摘要：We aimed to assess the feasibility of an affordable instrumentation, based on a non-dispersive infrared analyser, to obtain atmospheric CO2 mole fraction data for background CO2 measurements from a flux tower site in southern Finland. The measurement period was November 2006–December 2011. We describe the instrumentation, calibration, measurements and data processing and a comparison between two analysers, inter-comparisons with a flask sampling system and with reference gas cylinders and a comparison with an independent inversion model. The obtained accuracy was better than 0.5 ppm. The intercomparisons showed discrepancies ranging from –0.3 ppm to 0.06 ppm between the measured and reference data. The comparison between the analyzers showed a 0.1 ± 0.4 ppm difference. The trend and phase of the measured and simulated data agreed generally well and the bias of the simulation was 0.2 ± 3.3 ppm. The study highlighted the importance of quantifying all sources of measurement uncertainty.
