Rapid warming and increasingly stringent water allocations in arid Northwest China have exacerbated the inherent trade-offs among maize yield, water productivity (WPc) and economic returns. This study employed a statistical modeling-optimization pipeline (PLS-GA) and a machine learning-optimization pipeline (RF-GA) to build a framework for better maize production in the Hexi Corridor of Gansu Province, an arid region in northwest China. The framework uses yield and actual evapotranspiration (ETc act) prediction as well as multi-objective optimization calculations. The optimal irrigation, nitrogen application, and planting density were proposed for five production-demand scenarios and the consequent impacts on yield, WPc, and economic returns under future climate change were systematically assessed. Results showed that random forest (RF) outperformed partial least squares (PLS) regression in capturing non-linear relationships (R2= 0.80 vs. 0.51) for yield simulation, whereas PLS provided superior explanatory power for individual factors. Findings also showed that all scenarios in the historical period could have benefited from an increase in planting density by at least 13.1 % and precision planting, leading to improvements in yield, WPc, and economic returns of at least 20.2 %, 31.4 %, and 15.1 %, respectively, alongside reductions in nitrogen application and irrigation of at least 13.7 % and 6.3 %, respectively. During mid-century (2041–2050), planting density and irrigation were projected to decline 0.4–1.1 % and 0.1–3.5 %, respectively, while nitrogen application to increase by 4.7–9.9 %. These adaptive measures lead to enhanced yield (5.8–6.2 %) and economic returns (13.8–14.7 %), albeit with a decline in WPc (13.1–14.5 %). This study presents an integrated maize management strategy that simultaneously optimizes grain yield, WPc, and economic returns in the Hexi Corridor, while also contributing a scalable methodological framework for advancing climate-resilient agricultural practices in arid, irrigated agroecosystems of Northwest China and comparable regions.
扫码关注我们
求助内容:
应助结果提醒方式:
