The isotopic composition of surface water vapor flux (δE) is a quantity frequently used to investigate the local and regional water cycle. This study reports the results of a comparative evaluation of δE determined with the Keeling plot and the flux-gradient methods using high-frequency data collected at a cropland site and a lake site. Three regression models, ordinary least squares (OLS), York’s solution (YS), and geometric mean regression (GMR), were tested with the Keeling plot method. Results show that field characterization of measurement errors can improve the estimation of the YS regression. For both sites, broad agreement was achieved between the Keeling plot method with YS regression, the Keeling plot method with OLS regression and the flux-gradient method. For the lake site, OLS was the least biased of the three regression models in reference to the δE calculated by the Craig-Gordon model of isotopic evaporation of open water. These results favor the OLS over the YS regression for studies of isotopic evaporation when measurement errors in field conditions are unavailable.