Wind tunnel and CFD study of dust dispersion from pesticide-treated maize seed

Highlights

• The first CFD model of seed treatment dust dispersion is proposed.

• Dust particle trajectories are predicted using Lagrangian particle tracking.

• Model predictions were validated with wind tunnel data.

• Detailed description of physicochemical dust properties is necessary.

Abstract

Drilling of treated seeds with vacuum-based precision drills can cause emissions of pesticide-laden dust, which have been linked with declines of pollinator populations. Predicting the drift pattern of this type of dust is challenging because the particles are very irregular in terms of size, shape, density, internal porosity and chemical composition. In this work, a 3D Computational Fluid Dynamics (CFD) model of seed treatment dust drift was developed and validated with wind tunnel data. In the wind tunnel experiment, dust abraded from pesticide-treated seed was separated in three size fractions and released from a point source at a height of 0.7 m at three air velocities. Dust deposition was measured at six distances on the wind tunnel floor. The physicochemical properties of the dust samples were measured and implemented in the CFD model. Lagrangian tracking was used to calculate the dust particle trajectories. The simulated dust deposition patterns agreed with those observed in the wind tunnel trials. It was demonstrated that an accurate, particle size-dependent description of the shape, chemical content and internal porosity of the dust particles was necessary to achieve good validation results. The CFD model can be used as a basis for the simulation of dust drift in the field during sowing.

Introduction

Seeds of many crops are dressed with pesticides to protect the seedlings from pests and diseases. Maize seed is commonly treated with insecticides of the neonicotinoid family, including clothianidin, thiamethoxam and imidacloprid (Krupke and Long, 2015). These insecticides are highly toxic for honey bees (Apis mellifera) and other useful pollinators and can induce various sublethal effects (Godfray et al., 2015, Iwasa et al., 2004, Whitehorn et al., 2012, Stanley et al., 2015, Schneider et al., 2012). In recent years, multiple routes of exposure of neonicotinoids to pollinators have been identified (Krupke et al., 2012, Van Der Sluijs et al., 2015). Emission of abraded seed treatment dust during sowing is one of these routes (Nuyttens et al., 2013). In vacuum-based precision drilling of maize, a central fan generates a depression in the seeding elements of the seeder for seed singulation. If dust particles are abraded from the seed dressing due to friction, they are emitted along with the exhaust air from the fan into the environment. This has caused serious pollinator poisoning incidents in the past (Pistorius et al., 2009, Bortolotti et al., 2009, Cutler et al., 2014).

Since the correlation between pesticide dust drift and honey bee colony collapses was first observed, the phenomenon was studied in controlled conditions and in (semi-)field conditions. The number of true field experiments (Heimbach et al., 2014, Biocca et al., 2015, Girolami et al., 2012, Pochi et al., 2012, Tapparo et al., 2012), in which whole fields were sown with pesticide-treated seed by vacuum precision drills, has been very limited because field studies are expensive, complex, time-consuming and poorly reproducible due to variable wind conditions and dust properties. Furthermore, even when reproducible dust drift patterns are measured, the findings are likely limited to the specific setup and conditions of the study, as a result of the wide variety of soil and meteorological conditions, dust properties, seed drill designs, fan configurations and operational parameters such as the vacuum level.

Complementing experiments with simulations is a useful approach to deal with these limitations. Computational fluid dynamics (CFD) is a modeling technique that is increasingly used for the simulation of complex particle-laden flows (ERCOFTAC, 2008). It was successfully applied to the simulation of phenomena similar to dust drift, such as droplet drift during field crop spraying (Baetens et al., 2007) and orchard spraying (Endalew et al., 2010, Duga et al., 2014, Duga et al., 2015), allowing for a quantitative comparison of sprayer designs and wind conditions. Currently, no model of any kind is available for dust drift from sowing operations.

The aim of this study was to develop a 3D CFD model of the dispersion of seed treatment dust in an air flow, and validate this model with results of a wind tunnel experiment, in which three size fractions of seed abrasion dust were released at three air velocities and deposition was measured at six distances downwind. The physicochemical properties of the wind tunnel dust samples were characterized and implemented in the CFD model. Dust trajectories were predicted using Lagrangian particle tracking (LPT), in which the position of a number of representative particles is tracked over time by assessing the forces that act on the particles and calculating the resulting acceleration. In this case, only gravity and drag were considered. To calculate these forces and the deposition of active ingredient (a.i.) on the ground accurately, it is crucial that the irregular physicochemical properties (Devarrewaere et al., 2015, Foqué et al., 2014) of the seed treatment dust are accurately described. For gravity, these properties are particle size, the true density of the solid material, and the internal air porosity; for drag, this is particle shape. The results of the wind tunnel experiment and the CFD simulations will be discussed and, ultimately, the use of this model in the simulation of dust drift in field-realistic scenarios will be considered.