Evapotranspiration data from three commercial atmometers (ET 0ÀA) were compared over an irrigation season for instrumental precision, and compared against Penman-Monteith (ET 0ÀPM) data from an automatic weather station for accuracy. The effects of using contrasting ET 0 estimation methods on the timing and amounts of water application were then evaluated by using an irrigation-scheduling water-balance computer model. The ET 0ÀA data were statistically analyzed using linear regression, coefficients of determination (R 2), root-mean-squared error (RMSE), mean bias error (MBE), and the t-test. The study found that the ET 0ÀA data from individual atmometers were closely correlated both with each other and with the ET 0ÀPM data (R 2 ¼ 0:68–0:90). On the basis of water-balance modeling, using an atmometer for scheduling irrigation on potatoes would have resulted in a very similar distribution of irrigation events, but 15% more water would have been applied over the season compared to using ET 0ÀPM data. Day-to-day measured ET 0ÀA values may differ from ET 0ÀPM as the atmometer can only be read to AE0:5 mm accuracy, but when averaged over the period of a typical irrigation cycle these errors are reduced and the impact on the irrigation schedule is small. The study suggests that atmometers would be appropriate for scheduling deep-rooted irrigated crops in humid regions with long growing seasons in which the irrigation interval is not less than 5–7 days.