In which cropping systems can residual weeds reduce nitrate leaching and soil erosion?

Weeds are often solely considered with a negative viewpoint, but they may also provide services for agroeco-systems. Especially, the residual weed flora that is tolerated by integrated crop protection may contribute to a reduction of nitrate leaching and soil erosion during the summer and autumn fallow. To date, the determinants underlying these environmental weed benefits are largely unknown. The present study developed new indicators to account for the potential beneficial role of annual weed flora to reduce nitrate leaching and soil erosion at the field scale, and then calculated them from the outputs of a weed dynamics model. The aim was to analyse which cropping systems facilitate residual weed flora to reduce nitrate leaching and soil erosion, while minimizing negative weed impacts on crop production. When developing the indicators, the potential weed-based reduction of nitrate leaching was considered to increase with both the growth of the weed flora and the weed species potential to take up nitrogen; the potential weed-based reduction of soil erosion was assumed to increase with soil cover by weeds when soil cover by cash crops was low. Our simulation study included 259 arable cropping systems (covering a wide range of herbicide and tillage intensity, with each cropping system simulated over 28 years and repeated 10 times with 10 different weather series) in which the dynamics of 25 annual weed species was simulated. Simulations showed that the cropping systems promoting a high potential weed-based reduction of nitrate leaching were generally also those with a high potential weed-based reduction of soil erosion, pointing to a compatibility between these benefits provided by the weed flora. However, the cropping systems promoting these environmental benefits were generally also those that presented the highest crop yield losses. Tillage and crop rotation were identified as the cultural techniques with the greatest influence on the potential weed-based reduction of nitrate leaching and soil erosion, while herbicides had a more limited effect. Most of the studied cropping systems (representing "real situations" in which farmers or experimenters make the decisions about crop rotations and cultural techniques) tended to favour low crop yield losses rather than high weed-based environmental benefits. Interestingly, a few systems achieved both objectives. Systems that achieved relatively low crop yield losses and high weed-based environmental benefits mainly combined infrequent superficial tillage operations, with a low proportion of winter crops in the rotation and a very low tillage depth. Finally, only a few weed traits determined the role of the weed flora to potentially reduce nitrate leaching and soil erosion. They were seed traits (seed lipid content, seed area per weight and seed coat thickness), driving the early and fast appearance of the weed canopy after weed seed shed. This suggests that, for annual weed species, a high weed flora potential to reduce nitrate leaching and soil erosion is not restricted to specific weed species able to take up nitrogen and cover soil. Thus, our simulation study indicates that such a high potential to reduce nitrate leaching and soil erosion could therefore be reached in very different agroecosystems in terms of weed seed bank.