在当今全球变暖的大趋势下，气温的增高势必会影响到全球农作物的种植范围和产量。水稻作为全球三大粮食作物之一，对全球粮食安全的影响十分重大。目前关于全球变化对粮食影响的研究中，往往偏重气候变化影响下的作物单产和总产量的变化，而未考虑土地利用和作物需求带来的影响。利用社会经济模型研究的农作物需求量和产量的预测则多为全球或国家尺度的研究，难以与气候变化情景相结合。对气候变化下农作物风险研究则多假定农作物暴露范围不发生改变。本文试图综合考虑气候变化和社会经济变化两方面的因素，预测未来水稻在全球温度变化下的暴露变化，一方面可以探索农作物研究中气候变化和社会经济双情景综合的方法，另一方面，评估农作物在全球变化下的暴露变化是进行未来农作物风险评价的基础。 本文结合历史和未来的气候模式数据，土壤数据，历史水稻收获面积、产量、需求量等统计数据，历史和未来土地利用数据对全球温度变化下的水稻暴露进行预估。在气候变化层面，计算了历史时期（1970-2000年），未来近期（2016-2035年）和未来中期（2046-2065年）全球0.25°×0.25°栅格尺度的水稻自然种植适宜区，水稻生育期，温度趋势变化、温度波动和极端高、低温日数。在社会经济层面，计算了可比地理单元上的未来近期和未来中期水稻需求量和水稻平均单产。进一步建立水稻自然种植适宜区和水稻种植比例的关系构建了未来的水稻产量分配模型。得到RCPs和SSPs情景组合下的水稻收获面积和产量的预估。最后与温度趋势变化、波动和极端高、低温事件组合，得到未来温度变化下的水稻暴露并绘制图谱。研究结论如下： （1）全球温度变化方面，将未来的温度变化分解为温度趋势、波动和极端事件三个方面，分别构建并计算了反映温度趋势、波动和极端高、低温的指标。结果表明总体上全球均处于增暖趋势，其强度和范围在不同时间段内有所不同。历史时期到未来近期时间段仅有赤道地区有微弱的变暖趋势，未来近期到未来中期时间段增温趋势最明显。历史时期到未来中期时间段内增暖趋势的范围最大。RCP8.5的增暖趋势远高于RCP4.5，尤其是在未来近期到未来中期时间段，普遍的增温趋势达到0.05℃每年以上。温度波动变化方面在各个时期各个情景下有相似的空间分布情景，总体上体现为赤道波动最低，两极波动最高，沿海地区波动低于内陆。极端高温主要分布在非洲-亚洲沿线的亚热带地区，随时间变化逐渐增强，且RCP8.5情景下暴露日数远高于RCP4.5情景。极端低温的分布范围较高温更广，暴露日数也更长。且由于水稻的种植制度和生长期在不同气候条件下不同，低温日数在北半球呈现出沿纬度高低波动的趋势，南纬30°左右、北纬25°-30°和北纬50°-60°为极端低温暴露日数的高值带。随着时间推移，低温暴露日数的带状分布会整体向极地发生移动，且未来中期较未来近期更为明显，RCP8.5较RCP4.5更为明显。 （2）水稻分布范围、收获面积和产量方面，构建了一套综合考虑水稻自然适宜性，土地利用，水稻单产和水稻需求量的全球水稻分布范围、收获面积和产量的预估模型。结果表明，全球适宜水稻种植的自然条件分布非常广泛，在各个时期各个情景下变化不大，多源数据精度验证的结果表明水稻真实分布相对于水稻适宜区的总体精度在66%左右，Kappa系数在0.8-0.9之间，表明二者间一致性很高。在非洲撒哈拉以南，非洲东部，印度-东南亚-中国东部，美国东南部，均是非常适宜种植水稻的区域。未来水稻的需求量预估方面，印度，中国，东南亚各国，中亚和非洲个别国家仍是世界水稻需求量最大的地区。中国、俄罗斯的水稻需求量在减少，而非洲需求量的增加速度最快。全球未来水稻单产高的地区主要集中在经济发达的地区。双情景下的水稻种植范围和产量结果可以表明未来水稻的种植面积和产量仍集中在印度、东南亚国家和中国东部。欧洲东部，非洲东部，澳大利亚，阿根廷和美国在未来有较大的种植面积和产量的增加。 （3）水稻在温度变化下的暴露方面，构建了四种温度变化（趋势、波动、极端高温和极端低温）-两种未来情景（RCP4.5-SSP2和RCP8.5-SSP3）-三个时间段（历史时期，未来近期和未来中期）-四种暴露指标（暴露强度，暴露面积，暴露产量和暴露日数）四个维度组合的水稻温度变化暴露评价体系并绘制图谱。结果表明，水稻主产区的地区，在历史时期-未来中期和未来近期-未来中期均暴露在很强的温度趋势变化中。中国东部地区和美国中部地区在RCP8.5-SSP3情景下暴露在较强的温度波动下。各个时期的印度东部，未来中期RCP8.5-SSP3情景下的中国中部和美国中部有较长时间的极端高温高暴露。极端低温暴露最高的区域主要是中国和美国，且在各个时期都有较高的暴露。此外可以发现RCP8.5-SSP3情景下，全球范围内趋势，波动和极端高温影响的范围更大、强度更高，特别是中国中部和美国中部水稻高暴露区，相比于RCP4.5情景，水稻极端高温和温度波动风险大大增高。 （4）高强度温度变化下的水稻暴露变化方面，高温度趋势变化下，历史时期到未来时期，除RCP8.5-SSP2 历史时期-未来中期外暴露增量均较小，暴露面积增加低于150万公顷，暴露产量增加量低于0.2亿吨。高温度波动和高极端高温下的暴露总量和占比均随时间增高，极端低温暴露面积和产量同样随时间增高，但占比在未来中期较未来近期略有下降。RCP8.5-SSP3情景是总体暴露最高的情景，历史时期-未来中期高温度变化趋势下的暴露面积增加1744.5万公顷，暴露产量增加2.5亿吨；未来中期高温度波动、高极端高温日数和高极端低温日数下的暴露面积分别占全球水稻收获面积的22.0%，26.6%和19.4%，暴露产量分别占全球总产量的30.2%，20.7%和27.4%。高变暖趋势、高温度波动和高极端高温日数下，RCP4.5-SSP2情景下暴露量或暴露变化量均低于RCP8.5-SSP3情景，极端低温则相反，但两个情景相差不大。 本研究的创新之处主要体现在以下3个方面：1）结合气候变化情景RCPs和社会经济变化情景SSPs，预测了未来水稻的种植范围，种植面积和产量。相比使用单一情景的结果，兼顾了作物生长的自然条件要求和人类种植水稻的需求和历史因素。2）通过预测水稻未来生长期，预测了未来水稻极端高温和低温日数，完善了水稻暴露中暴露时间维度的评价。3）基于不同情景和不同时期，构建了水稻多温度变化指标-多情景-多时间-多暴露指标的四维暴露评价体系。完整详细地展示出水稻在温度变化下暴露的时空变化规律。 本研究为全球变化自然-社会双情景下的农作物暴露预估进行了方法上的探索，制作的暴露图谱是未来水稻全球温度变化风险的基础。分析了未来水稻在未来温度趋势、波动和极端温度变化下的暴露变化，为未来全球粮食安全战略决策提供依据。

Under the current global warming situation, the increasing temperature is bound to affect the planting range and production of global crops. As one of the three major cereals in the world, rice is closely related to global food security. In the current researches of crops response to the global change, many focus on changes in crop yields and total yields but hardly taking into account the effects of land use and crop demand. The predictions of crop demand and production by socio-economic models are mostly based on global or national scale and are difficult to integrate with climate change scenarios. Studies on crop risks under climate change have assumed that crop exposures will not change. Therefore, this paper attempts to combine the future climate change and socio-economic change scenario to predict the future rice exposure under global temperature changes. On the one hand, it can explore how climate change scenarios and socio-economic scenarios can be combined in crop exposure research. On the one hand, the exposure results are the basis for the risk of crop climate change. This research use climate model data, soil physical and chemical properties data, land use data, population, GDP, and rice statistical data. Based on these, at the climate change level, we calculate the natural suitability of rice, rice growth period, temperature trend change, temperature fluctuation and extreme hot and cold days in global 0.25°×0.25° grid scale in historical period (1970-2000), near