国际贸易引发的运输是长距离交通中的重要组成部分，也是难以实现碳中和的关键部门之一。这一部门面临需求持续增长、运输结构不断变化、能效持续提升但碳中和技术尚未成熟等问题，加上涉及跨境且责任主体不清，因此被认为是 “最难减排的部门” 之一。相比于研究较多的短途客运和货运，长距离交通的排放和驱动力的定量认识有限。要建立稳定、高效的减排机制，不仅需要全面掌握当前国际贸易运输的排放水平，更要追溯到各个主体，深入解析排放背后的驱动因素。本研究通过融合多源贸易流和运输模式分配数据库，结合机器学习方法，自主开发了国际贸易运输碳排放模型，涵盖航海、航空、公路和铁路四种运输模式，覆盖全球65个经济体、1221种货物类型。采用神经网络完成了对缺失数据的模拟，弥补现有研究中运输模式分配数据的不足。本研究利用此模型计算并解析了1995和2000两个历史参照年，2008~2009和2019~2021两段全球经济和疫情波动期，共计七年的海量排放数据和特征。量化了从运输模式、进出口国、贸易对、货物等维度的排放效率；通过112~200万个底层指标的量化，解析了排放与全球经济的复杂关联关系。研究发现，国际贸易运输碳排放从1995年的10.16亿吨，下降到了2021年的9.71亿吨，约占全球交通碳排放的12.6%。以2021年为例，海运碳排放达6.33亿吨，占国际贸易运输碳排放总量的65.1%；空运、公路和铁路分别贡献了1.82亿吨、1.44亿吨和0.11亿吨，各占18.8%、14.9%和1.2%。全球贸易（美元）在1995至2021年间增长了461%，但运输能效提升及贸易结构变化使得国际贸易运输在全球交通碳排放中的占比下降了6.6%。此外，金融危机和新冠疫情分别给2009年和2020年带来了10.4% 和4.3% 的排放减少，揭示了贸易运输排放的强韧性增加了减排的难度。碳排放对国家进、出口贸易同样会产生影响。研究发现，运输过程的碳税将对全球平均出口价格产生0.40%~1.24 %的上浮（98%置信区间），其中以位于非洲和拉美洲的新兴和发展中经济体的出口价格涨幅最大。各经济体的贸易对运输碳税不一致的敏感性可能会对公平贸易和公正转型产生影响。最后，本研究以“就近贸易”为原则，对2021年的国际贸易格局进行优化，发现可以减少41.6% 的贸易运输排放。
In the realm of international trade, transportation plays a crucial role in long-distance travel and poses a significant challenge in achieving carbon neutrality. This sector faces persistent demands for growth, fluctuating transportation structures, ongoing efforts to improve energy efficiency, and a lack of carbon-neutral technologies. Additionally, complexities in cross-border operations and unclear responsibility attributions make it one of the "hard-to-abate" sectors. Unlike short-distance passenger and freight transportation, there's limited quantitative understanding of emissions and driving forces in long-distance transportation. Comprehensive insights into current emission levels and stakeholders are essential for establishing stable and effective decarbonization mechanisms.With a self-developed international trade transportation CO2 emission model, this study integrated multiple sources of trade flow and transportation modal share databases with machine learning techniques. Encompassing four transportation modes—maritime, aviation, road, and rail—the model covered 65 economies and 1,221 types of commodities globally. Artificial neural networks were employed to simulate missing transportation modal share data, addressing the inadequacies observed in existing research. Utilizing this model, the study computed and analyzed extensive emission data and features spanning seven years, including historical years in 1995 and 2000, as well as periods marked by the global financial crisis in 2008 to 2009 and the pandemic in 2019 to 2021. Emission efficiency quantification was conducted across various dimensions such as transportation modes, importing and exporting countries, bilateral pairs, and commodities. Through the quantification of 1.12 to 2.05 million underlying indicators, the study elucidated the bidirectional relationship between the global economy and transportation emissions.The study revealed that international trade transportation CO2 emissions decreased from 1015.7 Mt in 1995 to 971.4 Mt in 2021, accounting for approximately 12.6% of global transportation CO2 emissions. For instance, maritime transportation emissions in 2021 amounted to 633 Mt, constituting 65.1% of total international trade transportation CO2 emissions; aviation, road, and rail transportation contributed 182, 144, and 11 Mt, respectively, accounting for 18.8%, 14.9%, and 1.2% each. Global trade (in USD) witnessed a 461% increase from 1995 to 2021. However, improvements in transport energy efficiency and changes in trade structures led to a 4.4% decrease in the share of international trade transportation in global transportation CO2 emissions. Additionally, the global financial crisis and the COVID-19 pandemic respectively resulted in emission reductions of 10.4% in 2009 and 4.3% in 2020, revealing the resilience of trade transportation emissions, which increases the difficulty of decarbonization.This study found that carbon tax on transportation emissions could lead to a 0.40% to 1.24% increase in global average export prices (98% confidence interval). Emerging and developing economies, notably in Africa and Latin America, experienced the greatest price hikes, indicating higher trade sensitivities to carbon tax. The varying trade sensitivities may have implications for fair trade and a just transition.Finally, adhering to the principle of "trade proximity," this study optimized the international trade landscape in 2021, revealing a potential 41.6% reduction in trade transportation emissions.
