As extreme temperature events become increasingly frequent, there is a growing need for daily CO2 emissions data to quantify their impacts. However, such data are available only from 2019 onward. To address this gap, we compiled over two million near-real-time observations of electricity generation, traffic activity, natural gas consumption or heating degree days (HDD), and industrial output since 2019, and used these high-frequency data to construct a daily CO2 emissions dataset for 2019–2024. We then applied machine-learning models and degree-day methods to disaggregate non-residential and residential monthly CO2 emissions for 1970–2018 to a daily basis. The historical dataset was then merged with the 2019–2024 dataset to produce a global daily CO2 emissions dataset spanning 1970 to 2024 for 14 countries and regions, covering four sectors: power, industry, residential, and transport (including ground transport and aviation). The resulting long-term dataset will enable robust analyses of extreme-temperature impacts on emissions and enhance the accuracy of chemical transport model inversions of carbon fluxes.
Tao et al. (Tue,) studied this question.