内容来源：https://research.google/blog/mapping-global-methane-emissions-from-space-with-deep-learning/

内容总结：

谷歌研究院近日发布一项基于深度学习的甲烷排放监测技术，可自动检测、量化并定位全球范围内的甲烷羽流，将卫星原始数据转化为可规模化的气候行动方案。相关成果已发表于《美国国家科学院院刊》（PNAS）。

甲烷是强效温室气体，其百年尺度增温潜力是二氧化碳的30倍，自工业时代以来贡献了约25%的人为气候变暖。由于甲烷在大气中寿命相对较短，快速减排被视为遏制全球升温的“快速行动”关键路径。目前，全球已有超125个国家签署《全球甲烷承诺》，力争到2030年减排30%。

为追踪全球排放源，科学家日益依赖星载成像技术。美国航天局（NASA）国际空间站上的“地球表面矿物尘埃源调查”（EMIT）仪器，原本用于绘制干旱地区矿物成分，但其高光谱能力可捕捉甲烷独特的光谱指纹。然而，地球地表背景复杂，部分物质光谱与甲烷相似，易造成误判。

谷歌研究院与NASA喷气推进实验室（JPL）合作，开发了名为“MAPL-EMIT”的端到端视觉变换器模型。该模型不再逐像素分析，而是结合光谱信息与空间上下文，通过分析气体扩散形态，区分真实的风吹甲烷羽流与光谱相似的地表物质，显著降低误报率。研究团队利用物理仿真框架生成了360万个合成甲烷羽流，直接注入真实EMIT影像中训练模型，使其适应多种大气与地理条件。

实测显示，MAPL-EMIT对专家标注的羽流识别召回率达84%，并能捕捉更多弱排放源，在约1100个EMIT影像中识别出的疑似羽流数量较现有方法增加约50%。在全球前25大垃圾填埋场中，模型成功绘制了其中24处的甲烷排放图。

为服务科学界，研究团队已在Google Earth Engine上发布全球甲烷羽流数据库及交互可视化应用，并在Kaggle平台开放训练模型与合成数据，在GitHub提供推理代码库。研究团队表示，随着NASA下一代成像光谱仪即将发射，其覆盖范围将扩大30至50倍，自动化高精度监测技术将愈发关键。该成果为地方决策者、研究人员和工业企业提供了快速识别甲烷排放的有力工具，有望推动全球气候目标实现。

中文翻译：

2026年9月1日

维沙尔·巴特楚，研究工程师，与米开朗基罗·康塞尔瓦，研究科学家，谷歌研究院

基于EMIT的甲烷分析与羽流定位模型是一个深度学习框架，可自动实现对全球甲烷羽流的检测、增强定量和源估算，将原始卫星数据转化为可规模化的气候行动。

甲烷是一种强效温室气体；以100年时间尺度计算，其增温潜力是二氧化碳的30倍。事实上，自工业时代开始以来，甲烷贡献了约25%的人为致暖效应。由于甲烷在大气中的寿命相对较短，及时减少这些排放为缓解全球气温上升提供了一条关键的“快速行动”路径。

这一紧迫性体现在《全球甲烷承诺》中，已有超过125个国家承诺到2030年将排放量减少30%。要实现这些目标，我们必须赋能各利益相关方，在废弃物、农业和能源部门中追踪局部点源（即在几十米量级的小空间范围内发生的排放）。最具成本效益的策略是减少油气基础设施、农业设施和垃圾填埋场的排放。

为了在全球范围内追踪这些排放，科学家们越来越依赖天基成像。一个典型例子是进驻国际空间站的NASA地球表面矿物粉尘源调查（EMIT）仪器。虽然该仪器最初设计用于绘制干旱地区的矿物组成，但NASA喷气推进实验室（JPL）的科学家以及更广泛的科学界已利用EMIT先进的高光谱能力来检测甲烷排放。通过为每个像素记录数百个不同波段的光，它使研究人员能够“看到”这些原本不可见气体的独特化学指纹。

在这些投入的基础上，我们在发表于《美国国家科学院院刊》（PNAS）的论文《基于EMIT高光谱辐射测量的深度学习全球甲烷点源监测》中，描述了一种将原始卫星数据转化为可规模化缓解行动的新方法。基于EMIT的甲烷分析与羽流定位（MAPL-EMIT）是一个深度学习框架，代表了向自动化检测、增强预测和全球甲烷羽流源估算迈出的重要一步。我们证明了MAPL-EMIT在专家标注的羽流上实现了84%的高召回率，并且与现有的基于匹配滤波的增强方法相比，具有更高的信噪比。为了支持更广泛的科学界，我们在Earth Engine上发布了全球羽流数据库，在Kaggle上发布了训练好的模型和合成羽流，并在Github上发布了推理库。

