Deficit irrigation (DI), where water is applied in amounts of less than what is required to meet crop water demand, seems to be one of the most effective ways to save irrigation water. However, DI practices are usually associated with higher levels of risk of yield penalizations. To minimize risk, accurate irrigation scheduling must be performed based on reliable soil water content (SWC) and crop water demand information. The goal was to determine whether a modelling approach based on the FAO-56 routine could be used to quantify maize evapotranspiration under drought and full irrigation conditions during three growing seasons in semiarid northeastern Colorado, USA. Maize ET estimations with FAO-56 were compared with actual values obtained running a root zone water balance where weather data, irrigation data, and values of SWC were used. Irrigation shortage (50, 54 and 19%) reduced ET in 2008, 2009, and 2010 (27, 34 and 14%). Irrigation water productivity increased as irrigation decreased. FAO-56 estimations were satisfactory but tended to overestimate ET. Averaged seasonal water use root mean square error never exceeded 0.89mmday(-1) for any treatment and season studied. Factors contributing to the performance of the FAO-56 procedure are also discussed. (c) 2018 John Wiley & Sons, Ltd.