Surface soil physical property fingerprint of piping influenced gully erosion

Understanding how soil physical properties differ between eroded and adjacent non-eroded microsites is essential for diagnosing early-stage rill–gully development and shallow piping in forested landscapes, yet paired microsite evidence remains limited. We quantified soil physical contrasts between gully/rill beds microsites (eroded) and adjacent banks (visually stable) across 13 cross-sections in a forested watershed in Serbia, while minimizing background variability in site conditions. We hypothesized that gully-bed soils differ from banks and that these contrasts are depth-dependent. Samples from two depth intervals (5–10 and 20–25 cm) were analyzed for soil moisture, density-based indices (bulk and dry bulk density, porosity, void ratio), particle-size distribution, Atterberg limits, and the clay activity index (AI). Microsite differences were strongest in the surface layer: gully-bed soils were wetter and showed lower clay, higher sand, higher AI, and higher liquid limit (wL) than banks. At 20–25 cm, microsite contrasts were weaker overall, but clay and sand contents remained significantly different, whereas silt content and particle density showed no meaningful variation across microsites or depths. Two-way ANOVA indicated that depth dominated density-derived indices and significantly affected moisture, clay, and plastic limit (wP), while sampling position significantly affected moisture, clay, AI, and wP; the depth × position interaction was significant only for moisture. An exploratory PCA-based composite score ( E -PCS) summarized coordinated texture–consistency/activity shifts and separated gully-bed from bank samples across both depths. Overall, even incipient piping-related incision produced measurable, depth-dependent microsite contrasts, supporting shallow-horizon prioritization and depth stratification in monitoring designs for similar forested watersheds.