Performance of soil and water conservation practices in the erosion evolution process, runoff dynamics and surface roughness

Soil and water conservation practices are used widely to prevent soil erosion and protect soil and water resources, which is significant for ecological restoration and food security. However, rill evolution processes, erosion and deposition characteristics and critical hydrodynamic parameters need more research. In order to investigate the effect of soil and water conservation practices on soil erosion dynamics, simulated rainfall experiments were undertaken in a laboratory on 15 degrees loess slopes with engineering measures (fish- scale pits, FSPs), tillage measures (artificial digging, AD; contour ploughing, CP) and bare slope (CK). The results showed that: (1) during rill erosion, hillslopes with FSPs, CP and AD were more likely to develop wide and shallow rills, while a bare slope (CK) was more likely to develop narrow and deep rills. At the end of the experiment (cumulative rainfall was about 150 mm), headward retreat erosion dominated the AD slope (maximum rill length: 3.27 m), side- wall expansion erosion dominated the CP slope (maximum rill width: 0.522 m) and bed incision erosion dominated the CK (maxi-mum rill depth: 0.09 m); (2) soil and water conservation practices reduced surface erosion and sediment transport and runoff velocity. However, the positive effects disappeared when rainfall amounts exceeded 82.5, 105 and 127.5 mm for FSPs, CP and AD, respectively; (3) for runoff kinetic energy and runoff shear strength of 3 J and 1.5 N/m(2), respectively, soil and water conservation measures had greater anti-erosion abilities than CK; (4) as rainfall duration increased, surface roughness, runoff rate and sediment concentration increased on the CK and FSP treatments, but decreased on the CP and AD treatments. This study has important implications for managing different soil and water conservation measures based on rainfall conditions and offers a deeper understanding of soil erosion processes.