Evaluation and comparison of pollination service indicators at agricultural landscape level

International audience

English. The current pollinator decline is a major issue, notably because one-third of the world agricultural production depends on pollination. However, quantifying whether the loss of some pollinators translates into less efficient pollination remains challenging, particularly at the scale of farmland territories, which is a relevant scale for action. On the one hand, clear links have been established between the abundance and diversity of pollinators and pollination efficiency only over relatively small areas, because time-consuming experimental methods prohibit studies over broader areas. On the other hand, nationwide or continentwide indices exist but they either measure potential pollination only or have not been validated against other metrics characterizing pollinators or pollination.The aim of our study was to assess the relevance of two pollination indicators developed at national or broader scale, to characterize pollination over a 950-ha farmland territory of field crops (Fénay platform near Dijon, France). We specifically evaluated the realized pollination service indicator, based on cropyields and their dependence on pollinators, and the potential pollination indicator simulated by the InVEST model, based on land use and landcover as proxies for potential floral and nesting resources.We examined the spatiotemporal variation of these indicators over the Fénay area between 2005 and 2019, their environmental correlates (farming practices, landscape, and weather) and how they were related to each other.Our results suggest that the two indicators designed for large spatial extent can prove useful within farmland territories: both indicators exhibit significant variation across the study area, some of which could be related with environmental variables known to influence pollinators. For example, the spatial variation in the realized pollination service index correlates with farming practices, e.g. less efficient pollination in plots with more herbicide use. No correlation with the distance to the nearest semi-natural element was found, which contradicts results from previous studies at the field scale. The temporal variation in the pollination service index correlates negatively with the duration and intensity of summer heat waves. The potential pollination indicator will be compared with the realized pollination indicator, and in the future with field measurements of pollination efficiency for cross validation.