A new instrument for stable isotope measurements of 13C and 18O in CO2 - instrument performance and ecological application of the Delta Ray IRIS analyzer

We used the recently developed commercially available Delta Ray isotope ratio infrared spectrometer (IRIS) to continuously measure the CO2 concentration c and its isotopic composition delta C-13 and delta O-18 in a managed beech forest in central Germany. Our objectives are (a) to characterize the Delta Ray IRIS and evaluate its internal calibration procedure and (b) to quantify the seasonal variability of c, delta C-13, delta O-18 and the isotopic composition of nighttime net ecosystem CO2 exchange (respiration) (RecoC)-C-13 and (RecoO)-O-18 derived from Keeling plot intercepts. The analyzer's minimal Allan deviation (as a measure of precision) was below 0.01 ppm for the CO2 concentration and below 0.03% for both ffi values. The potential accuracy (defined as the 1 sigma deviation from the respective linear regression that was used for calibration) was approximately 0.45 ppm for c, 0.24% for C-13 and 0.3% for O-18. For repeated measurements of a target gas in the field, the long-term standard deviation from the mean was 0.3 ppm for c and below 0.3% for both delta values. We used measurements of nine different inlet heights to evaluate the isotopic compositions of nighttime net ecosystem CO2 exchange (RecoC)-C-13 and (RecoO)-O-18 in a 3-month measurement campaign in a beech forest in autumn 2015. During this period, an early snow and frost event occurred, coinciding with a change in the observed characteristics of both (RecoC)-C-13 and (RecoO)-O-18. Before the first snow, (RecoC)-C-13 correlated significantly (p < 10(-4)) with time-lagged net radiation R-n, a driver of photosynthesis and photosynthetic discrimination against C-13. This correlation became insignificant (p > 0.1) for the period after the first snow, indicating a decoupling of delta C-13 of respiration from recent assimilates. For O-18, we measured a decrease of 30% within 10 days in (RecoO)-O-18 after the snow event, potentially reflecting the influence of O-18 depleted snow on soil moisture. This decrease was 10 times larger than the corresponding decrease in delta O-18 in ambient CO2 (below 3 %) and took 3 times longer to recover (3 weeks vs. 1 week). In summary, we conclude that (1) the new Delta Ray IRIS with its internal calibration procedure provides an opportunity to precisely and accurately measure c, delta C-13 and delta O-18 at field sites and (2) even short snow or frost events might have strong effects on the isotopic composition (in particular O-18) of CO2 exchange on an ecosystem scale.