Estimation of greenhouse tomato evapotranspiration under deficit irrigation based on SIMDualKc model

Abstract: The purpose of this study was to investigate the feasibility of SIMDualKc model in estimating greenhouse evapotranspiration and soil evaporation in northwest China. A deficit irrigation experiment was set up in tomato greenhouse in Yangling, Shaanxi from 2013 to 2015. Four treatments were designed including full irrigation, deficit irrigation with 50% full irrigation amount for only the development stage, both in the development stage and middle stage, and in the whole stage. Measurement items were meteorological data such as temperature, crop data (growing stage of tomato, root depth, plant height and ground surface coverage), soil data (moisture), irrigation depth, total evaporable water and easy evaporable water. Experiment date obtained in 2013-2014 was for model calibration and that in 2014-2015 was for model validation. The results showed that the SIMDualKc model could well simulate the evapotranspiration in 2013-2015 under the four treatments with the mean absolute error (MAE) of 0.22-0.33 mm/d, root mean square error (RMSE) of 0.26-0.48 mm/d, and coefficient of determination (R2) of 0.51-0.81. The model also simulated soil evaporation accurately with the MAE of 0.016-0.024 mm/d, RMSE of 0.013-0.034 mm/d and R2 of 0.63-0.84. After model calibration, the basal crop efficient of each stage was adjusted as 0.34 for the initial stage, 0.34 for the development stage, 1.16 for the mid season stage, and 0.63 for the later season stage. According to the result of simulation, the soil evaporation ratio of greenhouse tomato was highest in the initial stage (10.7%-17.7%), and lowest in the mid stage or later stage (3.1%-5.1%). When the tomato was planted with mulching material, the crop transpiration ratio was 94.6% in the whole growth stage, and the soil evaporation ratio was 5.4% in the whole growth stage. The coefficients of water deficit calculated based on the simulated evapotranspiration for the development, middle, and later stage were 0.8, 0.94 and 0.98 for the deficit treatments with 50% irrigation amount only in the development stage, 0.8, 0.69 and 0.91 for the deficit treatments with 50% irrigation amount both in the development stage and the middle stage, and 0.8, 0.7 and 0.63 for the deficit treatments with 50% irrigation amount in the whole stage. The water deficit coefficient decreased with the duration of deficit irrigation, but would increase after rehydration. The coefficient of water deficit was lowest when the deficit irrigation lasted for the whole stage. Therefore, the SIMDualKc model can accurately simulate evapotranspiration of greenhouse tomato under deficit irrigation in the northwest China, and it is reliable to analyze the response of tomato to deficit irrigation and its water compensation mechanism based on simulation results. The study provides data support for tomato cultivation in greenhouse under deficit irrigation.