Abstract:In the northern area of China, water supply is a major factor limiting crop yield. Maize is one of three major crops in China. The observation and simulation of evapotranspiration (ET) in maize fields are important processes in meteorology, hydrology, ecology and the other related fields. Thus studies on maize ET are critical for ensuring food security, saving irrigation water and increasing crop water use efficiency. A classical two-layer ET model, the Shuttleworth-Wallace (SW) model is appropriate for estimating ET in sparse vegetation conditions where soil evaporation and vegetation transpiration are significant. In this study, we adopted the Jarvis and Kelliher-Leuning canopy resistance models in relation to SW model to construct SW1 model and SW2 model, respectively. The SW1 and SW2 models were used to simulate ET in a summer maize field in Yucheng Agricultural Experimental Station of Chinese Academy of Sciences. Also experiments were conducted to measure daily ET in summer maize field via eddy covariance system during the main growing period of 20032004. ET simulated by the two models was validated using measured flux data. The results suggested that ET obtained by the two models were consistent with observed data. Correlation coefficients of the measured and simulated ET were above 0.85 ( P< 0.01) and the index of agreement of the measured and simulated data was over 0.92. The ratio of soil evaporation to ET decreased rapidly with increased leaf area index and that ratio for July was higher than those for August and September. At blossom and milk stage, both ET and soil evaporation reached maximum values. During this period, maize leaf growth was vegetative and with the largest canopy transpiration. Then ET and soil evaporation slowly decreased thereafter with gradual reduction in leaf area index, and with the ratio of soil evaporation to ET of 0.2. Sensitivity analysis showed that estimated ET by the SW model was most sensitive to the canopy resistance and the model sensitivity to canopy resistance increased with increasing leaf area index. At early growth stage of maize, the impact of soil surface resistance on ET was not negligible, especially with less vegetation cover. Among the parameters for canopy resistance calculation, estimated ET by SW1 model was most sensitive to change in field capacity. This was followed by minimum stomatal resistance and then effective leaf area index. SW2 model was most sensitive to maximum stomata conductance. Traditional SW model based on Jarvis’s equation for canopy resistance calculation had a complex calculation with several parameters. Then SW2 model based on Kelliher-Leuning equation only had half of SW parameters and therefore considerably simplified the ET model calculation. Compared with SW1 model, SW2 model was much more convenient in terms of application in ET calculation.