太阳辐射是作物生长发育的主要能量来源，对作物产量的形成具有极其重要的作用。近几十年来，全球变暗（global dimming，即地面接收的太阳总辐射量在年代际时间尺度上的降低现象）已经成为目前中国作物生产所面临的主要胁迫之一。全球变暗在降低太阳总辐射的同时，提高了散射辐射在太阳总辐射中所占的比例，而作物对散射辐射和直射辐射的响应不同。小麦和水稻是中国的两种主要粮食作物，明确全球变暗对小麦和水稻生长与产量的影响，对于保障未来气候变化背景下中国作物生产安全至关重要。为明确全球变暗背景下太阳总辐射下降和散射辐射比例提高对小麦和水稻生长与产量的影响，本研究将历史资料统计分析与田间试验相结合，首先利用1961-2016年长江中下游地区4个气象试验站点的气象（日平均、最低、最高温度，降水，太阳总辐射和散射辐射）和作物（生育期和产量）观测数据，分析了该区太阳总辐射与散射辐射比例的变化特征，以及基于历史数据的统计模型是否能够准确评估全球变暗对作物产量的影响。其次，以冬小麦（宁麦13）和单季水稻（南粳46）为研究对象，于2013-2016年在江苏省农业科学院试验田（32º03´N，118º87´E）开展田间遮阴试验，以聚乙烯（PE）薄膜作为覆盖材料来模拟全球变暗背景下的太阳辐射变化情况，每季试验设两个遮阴处理（T1和T2，覆盖不同厚度和层数的薄膜）和一个对照处理（CK，自然条件）；利用田间试验观测数据，验证了散射辐射比例提高对作物生长与产量具有“肥料效应”，探讨了散射辐射“肥料效应”产生的原因，明确了全球变暗背景下太阳辐射变化对作物生长与产量形成的影响。主要研究结果如下：

1．基于历史气象和作物产量数据的统计模型难以准确地评估长江中下游地区全球变暗对小麦和水稻产量的影响。长江中下游地区的太阳总辐射量自1961年开始呈下降趋势，虽然这一趋势在1990年后开始减缓甚至略有上升，但是并未恢复到1961年前后的太阳总辐射水平；而该区散射辐射比例自1961年以来呈现持续增加的趋势。基于历史气象和作物数据的统计回归模型分析结果表明，1961-2016年长江中下游地区太阳总辐射量降低会导致小麦和水稻减产，但水稻气候产量与散射辐射成负相关、小麦气候产量与温度成正相关，这与现有理论是相悖的。出现这种有悖于现有理论结果的主要原因是统计模型中各个变量之间（如太阳总辐射与温度之间）存在显著线性相关。因此，没有田间试验数据的支撑，仅仅依靠基于历史气象和产量数据的统计模型是难以准确地评估全球变暗对小麦和水稻产量的影响。

2．散射辐射比例提高对小麦和水稻产量和地上部生物量均具有“肥料效应”。散射辐射比例每提高10%，小麦产量和地上部生物量分别提高6.6%~7.8%和8.2%~9.3%，水稻产量和地上部生物量分别提高4.2%~4.5%和5.9%~6.2%；但本研究模拟的中度和重度污染下的散射辐射比例增加对产量和生物量的“肥料效应”不能完全弥补太阳总辐射下降造成的产量和生物量损失。遮阴处理下，小麦生长季的太阳总辐射下降11.8%~25.5%、散射辐射比例提高7.4%~20.4%，水稻生长季的太阳总辐射下降12.3%~29.9%、散射辐射比例提高7.9%~21.8%，很好地模拟了中度和重度污染下的太阳辐射情况；小麦和水稻产量分别降低8.31%~16.36%和7.81%~20.86%、地上部生物量分别降低6.39%~13.45%和5.63%~15.54%，其下降百分比均小于太阳总辐射的下降百分比。与CK处理相比，散射辐射比例每提高10%，各遮阴处理下小麦和水稻产量分别提高6.6%~7.8%和4.2%~4.5%，地上部生物量分别提高8.2%~9.3%和5.9%~6.2%；而太阳总辐射每降低10%，散射辐射比例分别需要提高14.0%~14.2%和11.5%~11.8%（小麦）、22.1%~23.9%和15.9%~17.7%（水稻）才能保持产量和地上部生物量不变。这些试验结果证实了全球变暗背景下散射辐射比例的提高对小麦和水稻产量和地上部生物量均具有“肥料效应”，而且散射辐射比例提高对产量的“肥料效应”小于对地上部生物量的“肥料效应”；但本研究模拟的中度和重度污染下的散射辐射比例增加对产量和生物量的“肥料效应”不能完全弥补太阳总辐射下降造成的产量和生物量损失。

