Effect of Advection on Evaporative Fluxes and Vapor Isotopic Ratios: The Lake Size Effect

Abstract:

It has been reported that advection of air from land can be identified hundreds of kilometers off shore. With advection, moisture builds up downwind, and the evaporative flux decreases and isotopic flux ratios increase with distance. If a lake is small relative to the equilibration distance, the fluxes of all water isotopologues averaged over the lake are different from those calculated using models without advection. The magnitude of the discrepancy depends on the lake size; we refer to this as the “lake size effect”.

In Kangerlussuaq, Greenland, we observed significant horizontal gradients in concentration, δD, and δ¹⁸O of vapor up to 5 km along the wind direction. Over a 0.5 km long lake, the observed average gradients were 1380 ppm/km for vapor content, 21‰/km for δD, 2.4‰/km for δ¹⁸O, and 5‰/km for d-excess. These gradients decreased with distance from the upwind shore. Over a stretch of another, much larger lake 4-5 km from the upwind shore, we observed gradients of 354 ppm/km, 1.5‰/km, 0.22‰/km and 0.3‰/km, for vapor concentration, δD, δ¹⁸O, and d-excess, respectively. These observations were modeled successfully using a two-dimensional (2-D, horizontal and vertical) steady state advection diffusion model. This model also computes evaporative fluxes.

Using the model results, we assess the magnitude of the lake size effect and its impact on water balance calculations. Under the condition of our field observations and for lakes less than 500 m along the wind direction, the mean flux δ¹⁸O and δD were at least 2‰ lower than the corresponding values from a 1-D model (vertical only). If using biased isotopic flux values for water balance calculations, the lake size effect would lead to an underestimation of the lake I/E (input to evaporation) ratio. For example, if the lake effect is 1‰, the corresponding underestimation of the I/E ratio is about 10% if using δ¹⁸O, and less than 2% if using δD for the computation. This argues for advantageous use of δD over δ¹⁸O in water balance and paleoclimate studies when the lake size is small or changes significantly over time. Still greater accuracy in water balance assessment can be achieved by using the 2-D model to correct for the lake size effect under the environmental conditions at the location of interest.