A common method to estimate the carbon isotopic composition of soil-respired air is to use Keeling plots (d13C versus 1/CO2 concentration). This approach requires the precise determination of both CO2 concentration ([CO2]), usually measured with an infrared gas analyser (IRGA) in the
field, and the analysis of d13C by isotope ratio mass spectrometry (IRMS) in the laboratory. We measured [CO2] with an IRGA in the field (n 637) and simultaneously collected air samples in 12 mL vials for analysis of the 13C values and the [CO2] using a continuous-flow isotope ratio mass spectrometer. In this study we tested if measurements by the IRGA and IRMS yielded the same results for [CO2], and also investigated the effects of different sample vial preparation methods on the [CO2] measurement and the thereby obtained Keeling plot results. Our results show that IRMS
measurements of the [CO2] (during the isotope analysis) were lower than when the [CO2] was measured in the field with the IRGA. This is especially evident when the sample vials were not treated in the same way as the standard vials. From the three different vial preparation methods, the one using N2-filled and overpressurised vials resulted in the best agreement between the IRGA and IRMS [CO2] values. There was no effect on the 13C-values from the different methods. The Keeling plot results confirmed that the overpressurised vials performed best. We conclude that in the cases where the ranges of [CO2] are large (>300 ppm; in our case it ranged between 70 and 1500 ppm) reliable estimation of the [CO2] with small samples using IRMS is possible for Keeling plot application. We also suggest some guidelines for sample handling in order to achieve proper results.