气候变化可导致作物严重减产，脆弱性加剧。社会经济发展则可为应对和适应气侯变化导致的不利影响提供支撑，减轻作物脆弱性。目前，对二者耦合作用下的作物脆弱性评估的研究尚不多见。小麦作为主要口粮作物，在保障全球粮食安全中占有重要地位。因此，如何定量评估气候变化与社会经济发展共同影响下的全球小麦脆弱性，既是灾害学领域急需解决的重要科学问题，也是应对气候变化与防范农业灾害风险的重大现实需求，具有重要的科学与实践价值。 本研究基于作物模型，以小麦为研究对象，采用多模型预测的方法对全球2020-2100年农田灌溉和施肥参数进行了预测，解决了社会经济因素在作物生长过程模拟中的参数化问题；采用SCE-UA方法对全球冬、春小麦生长模拟参数进行了率定和验证，构建得到未来气候变化与社会经济共同影响下的全球冬、春小麦生长模拟优化参数集；探讨了气候变化与社会经济发展情景下全球冬、春小麦脆弱性评价方法，评价并揭示了未来（2010-2100年）气候变化与社会经济发展情景下冬、春小麦脆弱性的空间分布格局和动态演变趋势。得到的结论如下： （1）构建了社会经济驱动下农田灌溉与施肥参数的集合预测模型，解决了社会经济因素在作物生长模拟中的参数化问题。筛选多元线性回归模型，决策树模型，自回归整合移动平均模型，支持向量机，多层感知器，径向基函数和随机森林模型，运用预测误差平方和最小原则，利用拉格朗日乘数法，构建了灌溉和施肥量集合预测模型。经检验，该集成预测模型具有较高精度，有效避免了单一模型预测不确性高的弊端。基于该模型，预估了共享社会经济路径情景下（SSP1、SSP2和SSP3）全球2020-2100年农田灌溉和施肥量，为模拟气候变化和社会经济共同影响下的小麦生长模型参数率定及验证提供了数据支撑。 （2）对全球冬、春小麦生长模拟参数进行了全局最优化率定和验证，构建了未来气候变化与社会经济共同影响下的全球冬、春小麦生长模拟优化参数集。采用EPIC模型对2000-2004年小麦生长过程和产量进行模拟，运用全局最优化参数优化SCE-UA算法进行参数调整，并依据小麦观测产量对模拟结果进行验证，得到了未来气候变化与社会经济共同影响下的全球冬、春小麦生长模拟优化参数集。验证结果表明运用EPIC模型模拟冬、春小麦生长过程和产量精度较高，误差在合理范围内，能够对未来小麦生长过程和产量进行较好的模拟。 （3）构建了EPIC模型支持下的小麦脆弱性动态评估方法，解决了气侯变化与社会经济发展耦合作用下的小麦脆弱性定量预估问题。本文将小麦脆弱性定义为未来由于气候变化和社会经济的影响而导致的小麦产量损失，产量损失越大，则小麦脆弱性越大。在参数优化后的EPIC模型支持下，以气候变化情景和社会经济情景作为模拟变量，模拟了RCP2.6、RCP4.5和RCP8.5，以及RCP2.6-SSP1、RCP4.5-SSP2和RCP8.5-SSP3情景下2010-2100年全球小麦生长过程和产量。通过对比模拟产量与历史时期正常年份小麦产量，计算得到了气候变化情景下的小麦自然脆弱性，和气候变化与社会经济发展共同影响下的小麦综合脆弱性。 （4）2020-2100年，除SSP3情景外，全球有灌溉装备的农田面积在SSP1和SSP2情景下均呈下降趋势。SSP3情景下预测末期（2100年）与预测初始期（2020年）相比，上升了5.6%，而SSP1情景下减少了7.3%，SSP2情景下则减少了8.9%。全球有灌溉装备的农田集中分布在东亚，南亚，西亚，和美国中部。2020-2100年，除SSP1情景外，全球农田氮肥和磷肥消耗量在SSP2和SSP3情景下均呈下降趋势。SSP1情景下预测末期与预测初始期相比氮肥和磷肥消耗量基本不变，而SSP2情景氮肥减少了16.7%，磷肥减少了9.1%，SSP3情景氮肥和磷肥都减少了2%。全球农田氮肥消耗集中分布在东亚，南亚，东南亚，欧洲，北美洲、南美洲东南部。全球农田磷肥消耗量集中在东亚、南亚、欧洲西部和南部、北美洲中部以及南美洲东南部。2020-2100年，全球农田钾肥消耗量在SSP1-SSP3情景下均呈增加趋势。预测末期与预测初始期相比，SSP1情景下增加17.8%，SSP2情景增加了9.7%，SSP3情景增加了8.8%。全球农田钾肥消耗集中分布在东亚、南亚、欧洲西部、北美洲中部以及南美洲东南部。 （5）2010-2100年，在气候变化影响下，从RCP2.6情景到RCP8.5情景，全球冬、春小麦产量都较基准期（1976-2005年）呈下降趋势。所有气侯情景下，春小麦产量下降较冬小麦更为显著，且RCP8.5情景下冬小麦下降幅度最大达17.1%，春小麦下降幅度最大达26.9%。气候变化影响下冬小麦脆弱性降低区域（即，减产率小于0区域）面积比例由2010年的3%提高到2100年的10%，脆弱性严重增加区域（即减产率在0.5-1之间区域）面积比例由11%提高到16%；春小麦脆弱性降低区域面积比例由3%提高到9%，脆弱性严重增加区域面积比例由15%提高到20%。可见，随着温室气体排放浓度的增加，小麦受到的不利影响越大。而春小麦产量损失较大的原因可能在于春小麦位于高纬度和低纬度地区，受温室气体影响更大。 （6）2010-2100年，与只考虑气候变化情景相比，在气候变化和社会经济共同影响下，从RCP2.6-SSP1情景到RCP8.5-SSP3情景下冬小麦产量提高了2.2%-2.9%，春小麦产量几乎不变。冬小麦综合脆弱性降低区域面积比例由2010年9%提高到2100年的14%，与只考虑气候变化相比，从RCP2.6-SSP1情景到RCP8.5-SSP3情景下提高了4%-6%。综合脆弱性降低区域主要集中在欧洲东部，西亚地区，中国华北平原地区，朝鲜，韩国，日本，美国西部。春小麦综合脆弱性降低区域面积比例由3%提高到10%，较只考虑气候变化情景相比提高了0-1%，主要集中在蒙古，中国东北地区，尼泊尔。可见，社会经济发展通过影响灌溉和施肥在一定程度上能够缓解小麦由于气候变化带来的产量损失，这意味着社会经济发展有助于人类提高适应气候变化的能力，从而抵消甚至逆转小麦由于气候变化带来的不利影响。 论文提出的耦合气侯变化与社会经济影响的小麦脆弱性预估方法、社会经济影响下的农田灌溉与施肥集成预测方法，为开展全球变化农作物脆弱性预估研究提供了方法论的支撑。论文的研究结果揭示了气候变化、耦合气侯变化与社会经济共同影响下的小麦自然脆弱性和小麦综合脆弱性时空演变规律。论文的研究发现，对加深气侯变化与社会经济发展影响下的全球小麦脆弱性认知，加强气候变化导致的小麦生产风险应对具有重要科学价值。

Climate change can lead to severe crop losses and increase vulnerability level. Socio-economic development can provide support to respond to and adapt to the adverse effects of climate change and reduce crop vulnerability. At present, there are few researches on crop vulnerability assessment under the coupling of the two. Wheat, as the main food crop, plays an important role in ensuring global food security. Therefore, how to quantitatively assess the global wheat vulnerability under the impact of climate change and socio-economic development is not only an important scientific problem that needs to be solved urgently in the field of disaster science, but also a major realistic demand for coping with climate change and preventing agricultural disaster risk, which has important scientific and practical value. Based on crop model, wheat was taken as the research object, the parameters of irrigation and fertilization in 2020-2100 were predicted by ensemble prediction model, which solved the problem of parameterization of socio-economic factors in the simulation of crop growth process. SCE-UA method was used to calibrate and verify the global winter and spring wheat growth simulation parameters, and to construct the optimization parameter