Abstract:Irrigation and fertilization scheduling significantly affect field water cycles. Root water uptake is the critical process that connects plant transpiration and soil water movements. It is necessary to quantify the crop water uptake sources under different irrigation and fertilization treatments for optimizing agricultural water management. The stable water isotopes of D and 18O are considered as ideal (natural) tracers for tracking water through the soil based on distinct isotopic signatures of water fluxes. The MixSIAR framework is the latest Bayesian stable isotope analysis mixing model in R that considers multiple sources of uncertainty and provides definite proportions of source contributions. In this study, dual stable isotopes of soil water and stem water were applied to determine seasonal water uptake patterns of winter wheat under different irrigation and fertilization treatments during 2013-2015 in Beijing, China. The contributions of the soil water at each depth to water uptake were quantified using the MixSIAR Bayesian mixing model. The direct inference method was also used to detect the potential root water uptake depth. Correlations between the water uptake patterns and soil moisture changes were further evaluated. The average contribution of soil water in the 0-20, 20-70, 70-150, and 150-200 cm layers was 35.6%, 27.6%, 23.1%, and 13.7%, respectively. The primary root water uptake depth was 0-20 cm (67.0%), 20-70 cm (42.0%), 0-20 cm (38.7%), and 20-70 cm (34.9%) during the greening-jointing, jointing-heading, heading-filling, and filling-harvest periods, respectively. Significant differences in crop water use appeared between the 2014 and 2015 growing seasons. The main root water uptake depth gradually increased from 0-20 cm (greening-jointing period) to 70-150 cm (heading-filling period) and was maintained at the 70-150 cm depth until the filling-harvest period in the 2014 season. However, winter wheat mainly took up soil water from the shallow layers (0-70 cm) over the 2015 season. In particular, the proportional soil water contribution in the 0-20 cm layer was remarkably higher (13.9%) than that in the 2014 season. Root water uptake patterns with soil moisture distributions were significantly influenced by different irrigation and fertilization treatments, especially in dry seasons. The main water uptake source was the soil water in the top layer (0-20 cm) under the T4 and T5 treatments during the jointing-heading period in 2015, because the root growth in the surface layer was stimulated by sufficient irrigation (80 mm) and abundant fertilization (≥ 210 kg/hm2 N) at early growth stages. Nitrogen deficiency with < 105 kg/hm2 N (T3) or less irrigation with 20 mm (T1 and T2) during the greening-jointing period promoted root growth in the deep soil layer (70-200 cm) and increased water adsorption by a mean of 29% during the jointing-filling period. Seasonal variations in the quantitative contribution of soil water at different depths were closely related to the soil moisture distributions. The large contribution of soil water in the 0-150 cm layer (86.3%) was consistent with its proportional consumption in soil water storage (92%) throughout the greening to harvest season of winter wheat. This study provides a simple and effective method for identifying crop water sources. The findings are of great significance for future fertilization and irrigation management.