从太空测量甲烷需要平衡三个关键因素：（1）视场（空间覆盖/重访频率），（2）空间分辨率，以及（3）光谱分辨率。

像TROPOMI这样的全球测绘仪器的设计目标是探测背景甲烷浓度的微小变化，其实现方式是将大覆盖范围（约2,600公里幅宽）、粗空间分辨率（约5.5公里×3.5公里）和精细光谱采样（0.1纳米）相结合。

相比之下，像EMIT这样的点源测绘仪器擅长在设施尺度上测量甲烷排放。它们通过将中等覆盖范围（80公里宽的视场）与非常高的空间分辨率（60米）以及中等光谱分辨率（7.4纳米光谱采样）相结合来实现这一点，足以在高信噪比下捕获甲烷的化学特征。

然而，在全球范围内充分释放这一丰富数据的潜力还面临额外挑战。地球多样化的地貌提供了复杂的背景，某些地表物质可能伪装成甲烷，使得识别较小或更弥散的源变得尤为困难。为了在EMIT团队基础性工作的基础上实现高通量的全球测绘，我们与他们合作，应用能够理解场景更广泛视觉上下文的深度学习模型。

这项工作与谷歌在Google Earth AI背后的更广泛努力相一致，这是我们的一系列地理空间模型和数据集，旨在将行星数据转化为可操作的智能。通过在大规模卫星图像上应用深度学习，我们旨在用针对性的环境监测专业工具来补充更广泛的行星AI计划。

我们使用端到端的视觉Transformer架构（Swin-S Transformer）构建了MAPL-EMIT。许多方法逐像素分析高光谱数据，而MAPL-EMIT则利用现代计算机视觉技术，将完整的光谱与其周围的空间上下文一起处理。通过分析气体在地表如何扩散，该模型能够更好地将真正的、随风飘散的甲烷羽流（从特定源扩散的甲烷气体轨迹）与仅仅具有相似光谱特征的地块区分开来——后者历来是导致甲烷误检的原因。

至关重要的是，这种空间感知能力使模型能够理清高度复杂的场景。在密集的工业区域，多个邻近设施的排放常常合并成一片云团。为了理解这些情景，MAPL-EMIT同时解决三个不同的任务：

基于Transformer的模型需要海量数据进行学习，但全球范围内数百万个真实甲烷排放的标注数据集根本不存在。为了克服这一障碍，我们开发了一个基于物理的模拟框架。我们生成了360万个合成甲烷羽流，并将其直接注入真实的EMIT场景中。通过使用拉格朗日烟团模型（该模型模拟粒子如何在空气中运动和扩散），我们能够重现真实气体排放的混沌和湍流现实。在这些高度逼真的模拟数据上训练，使MAPL-EMIT学会在各种大气和地理条件下识别甲烷。这种合成训练方法提供了几个关键优势：

部署在真实卫星数据上后，MAPL-EMIT展示了可规模化排放测绘的巨大潜力。在与NASA黄金标准的L2B甲烷羽流数据集进行基准测试时，该模型捕获了84%的专家标注羽流，并在约1,100个EMIT影像块中识别出约多50%的疑似羽流，展示了其从背景噪声中分离细微信号的能力。更多细节请参阅论文。

这种更高的灵敏度还使MAPL-EMIT能够可靠地捕获较弱的排放，改善了当前的检测极限。该模型在复杂环境中也被证明是稳健的，成功绘制了全球25个最大排放垃圾填埋场中24个的羽流图。

与许多高灵敏度模型一样，误报仍然是一个持续的挑战，尤其是在复杂地形中。为帮助缓解这一问题，输出结果配有基于物理的羽流置信度（光谱拟合）评分，该评分基于跨步推理中的检测次数进行评估，并使用多种其他属性进行验证，使用户能够根据自身需求在捕获真实羽流的能力和误报风险之间进行过滤和权衡。然而，这并不总是直截了当的，因此我们还根据这些属性为每个羽流标记“较低”或“较高”置信度，使用户可以直接使用这些数据。

MAPL-EMIT展示了一种强大的合作模式，将谷歌的机器学习专业知识与我们NASA JPL合作者的领域知识相结合。我们共同推进了设施尺度天基甲烷观测的全部潜力，为全球社区提供必要的工具，以推动减少温室气体排放的有意义行动。

