There has been extensive research on the effects of evaporation on the isotopic ratio of lacustrine and marine water bodies; however, there are limited data on how ablation or sublimation from lake or sea ice influences the isotopic ratio of the residual water body. This is a challenging problem because there remains uncertainty on the magnitude of fractionation during sublimation and because ablation can involve mixed-phase processes associated with simultaneous sublimation, melting, evaporation, and refreezing. This uncertainty limits the ability to draw quantitative inferences on changing hydrological budgets from stable isotope records in arctic, Antarctic, and alpine lakes. Here, we use in situ measurements of the isotopic ratio of water vapor along with the gradient diffusion method to constrain the isotopic ratio of the ablating ice from two lakes in the McMurdo Dry Valleys, Antarctica. We find that during austral summer, the isotopic fractionation of ablation was insignificant during periods of boundary layer instability that are typical during midday when latent heat is highest. This implies that the loss of mass during these periods did not yield any isotopic enrichment to the residual lake mass. However, fractionation increased after midday when the boundary layer stabilized and the latent heat flux was small. This diurnal pattern was mirrored on synoptic timescales, when following warm and stable conditions latent heat flux was low and dominated by higher fractionation for a few days. We hypothesize that the shifting from negligible to large isotopic fractionation reflects the development and subsequent exhaustion of liquid water on the surface. The results illustrate the complex and nonlinear controls on isotopic fractionation from icy lakes, which implies that the isotopic enrichment from ablation could vary significantly over timescales relevant for changing lake volumes. Future work using water isotope fluxes for longer periods of time and over additional perennial and seasonal ice-covered lake systems is critical for developing models of the isotopic mass balance of arctic and Antarctic lake systems.