3．全球变暗背景下散射辐射“肥料效应”产生的主要原因是作物冠层光能利用率的提高，而不是冠层光截获能力的提高；全球变暗背景下冠层光能利用率的提高不仅是因为叶片光合作用——光响应曲线的非线性，还因为植株对全球变暗产生了光合适应现象。遮阴处理下，小麦和水稻冠层光能利用率分别提高了0.48%~8.14%和1.38%~9.93%；从拔节期到灌浆期，小麦和水稻叶面积指数分别提高了0.57%~11.45%和2.55%~22.91%，但是冠层消光系数分别降低了0.64%~10.60%和2.91%~23.21%，叶面积指数和冠层消光系数的相反变化使得冠层光截获比例没有显著改变。因此，全球变暗背景下，散射辐射“肥料效应”产生的主要原因是作物冠层光能利用率的提高，而不是冠层光截获能力的提高。进一步分析影响作物冠层光能利用率的因子发现，遮阴处理下，从拔节期到灌浆期，小麦倒一叶和倒三叶饱和光强下的最大光合速率分别提高0.57%~11.15%和1.76%~16.18%、初始光能利用率分别提高1.54%~17.07%和1.43%~15.53%，水稻倒一叶和倒三叶饱和光强下的最大光合速率分别提高0.75%~13.08%（2013年拔节期降低1.71%~6.17%除外）和0.44%~7.83%（2015年T1处理拔节期降低1.72%除外）、初始光能利用率分别提高1.40%~13.69%和1.30%~8.02%；同时，植株C/N比降低，小麦和水稻绿叶氮素浓度分别提高0.94%~8.87%和0.08%~11.27%，且最大光合速率和初始光能利用率均与相应叶位的叶片氮素浓度显著正相关。因此，全球变暗背景下冠层光能利用率的提高不仅是由于现有研究中已经明确的叶片光合作用——光响应曲线的非线性使得较低光强下比高光强下叶片光能利用率高，而且还因为遮阴处理下降低的植株C/N比提高了叶片饱和光强下的最大光合速率和初始光能利用率，从而使得作物对全球变暗产生了光合适应现象。

4．全球变暗降低了小麦和水稻各产量构成因素和收获指数，特别是显著降低了每穗实粒数，导致散射辐射比例提高对产量的“肥料效应”小于对地上部生物量的“肥料效应”。遮阴处理下，小麦每平方米穗数降低3.81%~9.96%、每穗总粒数降低3.44%~7.53%、每穗实粒数降低4.13%~15.63%、千粒重降低2.25%~8.96%，最终导致小麦收获指数降低1.86%~3.66%；水稻每平方米穗数降低1.12%~4.50%、每穗总粒数降低4.19%~27.55%、每穗实粒数降低5.50%~30.54%、千粒重降低0.76%~8.13%，最终导致水稻收获指数降低1.94%~6.32%。遮阴处理造成小麦和水稻减产的主要原因是每穗实粒数的显著降低，而遮阴处理导致抽穗前地上部氮素吸收量的减少是每穗实粒数显著下降的原因。因此，全球变暗下每穗实粒数降低会削弱散射辐射对小麦和水稻产量的“肥料效应”，进而使得散射辐射比例提高对产量的“肥料效应”小于对地上部生物量的“肥料效应”。

综上所述，全球变暗背景下，由于冠层光能利用率的提高，散射辐射比例提高对作物地上部生物量具有“肥料效应”，而降低的每穗实粒数削弱了“肥料效应”对两种作物产量的影响；冠层光能利用率的提高不仅是因为叶片光合作用——光响应曲线的非线性使得较低光强下比高光强下叶片光能利用率高，同时还因为作物对全球变暗产生了光合适应现象。但本研究模拟的中度和重度污染下的散射辐射比例的增加对产量和生物量的“肥料效应”不能完全弥补太阳总辐射下降造成的产量和生物量损失。这些研究结果为小麦和水稻生产应对气候变化，以及进一步建立准确评估全球变暗对作物生产力影响的模拟模型提供了理论依据和实验数据支撑。

