6533b82ffe1ef96bd1296207

RESEARCH PRODUCT

Characterization of trophic links in an agrosystem : weed seed predation by carabid beetles

subject

description

Agroecology gives evidence of new perspectives in agriculture, and open doors for alternative weed management approaches through different regulation mechanisms. Seed predation is one of them, and the interest is growing to deepen knowledge, relative to this process. This Phd work aims (i) to describe weed seed predation, the dominant carabid predators and the available resource at a local scale in a common crop (winter wheat); (ii) to study predation profiles of main predators, and the pair-wise correlations between the biological components of the system; (iii) to bring new insights regarding potential regulation of weed communities, through seed predation by carabid beetles. Complementary approaches have been used to tackle these aspects: a long-term field study, some cafeteria tests, a simulation of annual predation rates from point-to-point estimates, as well as a large-scale field study of predators and seed resource. Study of spatio-temporal variations of predation, via field experiment during the main activity period of carabids, has shown a three-peak pattern, one of them corresponding to predators’ peak, just before harvest. Weed resource available on soil surface is high in the first and last peaks of predation, and no increase of resource has been observed after harvest. These dynamics are independent of the spatial scale (intra or inter-field) and carabid species. The different weed species do not show the same levels of predation, and, as a whole, small-seeded species are more eaten, in both lab and field conditions. However, carabids have different preferences among trophic guilds, but also within a same guild. Therefore, in order to study predation, it is essential to combine predators and seed data, as well as considering preferences of predators. A simulation of annual seed predation shows important rates of seed loss, reaching up to 80% for the most predated species, V.arvensis. A large scale study of seed bank shows a negative correlation between predators’ activity density and seed bank change, assuming that there is a seed bank regulation by carabid predators. All together, these results show that functional diversity of predators’ and weeds’ communities is a key factor in understanding predation.