A mass transfer evaporation model is proposed that uses MODIS water surface temperature data and land-based meteorological data, and employs a new approach to calibrate the transfer coefficient via closure of the long-term energy budget of the lake. Some of the longstanding issues of developing and applying lake evaporation models are reviewed, including the adequacy of using land-based meteorological data, the difficulty of applying transfer coefficients with fixed values calibrated elsewhere, and the need to estimate rates of change of stored enthalpy when the model involves energy budget concepts. Publicly available data from a 5-year measurement campaign at Lake Mead allow to quantify the effect of using land-based data, and subsequently to test the proposed model. We show that atmospheric stability effects are very important, and that their incorporation by means of existing stability functions in the literature produces good results with a one-parameter model that can be locally calibrated with the same input data used by the model, without the need of local evaporation measurements. The model is simple in its structure and data requirements, and can be widely applied.