Solar radiation is the primary energy source for crop growth and development, and plays a vital role in determining crop yield. A decadal decrease in incident global radiation, known as global dimming, has become one of the main threats with which China has to cope nowadays. Global dimming decreases the incident global radiation, but increases the fraction of diffuse radiation. Crop responses differently to diffuse radiation and direct radiation. Determining the impacts of global dimming on wheat and rice growth and yield, two main staple food crops in China, is of utmost importance for ensuring crop production in China under future changed climate. For clarifying the impacts of the combination of decreased global radiation and increased diffuse radiation fraction on wheat and rice growth and yield, this study combined statistical analysis using historical data and field experimental study. We started with statistical analysis of climate (including daily mean, maximum and minimum air temperature, precipitation, global and diffuse radiation) and crop (including phenology and yield) data in the middle and lower reaches of the Yangtze River from 1961 to 2016. We analysed the temporal changes in the global radiation and the fraction of diffuse radiation, and verified whether the global dimming impacts on crop yields can be accurately assessed by statistical models based on historical data. Secondly, field experiments with wheat (Ningmai 13) and rice (Nangeng 46) were conducted to mimic the solar radiation changes under global dimming by shading with polyethylene (PE) films in Nanjing, China, from 2013 to 2016. In each crop growing season, there were two shading treatments (T1 and T2, plots covered with different layers and thickness of films) and a control (CK, plots without cover). Based on the field experimental data, the fertilization effect of the increased diffuse radiation fraction was verified, the causes for the diffuse radiation fertilization effect were explored, and the impacts of solar radiation changes under global dimming on crop growth and yield were identified. The main research results are as follows:

1. The statistical analyses using historical climate and crop data cannot reliably assess the global dimming impacts on wheat and rice yields in the middle and lower reaches of the Yangtze River. The incident global radiation reduced between 1961 and 1990, and thereafter the reduction rate slowed down or even reversed to increase, in the middle and lower reaches of the Yangtze River. Even so, the global radiation has not recovered to the level around 1961. However, the fraction of diffuse radiation in this area continuously increased during 1961-2016. Our statistical results based on historical climate and crop data demonstrated that, for wheat and rice, the decreased global radiation during 1961-2016 led to losses in yields; however, the negative correlation between rice climatic yield and diffuse radiation, and the positive correlation between wheat climatic yield and temperature were in contrast with the established understanding. These results that contrary to the established understanding were due to the significant linear correlation among statistical variables (such as between global radiation and temperature). Therefore, without the support of experimental data, statistical analysis using historical data only was insufficient to accurately evaluate the impacts of global dimming on wheat or rice yields.

2. The increased diffuse radiation fraction had a fertilization effect on wheat and rice yield and aboveground biomass. The yield and aboveground biomass increased by, respectively, 6.6%~7.8% and 8.2%~9.3% in wheat, and 4.2%~4.5% and 5.9%~6.2% in rice, when the diffuse radiation fraction increased 10%, relative to CK treatment. However, the diffuse radiation fertilization effect under moderate and serve air pollution mimicked in our experiments cannot compensate completely for the losses in yield and biomass caused by the decreased global radiation. Under shading treatments, the global radiation decreased by 11.8%~25.5% and the diffuse radiation fraction increased by 7.4%~20.4% in wheat growing seasons, and the global radiation decreased by 12.3%~29.9% and the diffuse radiation fraction increased by 7.9%~21.8% in rice growing seasons, and the shading treatments mimicked well the solar radiation conditions under moderate and severe air pollution. Shading treatmens decreased yield by 8.31%~16.36% and 7.81%~20.86%, respectively, and decreased aboveground biomass by 6.39%~13.45% and 5.63%~15.54%, respectively, in wheat and rice. However, the percentages of the reductions in yield and aboveground biomass were lower than that in global radiation. The yield and aboveground biomass under shading treatments can increase by, respectively, 6.6%~7.8% and 8.2%~9.3% in wheat, and 4.2%~4.5% and 5.9%~6.2% in rice, when the diffuse radiation fraction increased 10%, relative to CK treatment. In addition, to keep the yield and aboveground biomass constant when the global radiation decreased 10%, the fraction of diffuse radiation that had to increase was, respectively, 14.0%~14.2% and 11.5%~11.8% for wheat, and 22.1%~23.9% and 15.9%~17.7% for rice, relative to CK treatment. These results confirmed that there was a fertilization effect of the increased diffuse radiation fraction under global dimming on yield and aboveground biomass, and the fertilization effect on aboveground biomass was greater than on yield, for both wheat and rice. However, the fertilization effect of the increased diffuse radiation fraction under moderate and serve air pollution mimicked in our experiments cannot compensate completely for the losses in yield and biomass caused by the decreased global radiation.

