Biology Meets Turbulence: Effects of Convective Eddies on Pollen Dispersion and Viability

We have combined a large-eddy simulation (LES) version of the WRF meteorological model with a Lagrangian particle dispersion model to predict maize pollen dispersion and viability. Our motivation is to assess the risk that traits in genetically modified (GM) crops can accidentally cross into conventional crops or wild relatives through transport of wind-borne pollen. In some cases, such as production of plant-made pharmaceuticals (PMPs), there is concern over even small probabilities of outcrossing. Assessment of this risk is challenging because it requires not only quantifying the transport and dispersion of pollen but also accounting for the fact that pollen is a living organism that responds to its environment. The Lagrangian approach is well suited to this challenge because it allows diagnosis of environmental conditions that pollen grains experience as they travel. WRF produces fields of wind, turbulence kinetic energy, temperature, and humidity which are then input to the Lagrangian dispersion model. The dispersion model in turn predicts transport of approximately 100,000,000 tracer particles that represent a statistical sample of a pollen cloud. We diagnose vapor pressure deficit at each point along the path of each tracer particle (i.e., virtual pollen grain) in order to calculate moisture content of the pollen grains and consequent loss of viability. Small amounts of viable pollen were predicted to be transported 5 km or more from the source. This contrasts with previous estimates of maize pollen transport using surface wind conditions, which have predicted transport only to a few tens of meters owing to the large terminal fall speed of maize pollen grains. The long-distance transport we predicted is the result of convective updrafts in the daytime boundary layer that lift maize pollen grains to heights of several hundred meters, so that the grains are transported long distances before settling to the ground. We also found that pollen lifted into the upper part of the boundary layer remains viable longer than has been inferred from surface observations of environmental conditions. This occurs because temperature in the daytime boundary layer decreases with height approximately as the dry adiabatic lapse rate while absolute humidity remains approximately constant with height. The upper part of the daytime boundary layer thus is an environment of low vapor pressure deficit which preserves the viability of pollen. These predictions are consistent with studies using piloted and remote controlled aircraft that have detected viable maize pollen through the entire depth of the convective boundary layer. Our results show that risk assessment for genetically modified crops must account for the complex interplay between atmospheric transport and biological processes.