Expert-based model of the potential for natural pest control with landscape and field scale drivers in intensively managed cereal-dominated agricultural landscapes

Related GLP Member: Pierre Chopin


  • We need indicators to identify landscapes where natural pest control can be enhanced.
  • Expert-based spatial models can integrate both field and landscape-scale knowledge.
  • Use in Sweden and validation with field data show promise for carabids predictions.
  • The model can be used to explore practices enhancing large-scale carabid pest control.
  • It can be used for large-scale ecosystem service assessment along with other tools.


Agricultural intensification has contributed to the loss of biodiversity and of the ecosystem services that it supports, such as natural pest control. Decision support tools are needed to understand and predict where natural pest control can be enhanced and pesticide applications decreased. While many studies have assessed the impact of field and landscape-scale management in a range of crops, few attempts have been made to synthesize this knowledge in a single model. In this study, we developed an expert-based moving window model of natural pest control potential. This model builds on the knowledge of 52 experts across Europe regarding the importance of herbaceous areas, forest interiors, and edges, and field scale agricultural management practices (i.e. farming system, field size and crop diversity) for the abundance of generalist predators (e.g. carabids, spiders), specialist predators (e.g. coccinelids) and parasitoid natural enemies. We assessed the model’s performance by comparing its predictions to field data on natural enemy abundance from 117 sites in Sweden. The natural pest control potential scores predicted by the model explained 11% of the variation in carabid field abundances. However, the model’s performance was less satisfactory for spiders and parasitoids. We provide guidance for improving this indicator, particularly by incorporating more ecological processes, such as accounting for the functional diversity of spiders and the density-dependent effects of parasitoid-host interactions. In addition, the model could be further refined by accounting for non-linear relations and potential threshold effects and interactions among field and landscape-scale management practices. In its current state, the developed indicator can be used to identify areas where further ecological intensification practices can be promoted to enhance natural pest control potential, especially for carabids.