TL;DR
This paper introduces the "Carbon Monitor," a novel near-real-time daily dataset of global CO2 emissions from fossil fuel combustion and cement production. This dataset provides unprecedented temporal resolution, allowing for the tracking of emissions variations influenced by daily activities, holidays, and major events like the COVID-19 pandemic, and revealed a significant global CO2 emission decline in early 2020.
What Problem Does This Paper Address?
Accurate and timely tracking of CO2 emissions is crucial for understanding climate change drivers and evaluating the effectiveness of mitigation policies. Traditional CO2 emission inventories are typically reported annually and often have a significant time lag (1-2 years or more) due to the time required for data collection, processing, and official reporting. This delay makes it difficult to monitor the immediate impacts of human activities, policy changes, or global events on emissions. The authors address this gap by developing a near-real-time daily dataset that can be frequently updated, providing a much more dynamic view of global CO2 emissions.
Methods
The Carbon Monitor dataset is constructed by integrating diverse, near-real-time activity data from various sectors. The core methodology involves:
- Power Sector: Utilizing hourly to daily electrical power generation data from 29 countries.
- Industrial Production: Employing monthly production data and indices for 62 countries/regions, focusing on cement, steel, and other energy-intensive industries.
- Road Transportation: Leveraging daily mobility data and indices from 416 cities worldwide.
- Aviation and Maritime Transport: Using individual flight location data and monthly data.
- Residential and Commercial Buildings: Incorporating monthly fuel consumption data, adjusted for daily air temperature, from 206 countries.
The approach assumes that daily emission variations are primarily driven by these activity data, while emission factors remain relatively constant over short timescales. For sectors with less granular data (e.g., industry, transport, residential), monthly or annual totals are downscaled to daily estimates using proxy data and established relationships.
Datasets
The construction of the Carbon Monitor dataset relies on a variety of publicly available and proprietary datasets:
- Activity Data:
- Hourly to daily electrical power generation data (e.g., from China's National Grid, India's POSOCO, US EIA's Hourly Electric Grid Monitor, EU's ENTSO-E).
- Monthly industrial production data and indices (e.g., from World Steel Association, National Bureau of Statistics of China, Ministry of Statistics and Programme Implementation of India).
- Daily mobility data and indices (e.g., TomTom congestion level data for 416 cities).
- Individual flight location data (FlightRadar24).
- Monthly fuel consumption data.
- ERA5 reanalysis of 2-meters air temperature data.
- Existing Emission Inventories (for baseline and scaling):
- EDGAR v4.3.2 and v5.0 (Emissions Database for Global Atmospheric Research).
- Global Carbon Budget.
- BP Statistical Review of World Energy.
- International Energy Agency (IEA) CO2 Emissions from Fuel Combustion.
- Third IMO GHG Study.
Key Insights and Ideas
- Dynamic Emissions Tracking: The most significant insight is the demonstration that near-real-time daily CO2 emissions can be reliably estimated by integrating diverse, high-frequency activity data. This moves beyond the limitations of traditional annual, lagged inventories.
- COVID-19 Impact Quantification: The dataset clearly captures the dramatic impact of the COVID-19 pandemic on global CO2 emissions, showing a significant decline in early 2020. This provides a quantitative measure of the pandemic's effect on human activity and emissions.
- Economic Recovery Signals: The dataset detected a re-growth in CO2 emissions by late April 2020, directly linked to China's economic recovery and the easing of lockdowns in other regions, highlighting the dataset's sensitivity to economic activity.
- Daily and Weekly Emission Patterns: The dataset reveals daily and weekly emission variations driven by workdays, holidays, and other cyclical human activities, offering a granular understanding of emission drivers.
- Sectoral Contributions to Variability: By breaking down emissions by sector, the dataset allows for the analysis of which sectors are most responsive to daily, weekly, and event-driven changes.
Key Results
- Global CO2 Emission Decline: The Carbon Monitor dataset showed a 7.8% decline in global CO2 emissions from January 1st to April 30th, 2020, when compared to the same period in 2019.
- Re-growth Detected: A regrowth of CO2 emissions was detected by late April 2020.
- Data Coverage: The dataset provides daily emissions for 9 countries/regions (China, India, US, EU27&UK, Russia, Japan, Brazil, ROW, and Globe) and covers the period from January 1st, 2019, to April 30th, 2020, with potential for frequent updates.
- Sectoral Breakdown: Daily emissions are provided for power, ground transport, industry, residential, aviation, and international shipping sectors.
Limitations and Future Work
- Data Availability and Quality: The accuracy of the dataset is inherently dependent on the availability and quality of the underlying activity data. For some regions or sectors, proxy data or downscaling methods were necessary, introducing potential uncertainties.
- Emission Factor Assumptions: The assumption that emission factors remain constant over short periods is a simplification. While justified for short-term variations, longer-term policy or technological shifts could alter these factors.
- Uncertainty Estimates: The paper provides detailed uncertainty estimates for each sector, with the overall uncertainty for the dataset estimated at ±6.8%. However, the authors acknowledge that uncertainties in aviation emissions are particularly difficult to assess due to data limitations.
- Future Work: The authors suggest that this daily updated CO2 emission dataset offers a range of opportunities for related scientific research and policy making. Future work could focus on expanding the temporal and spatial resolution, incorporating more detailed activity data, and refining emission factor estimations. The potential for frequent updates also allows for continuous monitoring of emission trends.
Summary updated