Simplification of CWSI empirical model and its application in water status diagnosis of winter wheat

In order to solve the problem that the traditional crop water stress index (CWSI) involves complicated calculation and cannot meet the demand for convenient water status estimation, this study took winter wheat as the research object to investigate the diurnal variations of canopy temperature and canopy-air temperature difference under different water treatments at key growth stages, develop a simplified empirical CWSI model, and verify the feasibility of predicting winter wheat water status using CWSI values. The experiment was carried out at the jointing, heading and filling stages of winter wheat on the North China Plain, and three irrigation treatments were designed with soil matric potential thresholds of 20, 50 and 80 kPa at 20 cm soil depth with tension meters, representing full irrigation, mild and severe water stress respectively, and each treatment had three replicates. Irrigation was initiated when the reading of tension meter dropped to the irrigation threshold, and corresponding leaf water content, soil water content and meteorological factors were simultaneously measured as well. The results showed that both canopy temperature and canopy-air temperature difference could effectively reflect differences in winter wheat water status, and higher soil water content corresponded to lower canopy temperature and canopy-air temperature difference, and vice versa. The lower baseline equation was established via regression between canopy-air temperature difference and vapor pressure deficit under full irrigation treatment, and the theoretical maximum canopy-air temperature difference was calculated with the lower baseline equation as the upper baseline, based on which a simplified empirical CWSI model was established. The diurnal dynamics of CWSI calculated by this simplified model were consistent with those derived from the Idso empirical model and the theoretical model with an extremely significant positive correlation (P<0.01), confirming the feasibility of calculating CWSI using the simplified empirical model. CWSI values calculated from both the simplified empirical model and the theoretical model increased with rising soil matric potential, while the simplified empirical model was more sensitive to variations in crop water status than the theoretical model. Regression analysis revealed that the CWSI from the simplified empirical model exhibited significant (P<0.05) or extremely significant negative correlations with winter wheat leaf water content and average soil water content within 0-60 cm soil depth, and that is, higher CWSI values corresponded to lower leaf and soil water content, and the corresponding regression model performed better in estimating the leaf water content with relative errors of 0.44%-4.82% than the soil water content with 2.10%-12.20%, indicating that the CWSI indicator can better characterize the crop water status than the soil water content. Based on the experimental results, an upper CWSI threshold of 0.6 is tentatively recommended for irrigation scheduling of winter wheat using the simplified empirical model, at which the winter wheat was kept under medium water stress stage. In conclusion, the simplified empirical CWSI model established for winter wheat in this study exhibits good reliability and sensitivity. The prediction model can satisfactorily estimate the crop and the soil water status, providing a scientific basis for water content monitoring and precision irrigation management of winter wheat.