Abstract. Stably stratified canopy flow in complex terrain has been considered a difficult condition for measuring net ecosystem-atmosphere exchanges of carbon, water vapor, and energy. A long-standing advection error in eddy-flux measurements is caused by stably stratified canopy flow. Such a condition with strong thermal gradient and less turbulent air is also difficult for modeling. To understand the challenging atmospheric condition for eddy-flux measurements, we use the renormalized group (RNG) k-ε turbulence model to investigate the main characteristics of stably stratified canopy flows in complex terrain. In this twodimensional simulation, we imposed persistent constant heat flux at ground surface and linearly increasing cooling rate in the upper-canopy layer, vertically varying dissipative force from canopy drag elements, buoyancy forcing induced from thermal stratification and the hill terrain. These strong boundary effects keep nonlinearity in the two-dimensional NavierStokes equations high enough to generate turbulent behavior. The fundamental characteristics of nighttime canopy flow over complex terrain measured by the small number of available multi-tower advection experiments can be reproduced by this numerical simulation, such as (1) unstable layer in the canopy and super-stable layers associated with flow decoupling in deep canopy and near the top of canopy; (2) sub-canopy drainage flow and drainage flow near the top of canopy in calm night; (3) upward momentum transfer in canopy, downward heat transfer in upper canopy and upward heat transfer in deep canopy; and (4) large buoyancy suppression and weak shear production in strong stability.

This study generated eddy covariance data to investigate atmospheric dynamics leeward of a small, forested hillside in upstate New York. The causes and effects of recirculation eddies were examined to support the larger goal of improving measurement of the exchange of energy, moisture, and trace gases between the terrestrial biosphere and the atmosphere over complex terrain. Sensors operated at five different altitudes on two separate towers—one at the top of the hill and one down the slope to the east—for approximately 8 weeks in the spring of 2013. During the experiment, the vertical potential temperature gradient was found to be the primary factor for determining whether winds interacting with the terrain features caused a recirculating eddy leeward of the hill. The study found evidence that the recirculation influenced carbon dioxide flux and caused the air column to be vertically well mixed.