3. The diffuse radiation fertilization effect under global dimming was mainly attributed to the enhanced canopy light use efficiency but not to an improved canopy light interception; the increased canopy light use efficiency was explained not only by nonlinearities of the photosynthetic light response curves, but also by photosynthetic acclimation to global dimming. Under shading treatments, for wheat and rice, the canopy radiation use efficiency increased by 0.48%~8.14% and 1.38%~9.93%, respectively; the leaf area index increased by 0.57%~11.45% and 2.55%~22.91%, respectively; while the canopy light extinction coefficient decreased by 0.64%~10.60% and 2.91%~23.21%, respectively. Due to the opposite changes of leaf area index and light extinction coefficient, shading hardly altered the fraction of the incident radiation intercepted by canopy. Therefore, the diffuse radiation fertilization effect was mainly attributed to the increased canopy light use efficiency but not any improved canopy light interception. We further found that, under shading treatments, the light-saturated gross photosynthetic rate of the first and third leaves counted from the top increased by, respectively, 0.57%~11.15% and 1.76%~16.18% for wheat, and 0.75%~13.08% (except shading treatments decreased by 1.71%~6.17% at stem-elongating stage in 2013) and 0.44%~7.83% (except T1 treatment decreased by 1.72% at stem-elongating stage in 2015) for rice; the initial light-use efficiency of the first and third leaves increased by, respectively, 1.54%~17.07% and 1.43%~15.53% for wheat, and 1.40%~13.69% and 1.30%~8.02% for rice. Shading also decreased the C/N ratio in the shaded plant, but increased green leaf nitrogen concentration by 0.94%~8.87% for wheat and 0.08%~11.27% for rice. We also found that both light-saturated gross photosynthetic rate and initial light-use efficiency were significantly correlated with leaf nitrogen concentration. Therefore, the increased canopy light use efficiency arose not only from the saturating shape of the photosynthetic light response curves but additionally due to crop acclimation to global dimming that caused by the decreased C/N ratio in shaded plant.

4. Global dimming reduced wheat and rice yield components and harvest index, especially significantly reduced the number of filled grains per ear, resulting in a smaller fertilization effect of the increased diffuse radiation fraction on yield than on aboveground biomass. For wheat, shading decreased ear number per unit area by 3.81%~9.96%, total grain number per ear by 3.44%~7.53%, filled grain number per ear by 4.13%~15.63% and thousand grain mass by 2.25%~8.96%, and ultimately decreased harvest index by 1.86%~3.66%; for rice, shading decreased ear number per unit area by 1.12%~4.50%, total grain number per ear by 4.19%~27.55%, filled grain number per ear by 5.50%~30.54% and thousand grain mass by 0.76%~8.13%, and ultimately decreased harvest index by 1.94%~6.32%. For both crops, yield loss was mainly ascribed to the reduction in the filled grain number per ear, which was predominantly caused by the decreased nitrogen uptake before heading under shading treatments. Therefore, the decreased number of filled grains per ear under shading discounted the diffuse radiation fertilization effect on wheat and rice yield, resulting in a smaller fertilization effect on yield than on aboveground biomass.

In summary, there was a fertilization effect of the increased diffuse radiation under global dimming on crop productivity, due to the enhanced radiation use efficiency. However, the decreased filled grain number per ear under global dimming discounted the fertilization effect on the yield of both crops. Moreover, the enhanced canopy light use efficiency arose not only from the saturating shape of the photosynthetic light response curves but additionally due to crop acclimation to global dimming. However, the fertilization effect of the increased diffuse radiation fraction under moderate and serve air pollution mimicked in our experiments cannot compensate completely for the losses in yield and biomass caused by the decreased global radiation. The results in the present study could supply a theoretical basis and experimental data support for wheat and rice production to cope with climate change and further establishing crop models that could be used for precisely assessing the global dimming impact on crop productivity.