No-tillage Straw Mulching Alters Fallow Rainfall Use and Improves Long-Term Water Balance and Yield Stability of Winter Wheat: A 38-Year Simulation on the Loess Plateau

Understanding how tillage-residue practices regulate water balance is essential for sustaining rain-fed agriculture winter wheat production on the Loess Plateau. This study integrated a three-year field experiment (2020−2023) with a 38-year (1978–2016) WHCNS simulation to compare the effects of no-tillage with surface straw mulching (NT) and rotary tillage with straw incorporation (RT) on soil water dynamics and winter wheat yield. The WHCNS model achieved acceptable performance for soil moisture, leaf area index, and biomass, but failed to accurately capture yield losses triggered by extreme pre-sowing-to-early-seedling-stage waterlogging. Compared with RT, NT suppressed soil evaporation, promoted transpiration and infiltration, and improved water use efficiency, grain yield and inter-annual yield stability. Correlation and covariance-based structural equation modelling (CB-SEM) revealed divergent fallow-rainfall utilization mechanisms between the two systems. Under RT, high bare-soil evaporation during hot July–August largely dissipated early fallow rainfall, so pre-sowing initial soil-water storage (ISW) and yield strongly relied on late-fallow September precipitation. By contrast, surface straw mulching under NT reduced non-productive evaporation, enabling early-season July–August fallow rainfall to effectively recharge soil water; September rainfall still exerted positive effects, yet August rainfall made meaningful contributions to ISW. For both treatments, fallow rainfall regulated grain yield mainly via the mediation chain: fallow rainfall → pre-sowing ISW → water stress → wheat yield. This study provides theoretical insights for adopting NT as a sustainable water-saving management practice for semi-arid rain-fed wheat regions, though simulation-derived findings require further field validation.