A near-real-time monthly dataset of CO₂, CH₄, and N₂O emissions from fossil fuel combustion

Abstract. Fuel-combustion greenhouse gas emissions account for about 67 % of the total anthropogenic emissions, and thus reliable estimates lay the foundation for the Global Stocktake, the evaluation of nationally determined contributions (NDCs), and the quantification of remaining carbon budgets. However, to fully understand the current status of greenhouse gas emissions and to formulate policies for effective control of greenhouse gas emissions, a dataset that provides a comprehensive, up-to-date, and sector-specific record of greenhouse gases is needed. This paper presents a global dataset covering CO₂, CH₄, and N₂O from 1960 to 2025. The dataset reports emissions classified by country, sector, fuel, and gas at monthly resolution, covering emissions from fuel combustion. It covers 153 reporting entities, which consist of 150 individually reported countries, one residual entity covering 48 remaining countries and territories, international aviation, and international shipping. Historical emissions from 1960–2023 are reconstructed using emission factors and statistical data of energy consumption derived from International Energy Agency (IEA) energy activity dataset. For the latest two years without available IEA activity data, we constructed a framework for predicting emissions based on historical emissions data and monthly activity indicators from CEIC. We evaluate six individual prediction methods and their ensembles using historical emissions from 2019 to 2023 and select the best prediction method for each specific country, sector, and gas combination. The best method is applied to estimate annual emissions for 2024–2025. In addition, we distribute the annual emission values to monthly values by using monthly activity patterns. The dataset demonstrates that CO₂ and N₂O show overall increasing trends with CO₂ reaching 37.7 Gt CO₂ yr^-1 in 2025 and N₂O reaching 1,057 kt N₂O yr^-1 in 2025. In contrast, CH₄ reached its peak around 1990 and then declined and remained close to 11 Mt CH₄ yr^-1. Emissions of all three gases also follow a seasonal pattern with emissions being highest in winter and lowest in summer. For validation purposes, we compare our dataset with existing datasets and find good overall agreement. This dataset supports carbon budget tracking, NDC and Global Stocktake assessment, and near-real-time monitoring of sectoral greenhouse gas emissions.