Comparison of high-frequency spectral correction methods for eddy-covariance fluxes over a central German cropland: effects on energy balance closure

The persistent lack of energy balance closure in single-tower eddy-covariance measurements remains a major source of uncertainty in surface–atmosphere exchange studies. In most eddy-covariance studies, turbulent fluxes (sensible and latent heat) underestimate available energy (net radiation minus ground heat flux), potentially affecting evapotranspiration estimates used in irrigation management, and propagating uncertainties into land-surface model evaluation and flux upscaling. One important contributor to this imbalance can be the choice of data processing steps, particularly corrections for high-frequency spectral losses, which are known to significantly influence eddy-covariance flux estimates. However, their impact on energy balance closure has not yet been sufficiently quantified for long-term cropland observations.

Here, we investigate how different high-frequency spectral correction methods affect turbulent fluxes and energy balance closure at a managed cropland site in Reinshof, central Germany. Three years of eddy-covariance data collected over rotating crops (winter wheat, winter barley, and sugar beet) during 2022–2024 were processed using EddyPro, applying both analytical (Moncrieff et al., 1997; Massman, 2000; Horst, 1997) and in situ (Ibrom et al., 2007; Fratini et al., 2012) spectral correction methods.

Results show that the choice of spectral correction methods led to differences of up to 3.5% in annual energy balance closure estimates for years using open-path gas analyzers and up to 9.4% for years using closed-path gas analyzers. The in situ correction by Fratini et al. (2012) consistently resulted in the highest energy balance closure across all years, whereas differences among analytical corrections were minor, with a maximum difference of 0.8% in 2023. These effects were driven exclusively by changes in latent heat flux, which increased by 5-15% for open-path systems and by 38% for closed-path systems at the annual scale after spectral correction.

Overall, this study demonstrates that the choice of high-frequency spectral correction methods critically affects energy balance closure estimates in long-term eddy-covariance measurements, with effects varying in magnitude between open- and closed-path systems.