Modeling population dynamics of pests and beneficials: a basis for an IPM decision support system in protected tomato and pepper crops.

The main applicant, Biobest Group N.V., is worldwide leader in biological pollination and biological control in protected crops. An important pillar within Biobest is the free technical service that is offered when a grower buys products. A technical advisor of Biobest will visit the grower weekly or biweekly to discuss and optimize the IPM strategy. Due to the yearly growth of 15 to 20%, Biobest faces many challenges concerning recruitment, training and logistics. An accurate, reliable and trustworthy advice starts with a good monitoring of pests and beneficials in time and space. Unfortunately, monitoring comes with a cost, which is often a hurdle for many growers. Therefore, Biobest is searching the market for affordable alternatives and invested a lot of resources in new innovative technology. Crop-Scanner is Biobest's monitoring tool for data input. Flying insects can be caught with yellow sticky traps. For this purpose, Biobest developed the Crop-Scanner add-on Trap-Scanner, which is an algorithm that identifies six different pests and beneficials on smartphone pictures. Recently, this system was optimized in collaboration with the company PATS (The Netherlands) to a fully automatic system with a fixed camera (Trap-eye). March 2021 Biobest invested 7.5 million euro in the Canadian company Ecoation. This company offers a visual attractive dashboard where growers can enter real-time monitoring data and visualize heatmaps. The ultimate goal of Biobest is a self-learning prescriptive DSS that will 1) give automatic IPM advice, and 2) increase the efficacy of the released beneficials. The roadmap towards this goal started with detailed monitoring data (descriptive analytics) followed by a rule-based decision support system (DSS) (IF THEN ELSE model). This project will focus on the next step, the predictive analytics. In short, this step describes population models that 1) predict whether a pest is under control or not based on the number of beneficials on the plants, or/and 2) predict population densities of predator and prey in the near future. Simple one on one predator-prey models were already developed in the LA-Traject PeMaTo (nr: 140948) (Moerkens et al., 2021, Brenard, 2020). Mathematical models based on differential equations are the standard in population ecology (Gause et al., 1936). However, the scientific novelty lies in the complex ecological dynamics that occur once multiple pests and predators are present. The added value of artificial intelligence within the field of population ecology will be investigated.

Keywords:HORTICULTURE, INTEGRATED PEST MANAGEMENT

Disciplines:Auto-ecology, Agricultural plant protection

Project type:Collaboration project