Simulation of honeybee nectar foraging for determining effects on local flora

Abstract

This work presents a simulator of honeybee (Apis mellifera) colony foraging behavior. Honeybees are social insects that show a high level of organisation and very complex behaviour. This includes the coordination of foraging efforts thanks to their sophisticated mechanism of advertising and recruitment, which includes the well-known waggle dance. This foraging process is needed for general pollination and makes honeybees crucial for a sustainable environment. There exists previous work in the area of honeybee modelling and simulation. Some simulator models use such a fine resolution that can only run for real-time hours. Other work is focused on simply replicating a published observation. Instead, we are interested in simulating long time periods (e.g. months) and learning how the foraging spread patterns and distributions can affect the local flora or contribute to the transmission of diseases or parasites. The main motivation for this work is to support the decision making process in the Biosecurity context. Biosecurity plays a key role in the economic viability of Australia's plant industries due to the devastating effects that plant pests and diseases can cause. Hence, the importance of preventing the spread of diseases between apiaries or to a disease free area (e.g. Fire Blight (Erwinia amylovora) is a serious disease in many fruits and uses honeybees as one of its main vectors for transmission). The simulator is Web based1 so that users do not have to install or update software, or worry about enough computational power or platform compatibility. The user interface allows the introduction of the simulation parameters (e.g. spatial resolution and size, model configuration settings, visual specification of the simulation environment on the map) and receive a detailed results report. A simulation of several weeks usually takes a few seconds to complete. The simulation model is stochastic and discrete in time with 1-trip time steps, which is approximated by difference equations. It is also explanatory, meaning that its behavioural rules depend on well-established entomological knowledge. It can support any number of beehives and floral patches over a two dimensional landscape of unrestricted size and varying resolution. The main simulation model has four interrelated models that include i) a spatial model that represents the landscape to simulate by a grid of cells; ii) a flora model that represents a patch with a number of flowers and seasons defined by nectar quality and quantity; iii) a weather model that considers time periods of temperature and daily hours of light; iv) a foraging model, which represents the foraging behaviour of a honeybee colony.