Background Karst regions experience severe soil erosion and water loss due to their unique hydrogeological characteristics and frequent natural disasters. Biocrusts, as extensively developed surface coverings, play a crucial role in regulating soil erosion. However, the impact of moss biocrust coverage on the distribution characteristics and stability of soil aggregates under different lithological conditions remains unclear. This study aims to investigate the effects of lithology and moss crust coverage on the distribution characteristics and stability of soil aggregates, thereby providing theoretical support for preventing and controlling soil erosion in karst regions.

Methods This study selected crustose moss sites dominated by mosses on dolomite and clastic rocks, establishing five levels of crustose moss cover (1%−20%, 20%−40%, 40%−60%, 60%−80%, and 80%−100%), with bare soil without crust cover as the control. Using the Le Bissonnais method, three treatments were applied: fast wetting (FW), pre-wetting oscillation (WS), and slow wetting (SW) treatments. Aggregate distribution, mean mass diameter, geometric mean diameter, erodibility K value, and soil fractal dimension D were compared across coverage levels and lithologies. Pearson correlation and path analysis were to identify key factors governing aggregate stability.

Results 1) Lithology and moss crust coverage significantly influenced soil properties (P < 0.05). In dolomite rock areas, moss-crust-covered soils showed average increases of 3.20%, 3.44%, and 27.99% in organic matter, silt, and clay content, respectively, compared to bare soil. In clastic rock areas, moss crust coverage increased organic matter, silt, and clay content by 7.70%, 8.19%, and 86.03% on average, compared to bare soil. As crust coverage increased, sand content decreased while silt and clay content increased in both rock types. 2) Under the LB method treatment, the fast wetting (FW) treatment caused the most severe disruption to soil aggregates, followed by the pre-wetting oscillation (WS) treatment, with the slow wetting (SW) treatment causing the least disruption. After all three treatments, soil aggregates predominantly consisted of particles > 2 mm in diameter. 3) For both rock types, the mean weight diameter (MWD) and geometric mean diameter (GMD) followed the order SW > WS > FW, while RSI > RMI. The fractal dimensions of erodibility, K and D values for both rock types, also followed the order FW > WS > SW. Under 80%−100% crust coverage, both MWD and GMD were higher in the two study areas than in other coverage levels, while K and D values were the lowest, indicating the highest aggregate stability and strongest erosion resistance. 4) The contents of silt, clay, and organic matter significantly influenced soil aggregate stability, with mechanical composition being a particularly critical factor.

Conclusions The development of biocrusts significantly enhances aggregate stability and erosion resistance, with lithology and biocrust coverage being key factors influencing these properties. The findings may provide fundamental data and a scientific basis for ecosystem restoration and soil erosion prediction in karst regions of Southwest China.