Evaluating a Virtual Tall Tower approach for estimating atmospheric CO₂ mixing ratios from near-surface measurements

Abstract. Atmospheric CO₂ measurements provide essential constraints for atmospheric modelling and carbon-budget estimates. Measurements made at tall towers, typically at heights of approximately 100m and above, are more representative of well-mixed atmospheric conditions than near-surface observations, however their spatial coverage is limited. Virtual Tall Tower (VTT) methods offer a potential approach to scale CO₂ mixing ratios from eddy-covariance stations, where measurements are made closer to the surface, typically 2–50m above ground, to tall-tower heights, thereby expanding the spatial coverage of the existing sparse atmospheric observation network. In this study, we evaluated the performance of an existing VTT approach
for estimating CO₂ mixing ratios from 35m to 150m at the combined ecosystem–atmosphere station at Svartberget, a forest-dominated site in northern Sweden. The dataset covered five years (2020–2024) of eddy-covariance fluxes, tall-tower CO₂ mixing-ratio measurements, and planetary boundary layer height from ERA5 reanalysis. The baseline VTT set-up estimated CO₂ mixing ratios with low bias (0.03 μmolmol⁻¹), low RMSE (0.75 μmolmol⁻¹), and a high squared Pearson correlation coefficient (r² = 0.99). The largest discrepancies between estimated and measured mixing ratios occurred under atmospheric conditions with weakened vertical mixing, particularly during winter months and in the early morning. Moreover, surface CO₂ flux, sensible heat flux, planetary boundary layer height, and displacement height were among the most influential variables in the approach, as perturbations in these variables led to changes in performance metrics. The underlying VTT assumption of well-mixed conditions is an important constraint that limits the applicability of the method under stable atmospheric conditions, such as during night-time and cold days. In addition, well-calibrated eddy- covariance CO₂ mixing-ratio measurements are crucial for good VTT performance. Therefore, although the VTT approach is a promising tool for expanding atmospheric CO₂ mixing-ratio information from near-surface measurements, further investigations across different sites and environmental conditions are needed to improve and generalize the approach.