Biological control provided by entomophagous arthropods is an ecosystem service with the potential to reduce pesticide use in agriculture. The distribution of entomophagous arthropods and the associated ecosystem service over crop fields is affected by their dispersal capacity and landscape heterogeneity. Current knowledge on entomophagous arthropod distribution and movement patterns, in particular for soil dwelling predators, is insufficient to provide advice on how a production landscape should be re-arranged to maximally benefit from biological pest control. Movement has mainly been measured in single habitats rather than in habitat mosaics and as a consequence little information is available on behaviour at habitat interfaces, i.e. the border between two habitats.
This study contributes to insight into movement patterns of the entomophagous arthropod Pterostichus melanarius (Illiger) in an agricultural landscape as a knowledge basis for redesign of landscapes for natural pest control. Movement patterns were studied with video equipment in experimental arenas of 5 m2 and with mark-recapture at much larger scales in the field. Interpretation of the results was supported by diffusion models that accounted for habitat specific motility µ (L2 T−1), a measure for diffusion of a population in space and time, and preference behaviour at habitat interfaces.
Movement of carabids has mostly been quantified as movement rate, which cannot be used for scaling-up. Available information on movement rate of carabids was made available for scaling-up by calculating motility from published data and looking for patterns through meta-analysis of data from thirteen studies, including 55 records on twelve species. Beetles had on average a three times higher motility in arable land than in forest/hedgerow habitat. The meta-analysis did not identify consistent differences in motility at the individual species level, and a grouping of species according to gender or size did not demonstrate a significant gender or size effect.
A methodology to directly estimate motility from data using inverse modelling was evaluated on data of a mass mark-recapture field experiment in a single field of winter triticale (x Triticosecale Wittmack.). Inverse modelling yielded the same result as motility calculated from squared displacement distances. In the first case, motility was calculated as an average over motility of individuals, in the second case motility was estimated from a population density distribution fitted to the recapture data. The similarity in motility between these two very different approaches strengthens the confidence in motility as a suitable concept for quantifying dispersal rate of carabid beetles, and in inverse modelling as a method to retrieve movement parameters from observed patterns.
Effect of habitat heterogeneity on movement behaviour
The effect of habitat heterogeneity on movement behaviour was studied for P. melanarius across adjacent fields of oilseed radish (Raphanus sativus) and rye (Secale cereale) in a mark-recapture experiment. The field study was complemented by observations on movement behaviour in the experimental arena. Motility was neither significantly different between the crop species in the field nor in the arena. Overall movement in the field was significantly affected by behaviour at the interface between the crops. Beetles moved more frequently from rye to oilseed radish than in the opposite direction. The arena data indicated greater frequency of habitat entry into oilseed radish as compared to rye. Analysis of video tracking data from the arena resulted in estimates of motility that, when scaled up were close to those obtained in the field. Thus, the studies at the smaller and larger scales gave qualitatively and quantitatively similar results.
The effect of habitat heterogeneity on within-season dispersal behaviour was further explored in an agricultural landscape mosaic comprising perennial strips and different crop species with distinct tillage management. Semi-natural grass margins were functionally different from the crop habitats. Motility was lower in margins than in crop habitats, and at the crop-margin interface more beetles moved towards the crop than to the margin. Margins thus effectively acted as barriers for dispersal. In the crop habitats motility differed between fields but no consistent relations were found with crop type, food availability or tillage. Based on the motility in crop habitats P. melanarius was predicted to disperse over a distance of about 100 – 160 m during a growing season in a landscape without semi-natural elements. Given this range little redistribution of beetles is expected between fields within a growing season, even more when fields are surrounded by grass margins or hedgerows, meaning that the success of biological control by this species is more dependent on field management affecting local population dynamics than on habitat heterogeneity.
Predictive dynamic models for population spread
This thesis has resulted in a methodological approach to quantify dispersal behaviour of ground-dwelling insects from mark-recapture data in heterogeneous environments using inverse modelling. The combination of models and data proved to be powerful for studying movement and contributes to the development of predictive dynamic models for population spread of entomophagous arthropods. These models for population spread may be used as part of multi-objective assessment of alternative landscape configurations to find spatial arrangements of land use that maximize the ecosystem service of biological control as part of a wider set of landscape functions.