Splash erosion potential under tree canopies in subtropical SE China

Sand-filled splash cups were used to study the erosivity of rainfall and throughfall in the humid subtropics of southeast China. Our results showed that the splash cup measurements yielded precise and reproducible results both under open field conditions and under forest vegetation. The splash cups were exposed to forest stands of different age and to selected species ( Schima superba , Castanopsis eyrei ) in the Gutianshan National Nature Reserve (GNNR). The measurements in the open field revealed a close relationship between unit sand loss (g m − 2 ), rainfall amount (mm) (R 2 = 0.94) and maximum rainfall intensity (mm h − 1 ) (R 2 = 0.90). The highest correlation was obtained between unit sand loss (g m − 2 ) and the average of the five highest five minute interval rainfall intensities (mm h − 1 ) (R 2 = 0.96). This underlines the reliability of the splash cups used. The best results on the relationship between variables related to precipitation and sand loss were obtained with the Vaisala sensor compared to a standard tipping-bucket rain gauge. This is mainly due to the fact that the drop impact-based Vaisala sensor (0.8–5.0 mm drops are recorded assuming their terminal velocity) does not measure drops with low kinetic energy, whereas drop size distribution plays no role for the measurements with the tipping-bucket rain gauge. The results obtained under forest vegetation show that the erosive power of throughfall drops is 2.59 times higher compared to the open field, which accentuates the importance of shrub, herb and litter layers in forest ecosystems to protect the soil against erosion. Coalescing drops from leaves and branches (drips) are responsible for this enormous gain in erosive power. Moreover, the results show that the erosion potential under forest is related to the forest structure, especially height and canopy cover. The erosion potential of medium and old grown forests is 1.53 times higher than of young forests. Further, differences in sand loss between Schima superba and Castanopsis eyrei indicate that the erosion potential and the spatial heterogeneity of throughfall is species-specific, highlighting the importance of selecting species for afforestation projects considering soil erosion potential.