An Eulerian model for pollen dispersal over heterogeneous landscapes

With the introduction of genetically modified (GM) crops, predictive tools modelling dynamics, turbulence and pollen concentrations over heterogeneous canopies are required to assess the cross-pollination rates between GM and conventional crops. Pollen dispersal is indeed closely linked to flow dynamics and turbulence within and above crops, themselves depending on the landscape heterogeneities such as roughness changes, field discontinuities, gaps, roads, tree lines, forest plots, etc. It is therefore of primary importance to assess the influence of these heterogeneities on pollen dispersal. To address this problem we have adapted a CFD type model, Aquilon, to canopy flow and added an advection-diffusion conservation equation for pollen particles. The flow equations in the canopy are modified to account for the drag forces and the production of turbulent kinetic energy by the vegetation. Turbulence is modelled statistically with a k-ε closure scheme. The relative velocity between air parcels and particles is represented through the addition of a particle settling velocity, and deposition is represented by a sink term accounting for impaction and sedimentation. The dynamic part of the model has been previously validated in 2D cases (continuous and discontinuous vegetation canopies) against wind-tunnel and in-situ measurements as well as in a 3D heterogeneous urban park. This model is validated against two field experiments where the airborne concentration and the deposition rate of pollen were measured downwind of maize plots. The model simulates correctly concentration profiles but underestimates the maximum in the deposition rate observed just behind the source plot. This discrepancy seems to be due to an underestimation of the average maize pollen settling velocity which may be increased by the turbulence in the plot wake, as compared to its value in an undisturbed flow.