Evolution of the nocturnal decoupled layer in a pine forest canopy

Estimates of the biosphere-atmosphere exchange rates measured using the eddy-covariance (EC) technique are often found to deviate from their expected values. The mismatch is caused by a variety of poorly known and quantified processes, such as storage, subcanopy advection, drainage flows and other non-turbulent air motions, which become particularly important at nighttime. In many forest sites, clear and calm nights favor the generation of the phenomenon commonly known as decoupling, when the above- and sub-canopy air layers relation can be significantly weakened. The data obtained above and within a Scots pine forest at the SMEAR II-station in Hyytiälä, southern Finland, were used to study the decoupling conditions. Certain features of the site (sloping terrain, tall pine trees, sharp separation of the canopy and trunk spaces, insignificant undergrowth) facilitate and augment the development of decoupling conditions. As a result, the EC measurement data contain a multitude of severe CO2 flux loss cases biasing the carbon balance estimates. The results concerning decoupling cases are presented and the mechanisms of generation and alteration of the decoupling conditions are discussed. Different regimes of decoupling were detected based on the vertical profile of the mean wind direction. Decoupling was defined as the periods when the wind directional shear in the canopy or trunk space exceeded predefined thresholds. In at least 18.6% of all nighttime periods, decoupling conditions were identifiable by high wind directional shear in the canopy sublayer. A close relation between the stability estimated by the Richardson number and the decoupled layer thickness was observed. Decoupling interface tended to move to a higher level as stability increased. Drainage flow was detected near the ground in the cases of maximum decoupled layer thickness. Finally, the difference was drawn between the drainage flow and the decoupled layer based on observational evidence.