Comparison of standard and actual crop evapotranspiration estimates derived from different evapotranspiration methods on the Canadian Prairies

Evapotranspiration (ET) models have become essential tools in areas such as climate modelling, weather forecasting, crop yield forecasting and irrigation planning. The modified second-generation Prairie Agrometeorology Model (PAMII1+2) estimates standard ET (ETc) (i.e. evapotranspiration from disease free, well fertilized crops, grown under optimum soil water conditions) by simulating crop development and the soil water balance using daily minimum and maximum surface air temperature, rainfall and upper air conditions to simulate the depth of the planetary boundary layer. The FAO56 Penman-Monteith (FAO56 P-M) method uses a reference surface/combination approach to calculate reference ET (ETo) and then multiply it by a crop coefficient (Kc) to derive ETc. The simplified water balance (WB) method estimates actual ET (ETa) as the residual of the water balance where precipitation and soil water variation are known. The objective of this study was to compare both ETc and ETa estimates from PAMII1+2 to those derived using the FAO56 P-M and the water balance models. Comparison of daily ET rates between PAMII1+2 and the FAO56 P-M method indicated significant differences in the ETc (RMSD = 1.88 mm d-1, r2 = 0.45) and ETa (RMSD = 1.65 mm d-1, r2 = 0.45). When compared to the WB derived weekly and biweekly total ETc, PAMII1+2 produced similar rates of ETa (r2 = 0.62, slope = 0.65), as the FAO56 P-M (r2 = 0.61, slope = 0.50) model. Both models produced overall ETa rates that fell within the range of the measurement error associated with the WB method. Since PAMII1+2 produced similar ETc and ETa rates compared to the FAO56 P-M method, for any potential user the choice between the PAMII1+2 and FAO56 P-M is less a matter of accuracy and more a function of the type of input data available. Copyright (c) 2011 John Wiley & Sons, Ltd.