Rare-metal pegmatite–greisen systems are commonly small, structurally controlled, and difficult to delineate using conventional mapping alone. This study proposes a multiscale remote-sensing workflow for prospecting Li–Nb–Ta–Cs mineralisation in the Kalba–Narym rare-metal belt (East Kazakhstan) by integrating Sentinel-2 multispectral imagery, UAV-derived centimeter-scale orthomosaics, structural (lineament) analysis, and field-based mineralogical–geochemical validation. Sentinel-2 responses were first calibrated using known occurrences to derive alteration proxies related to greisenisation, silicification, Na-metasomatism, and oxidation. These proxies were combined into an Integrated Hydrothermal Alteration Index (IHAI) to highlight areas where multiple alteration processes overlap. Lineament mapping from Sentinel-2 and DEM products indicates dominant NW–SE and NE–SW structural trends, zones of elevated lineament density and intersection systematically coincide with high IHAI values. UAV orthomosaics refine satellite-scale anomalies by resolving quartz-vein networks, fracture corridors, and surface-alteration textures that are not detectable at 10–20 m resolution. Mineralogical and geochemical data confirm that high-IHAI targets correspond to albitised pegmatites and greisenised rocks enriched in Li, Nb, Ta, and Cs. The results demonstrate that combining freely available Sentinel-2 data with UAV observations and targeted ground validation provides a cost-effective and transferable framework for reducing false positives and prioritising exploration targets in structurally complex granitoid terranes.

Remote sensing can support mineral exploration in geologically complex terrains, but distinguishing mineralisation-associated spectral responses from lithological and surficial variability remains challenging. This study presents and evaluates an integrated workflow for delineating and prioritising prospective zones for gold mineralisation at the Suzdal deposit, Kazakhstan. Sentinel-2 imagery was analysed using diagnostic band ratios, unsupervised spectral clustering, principal component analysis (PCA), and reference-signature-based spectral similarity classification. The resulting spectral patterns were integrated with lithological contacts, structural–geomorphological features, residual structures, denudation patterns, and lineaments; the delineated prospective zones were subsequently evaluated using independent exploration-drilling data and supporting geophysical information. PCA highlighted broad multispectral variability, whereas spectral similarity classification provided more spatially explicit delineation of mineralisation-associated surface responses. Higher-priority zones were identified where complementary remotely sensed responses converged within favourable geological–structural domains. Quantitative validation showed that 6 of 7 assay-confirmed Au-positive drill holes occurred within the delineated prospective zones, corresponding to a Spatial Capture Rate (SCR) of 85.7%. The novelty lies in combining deposit-specific geological criteria with complementary Sentinel-2 analyses within a unified targeting framework and subsequently evaluating the resulting prospectivity pattern against independent exploration evidence. The workflow provides a reproducible approach for exploration-target prioritisation at Suzdal and potentially in comparable gold-mineralised settings in Kazakhstan.