The Muji spring travertines, located in the Muji Basin in the eastern Pamirs Plateau, represent a typical spring deposit found on plateaus that is characterized by arid and semi-arid climatic conditions. However, its formation mechanisms remain poorly understood. This study aims to explore the recharge processes of the spring, the sedimentary environment, and the genetics of Muji spring travertines through a comparative analysis of conventional hydrochemistry, H-O stable isotope analysis of both spring and river water, and petrographic observation, as well as in situ analysis of major and trace elements present in calcite within travertines. The basin is surrounded by mountains with a topography that facilitates groundwater convergence within it. Carbonate-bearing strata are extensively developed around the basin, which serves as a crucial material foundation for travertine development. It infiltrates underground through fractures and faults, interacting with carbonate rocks to produce significant amounts of HCO₃⁻, Ca²⁺, and Mg²⁺. The observed range of isotopic compositions (δ²H, −102.27‰ to −96.43‰; δ¹⁸O, −14.90‰ to −14.36‰) in water samples suggests that their primary origin was from glacial and snowmelt sources. The concentration of HCO₃⁻ in spring water samples exhibits significant variability, with the highest value being 1646 mg·L⁻¹, which deviates significantly from the typical composition of karst groundwater. During its migration, groundwater undergoes the dissolution of gaseous CO₂ derived from deep metamorphic processes, leading to variable degrees of mixing with geothermal groundwater containing elevated concentrations of dissolved components that enhance the dissolution potential of carbonate rocks. Eventually, upwelling occurs along the Southwestern Boundary Fault of Muji, resulting in the formation of linear springs characterized by CO₂ escape. The Muji laminated travertines exhibit distinct white and dark laminae, and radial coated grains consisting of micritic and sparry layers. Chemical composition analyses reveal significant differences in the trace and rare-earth element composition, as well as the Mg/Ca ratio, of the two types of travertines. Specifically, the micritic laminae of the pisoid (Mg/Ca = 0.019; Sr = 530 × 10⁻⁶; Ba = 64.6 × 10⁻⁶) and the dark laminae of the laminated travertine (Mg/Ca = 0.014; Sr = 523 × 10⁻⁶; Ba = 48.1 × 10⁻⁶) exhibit generally higher Mg/Ca ratios and Sr, Ba contents than the neighboring sparry laminae (Mg/Ca = 0.012; Sr = 517 × 10⁻⁶; Ba = 36.6 × 10⁻⁶) and white laminae (Mg/Ca = 0.006; Sr = 450 × 10⁻⁶; Ba = 35.6 × 10⁻⁶). The development of laminated travertines and radial coated grains here is attributed to periodic changes in groundwater recharge induced by seasonal temperature fluctuations, as evidenced by the structural characteristics of the two types of travertines and the trace element analysis of different layers. Algae play a role in forming the dark laminae of laminated travertines and the micritic laminae of pisoids.

Traditional geological mapping is often time-consuming, labor-intensive, and restricted by rugged terrain. This study addresses these challenges by proposing a novel methodology for automated lithological identification in the Ququleke area of the eastern Kunlun Mountains, which pioneers the integration of portable UAV oblique photogrammetry with a Coordinate Attention-enhanced DeepLabV3+ (CA-DeepLabV3+) semantic segmentation framework for geological mapping. Using a DJI Mavic 3M quadcopter, high-resolution oblique photogrammetric orthophotos were captured to build a pixel-level lithology dataset containing four classes: sandstone, diorite, marble, and Quaternary sediments. The CA-DeepLabV3+ model, adapted from the DeepLabV3+ encoder–decoder framework, integrates a lightweight MobileNetV2 backbone and a Coordinate Attention mechanism to strengthen spatial position encoding and fine-scale feature extraction, crucial for detailed lithological discrimination. Experimental evaluation demonstrates that the proposed model achieves an overall accuracy of 97.95%, mean accuracy of 97.80%, and mean intersection over union of 95.71%, representing a 5.48% improvement in mean intersection over union (mIoU) over the standard DeepLabV3+. These results indicate that combining UAV oblique photogrammetry with the CA-DeepLabV3+ network enables accurate lithological mapping in complex terrains. The proposed method provides an efficient and scalable solution for geological mapping and mineral resource exploration, highlighting the potential of low-altitude UAV remote sensing for field-based geological investigations.