CONSTRAINTS ON LAKE EVAPORATION AND ASSOCIATED KINETIC ISOTOPE FRACTIONATION VIA WATER VAPOR AND VAPOR ISOTOPE MASS BALANCE

Recycled moisture from lakes and the land surface is an important component of the terrestrial water budget but is difficult to measure. Existing measurement methods include water-energy balance modeling, eddy covariance, and vapor isotope mass balance (wherein downwind vapor = upwind vapor + evapotranspiration). In examining lake evaporation, the isotope mass balance approach has traditionally relied on a system of equations with hydrogen isotopes (δD), oxygen isotopes (δ¹⁸O), and the derived d‑excess parameter (d‑excess = δD – 8×δ¹⁸O). However, the system is generally underconstrained and results can be significantly hampered by imprecise knowledge of critical parameters. We developed a method to estimate three important parameters: the fraction of recycled moisture (F), evaporation rate (E), and the magnitude of kinetic isotope fractionation during evaporation (n, the diffusion-turbulence factor). The model’s Monte Carlo simulations leveraged daily precipitation isotope data (δ¹⁸O d‑excess, and the emerging triple oxygen isotope parameter, Δʹ¹⁷O), lake water isotope data, and weather data. We tested this approach in the Lake Michigan (USA) region, a locus of previous work. The model identified 24 of 94 precipitation events as clearly influenced by lake evaporation (minimum F > 0.1). These events primarily occurred in the cold season, in agreement with prior work, and could be substantial (up to ≈60 % of total vapor). Estimated evaporation rates ranged between 0–10 mm/day and largely overlapped with independent estimates from a hydrodynamical modeling approach. In terms of kinetic fractionation, most events required lower values for the diffusion-turbulence factor (0.05 ≤ n ≤ 0.5), in agreement with values used for open ocean conditions but in contrast to values of ≥0.5 commonly applied in lake studies. Parameter n also likely varied through time due to variability in environmental conditions like the lake-air temperature gradient. We conclude this isotope-based method is valuable in providing a new perspective on moisture recycling and fulfilling previously unmet needs to estimate evaporation rate and robustly assess error. Investigations of modern lake evaporation via stable isotopes will benefit from the use of vapor isotope data, modeling seasonal or shorter timescales, and incorporating lower, variable values for the diffusion-turbulence factor.