From Coherent Motions to Extreme Values: Decoupling Scalar Dispersion in the Roughness Sublayer

The exchange of passive scalars between complex vegetation canopies and the atmosphere is a critical process governing biosphere-atmosphere exchange, fire propagation, and pollution dispersion. However, a central modeling challenge in the Roughness Sublayer (RSL) is the failure of eddy-diffusivity (K-theory) models, which cannot account for transport against the local mean gradient. This study investigates the mechanisms of counter-gradient (CG) transport and their link to extreme scalar events using wind tunnel experiments of a passive scalar released from a localized source at the top of a two-height canopy.

Simultaneous high-resolution velocity and concentration measurements reveal distinct regions of CG flux in the RSL. Using Quadrant Analysis, we conditionally sample the turbulent scalar flux, distinguishing events based on streamwise and vertical velocity fluctuations. We focus on two primary types of coherent motion: sweeps (high-speed downward motion, u'>0, w'<0) and ejections (low-speed upward motion, u'<0, w'>0). We identify sweeps as the primary drivers of CG transport, entraining low-concentration ambient fluid downwards against the local gradient. Conversely, ejections are found to contribute mainly to down-gradient transport. To quantify the interplay between these coherent motions and the statistical distribution of the scalar, we formulate a semi-analytical closure model employing an orthogonal series expansion of the three-component joint Probability Density Function (PDF). This approach allows us to rigorously link the imbalance of sweep and ejection events to the non-Gaussian tails of the scalar PDF. We demonstrate that the breakdown of gradient diffusion is not a random error, but a deterministic consequence of these extreme, intermittent events. These results provide a mechanistic and statistical basis for improving scalar-transport models at the canopy–atmosphere interface.