Mechanistic links between underestimated CO2 fluxes and non-closure of the surface energy balance in a semi-arid sagebrush ecosystem

Abstract

The surface energy balance non-closure problem in eddy covariance (EC) studies has been largely attributed to the influence of large-scale turbulent eddies (hereafter large eddies) on latent and sensible heat fluxes. However, how such large eddies concurrently affect CO _2 fluxes remains less studied and mechanistic links between the energy balance non-closure and CO _2 fluxes are not well understood. Here, using turbulence data collected from an EC tower over a sagebrush ecosystem during two growing seasons, we decomposed the turbulence data into small and large eddies at a cutoff frequency and analyzed their contributions to the fluxes. We found that the magnitude of CO _2 fluxes decreased concurrently with decreased sensible and latent heat fluxes (and thus increased energy balance non-closure), primarily caused by large turbulent eddies. The contributions of such large eddies to fluxes are dependent not only upon their magnitudes of vertical velocity ( w ) and scalars (i.e. temperature, water vapor density, and CO _2 concentration), but also upon the phase differences between the large eddies of w and scalars via their covariances. Enlarged phase differences between large eddies of w and these scalars simultaneously led to reductions in the magnitudes of both CO _2 and heat fluxes, linking the lower CO _2 fluxes to energy balance non-closure. Such increased phase differences of large eddies were caused by changes in the structures of large eddies from unstable to near neutral conditions. Given widespread observations in non-closure in the flux community, the processes identified here may bias CO _2 fluxes at many sites and cause upscaled regional and global budgets to be underestimated. More studies are needed to investigate how landscape heterogeneity influences CO _2 fluxes through the influence of associated large eddies on flux exchange.

Keywords

• CO2 flux
• eddy covariance method
• ensemble empirical mode decomposition
• sensible and latent heat fluxes
• surface energy balance closure