Eddy covariance measurements of the forest floor CO2 exchange in two contrasting forest stands in boreal Sweden

Boreal forests cover a large portion of land surface area in the northern hemisphere and greatly affect the global carbon (C) cycle and climate. Since these forests exchange carbon dioxide (CO₂) with the atmosphere in different vertical layers, many different CO₂ sources and sinks exist within the complex forest stand. The forest floor (soil and understory vegetation) may act as an important component of the C budget in a forest stand, however its contribution may vary from negligible to determining the inter-annual variability of ecosystem C balance. To date, there are a limited number of studies that have directly quantified the CO₂ fluxes over a forest floor using the eddy covariance (EC) method primarily due to challenges and potential violation of underlying assumptions when applying this method in the trunk space where turbulence characteristics are complicated, intermittent, and not in accordance with universal theories.

In this study, we installed two identical EC flux systems at two contrasting boreal forests (sparse pine stand vs. a dense mixed pine-spruce stand) in Sweden to measure the forest floor CO₂ exchange. We developed site-specific ideal cospectral models for the below-canopy fluxes in the trunk space under the well-mixed condition as defined by the standard deviation of the vertical wind speed. Spectral correction to the half-hourly fluxes was performed based on the newly fitted cospectral models at each site. Chamber measurements of CO₂ fluxes during the growing season were conducted to compare with the estimates from the below-canopy EC data.

Our below-canopy cospectral models show that more high-frequency signals (small eddies) occurred in the forest trunk space compared to the ideal above-canopy cospectral model. The high-frequency contribution was greater in the dense pine-spruce forest compared to the open pine stand. The spectral corrected CO₂ fluxes measured by the EC method agreed well with the concurrent chamber measurements. The EC results revealed that the forest floor of the two contrasting stands acted as net CO₂ sources during the 3-year period (2017-2019). This study highlights that by applying a data correction based on site-specific below-canopy cospectral models, the EC method can be used in the trunk space to accurately measure the net CO₂ exchange between the forest floor and the atmosphere and thus to improve our understanding of the role of the forest floor in the ecosystem-scale C budget in the boreal forest region.