Estimating Annual CO2 Flux for Lutjewad Station Using Three Different Gap-Filling Techniques

Long-term measurements of CO2 flux can be obtained using the eddy covariance technique, but these datasets are affected by gaps which hinder the estimation of robust long-term means and annual ecosystem exchanges. We compare results obtained using three gap-fill techniques: multiple regression (MR), multiple imputation (MI), and artificial neural networks (ANNs), applied to a one-year dataset of hourly CO2 flux measurements collected in Lutjewad, over a flat agriculture area near the Wadden Sea dike in the north of the Netherlands. The dataset was separated in two subsets: a learning and a validation set. The performances of gap-filling techniques were analysed by calculating statistical criteria: coefficient of determination (R2), root mean square error (RMSE), mean absolute error (MAE), maximum absolute error (MaxAE), and mean square bias (MSB). The gap-fill accuracy is seasonally dependent, with better results in cold seasons. The highest accuracy is obtained using ANN technique which is also less sensitive to environmental/seasonal conditions. We argue that filling gaps directly on measured CO2 fluxes is more advantageous than the common method of filling gaps on calculated net ecosystem change, because ANN is an empirical method and smaller scatter is expected when gap filling is applied directly to measurements.

Continuous CO2/CH4/CO measurements (2012–2014) at Beromünster tall tower station in Switzerland

The understanding of the continental carbon budget is essential to predict future climate change. In order to quantify CO₂ and CH₄ fluxes at the regional scale, a measurement system was installed at the former radio tower in Beromünster as part of the Swiss greenhouse gas monitoring network (CarboCount CH). We have been measuring the mixing ratios of CO₂, CH₄ and CO on this tower with sample inlets at 12.5, 44.6, 71.5, 131.6 and 212.5 m above ground level using a cavity ring down spectroscopy (CRDS) analyzer. The first 2-year (December 2012–December 2014) continuous atmospheric record was analyzed for seasonal and diurnal variations and interspecies correlations. In addition, storage fluxes were calculated from the hourly profiles along the tower. The atmospheric growth rates from 2013 to 2014 determined from this 2-year data set were 1.78 ppm yr⁻¹, 9.66 ppb yr⁻¹ and and -1.27 ppb yr⁻¹ for CO₂, CH₄ and CO, respectively. After detrending, clear seasonal cycles were detected for CO₂ and CO, whereas CH₄ showed a stable baseline suggesting a net balance between sources and sinks over the course of the year. CO and CO₂ were strongly correlated (r² > 0.75) in winter (DJF), but almost uncorrelated in summer. In winter, anthropogenic emissions dominate the biospheric CO₂ fluxes and the variations in mixing ratios are large due to reduced vertical mixing. The diurnal variations of all species showed distinct cycles in spring and summer, with the lowest sampling level showing the most pronounced diurnal amplitudes. The storage flux estimates exhibited reasonable diurnal shapes for CO₂, but underestimated the strength of the surface sinks during daytime. This seems plausible, keeping in mind that we were only able to calculate the storage fluxes along the profile of the tower but not the flux into or out of this profile, since no Eddy covariance flux measurements were taken at the top of the tower.