中文摘要：

研究作物产量对气候变化的响应,对于指导区域农业生产,保障粮食安全和生态安全具有一定的理论指导意义。结合大田试验与农业生产系统模拟模型（Agricultural Production Systems Simulator,APSIM）,在验证模拟研究区冬小麦、玉米和紫花苜蓿产量可靠性的基础上,分析5个降水变化梯度（降水量不变、降低10%和20%、升高10%和20%）和5个气温变化梯度（不变、降低1.5和1℃、升高1.5和1℃）组合情景下3种作物的产量变化趋势。结果表明：APSIM模型在试验点对3种作物籽粒产量和生物量的模拟精度较高,决定系数R2在0.80～0.93之间,归一化均方根误差在11.35%～22.48%之间,模型有效系数在0.53～0.91之间。冬小麦、玉米和紫花苜蓿在气温升高、降水量减少的情景下减产,减产的最大幅度分别为38.7%、40.3%和41.8%;冬小麦、紫花苜蓿的在气温降低、降水量增加时增产,增产的最大幅度分别为29.8%和51.7%;玉米在降水量增加、温度不变的情景下增产幅度最大,为22.0%。总之,在研究范围内,3种作物的产量随降水的增加而增高;玉米的产量随气温升高先增高后降低,另2种作物的产量随气温的升高而降低;紫花苜蓿适应气候变化的能力最强。结果对明确黄土高原地区主要作物的生产走势,制订农业布局、管理措施等具有一定意义。

英文摘要：

Investigating the response of crop production to climate change can help to optimize local agricultural practices, and then ensure food and ecological security. Crop models can provide a useful way to examine the effects of a range of climatic condition, management or crop cultivar on crop growth and yield in field and pasture. This work investigated the effects of precipitation and air temperature changes on the production of winter wheat, maize and lucerne in rain-fed agriculture area located in the central and western Loess Plateau by field experiment and crop simulation model. The field experiment was conducted at Qingyang Loess Plateau Experimental Station of Lanzhou University through 2001 to 2010, and the Agricultural Production Systems Simulator（APSIM） was applied in this study to simulate the growing process of winter wheat, maize and lucerne. The APSIM was validated with the experimental data firstly, and then the APSIM was applied to simulate the yield variability of the crops under the combinations 5 precipitation levels and 5 air temperature levels based on historical climatic data from 1961 to 2010. Temperature levels were： 1）-1.5℃ decrease in daily mean temperature（T1）; 2）-1℃ decrease in daily mean temperature（T2）; 3） historical daily temperature（T2）; 4） 1℃ increase in daily mean temperature（T4）; and 5） 1.5℃ increase in daily mean temperature（T5）. Precipitation levels were： 1） 20% decrease in daily precipitation（P1）; 2） 10% decrease in daily precipitation（P2）; 3） historical daily precipitation（P3）; 4） 10% increase in daily precipitation（P4）; and 5） 20% increase in daily precipitation（P5）. Results showed that the APSIM can predict the grain yield and biomass of the 3 crops accurately with the determination coefficients varied between 0.80-0.93, the normalized root mean square errors varied between 11.35%-22.48%, and the model efficiency varied between 0.53-0.91; Overall, APSIM was powerful to simulate the crop grain yield an