An individual-based model of Trichogramma foraging behaviour: parameter estimation for single females.
Abstract
The aim of this work was to estimate spatio-temporal parameters of an individual based model of Trichogramma female foraging behaviour in a mass rearing system. The system consisted of Trichogramma brassicae and the host Ephestia kuehniella. The foraging behaviour of single females was examined using direct visual observation for building an ethogram with the following behavioural states: searching, drumming, drilling, oviposition and host-feeding. The distribution of the time spent in the behavioural states was estimated using the gamma distribution and a transition probability matrix between the behavioural states was built. The mean time duration for searching, drumming, drilling and oviposition was similar for all behavioural states in the experiments and amounted to half of a minute. Host feeding, however, lasted more than 2 min. An image analysis system was used to record the paths of single Trichogramma females foraging on a spatial lattice arrangement of host eggs. Paths were analysed using a vector autocorrelation function. The velocity vector was analysed separately for the speed and directional components. The analysis showed a short time memory for the directional process and a negligible memory for the speed component. A stochastic differential equation, an Ornstein Uhlenbeck-process, was used for modelling the directional changes. The data were then combined with the information about speed changes and simulated paths were obtained. Visual assessment with recorded paths showed good agreement; the characteristic loops and circlings, as well as the longer meanders were faithfully represented. The theoretical and simulated processes overestimated the mean squared displacement in the first few hundred seconds but provided a good fit thereafter.