通过扩展我们在完整EMIT数据目录中检测羽流的能力，MAPL-EMIT为当地利益相关者、研究人员、政策制定者和行业提供了一种强大的新工具，使其能够比以往更快地识别甲烷排放。随着NASA准备发射下一代成像光谱仪（将覆盖率提高30至50倍），稳健且自动化的技术比以往任何时候都更加重要。我们邀请更广泛的科学界探索我们在Earth Engine上新发布的全球羽流数据库，以及用于交互式可视化的Earth Engine应用程序，从Kaggle下载训练好的模型和合成羽流，并访问我们在Github上的推理库。我们希望这些数据为促进有针对性的减排提供坚实的基础，并使我们所有人向实现全球气候目标更近一步。

我们要感谢谷歌和NASA JPL的各位同仁，他们完成了这项工作并促成了此次发布，包括（按字母顺序排列）：亚历克斯·威尔逊、安娜·M·米哈拉克、瓦伦·古尔尚、菲利普·G·布罗德里克、安德鲁·K·索普、克里斯托弗·V·阿斯代尔、布拉克·埃基姆、卡尔·埃尔金、塔尔·盖勒、奥姆里·吉隆、尼塔·戈亚尔、曼西·坎萨尔、罗伊·纳德勒、约翰·普拉特、谢尔盖·沙梅斯、比乔伊·谢蒂、亚伦·索纳本德、迪皮卡·苏基贾、沙哈尔·蒂姆纳特、马克西姆·诺伊曼、安东·赖丘克、弗朗西斯·鲁兰、亚当·R·布兰特、大卫·R·汤普森、罗伯特·O·格林、杰伊·拉津斯基、维沙尔·V·巴特楚和米开朗基罗·康塞尔瓦。

