Spatially Explicit Individual-Based Modelling Of Insect-Plant Interactions: Effects Of Level Of Detail In Queensland Fruit Fly Models

A form of agent-based modelling known as individual-based modelling has been used widely to simulate hypothesised mechanisms underlying processes in the real world, especially in the field of behavioural ecology. To apply this approach to insect-plant interactions the Queensland fruit fly (Qfly), Bactrocera tryoni (Froggatt) (Diptera: Tephritidae), a major native insect pest of horticulture in Australia, will be examined. Due to restriction on use of some chemicals, currently there is no suitable management system for control of Qfly. However, better Qfly management can be potentially advanced by targeted behavioural, physiological and ecological research on this species. Many control options depend on behavioural fundamentals of Qfly such as foraging and movement patterns (e.g. trapping control technique), and the use of resources like protein and food sources (e.g. bait spray control technique).In the present study, two 3-dimensional individual-based models with different levels of detail have been developed to simulate Qfly behaviour and movement patterns on host plants, one with NetLogo, and the other with L-studio. Both provide a dynamic platform for simulating insect movement decisions underlying searching and feeding behaviour. The spatial unit in the NetLogo model is based on 'vegetation cubes', while the Lindenmayer System (L-System) formalism-based approach underlying L-studio allows the model to have more detailed plant architecture, with individual leaves and stem segments. Both models predict that Qflies spend more time in the mid to upper canopy, which shows a good agreement with published literature. The overall objective is to test whether it is possible to predict fruit fly movement and distribution based on hypothesized behaviour and also assess whether such a model can provide insights that may have been missed in past experimental studies, investigate what level of detail is best used for addressing different types of questions from entomologists, and to show the advantages and disadvantages of both models.Our study suggests that, the NetLogo model can be better used to investigate scientific questions like insect spatial population distribution on plant canopies and how different tree architectures affect their behaviour, while the L-system model is better to use to look at how foliage density and foliage position affect fruit flies behaviour and to simulate landscape scales such as orchards including multiple trees. Using the model with the right level of detail to inform, develop and test research allows new insights into insect-plant interactions and can inform experiments carried out in the field that have application in better pest management.