94) and delineating high-prospectivity zones that coincide with known occurrences and highlight additional targets. The proposed workflow provides a practical, transferable reference for gold prospectivity mapping in complex orogenic belts worldwide.

The Gangdese metallogenic belt in Tibet is an important polymetallic metallogenic belt formed during the subduction of the Neo-Tethys Ocean and subsequent India–Asia collision. Adakitic rocks are widely distributed in this belt and are considered to be closely related to porphyry–skarn Cu-Mo polymetallic mineralization. However, the petrogenesis and geodynamic setting of the Late Cretaceous adakites in the Gangdese belt remain controversial. In this study, we focus on the quartz diorite in the Nuri Cu-W-Mo deposit along the southern margin of the eastern Gangdese belt. LA-ICP-MS zircon U-Pb dating yields a Late Cretaceous age of 93.6 ± 0.4 Ma for the quartz diorite. Whole-rock geochemistry shows that the quartz diorite possesses typical adakitic signatures, with high SiO₂, Al₂O_3, and Sr contents, but low Y and Yb contents. The relatively low K₂O content and high MgO, Cr, and Ni contents, as well as the positive zircon εHf(t) values (+6.58 to +14.52), suggest that the adakites were derived from the partial melting of the subducted Neo-Tethys oceanic slab, with subsequent interaction with the overlying mantle wedge. The Late Cretaceous magmatic flare-up and coeval high-temperature granulite-facies metamorphism in the Gangdese belt were likely triggered by Neo-Tethys mid-ocean ridge subduction. The widespread occurrence of Late Cretaceous adakitic intrusions and associated Cu mineralization in the Nuri ore district indicate a strong tectono-magmatic-metallogenic event related to the Neo-Tethys subduction during this period. This study provides new insights into the petrogenesis and geodynamic setting of the Late Cretaceous adakites in the Gangdese belt, and has important implications for Cu polymetallic deposit exploration in this region.

As a key target area in China’s new round of strategic mineral exploration initiatives, Tibet possesses favorable metallogenic conditions shaped by its unique geological evolution and tectonic setting. In this paper, the Saga region of Tibet is the research object, and Level-2A Sentinel-2 imagery is utilized. By applying mixed pixel decomposition, interfering endmembers were identified, and spectral unmixing and reconstruction were performed, effectively avoiding the drawback of traditional methods that tend to remove mineral alteration signals and masking interference. Combined with band ratio analysis and principal component analysis (PCA), various types of remote sensing alteration anomalies in the region were extracted. Furthermore, the fractal box-counting method was employed to quantify the fractal dimensions of the different alteration anomalies, thereby delineating their spatial distribution and fractal structural characteristics. Based on these results, two prospective mineralization zones were identified. The results indicate the following: (1) In areas of Tibet with low vegetation cover, applying spectral mixture analysis (SMA) effectively removes substantial background interference, thereby enabling the extraction of subtle remote sensing alteration anomalies. (2) The fractal dimensions of various remote sensing alteration anomalies were calculated using the fractal box-counting method over a spatial scale range of 0.765 to 6.123 km. These values quantitatively characterize the spatial fractal properties of the anomalies, and the differences in fractal dimensions among alteration types reflect the spatiotemporal heterogeneity of the mineralization system. (3) The high-potential mineralization zones identified in the composite contour map of fractal dimensions of alteration anomalies show strong spatial agreement with known mineralization sites. Additionally, two new prospective mineralization zones were delineated in their periphery, providing theoretical support and exploration targets for future prospecting in the study area.