Evaporation and transport of water isotopologues from Greenland lakes: The lake size effect

Isotopic compositions of evaporative flux from a lake are used in many hydrological and paleoclimate studies that help constrain the water budget of a lake and/or to infer changes in climate conditions. The isotopic fluxes of evaporation from a water surface are typically computed using a zero dimensional (0-D) model originally conceptualized by Craig and Gordon (1965). Such models generally have laminar and turbulent layers, assume a steady state condition, and neglect horizontal variations. In particular, the effect of advection on isotopic variations is not considered. While this classical treatment can be used for some sections of large open surface water bodies, such as an ocean or a large lake, it may not apply to relatively small water bodies where limited fetch does not allow full equilibration between air from land and the water surface. Both horizontal and vertical gradients in water vapor concentration and isotopic ratios may develop over a lake. These gradients, in turn, affect the evaporative fluxes of water vapor and its isotopic ratios, which is not adequately predicted by a OD model. We observed, for the first time, the vertical as well as horizontal components of vapor and isotopic gradients as relatively dry and isotopically depleted air advected over the surfaces of several lakes up to a 5 km fetch under winds of 1-5 m/s in Kangerlussuaq, Greenland. We modeled the vapor and isotopic distribution in air above the lake using a steady state 2-D model, in which vertical diffusive transport balances horizontal advection. The model was verified by our observations, and then used to calculate evaporative fluxes of vapor and its isotopic ratios. In the special case of zero wind speed, the model reduces to 1-D. Results from this 1-D model are compared with those from the 2-D model to assess the discrepancy in isotopic fluxes between advection and no advection conditions. Since wind advection above a lake alters the concentrations, gradients, and evaporative fluxes of water isotopes, it alters the water balance and isotope ratios of the lake and the relationship between them. These effects are greatest for small lakes. If wind advection is neglected in the inference of water balance from lake isotopes, an error is thus introduced, the magnitude of which depends on lake size. We refer to this as the lake size effect. For lakes less than 500 m in length along the wind direction, the average delta O-18 and delta D of vapor flux are at least 2 parts per thousand lower than the corresponding flux values from the 1-D model. The magnitude of the resulting relative error in water balance calculations is much greater if using delta O-18 than delta D in mass balance calculations; the former is about eight times the latter. This result argues that water balance calculated with delta D is less sensitive to the difference in lake size and/or its change over time. The 1-D model result is also compared with that from a comparable 0-D model. Since vertical vapor and isotope gradients always exist (even under no advection conditions), one may not obtain correct flux values if the relative humidity and isotopic ratios in ambient air measured at an arbitrary height are used for the O-D model calculation. Typically, the standard meteorological measurements at 2 or 10 m would result in an underestimate of the delta O-18 and delta D values of the vapor flux. This work has provided the first quantification on the effect of advection on isotopic fluxes of evaporation. The method of mobile vapor analysis combined with 2-D modeling can be applied to other environmental settings, in which the size of advection effect on isotopic fluxes depends upon relationships among local meteorological and hydrological variables. Our results also suggest that incorporating isotopic vapor measurements can help constrain modeled evaporation rates, which is worth exploring further in future studies. (C) 2015 Elsevier Ltd. All rights reserved.