future (2016-2035) and future mid-term (2046-2065). At the socio-economic level, the demand and the average yield of rice on the comparable geographic unit scale were estimated. Further, the future rice production distribution model was constructed by establishing the relationship between the natural suitable area of rice and the ratio of rice planting area to cultivated land area and estimate the rice harvested area and production under the scenario combination of RCPs and SSPs. Finally, combined with temperature trend, fluctuations and extreme hot and cold days, the rice exposure map under future temperature changes is obtained. The research conclusions are as follows: (1) In terms of global temperature change, the future temperature changes are decomposed into three aspects: temperature trend, fluctuation and extreme events. The indicators reflecting temperature trend, fluctuation and extreme high and low temperature are constructed and calculated. The results show that the global warming trend is generally, and its intensity and range varies in different time periods. From the historical period to the near future, only the equatorial region has a slight warming trend, and the trend of temperature increase is most obvious in the near future to the mid-term. The range of warming trends from the historical period to the mid-term period is the largest. In RCP8.5 scenario, the warming trend is much higher than RCP4.5, especially in the near future to the future mid-term period in which the general warming trend reaches 0.05 °C per year or more. In terms of temperature fluctuations, there are similar spatial distribution in every scenario and period combination. The overall distribution is that the equatorial fluctuations are the lowest and highest in polar, and the coastal areas are more stable than inland. The extreme high temperature is mainly distributed in the subtropical area in Africa-Asia region, which gradually increases with time, and the exposure days under the RCP8.5 scenario are much higher than the RCP4.5 scenario. The extremely low temperature distribution range is higher than extremely high temperature, and the number of exposure days is longer. Because the rice planting system and growing season are different under different climatic conditions, the low temperature days show a trend of fluctuation along the latitude in the northern hemisphere. The high value bands for low temperature exposure days are distributed in about 30° south latitude, 25°-30° north latitude and 50°-60° north latitude. As time goes by, the band-like distribution of the low-temperature exposure days will shift to pole, especially in the future medium-term period and on RCP8.5 scenario. (2) In terms of rice distribution range, harvested area and production, a model to evaluate these were constructed to take into account the natural suitability of rice, land use, rice yield and rice demand. The results show that the natural conditions suitable for rice planting are widely distributed, and there is little change under various scenarios in each period. The results of multi-source data accuracy verification show that the overall accuracy of rice suitable area relative to rice true distribution is about 66%, Kappa coefficient between 0.8 and 0.9 indicating a high consistency. The very suitable regions for growing rice are mainly distributed in sub-Saharan Africa, eastern Africa, India-Southeast Asia-east China, and the southeastern United States. In terms of future demand for rice, India, China, Southeast Asian countries, Central Asia and several African countries are still the regions with the largest demand for rice in the world. Rice demand in China and Russia is decreasing, while it is growing fastest in Africa. The regions with high yields are mainly concentrated in developed regions. The rice planting range and production results under the two scenarios indicate that the future planting area and production of rice are still concentrated in India, Southeast Asian countries and eastern China. Eastern Europe, Eastern Africa, Australia, Argentina and the United States have larger planting areas and increased production in the future. (3) In terms of rice exposure under temperature changes, an exposure analysis and mapping system combined