Agroecological food web modelling to evaluate and design organic and conventional agricultural systems

Abstract

The question of whether organic or conventional agriculture is most suitable for meeting world food requirements and improving resilience to climate change is the topic of much current research. Most comparisons, however, focus either on output efficiency such as yields or on impacts of different nutrient management strategies on the sustainability of the agrosystems, and the impacts of each on the agricultural food webs and beneficial insects responsible for pest control – or outbreaks – has been often overlooked. While standard cropping models can explain if, why and how organic nutrient and crop management work and how they should be adapted to climate change, the lack of mechanistic models of agroecology prevents us from explaining why and how passive and active biocontrol and integrated pest management function, when they do not function, and what optimal management strategies could be employed.In this research, we show that agroecological food web models calibrated with field population dynamics data can be used to demonstrate the mechanisms behind food web dynamics that have been previously observed in the field. Results of scenario simulations show that chemical control provides immediate relief from pest pressures, but at high risk of later pest resurgence if control is not repeated; on the contrary, biological control requires more time to reduce pest populations to acceptable levels but with minimal risk of causing resurgence. In all cases, success of pest control measures is highly dependent on the date of action. In addition, the use of modelling tools to optimise biological control application dates led to much better control than either fixed date or pest population threshold-based applications. These analyses and resulting integrated pest management intervention recommendations are only possible with agroecological food web population dynamics models.We encourage future studies to examine more complex food webs from a variety of agroecosystems to test whether functional responses differ significantly, and hope that this approach will succeed in bringing agroecological food web predictive modelling to the level where it can routinely be used as a decision-making support tool, as hydrological and crop models are employed today.