set under the influence of climate change and socio-economic development. The spatial distribution pattern and dynamic evolution trend of winter and spring wheat vulnerability under climate change and socio-economic development (from 2010 to 2100) were evaluated and revealed. The results are as follows: （1）An ensemble prediction model of farmland irrigation and fertilization parameters driven by socio-economic development was built to solve the problem of parameterization of socio-economic factors in crop growth simulation. Multiple linear regression model, Decision tree model, Autoregressive integrated moving average model, Support vector machine, Multi-layer perceptron, Radial basis function and Random forests were selected, and based on the principle of minimum sum of squares of prediction error and Lagrange multiplier method, the ensemble prediction model of irrigation and fertilizer was established. It has been proven that the ensemble prediction model has high accuracy and can avoid the disadvantages of low accuracy of single model. Based on this model, the global irrigation and fertilization in 2020-2100 under SSP1, SSP2 and SSP3 was predicted, which provided data support for parameter calibration and verification of wheat growth model under the influence of climate change and socio-economic development. （2）The global winter and spring wheat growth simulation parameters were calibrated and verified, and the optimization parameter set was constructed under the influence of climate change and socio-economic development. EPIC model was used to simulate the growth process and yield of wheat from 2000 to 2004, SCE-UA algorithm was used to calibrate the parameters, and the simulation results were verified according to the observed yield of wheat, and the global optimal parameter set of winter and spring wheat growth simulation under the influence of climate change and socio-economic development was obtained. The results showed that using EPIC model to simulate the growth process and yield of winter and spring wheat is of high accuracy, and the error is within a reasonable range. （3）The dynamic evaluation method of wheat vulnerability supported by EPIC model was established to solve the quantitative estimation problem of wheat vulnerability under the influence of climate change and socio-economic development. In this paper, wheat vulnerability is defined as wheat yield loss caused by climate change and socio-economic development. The greater the yield loss, the higher the vulnerability level of wheat. With the support of the parametric EPIC model, the global wheat growth process and yield from 2010 to 2100 under RCP2.6, RCP4.5 and RCP8.5, as well as RCP2.6-SSP1, RCP4.5-SSP2 and RCP8.5-SSP3 scenarios were simulated using climate change scenario and socio-economic scenario as simulation variables. By comparing simulated yields with wheat yields in normal years in historical periods, the physical vulnerability of wheat under climate change scenarios and the integrated vulnerability of wheat under the influence of climate change and socio-economic development were calculated. （4）From 2020 to 2100, except for the scenario of SSP3, the global farmland area shows a decreasing trend under the scenarios of SSP1 and SSP2. Compared with the initial prediction period (in 2020), the final prediction period (in 2100) increases by 5.6% in SSP3, while the SSP1 scenario decreases by 7.3%, and the SSP2 scenario decreases by 8.9%. The global farmland is concentrated in east Asia, South Asia, west Asia, and America. From 2020 to 2100, except for the SSP1 scenario, the global consumption of nitrogen fertilizer and phosphorus fertilizer under the SSP2 scenario and SSP3 scenario shows a decreasing trend. In the SSP1 scenario, the consumption of nitrogen and phosphate fertilizer is basically unchanged at the end of the prediction period compared with the initial prediction period, while in the SSP2 scenario, nitrogen fertilizer reduces by 