Runoff generation related to intra-field soil surface characteristics variability: application to conservation tillage context.

Abstract Conservation tillage (CT) constitutes one way to limit on- and off-site impacts of soil erosion. Literature analysis suggests that CT is more effective in reducing soil erosion than surface runoff. The effect on runoff mitigation may be rather limited and in some cases, CT causes runoff to increase compared to conventional tillage systems. From the review of infiltrometry and runoff studies, we have concluded that runoff mitigation can be explained by infiltrability improvement of the soil profile and limitation of surface structural degradation. As data on soil profile infiltration are now better documented, we choose to study how the surface may limit runoff generation. To describe soil surface ability to generate runoff, soil surface characteristics (SSC) descriptors were used. This study investigates small-scale spatial variability of SSC which is usually smoothed by runoff measurements at higher scale. Two patterns of SSC were especially studied. One is the pattern created by tillage in maize field, made up of three features (seedbed, seed line and wheel track) which was investigated through small plots SSC surveys. The other is the pattern created by the alternation of mulch and bare patches which was investigated by rainfall simulations. SSC surveys reveal that mulch and crust cover varied amongst features created by tillage. Crust extent, on the whole, is higher under conventional tillage but presents differences between tillage features. Under reduced tillage (RT), seed lines and wheel tracks present a lower mulch cover and a higher crust extent. No-tillage (NT) is characterized by the absence of the tillage pattern and a lower crust extent. Concerning the influence of the mulch pattern, no effect was observed on runoff amounts under simulated rainfall. Further research is required to assess pattern effects in CT context, especially measuring which features mostly contribute to runoff. There is also a clear need to adapt infiltrometry and rainfall simulation procedures to correctly take CT specificity (e.g. mulch presence) into account.