A Rock Geochemical and Sentinel-2 Alteration-Mapping Dataset with Anomaly Assessment for the Tozbulak-Aktosty Area, Bukhara Region, Uzbekistan

Abstract

The Tozbulak-Aktosty Area in the Bukhara Region of Uzbekistan is a granite- and pegmatite-hosted rare-metal terrane in which rare-earth elements (REE) and lithium (Li) occur within structurally controlled zones. This data article presents an openly available geochemical and remote-sensing dataset comprising 45 rock samples analyzed for total REE (TREE) and TREE plus yttrium (TREY), 15 individual REE (La-Lu and Y), and Li, together with seven Sentinel-2 Level-2A alteration layers and three Copernicus DEM derivatives covering an area of approximately 44×23 km2 (precise coordinates withheld at the request of the data provider). The satellite products include iron-oxide (B4/B2), clay (B11/B12) and hydroxyl (B8A/B11) band ratios, Crósta principal- component (PCA) iron and hydroxyl indices, a lineament-enhancement layer, and a SWIR lithological composite. TREE concentrations span three orders of magnitude (3-5,614 ppm; median, 100 ppm), with three samples exceeding 1,000 ppm and a mean light-to-heavy REE ratio of 4.8, reflecting a spectrum from non-mineralized host rocks to zones significantly enriched relative to the host rock. Because hard-rock REE ore typically contains several weight-percent TREE, the maximum value observed here (0.56 wt%) is interpreted as a host-rock-relative anomaly rather than ore-grade mineralization. To demonstrate reusability, alteration indices were extracted at each sample site and correlated with geochemistry; pixel-scale Spearman correlations were weak because high-concentration REE occurs as point-source enrichment within 20 m spectral pixels, yet a clear area-scale association emerged. The Northern Tozbulak occurrence combines the highest mean TREE, the strongest lineament enhancement, and an elevated Crósta hydroxyl response. An integrated model fusing an inverse-distance geochemical anomaly surface with weighted alteration and structural indices delineated candidate anomaly zones for REE, Li, and Au-Cu that warrant further evaluation. The openly available dataset, structured following FAIR principles, provides a reproducible baseline for anomaly mapping, remote-sensing studies, and machine- learning applications in arid rare-metal terranes.