3.3 Isotope measurement
Water samples were filtered (0.2 µm cellulose acetate) prior to analysis of stable isotopes (δ18O and δ2H values) in the water isotope lab at IGB Berlin, using a Picarro (Santa Clara, CA, USA) L2130-i cavity ring-down spectrometer (CRDS). Measurements were routinely checked for organic contamination using the Picarro ChemCorrect software.
Isotope values and standard deviations are based on six replicate measurements of each sample, discarding the first three measurements to account for memory effects. In house criteria excluded all injections with a water SD higher than 400 ppm or water amount deviation greater than 3000 ppm to furthermore improve precision.
For instrument calibration we used 3 laboratory standards for each group of 24 samples: L (δ18O -17.86 ‰ and δ2H -109.91 ‰), DEL (δ18O -10.03 ‰ and δ2H -72.81 ‰), H (δ18O 2.95 ‰ and δ2H 0.29 ‰). A fourth lab standard, M (δ18O -7.68 ‰ and δ2H -56.70 ‰), was used as quality and drift control after every 6 samples. All lab standards were referenced against primary measurement standards: VSMOW2 (Vienna Standard Mean Ocean Water 2), GRESP (Greenland Summit Precipitation), and SLAP2 (Standard Light Antarctic Precipitation 2) from the IAEA (International Atomic Energy Agency, Vienna International Centre, A-1400 Vienna, Austria).
The measurement uncertainty was quantified by error propagation, including the parameters a) uncertainties of lab standards; b) errors of standard calibration; c) average standard deviation of replicate measurements. Based on this, measurement uncertainty was estimated to account 0.16 ‰ for δ18O and 0.57 ‰ for δ2H.