Search

• Different scenarios of soil water stress remarkably influenced the plant height and aboveground biomass of winter wheat. • The returning-green stage was a critical stage for winter wheat irrigation. • The current CERES-Wheat model had large simulation errors under soil water stress at early vegetative stages. • The current CERES-Wheat model was unable to simulate the phenology discrepancy caused by soil water stress. • The uncertainties in permanent wilting point could introduce large simulation errors under severe stress conditions. Soil parameters related to soil water holding capacity could play an important role in simulating winter wheat growth under severe soil water stress, which could heavily influence the simulated soil water contents, and then biomass and final yield. In this study, a field experiment of winter wheat (Triticum aestivum L.) was conducted in two consecutive growing seasons (2012–2014) under rainfall shelter in arid areas of China, with the purpose to identify to what extent the soil parameters could influence the simulated output variables in the DSSAT-CERES-Wheat model under water stress conditions. The permanent wilting point (PWP), which were initially indirectly measured based on soil sampling, were manually tuned through a trial-and-error method based on field observations of soil water content and aboveground biomass. The results showed that the maximum advancing of maturity date was about five days under water stress conditions. The stages of returning green and grain-filling were critical periods for agricultural water management of winter wheat in arid areas. The relative mean absolute error (RMAE) of simulated and observed variables were almost all less than 20% when water stress occurred at the heading and grain-filling stages. However, there were relatively large simulation errors when water stress occurred at the wintering and returning green stages. In addition, the CERES-Wheat model did not correctly simulate the discrepancies in phenology dates of winter wheat. The overall averaged root mean square error of all treatments for total water storage in 0−100 cm soil layer and winter wheat biomass decreased to 0.3 mm and 750 kg ha−1 after manually tuning the initially indirectly measured value of PWP. In general, the CERES-Wheat model showed some limitations to simulate winter growth under complicated arid conditions. Meanwhile, the measurement uncertainty in soil parameter PWP could introduce large simulation errors in simulating crop growth response to severe water stress conditions. We suggest the soil parameter PWP should be more carefully calibrated based on relevant field observations since it is related to crop type and difficult to be directly and precisely measured. In general, it is necessary to further research into the responding mechanism and simulation of winter wheat growth to water stress in liable-drought areas. [ABSTRACT FROM AUTHOR]