Sound observations are essential for understanding lake-atmosphere energy exchanges, which govern lake water levels and energy budgets and are critical for weather forecasting, water quality assessment and navigation. While advances have been made with observation programs on large lakes, such as North America’s Laurentian Great Lakes, current fixed-location platforms measure relatively little of the spatial variability in turbulent fluxes across these large systems. Here, we apply ship-borne eddy covariance derived turbulent flux measurements that reveal the presence of evaporation zones separated by turbulent flux boundaries that first form near shore and expand outward. Boundaries continue to migrate with time, with atmospheric conditions and the lake thermal state, but will collapse as the lake warms through summer. These results have implications for how evaporation should be estimated, which would be improved with distributed approaches. Coupled lake-atmosphere representation in predictive platforms are likely necessary to capture these spatial dynamics which would improve water level prediction, lake thermal and ecosystem characterization, and weather forecasting. Lastly, the date of turbulent flux boundary collapse could be used as a new metric of lake function and state that augments existing metrics such as date of stratification.