Addressing forest canopy decoupling in eddy covariance flux measurement networks

The eddy covariance (EC) method is the standard technique for determining forest ecosystem-atmosphere turbulent exchange, however, it encounters a significant challenge: the air masses below the canopy often become decoupled from the air masses above it. Consequently, the EC measurements of scalar fluxes (e.g. H₂O and particularly CO₂) above the canopy can be biased due to missing signals from below-canopy processes. This decoupling is strongly site dependent and influenced by atmospheric conditions, canopy properties and tower-surrounding topography. Multiple approaches have been developed in the recent decades to address decoupling (e.g. u_* filtering, quality flags for flux measurements, storage change evaluations, direct advection measurements), however, all of them appeared to be insufficient to fully tackle the problem. A promising additional approach is based on subsequent EC measurements below and above the canopy. Specifically, examining the correlation of the standard deviation of vertical wind (obtained via sonic anemometers) below and above the canopy provides insight into the coupling state, as this correlation remains linear during fully coupled periods.

To date, there is no standardized approach to address decoupling yet, hence, it is commonly not explicitly considered in EC measurement networks and infrastructures such as FLUXNET or ICOS (Integrated Carbon Observation System). A specialized working group within ICOS strives for addressing this by initiating an extensive multi-site experiment. This multi-site experiment aims to i) evaluate the performance of different types of sonic anemometers below canopy for decoupling investigations, ii) explore the spatial heterogeneity of below canopy processes in relation to decoupling, iii) develop a robust procedure to integrate decoupling investigations in the standard processing of EC measurement networks.

The anticipated experimental design involves three testing sites chosen to represent a broad range of canopy characteristics. These sites consist of a deciduous broadleaf forest in flat terrain (Lanžhot, Czech Republic), a coniferous forest in mountainous terrain (Renon, Italy), and a tall evergreen needleleaf forest in moderately complex mountain-valley terrain (Wind River, USA). The working group, in collaboration with industry partners, plans to deploy approximately 30 sonic anemometers across these sites. While around 10 sonic anemometers of the same type will be installed at the Wind River site, the rest, comprising different types, will be set up at Lanžhot and Renon. Installations, arranged in an array below the canopy around the primary EC measurement tower, are scheduled to commence in spring 2024, with the goal of year-long data collection to cover all seasons.

This presentation will set the proposed experiment on a solid theoretical background, introduce the measurement design and discuss the experiment aims.