Effects of road and river networks on sediment connectivity in mountainous watersheds

Road and river networks in mountainous watersheds play an important role in transporting eroded sediments. However, the underlying transport mechanisms remain poorly understood, particularly in terms of how alterations to flow paths caused by road and river networks influence sediment connectivity. Therefore, using data from the Dongshuanghe (DSH) watershed in the Dabie Mountain in Central China, this study investigated the spatial relationships between road and river networks, analyzed the effects of road and river networks on the spatial distribution of sediment connectivity, and determined the primary factors influencing sediment connectivity. The primary factors were identified using the index of connectivity (IC), buffer analysis, a random forest (RF) model, and a geographical detector model (GDM). The results indicated that road and river networks were spatially closely associated: closer to rivers, the density of high-grade roads increased, while farther from road-river crossings, the length and density of rivers and high-grade roads decreased. Both road and river networks affected the IC. In particular, for permanent drainage lines or local sinks, the average value of IC_R (i.e., IC affected by road networks, -0.97) was higher than that of IC (-2.17). Thus, values of IC_R decreased substantially with increasing distance to the closest roads (R² = 0.73). However, beyond a threshold of 150 m, the effect of roads on IC_R gradually diminished. In addition, the structural characteristics of road networks, particularly slope (Road_S), had greater explanatory power for spatial variation in the IC_R. In conclusion, compared to the river networks, the effect of road networks on IC_R was more important, which was not only reflected in the spatial distribution of IC_R, but also in the factors influencing IC_R.