Abstract
With the urgent need to implement EU country pledges and monitor the effectiveness of the Green Deal, robust Monitoring, Reporting, and Verification (MRV) tools are essential to track sectoral emission changes. Current official inventories provide annual estimates with a lag of over a year, failing to capture rapid variations caused by recent shocks such as COVID-19 lockdowns, economic rebounds, and the energy crisis. This paper presents Carbon Monitor Europe (CM-EU), a near-real-time dataset of daily fossil fuel and cement emissions from January 2019 through December 2021 for 27 EU countries and the United Kingdom. By estimating emissions across six sectors—power, industry, ground transportation, domestic aviation, international aviation, and residential—this dataset improves the timeliness and temporal resolution of European emissions data to better inform policymakers and the public.
Introduction: The Latency Challenge
Traditional greenhouse gas inventories are the "gold standard" for climate policy, yet they suffer from a significant latency problem. Official reports to the UNFCCC or Eurostat typically lag reality by one to two years. While sufficient for long-term planning, this delay is inadequate for monitoring rapid socio-economic shifts. Recent global shocks—specifically the COVID-19 pandemic and the energy volatility sparked by the war in Ukraine—demonstrate the need for faster feedback loops. Without daily or monthly data, it is difficult to discern how specific sectors, such as aviation or residential heating, respond to crises or policy interventions in real-time.
Methodology
The Carbon Monitor Europe (CM-EU) system estimates daily emissions by combining a static baseline with dynamic, high-frequency activity data. The core calculation follows the formula:
$$Emis = \sum AD \times EF$$
Where $AD$ represents activity data and $EF$ represents the emission factor. The researchers utilized the 2019 annual emissions from the Emissions Database for Global Atmospheric Research (EDGAR) as the baseline. To "nowcast" emissions before official fuel consumption statistics are available, the system employs several sector-specific proxies:
- Power: Hourly electricity generation data sourced from ENTSO-E and the Balancing Mechanism Reporting Service (BMRS).
- Industry: Monthly industrial production indices (IPI), disaggregated to a daily scale using power generation as a proxy.
- Ground Transportation: A novel approach using TomTom congestion levels. A sigmoid regression model, calibrated with vehicle counts in Paris, converts "extra travel time" into actual vehicle activity.
- Aviation: Individual commercial flight data from FlightRadar24 to calculate actual distance flown per day.
- Residential: Daily population-weighted "heating degree days" (HDD) derived from ERA5 air temperature data to estimate heating demand.
Key Insights and Sectoral Trends
The "V-Shaped" Recovery
The CM-EU data captures a dramatic 28.13% drop in emissions during the peak lockdown months of April and May 2020. This was followed by an 8.82% rebound in 2021 as economies reopened. Despite this rebound, 2021 emissions remained 2.71% lower than 2019 levels.
Sectoral Divergence
While most sectors began returning to 2019 levels by late 2021, international aviation remained a significant outlier, staying 48.3% below pre-pandemic levels. In the residential sector, the study found that temperature fluctuations (weather) dominate emission changes more than behavioral shifts, except during the strictest confinement periods in countries like France and Italy.
National Variance
Emission trends varied significantly across borders. Germany saw the largest absolute decrease in 2020 (-65.5 Mt CO$_2$). Estonia experienced the most volatile swing, with a 30% decrease in 2020 followed by a 34% rebound in 2021. These variations highlight the importance of country-level monitoring to understand the efficacy of local climate policies.
Validation and Uncertainty
The authors estimate the overall uncertainty of the daily dataset at ±13.6%. While higher than the uncertainty found in annual inventories, this margin is considered acceptable for near-real-time monitoring. The results were validated against monthly and quarterly estimates from national agencies (such as Citepa in France and CBS in the Netherlands) and satellite-derived data from private firms like Kayrros. The high correlation ($R^2 \approx 0.79$ for ground transport) demonstrates that non-traditional data, such as GPS-based congestion, can accurately proxy fossil fuel combustion.
Limitations and Future Outlook
Several constraints remain in the current model. The system assumes constant emission factors, meaning it does not yet fully account for the gradual greening of vehicle fleets or shifts in industrial efficiency between 2019 and 2021. Additionally, the ground transport model relies heavily on city-level congestion, which may not perfectly capture rural driving or long-haul trucking trends. Future iterations of Carbon Monitor Europe aim to integrate satellite-based atmospheric observations to further refine these bottom-up proxy estimates and provide an even more granular view of Europe's path toward carbon neutrality.
Summary updated