The study investigates the unprecedented variability in CO₂ emissions from fossil fuel combustion and cement production during the COVID-19 pandemic. Existing methods for estimating fossil CO₂ emissions typically use activity data and emission factors, with relatively low uncertainty compared to other pollutants. The pandemic's impact, however, necessitated new approaches due to the lack of low-latency, direct activity data at a global scale. Previous research attempted to estimate CO₂ reduction using proxies like government policies or mobility data. This study aims to provide a high-resolution, near real-time global dataset of daily CO₂ emissions for 2020. The purpose is to analyze the pandemic’s impact on emissions and understand the underlying drivers behind the observed changes, improving our ability to monitor and respond to future emissions changes. The importance lies in understanding the potential for rapid emission reductions and the implications for climate change mitigation strategies.

Existing literature highlighted the significant decline in air pollution and CO₂ emissions during the initial COVID-19 lockdowns. Studies utilized various proxies, such as government lockdown stringency, mobility data, and energy consumption data, to estimate the impact on greenhouse gas emissions. However, many studies had limitations in spatial or temporal resolution, lacking the daily global-scale dataset needed for a comprehensive understanding of the dynamics of CO2 emission changes. This study builds upon previous work by providing a more comprehensive and high-resolution daily dataset of global CO2 emissions, using a near real-time monitoring system, allowing for a more accurate assessment of the pandemic's impact and the complex interplay of various factors influencing emission reductions.

The study utilized a daily CO₂ emissions dataset for 2020, derived from inventories and near real-time activity data from the Carbon Monitor project. This dataset incorporates information from various sectors including power generation (29 countries), industry (73 countries), ground transportation (406 cities), aviation and maritime transportation, and residential fuel use (206 countries). Daily emissions (Emis) were estimated using the formula: Emis = ΣΣΣ AD × EF, where AD represents activity data (e.g., energy consumption) and EF represents emission factors (CO₂ emissions per unit of activity). The study assumed that emissions factors remained constant between 2019 and 2020, thus daily emissions were proportional to daily activity data. The methodology involved disaggregating annual emissions from 2019 to daily levels and then calculating daily 2020 emissions based on daily activity changes. Specific methods varied by sector. For instance, the power sector used daily national thermal production data, while the industry sector relied on industrial production data or indices. Ground transportation utilized a sigmoid model linking traffic congestion levels to car counts. Aviation and international shipping emissions were estimated using flight distance data and the Baltic Dry Index respectively. The methodology also included a baseline simulation for 2020, accounting for historical emission trends. Uncertainty estimations were performed sector-by-sector, combining uncertainties from activity data and emission factors. A detailed uncertainty analysis was provided in the supplementary material.

The study found a global decrease of 6.3% (2,232 MtCO₂) in CO₂ emissions in 2020 compared to 2019. The largest weekly decline was 17%, occurring during week 15 (April 6-12). Emissions gradually recovered from late April, with further reductions in late 2020 due to subsequent lockdown waves. The reduction was largest in April (16.3%), declining only 0.5% by December compared to December 2019. Nationally, emissions fell by 9.7% in Brazil, 9.5% in the US, 8.0% in Russia, 7.9% in India, 7.3% in the EU and UK, and 4.7% in Japan, while China saw a slight 0.9% increase. Sectoral analysis revealed the largest contribution to the global decrease came from ground transportation (10.9%, 709 MtCO₂), followed by international bunkers (36.9%, 503 MtCO₂), power (4.1%, 554 MtCO₂), and industry (2.6%, 265 MtCO₂). Correlation analysis showed a strong relationship between CO₂ reduction and COVID-19 related factors (daily deaths, government response stringency, human mobility, and energy demand) during the initial lockdown period (March-May). However, this correlation weakened during the later waves (October-December), suggesting that the response to later waves involved less stringent measures. The 2020 decline was significantly larger than previous emission drops caused by the 2008 financial crisis (380 MtCO₂) and the end of World War II (814 MtCO₂).

The study's findings demonstrate the significant impact of COVID-19 lockdowns on global CO₂ emissions, highlighting the potential for rapid reductions. The substantial decrease in 2020, comparable to the reductions needed to meet 1.5°C climate targets, underscores the need for aggressive climate action. The disparity in emission changes between countries and sectors reflects the varying economic structures and responses to the pandemic. The weakening correlation between CO₂ reduction and pandemic-related factors in later waves points to a complex interplay between public health measures, economic recovery, and emission reduction strategies. The study's findings emphasize the need for sustainable economic recovery strategies to avoid a rebound in emissions and promote a green transition. This requires integrating climate objectives into economic stimulus packages and strengthening international cooperation on climate change.

This study presents a high-resolution dataset revealing a substantial global CO₂ emission reduction in 2020 due to COVID-19. The results highlight the potential for rapid emissions reductions and reinforce the urgent need for a swift and sustained energy transition. Future research should focus on monitoring post-pandemic emission trajectories, investigating the long-term impacts of policy responses, and exploring strategies to decouple economic growth from carbon emissions. Improving real-time global-scale data collection and analytical tools is essential for more effective climate change mitigation.

The study's reliance on proxy data for some sectors and regions introduces uncertainty into the estimations. The assumption of constant emission factors might oversimplify the actual variations in emission intensities. The analysis focuses primarily on the first year of the pandemic and might not fully capture the long-term effects of the pandemic on emissions patterns and the recovery period. Further investigation into the interplay of different economic sectors and their individual contributions to both the initial emission reductions and the subsequent rebound is necessary to inform climate policy.