with four temperature changes (trends, fluctuations, extreme high temperatures and extreme low temperatures), two future scenarios (RCP4.5-SSP2 and RCP8.5-SSP3), three time periods (historical period near future and mid-term future), and four exposure indicators (exposure intensity, exposed area, exposure production and number of exposure days) were constructed. The results show that the main rice producing region are exposed to strong temperature trend changes during the historical period to the mid-term and the near future to the medium and future. Eastern China and the central United States are exposed to strong temperature fluctuations under the RCP8.5-SSP3 scenario. In the eastern part of India in various periods, central China and central United States under the RCP8.5-SSP3 scenario in the medium term, high rice production will be exposed to extreme high temperature in a long time. The areas with the highest exposure to extreme low temperatures are mainly in China and the United States in every period of this study. In addition, it can be found that under RCP8.5-SSP3 scenario, global temperature trends, fluctuations and extreme heat have greater range and intensity, especially in high-exposure areas of central China and central US, where the rice production risk of extreme high temperature and temperature fluctuations is higher than in the RCP4.5 scenario. (4) This paper calculates the changes of rice exposure under high intensity temperature changes. Under the change of high temperature trend, from the historical to the future period, except for the historical to mid-term period of RCP8.5-SSP2, the warming trend is gentle and the exposed area and production is less than 1.5 million hectares and 200 million tons respectively. The total exposure and its proportion under high temperature fluctuations and high extreme temperatures are increasing with time, and the extreme low temperature exposure area and output are also increasing with time, but the proportion is slightly lower in the mid-term future near future. The RCP8.5-SSP3 scenario has the highest overall exposure. During the historical to mid-term period, the exposure area and production under the high temperature trend is increased by 17.45 million hectares and 250 million tons. In the mid-term future, the exposed area under high temperature fluctuations, long extreme high temperature days and high extremes low temperature days accounted for 22.0%, 26.6% and 19.4% of the global rice harvest area, respectively, and the exposure production accounted for 30.2%, 20.7% and 27.4% of the global total production, respectively. Under the high warming trend, high temperature fluctuation and long extreme hot days, the exposure and exposure change in the RCP4.5-SSP2 scenario is lower than the RCP8.5-SSP3 scenario, while the long extreme cold days is the opposite, but the gap between two scenarios are vary small. The innovations of this research are mainly reflected in the following three aspects: 1) Combining climate change scenarios RCPs and socio-economic change scenarios SSPs, predicting the future planting range, planting area and production of rice. Compared to the results of using a single scenario, factors such as suitable environment for crop growth, food demand and land use are taken into account. 2) By predicting the future growth period of rice, and predicting the extreme high temperature and low temperature days of rice in the future, the time aspect of rice exposure is further improved. 3) Based on different scenarios and different periods, constructed evaluate system of rice exposure of multiple temperature changes indicator (trends, fluctuations, extremes), multiple scenarios (RCP4.5-SSP2, RCP8.5-SSP3), multiple time period (historical period, near future, mid-term future). The temporal and spatial variation of rice exposure under temperature changes is demonstrated as completely as possible. This study explored the natural-social dual scenarios for crop exposure predictions under global change. The exposure maps are the basis for future global temperature change risks for rice. The future changes of rice exposure in the temperature trends, fluctuations and extreme temperature changes can be the basis for global food security strategic decisions in future.