Measurements and Langevin simulations of mean tracer concentration fields downwind from a circular line source inside an alfalfa canopy

Two Langevin simulations of trajectories of marked fluid elements in inhomogenous turbulence, where the Lagrangian length and vertical velocity scales are height dependent, were compared with field data. A CO₂ tracer was released from a circular line source and the concentration profiles were measured for diffusion distances of 50 and 100 cm inside and above an alfalfa canopy.One of the simulations, suggested by Wilson et al. (1983), biases the vertical velocities by adding a mean upward drift. The second simulation proposed here by-passes this difficulty by reflecting marked particles according to a probability calculated from the gradient in vertical velocity variance between the beginning and the end of each step. This simulation also makes use of a constant time-scale within the canopy, following preliminary results from a turbulence experiment within a forest (Leclerc, 1987).Comparing the results of these simulations with the field data shows that the simulation proposed by Wilson et al. (1983) does not correctly reproduce the difusion for the larger fetch in systems exhibiting strong gradients in vertical velocity variance. Instead, the modelled plumes exhibit a bulge at the source height whereas the field data show smooth profiles. In addition, the modelled plumes overestimate the vertical spread of the plumes, which is possibly due to the inadequacy of the approach in severely inhomogeneous systems. In contrast, the results from the tracer experiments indicate that the diffusion can be better reproduced with the use of a reflection probability calculated at each step. The discrepancies between the experimental results and the simulation using a reflection probability are attributed to stability effects.