Over-exploitation of groundwater for agricultural irrigation has attracted worldwide attention because of the rapid groundwater depletion and environmental consequences induced by the current intensified cropping system. The combination of cropping system modeling with field experimentation is a powerful tool to evaluate the long-term consequences of changing water management and cropping systems under inter-annual climate variability. Here, we conducted two field experiments for early maize (EM, 2012–2018) single cropping system and winter wheat – summer maize (WWSM, 2015–2018) double cropping system in the North China Plain (NCP). Field observations were carried out for crop growth, yield, and soil volumetric moisture content (VMC) calibration with the Agricultural Policy/Environmental eXtender (APEX) model. The calibrated APEX model was used to assess the long-term (35 years, 1985–2019) impacts of EM and WWSM cropping systems under four irrigation scenarios on sustainable water and crop productivity using historical weather data. Results showed that the APEX replicated growth, yield, and VMC all reasonably well with R² > 0.5, percent bias < 25%, and index of agreement > 0.8. The model realistically simulated VMC under full irrigation (FI), while model simulations under reduced irrigation were not that accurate. Neglectable yield losses of EM were observed by changing EM-FI to rainfed (EM-RF) (7.0 vs. 6.8 Mg ha⁻¹) and lower inter-annual yield variability. Shifting WWSM-FI to EM-RF saved 47% evapotranspiration (ET) and 115% net water use (NWU) regardless of the precipitation category years, but caused 54% reduction in yield. Critical irrigation (CI) significantly reduced ET (7%) and NWU (28%) of WWSM, without yield loss, compared to FI, indicating great potentials of water-saving through optimized irrigation strategy. Minimum irrigation (MI) saved ET by 17% and NWU of WWSM by 62%, but led to 15% EY reduction compared with FI. In conclusion, the WWSM cropping system under CI has the potential to maintain yield with less water consumption, while rainfed EM is the best alternative option for mitigating groundwater over-exploitation with a certain extent risk in crop yield losses.

由于目前集约化的耕作制度导致地下水的快速消耗和环境后果，农业灌溉用水的过度开发引起了全世界的关注。种植系统建模与田间试验相结合是评估在年际气候变化下改变水管理和种植系统的长期后果的有力工具。在这里，我们对华北平原的早玉米（EM，2012–2018）单季种植系统和冬小麦–夏玉米（WWSM，2015–2018）双季种植系统进行了两次田间试验。使用农业政策/环境排放模型（APEX）对农作物生长，产量和土壤体积含水量（VMC）校准进行了实地观察。校准的APEX模型用于使用历史天气数据评估四种灌溉情景下EM和WWSM种植系统对可持续水和作物生产力的长期（35年，1985-2019年）影响。结果表明，APEX可以很好地复制生长，产量和VMC。² > 0.5，偏差百分比<25％，一致性指数> 0.8。该模型在完全灌溉（FI）下逼真地模拟了VMC，而在减少灌溉条件下的模型仿真却不那么准确。通过将EM-FI改为雨养（EM-RF），观察到EM的产量损失可忽略不计（7.0 vs. 6.8 Mg ha ^-1）和较低的年间收益率差异。将WWSM-FI转换为EM-RF可以节省47％的蒸散量（ET）和115％的净用水量（NWU），而与降水类别的年份无关，但是却导致产量降低54％。与FI相比，关键灌溉（CI）显着降低了WWSM的ET（7％）和NWU（28％），而没有产量损失，表明通过优化灌溉策略可以节水。与FI相比，最低灌溉（MI）节省了ET的17％和WWSM的NWU的62％，但使EY减少了15％。综上所述，CI下的WWSM种植系统有可能以较少的耗水量维持产量，而雨育EM是减轻地下水过度开发的最佳替代方案，在一定程度上会降低作物的产量损失风险。