Resolving combined influences of inflow and evaporation on western Greenland lake water isotopes to inform paleoclimate inferences

Abstract

Stable isotopes of oxygen (δ18O) and hydrogen (δ2H) in precipitation are widely employed tracers of the global hydrologic cycle, and are frequently inferred from lake-water-derived proxies in sediments of high-latitude lakes. Lake-water isotope proxies archive precipitation δ18O and δ2H values, modulated by lake hydrological processes, which may be functionally classified into processes that affect source water isotope values (i.e. inflow δ18O and δ2H) and catchment-integrated evaporation. Respectively, these controls form the basis of interpretations of precipitation isotope and effective precipitation signals from lake-water isotope proxy records. Conventionally, a single control on lake water isotope variability is assumed for a given record. Yet sensitivity to these controls depends on regional hydroclimate and local hydrology, which may change through time. We quantified the relative impacts of variations in inflow δ18O and evaporative 18O enrichment on lake water δ18O in response to spatially variable aridity, using measurements of lake water δ2H and δ18O from 140 western Greenland lakes located between the Labrador Sea and western Greenland Ice Sheet margin. We calculated source water δ18O of lake waters (δI) using a recently developed Bayesian method and quantified evaporation-to-inflow ratios (E/I) using a modified Craig-Gordon model. δI varied by 11.1‰ across the study region, superimposed by evaporative 18O enrichment of up to 20.0‰ and E/I ranging from nearly no evaporative loss (E/I < 0.10) to desiccation (E/I > 1). Lakes can be broadly classified as predominantly sensitive to inflow or evaporation, corresponding to their location along the aridity gradient, and there are significant trends in both δI and E/I across the study area. Substantial local variability in δI and E/I suggests catchment hydrology determines the sensitivity of δI and E/I to changes in aridity, and implies that hydrological end-member lakes within a small region may provide complementary records of seasonal precipitation isotope values and ice-free-season evaporation. Deconvolving modern controls on lake water isotope values provides essential support for quantitative and seasonal paleoclimate inferences from paleolimnological isotope data, which will improve constraints on the long-term variability of the Arctic hydrologic cycle.