Estimating greenhouse gas emissions using the next generation of satellite observations

我们的团队已经证明，基于使用气象局NAME模型和贝叶斯方法模拟大气气体输送，可以从上一代卫星观测推断出国家尺度上的温室气体排放（例如Ganesan等人，2017年）。然而，由于所涉及的数据集的规模，将现有方法扩展到新一代卫星数据具有挑战性。在这里，我们建议使用新的数据科学方法来探索如何从大型大气浓度数据集中推断温室气体通量，并利用这些信息来解决一系列紧迫的挑战，例如：a)是什么推动了当前大气甲烷的快速增长？B)陆地碳汇如何随时间变化？c)国家温室气体排放报告是否可靠？

Our team has shown that GHG emissions can be inferred at national scales from the previous generation of satellite observations, based on simulations of atmospheric gas transport using the Met Office NAME model and Bayesian methods (e.g. Ganesan, et al., 2017). However, the extension of current approaches to the new generation of satellite data are challenging due to the size of datasets involved. Here, we propose to use novel data science approaches to explore how to infer GHG fluxes from large atmospheric concentration datasets and use this information to tackle a range of pressing challenges, such as: a) what is driving the current rapid growth in atmospheric methane? b) how is the terrestrial carbon sink changing with time? c) are national GHG emissions reports reliable?