Baffin Bay sea ice extent and synoptic moisture transport drive water vapor isotope (δ18O, δ2H, and deuterium excess) variability in coastal northwest Greenland

At Thule Air Base on the coast of Baffin Bay (76.51 N, 68.74W), we continuously measured water vapor isotopes (δ18O, δ2H) at a high frequency (1 s−1) from August 2017 through August 2019. Our resulting record, including derived deuterium excess (dxs) values, allows an analysis of isotopic–meteorological relationships at an unprecedented level of detail and duration for high Arctic Greenland. We examine isotopic variability across multiple temporal scales from daily to interannual, revealing that isotopic values at Thule are predominantly controlled by the sea ice extent in northern Baffin Bay and the synoptic flow pattern. This relationship can be identified through its expression in the following five interacting factors: (a) local air temperature, (b) local marine moisture availability, (c) the North Atlantic Oscillation (NAO), (d) surface wind regime, and (e) land-based evaporation and sublimation. Each factor’s relative importance changes based on the temporal scale and in response to seasonal shifts in Thule’s environment. Winter sea ice coverage forces distant sourcing of vapor that is isotopically light from fractionation during transport, while preventing isotopic exchange with local waters. Sea ice breakup in late spring triggers a rapid isotopic change at Thule as the newly open ocean supplies warmth and moisture that has ∼ 10 ‰ and ∼ 70 ‰ higher δ18O and δ2H values, respectively, and ∼ 10 ‰ lower dxs values. Sea ice retreat also leads to other environmental changes, such as sea breeze development, that radically alter the nature of relationships between isotopes and many meteorological variables in summer. On synoptic timescales, enhanced southerly flow promoted by negative NAO conditions produces higher δ18O and δ2H values and lower dxs values. Diel isotopic cycles are generally very small as a result of a moderated coastal climate and the counteracting isotopic effects of the sea breeze, local evaporation, and convection. Future losses in Baffin Bay’s sea ice extent will likely shift mean annual isotopic compositions toward more summer-like values, and local glacial ice could potentially preserve isotopic evidence of past reductions. These findings highlight the influence that the local environment can have on isotope dynamics and the need for dedicated, multiseason monitoring to fully understand the controls on water vapor isotope variability.


S1 Quality checking and instrument stability issues
The raw data from the L2130-i was passed through a series of quality checks prior to humidity response correction to remove readings that were well outside typical values observed at Thule. Most of these erroneous readings appear due to liquid water contamination of the intake tubing, likely from precipitation during intense cyclones. These quality checks identified and removed data where: 5 • L2130-i diagnostic data (e.g., chamber temperature, pressure, status, etc.) were unstable or out of typical ranges • Water vapor mixing ratio >15000 ppmv An additional quality check was performed after humidity response calibration. Visual inspection was used to identify clearly 10 abnormal isotopic or mixing ratio values (e.g., very large and/or abrupt changes not supported by meteorological data), and these observations were removed.
The machine was initially used at a tundra-based field site in Thule for a summer project in 2015, and it was installed its present location in October 2016. However, there were issues with cavity pressure stability and irregular isotopic readings which culminated in a full systems crash in May 2017. The system was restored on 04 Aug 2017 with stable cavity pressure 15 that has continued through present. Data from before the system restoration has poor correlation in water vapor mixing ratio between the L2130-i and the SMT weather station. Winter isotopic values and mixing ratios are also much higher in the prerestoration data than the next two winters despite generally similar winter weather. Out of caution, we have restricted our analyses and discussion to only post-restoration data.

S2 Humidity response calibrations 20
To correct for d 18 O and δ 2 H accuracy and precision bias at low water vapor mixing ratios, we injected standard waters for ten minutes at ten different flow rates. The last 200 observations of each injection were saved, and a nonlinear regression was performed on the δX vs. mixing ratio relationship, where δX is either d 18 O or δ 2 H, to determine accuracy corrections. The nonlinear regression was of the form: where δXcorrection is the difference between the observed isotopic value and the actual standard isotopic value, q is the water vapor mixing ratio, and a and b are constants. Calculated values for regression parameters are given in Table S4. Confidence intervals for predicted humidity response corrections were estimated using the predictNLS function from the propagate package in R.
Changes in analytical precision at low water vapor mixing ratios were calculated with a nonlinear regression of the form: 30 where δXprecision is the standard error of the mean isotopic value for a given flow rate, q is the water vapor mixing ratio, and a and b are constants. Calculated values for regression parameters are given in Table S5.

S3 Isotope-isotope relationships
Over the full dataset, d 18 O and δ 2 H have a strong linear relationship with low parameter standard error: δ 2 H = 6.959 ± 0.003 * 35 d 18 O -18.07 ± 0.09‰ (r 2 = 0.98, n=111138, 10 min data). Overall, this value is comparable to other slopes observed at other high latitude sites, such as 6.8 at Ivittuut, Greenland, (Bonne et al., 2014), 6.5 at NEEM, Greenland, (Steen-Larsen et al., 2013, 6.0-6.5 at Dome C, Antarctica (Casado et al., 2016), and 6.95 from the vapor mixing line at Kangerlussuaq, Greenland (Kopec et al., 2014). Changes in d 18 O are thus closely mirrored in δ 2 H, and most differences are only detectable on very short timescales (i.e., less than hourly) when some minor lead-lag between relative maxima and minima may occur. The dxs at Thule is 40 negatively correlated with both d 18 O and δ 2 H (r10min = -0.78 and -0.70, respectively).       Table S2. Standard waters calibration results after dry air system installation, performed roughly every 25 hours. Mean values and standard deviations of the last 200 observations for each calibration are given here. Some days do not have data due to failed calibration from a clogged injection needle or stuck injection piston. Some of the calibrations included here extend beyond the limit of ambient data discussed in the study, but station operation has continued without interruption.