Stable isotopes provide robust information on the structure, function, and processes of ecological systems and can be powerful tools to address continental scale ecological issues. For example, δ18O and δD undergo predictable transformations in the hydrological cycle that allow tracing of water sources throughout the biosphere and climate system, and facilitate partitioning of energy balance and evaporation components. Therefore, the H2O isotopic atmospheric measurements are a key data product in the National Ecological Observatory Network (NEON). However, maintaining long-term (30 years) consistency and traceability to IAEA reference standards, and minimizing the uncertainties in measurements made over a large range of humidity conditions, are a great challenge in the NEON design. Here, we present the NEON strategy for both laboratory and field calibration of laser based cavity ring-down spectroscopic (CRDS) instruments for water isotope measurements. The overall strategy includes: preparing and maintaining multiple in-house water standards for lab and field calibration, which are traceable to IAEA reference standards; performing Allan variance tests to characterize the stability of the instruments and determine the steady state for measurement; quantifying the specific humidity dependence of individual instruments; distinguishing the affects of memory and instrument drift. All of the above determine the frequency of field calibrations necessary.
Results/Conclusions
In-house water standards have been created from a variety of filtered meteoric waters over a range of ~40 per mil δ 18O, which are traceable to IAEA reference standards and properly stored to supply for the lab and field calibrations for decades. Allan variance tests indicate that different units require different length of time to reach steady state condition for valid measurements and should be characterized prior to and during field deployment. Humidity dependency tests show clear non-linear response function of δ 18O and δD measurements to humidity when water vapor concentration is < 5000 ppm, thus requiring careful post-measurement corrections. Due to sticky nature of the water, memory effects in the instrument are observed, and memory factors are calculated to apply the necessary corrections. Because of the instrument drift, multiple automated calibrations are recommended to minimize the measurement uncertainties and normalize to the VSMOW scale.