英文来源：

September 1, 2026
Vishal Batchu, Research Engineer, and Michelangelo Conserva, Research Scientist, Google Research
The Methane Analysis and Plume Localization with EMIT model is a deep-learning framework that automates the detection, enhancement quantification, and source estimation of methane plumes globally, turning raw satellite data into scalable climate action.
Methane is a potent greenhouse gas; over a 100-year timeframe, its warming potential is 30 times greater than that of carbon dioxide. In fact, it has driven approximately 25% of human-induced warming since the start of the industrial era. Because methane has a relatively short atmospheric lifespan, promptly reducing these emissions offers a critical "fast-action" pathway to mitigating global temperature rise.
This urgency is reflected in the Global Methane Pledge, where over 125 countries have committed to a 30% emissions reduction by 2030. To hit these targets, we must empower stakeholders to track localized point sources (emissions occurring from a small spatial footprint on the order of a few tens of meters) across the waste, agriculture, and energy sectors. The most cost-effective strategies are to mitigate emissions from oil and gas infrastructure, agricultural facilities, and landfills.
To track these emissions on a global scale, scientists increasingly rely on space-based imaging. A prime example is NASA’s Earth Surface Mineral Dust Source Investigation (EMIT) instrument on the International Space Station. While originally designed to map mineral composition in arid regions, scientists at NASA’s Jet Propulsion Laboratory (JPL) and the broader scientific community have leveraged EMIT's advanced hyperspectral capabilities to detect methane emissions. By recording hundreds of distinct bands of light for every pixel, it allows researchers to "see" the unique chemical fingerprints of these otherwise invisible gases.
Building on these investments, in “Global monitoring of methane point sources using deep learning on hyperspectral radiance measurements from EMIT”, published in Proceedings of the National Academy of Sciences (PNAS), we describe a new approach that turns raw satellite data into scalable mitigation action. Methane Analysis and Plume Localization with EMIT (MAPL-EMIT) is a deep-learning framework that represents a significant step toward automating the detection, enhancement prediction, and source estimation of methane plumes globally. We demonstrate how MAPL-EMIT achieves a high recall of 84% on expert annotated plumes and has a high signal to noise ratio compared to existing matched-filter-based enhancement methods. To support the broader scientific community, we're releasing our global plume database on Earth Engine along with the trained model and synthetic plumes on Kaggle and an inference library on Github.
Measuring methane from space requires balancing three key factors: (1) field of view (spatial coverage/revisit), (2) spatial resolution, and (3) spectral resolution.
Global mappers like TROPOMI were designed to detect small changes in background methane concentrations by integrating high coverage (approximately 2,600 km swath width), coarse spatial resolution (around 5.5 km x 3.5 km), and fine spectral sampling (0.1 nm).
In contrast, point source mappers like EMIT excel at measuring methane emissions at the facility scale. They achieve this by combining moderate coverage (an 80 km wide field of view) with very high spatial resolution (60 meters) and a moderate spectral resolution (7.4 nm spectral sampling), sufficient to capture the chemical signature of methane at a high signal to noise ratio.
However, fully unlocking the potential of this rich data at a global scale presents additional challenges. The Earth's varied landscapes provide a complex backdrop, and some surface materials can masquerade as methane, making the identification of smaller or more diffuse sources particularly challenging. To build on the EMIT team's foundational work and enable high-throughput global mapping, we collaborate with them to apply deep-learning models that can understand the broader visual context of the scene.
This work aligns with Google’s broader effort behind Google Earth AI, our collection of geospatial models and datasets to turn planetary data into actionable intelligence. By applying deep learning to satellite imagery at scale, we aim to complement broader planetary AI initiatives with specialized tools for targeted environmental monitoring.
We built MAPL-EMIT using an end-to-end vision transformer architecture (Swin-S transformer). While many approaches analyze hyperspectral data on a pixel-by-pixel basis, MAPL-EMIT leverages modern computer vision techniques to process the complete spectrum of light alongside its surrounding spatial context. By analyzing how gas disperses across the landscape, the model is better equipped to distinguish a true, wind-blown methane plume (a trail of methane gas dispersing from a specific source) from a patch of ground that simply shares a similar spectral signature, which has historically caused false methane detections.
Crucially, this spatial awareness empowers the model to untangle highly complex scenes. In dense industrial regions, emissions from multiple neighboring facilities often merge into a single cloud. To make sense of these scenarios, MAPL-EMIT simultaneously solves three distinct tasks:
Transformer-based models require massive amounts of data to learn, but a global, labeled dataset of millions of real-world methane emissions simply doesn't exist. To overcome this, we developed a physics-based simulation framework. We created 3.6 million synthetic methane plumes and injected them directly into real EMIT scenes. By using Lagrangian puff models, which simulate how particles move and disperse through the air, we were able to recreate the chaotic, turbulent reality of actual gas emissions. Training on these highly realistic simulations allowed MAPL-EMIT to learn to spot methane under a vast variety of atmospheric and geographic conditions. This synthetic training approach provided several key advantages:
Deployed on real-world satellite data, MAPL-EMIT demonstrates strong potential for scalable emissions mapping. Upon benchmarking against NASA's gold-standard L2B methane plumes dataset, the model captures 84% of expert-annotated plumes and identifies around 50% more plausible plumes across ~1100 EMIT granules, showcasing its ability to separate subtle signals from background noise. See the paper for more details.
This increased sensitivity also allows MAPL-EMIT to reliably capture weaker emissions, improving on current detection limits. The model also proved robust in complex environments, successfully mapping plumes at 24 of the world's 25 top-emitting landfills.
As with many highly sensitive models, false positives remain an ongoing challenge, particularly in complex terrain. To help mitigate this, outputs are paired with physics-based plume confidence (spectral fit) scores, assessed based on the number of detections over strided inference, and evaluated using multiple other properties, enabling users to filter and trade off between the ability to capture real plumes and the risk of false positives as they see fit. However this isn’t always straightforward, which is why we also tag each plume with a “lower” or “higher” confidence based on these properties, allowing users to directly use the data.
MAPL-EMIT showcases a powerful collaboration, bringing together Google's machine learning expertise with the domain knowledge of our collaborators at NASA JPL. Together, we are advancing the full potential of space-based methane observations at the facility scale, providing the global community with the tools necessary to enable meaningful action on reducing greenhouse gas emissions.
By expanding our ability to detect plumes across the full EMIT data catalog, MAPL-EMIT provides a powerful new tool for local stakeholders, researchers, policymakers, and industries to identify methane emissions faster than ever. As NASA prepares to launch the next generation of imaging spectrometers that will increase coverage by a factor of 30–50 times, robust and automated techniques are more important than ever. We invite the broader scientific community to explore our newly released global plume database on Earth Engine along with an Earth Engine App for interactive visualization, download the trained model and synthetic plumes from Kaggle, and access our inference library on Github. We hope this data provides a robust foundation for facilitating targeted mitigation and brings us all one step closer to meeting our global climate goals.
We would like to thank individuals across Google and NASA JPL who carried out this work and made the launch possible, including (in alphabetical order): Alex Wilson, Anna M. Michalak , Varun Gulshan, Philip G. Brodrick, Andrew K. Thorpe, Christopher V. Arsdale, Burak Ekim, Carl Elkin, Tal Geller, Omry Gillon, Nita Goyal, Mansi Kansal, Roy Nadler, John Platt, Sergei Shames, Bijoy Shetty, Aaron Sonabend, Deepika Sukhija, Shahar Timnat, Maxim Neumann, Anton Raichuk, Frances Reuland, Adam R. Brandt, David R. Thompson, Robert O. Green, Jay Radzinski, Vishal V. Batchu, and Michelangelo Conserva.