A spatiotemporal model for predicting grain aphid population dynamics and optimizing insecticide sprays at the scale of continental France

We expose here a detailed spatially explicit model of aphid population dynamics at the scale of a whole country (Metropolitan France). It is based on convection–diffusion-reaction equations, driven by abiotic and biotic factors. The target species is the grain aphid, Sitobion avenae F., considering both its winged and apterous morphs. In this preliminary work, simulations for year 2004 (an outbreak case) produced realistic aphid densities, and showed that both spatial and temporal S. avenae population dynamics can be represented as an irregular wave of population peak densities from southwest to northeast of the country, driven by gradients or differences in temperature, wheat phenology, and wheat surfaces. This wave pattern fits well to our knowledge of S. avenae phenology. The effects of three insecticide spray regimes were simulated in five different sites and showed that insecticide sprays were ineffective in terms of yield increase after wheat flowering. After suitable validation, which will require some further years of observations, the model will be used to forecast aphid densities in real time at any date or growth stage of the crop anywhere in the country. It will be the backbone of a decision support system, forecasting yield losses at the level of a field. The model intends then to complete the punctual forecasting provided by older models by a comprehensive spatial view on a large area and leads to the diminution of insecticide sprayings in wheat crops.     

Tracing individual movements of aphids reveals preferential routes of population transfers in agroecosystems.

Agricultural pests are not restricted to crops, but often simultaneously or successively use different cultivated and uncultivated hosts. Nevertheless, the source-sink role of cultivated and uncultivated habitats in the life cycle of crop pests remains poorly understood. This is largely due to the difficulty of tracking displacements of small organisms in agricultural landscapes. We used stable-isotope ratios in order to infer the natal host plant of individuals of the English grain aphid Sitobion avenae colonizing wheat fields in autumn. We showed that among the numerous plant sources of S. avenae, maize, which has been intensively grown in western France since the 1960s, provided most aphids that attack wheat fields early in autumn. This study illustrates how insect pests respond to land-use changes within a relatively short period of time, rapidly acquiring a new host that in turn affected their population biology considerably by playing a pivotal role on their annual life cycle.     

Hierarchical Bayesian Modelling of plant colonisation by winged aphids: Inferring dispersal processes by linking aerial and field count data

Understanding and modelling insect pest dispersal is an important prerequisite for designing integrated pest management programs. Nevertheless, studies investigating the dispersal of small insects in natural conditions remain scarce mainly because of the difficulty of tracking the movements of these organisms. Here we propose to use Hierarchical Bayesian Modelling (HBM) framework to gain knowledge on hidden processes that cannot be observed directly in natura, such as insect landing and insect mortality, through the definition of latent variables. An HBM describing crop colonization by winged aphids was fitted to a large dataset of field observations issued from a long term survey at a wide scale of both aerial and field densities of the bird cherry-oat aphid Rhopalosiphum padi. This study provides the first evidence that suction trap data are reliable proxies of aphid colonizing rates in cereal fields in autumn and can be a nice alternative to the very time-consuming crop sampling. The proportion of winged aphids landing in cereal fields is shown to vary between regions according to the degree of investment of local R. padi population in sexual reproduction. Results also indicate that under autumnal field conditions, less than 5% of winged aphids survive more than 10 days after landing. This HBM provides the basis of a predictive model for aphid crop colonization that fully accounts for all sources of uncertainty. It should be of great value to improve the trust of users in any decision making systems.