Publications

Dispersal of a carabid beetle in farmland is driven by habitat-specific motility and preference at habitat interfaces

Allema, Bas; Hemerik, Lia; Rossing, Walter A.H.; Groot, Jeroen C.J.; Lenteren, Joop C. van; Werf, Wopke van der

Summary

Carabid beetles are common predators of pest insects and weed seeds in agricultural systems. Understanding their dispersal across farmland is important for designing farms and landscapes that support pest and weed biological control. Little is known, however, about the effect of farmland habitat discontinuities on dispersal behaviour and the resulting redistribution of these beetles. We released 1,985 well-fed and 1,680 food-deprived individuals of the predatory carabid beetle Pterostichus melanarius (Illiger) (Coleoptera: Carabidae) on a farm in Wageningen, The Netherlands. We recaptured 23.6% of those beetles over a period of 23 days in 2010. The farmland comprised agricultural fields with various crop species and tillage, separated by strips of perennial vegetation. We developed discrete Fokker-Planck diffusion models to describe dispersal based on motility (m2 day−1) and preferential behaviour at habitat interfaces. We used model selection and Akaike’s information criterion to determine whether movement patterns were driven by variation in motility between habitats, preferential behaviour at habitat interfaces, or both. Model selection revealed differences in motility among habitats and gave strong support for preferential behaviour at habitat interfaces. Behaviour at interfaces between crop and perennial vegetation was asymmetric, with beetles preferentially moving towards the crop. Furthermore, beetles had lower motility in perennial strips than in arable fields. Also between arable habitats movement was asymmetric, with beetles preferentially moving towards the habitat in which motility was lowest. Neither crop type nor tillage explained differences in motility between crop habitats. Recapture data representing dispersal patterns of beetles were best described by a model that accounted for differences in motility between farmland habitats and preferential behaviour at habitat interfaces. Motility in farmland and behaviour at interfaces can also be estimated for other organisms and farmland habitats to support design of farmland conducive to natural pest suppression. Landscape design for early recruitment of carabids into arable fields should take into account the quantity and quality of resource habitats in the landscape, their proximity to crop fields, movement rates, and the possibility of movement responses at interfaces between landscape elements.