16.7%, phosphorus fertilizer reduces by 9.1%, and both nitrogen and phosphate fertilizer reduce by 2% in the SSP3 scenario. The global nitrogen fertilizer is concentrated in east Asia, South Asia, southeast Asia, Europe, North America and southeast America. Global phosphate fertilizer is concentrated in east and South Asia, western and southern Europe, central North America and southeastern South America. From 2020 to 2100, the consumption of potash fertilizer will increase under the scenario of SSP1-SSP3. Compared with the initial prediction period, the final prediction period increased by 17.8% under the SSP1 scenario, 9.7% under the SSP2 scenario, and 8.8% under the SSP3 scenario. The global potash is concentrated in east Asia, South Asia, Western Europe, central North America and southeast South America. （5）From RCP2.6 to RCP8.5, global winter and spring wheat yields showed a downward trend from the base period (from 1976 to 2005) from 2010 to 2100 under the influence of climate change. The yield decline of spring wheat was more significant than that of winter wheat under all climate change scenarios. Under RCP8.5, winter wheat falls by 17.1% and spring wheat falls by 26.9%. Under the influence of climate change, the proportion of areas with reduced vulnerability of winter wheat (i.e., areas with reduced yield rate less than 0) increased from 3% in 2010 to 10% in 2100, and the proportion of areas with severe increase in vulnerability (i.e., areas with reduced yield rate between 0.5 and 1) increased from 11% to 16%. The proportion of areas with reduced vulnerability of spring wheat increased from 3% to 9%, and the proportion of areas with increased vulnerability increased from 15% to 20%. It can be seen that with the increase of greenhouse gas emission concentration, wheat is more adversely affected. The reason that spring wheat yield suffers from greater losses is that spring wheat is located at high and low latitudes, which are more affected by greenhouse gases （6）From 2010 to 2100, compared with the climate change scenario, the winter wheat yield increases by 2.2% to 2.9% from RCP2.6-SSP1 to RCP8.5-SSP3, the spring wheat yield is almost unchanged. The integrated vulnerability reduction area of winter wheat increases from 9% in 2010 to 14% in 2100, and compared with climate change, it increases by 4% to 6% from RCP2.6-SSP1 to RCP8.5-SSP3 scenario. The integrated vulnerability reduction areas are mainly concentrated in eastern Europe, west Asia, north China plain, north Korea, South Korea, Japan and western America. The integrated vulnerability reduction area of spring wheat increases from 3% in 2010 to 10% in 2100, and compared with climate change, it increases 0% to 1% from RCP2.6-SSP1 to RCP8.5-SSP3 scenario, which mainly in Mongolia, northeast China, Nepal. It can be seen that the increase of irrigation and fertilization brought about by socio-economic development can alleviate the loss of wheat yield due to climate change to a certain extent, which means that socio-economic development can help humans improve their ability to adapt to climate change, thereby offsetting or even reversing adverse effects of climate change on wheat. The prediction method of wheat vulnerability under climate change and socio-economic development and the ensemble prediction method of farmland irrigation and fertilization under socio-economic development proposed in this paper provide methodological support for the research of global crop vulnerability prediction. The results of this paper reveal the temporal and spatial evolution trend of wheat physical vulnerability and integrated vulnerability under climate change coupled with climate change and socio-economic development. The findings in this paper is of great scientific value to deepen the understanding of global wheat vulnerability under the influence of climate change and socio-economic development and to strengthen the risk response of wheat production caused by climate change.