Combined measurement of roots, δ18O and δ2H, and a Bayesian mixed model capture the soil profiles of wheat water uptake in a deep loamy soil

The profile of crop water uptake from the soil depends on rainfall regime (amount, seasonality, frequency distribution of rainfall event size), soil, crop, and management. This study, with a focus on winter wheat in a wheat-fallow system, combines measurements of hydrogen (δD) and oxygen (δ¹⁸O) isotopes with a Bayesian mixing model (MixSIAR), and measurements of root length density to (i) quantify crop water uptake from soil down to 3 m depth, (ii) to assess the influence of soil water at sowing, soil mulching, seasonal conditions and their interaction on the profiles of soil water uptake, and (iii) to probe for relations between yield and the profiles of soil water uptake. Across treatments and seasons, water uptake at jointing featured a ratio 2.1: 1.0: 1.8: 2.2 in four soil layers, top 0.2 m, 0.20.4 m, 0.41.2 m, and 1.2–3.0 m. At anthesis, the ratios shifted to 5.2: 1.0: 1.7: 2.0. Water uptake at jointing was higher from top-soil in dry (∼60 %) than in wet condition (∼30 %), and the opposite was true in deeper layers; water supply had a smaller effect on the profiles of water uptake at anthesis. Compared to bare ground, mulch favored root proliferation and water uptake in 0.42.0 m soil layer. For a given soil layer, soil moisture correlated negatively with root length density. Yield correlated positively and linearly with water uptake from 0.43.0 m soil at jointing, indicating that faster root development at early stages favors water uptake from deep soil in the critical period of grain yield formation. We discuss the implications of our findings for agronomic management and breeding.