Using Large-Eddy Simulations and experiments to quantify long-range tree seed dispersal

One of the key challenges to quantifying long-distance seed dispersal by wind is describing the complex and three-dimensional eddy motion inside and above canopies. The Large-Eddy Simulations (LES) option of the Regional Atmospheric Modeling System (RAMS), a state-of-the-art, three-dimensional, non-hydrostatic atmospheric numerical model was modified to simulate atmospheric turbulence within and above tree canopies. The primary modification included a new coordinate system designed to contour the shape of trees, thereby permitting a more faithful descriptor of the interaction between canopy morphology and the "energetic" eddies. A detailed energy balance of the tree canopy was also introduced in the model. The model was set up to simulate turbulence within a second-growth 180-year old hardwood stand at Duke Forest having a mean canopy height of 30 m and a maximum leaf area index of 6. The LES was initialized with measured hourly mean wind speed, temperature, and humidity profiles near a 45-m tower in this stand. The various statistical moments needed to drive seed trajectories were compared with the corresponding moments measured at the tower. A Lagrangian Dispersion Particle Model (LDPM) driven with the turbulent flow simulated with the RAMS-LES was then used to quantify seed dispersal at that site. The combined RAMS-LES and LDPM simulations adequately match the patterns of seed dispersal as quantified by observations conducted near the tower.