Abstract:Against the backdrop of increasingly frequent extreme rainfall events, preferential flow has emerged as a critical mechanism affecting soil water movement and slope stability in landslide-prone areas, necessitating in-depth investigation. This study focuses on shallow landslide zones in granite residual soils, selecting three typical soil types—Lithocarpus-covered residual soil (MH), exposed residual soil (BR), and completely weathered granite (QW). Through field dye tracing experiments and image processing techniques, the spatial distribution characteristics and parameters of preferential flow within soil profiles were analyzed. The results reveal the evolution patterns of preferential flow across different soil structures and explore the intrinsic relationship between preferential flow and soil structural features, thereby providing theoretical support for understanding landslide formation mechanisms. The study quantitatively identifies differences in preferential flow characteristics and structural control mechanisms among different types of granite residual soils:1.Preferential flow was detected in all three soil profiles, but with significant differences in spatial distribution and intensity;2.The proportion of preferential flow area decreased with depth, and its horizontal distribution within profiles showed high heterogeneity, indicating a strong structural control over water movement pathways;3.The preferential flow index (PFI) varied significantly among the soil types, with the highest in MH (0.5969), suggesting its structure favors strong preferential flow, and the lowest in BR (0.3254), where water infiltration is mainly limited by the degree of pore development;4.PFI effectively quantified the intensity and infiltration depth of preferential flow, revealing how structural differences drive rapid shallow water infiltration.This study expands the quantitative characterization methods of preferential flow in granite residual soil landslide areas, offering new insights for hydrological modeling and landslide risk assessment.