TL;DR
The COVID-19 pandemic caused a significant, unprecedented drop in air pollutant and greenhouse gas emissions. This study compares near-real-time emission estimates derived from novel data sources (like mobility data) with official, lagged inventories for Europe. While overall annual trends align, significant discrepancies emerge at the sector and country level, highlighting the need for caution and further validation of these emerging near-real-time methods.
What Problem Does This Paper Address?
Official national emission inventories, crucial for tracking progress towards climate and air quality goals, have a significant time lag (typically 2 years). This delay makes them unsuitable for timely monitoring and modeling, especially during major, unexpected events like the COVID-19 pandemic. The pandemic's drastic reduction in human activity created an urgent need for "near-real-time" emission estimates. This paper addresses the challenge of evaluating the reliability and comparability of these emerging near-real-time emission datasets against established official inventories, particularly for the European region. It aims to understand the discrepancies and provide insights for their future use in modeling and policy.
Methods
The study employs an intercomparison approach, comparing emission changes from four different near-real-time datasets with official emission inventories. The near-real-time datasets utilize various proxies, including:
- Mobility data: Derived from smartphones (e.g., Google COVID-19 Community Mobility Reports) and navigation apps (e.g., TomTom).
- Energy data: Electricity demand and generation statistics from sources like ENTSO-E.
- Industrial production indices: Data from Eurostat.
- Traffic counts: Traditional traffic data.
- Flight and shipping data: Information from aviation authorities and maritime trade databases.
These proxies are used to derive "adjustment factors" that quantify the reduction in emissions compared to a business-as-usual scenario. The study analyzes these changes at annual, quarterly, and monthly resolutions, focusing on various air pollutants (NOx, NMVOC, CO, SO2, NH3, PM2.5) and greenhouse gases (CO2, CH4) across different sectors and countries in the EU27 + UK.
Datasets
The study utilizes the following datasets:
- Official Emission Inventories:
- Air Pollutants (APs): Reported under the Convention on Long-Range Transboundary Air Pollution (CLRTAP) and managed by the EMEP Centre on Emission Inventories and Projections (emep_ceip).
- Greenhouse Gases (GHGs): Reported under the United Nations Framework Convention on Climate Change (UNFCCC) (unfccc).
- Near-Real-Time Emission Estimates:
- guevaraetal: Copernicus Atmosphere Monitoring Service (CAMS) COVID-19 emission adjustment factors (European coverage).
- doumbiaetal: COvid-19 adjustment Factors fOR eMissions (CONFORM) dataset (global coverage).
- forsteretal: COVID-19 emission estimates developed by Forster et al. (global coverage).
- liuetal: Carbon Monitor initiative (European version available for CO2 emissions).
The study also references data from sources like Google COVID-19 Community Mobility Reports, TomTom, Flightradar24, UN COMTRADE, ENTSO-E, Eurostat, and national inventory submissions (e.g., CITEPA for France, BEIS for the UK, CBS for the Netherlands) for specific comparisons.
Key Insights and Ideas
- The "Great Equalizer" of COVID-19: The pandemic provided a unique, albeit tragic, opportunity to observe the impact of drastic human activity reduction on emissions. This event underscored the limitations of traditional, lagged emission reporting.
- Near-Real-Time Data is Promising, But Not Perfect: Novel data sources like mobility data can offer valuable near-real-time insights, but they are not a silver bullet. Their effectiveness varies significantly by sector and country.
- Sectoral and Country-Level Discrepancies are Key: While overall European emission drops for pollutants like NOx and CO2 showed general agreement between near-real-time and official estimates, the devil is in the details. Significant differences emerge when drilling down into specific sectors (e.g., manufacturing industry, road transport) and individual countries.
- Non-COVID Factors Still Matter: For sectors like residential combustion, shipping, and public energy, changes in emissions between 2019 and 2020 were often driven by pre-existing trends like meteorology (warmer winters), decarbonization policies (coal phase-outs), and regulatory changes (Global Sulphur Cap), rather than solely by pandemic restrictions.
- Data Heterogeneity is a Challenge: The study highlights how different countries report emissions for certain activities (like solvent use from pandemic products) using varying levels of detail, leading to inconsistencies even within official reporting.
Key Results
- Overall Trends Align: For most species, annual total emission changes in 2020 reported by official and near-real-time estimates were "fairly in line." NOx and fossil fuel CO2 showed the largest reductions across Europe.
- Significant Sectoral/Country Discrepancies:
- Manufacturing industry NOx: Relative changes ranged from -21.4% to -5.4% between official and non-official estimates.
- Road transport CO2: Relative changes ranged from -29.3% to -5.6%.
- Monthly/Quarterly Variations: Discrepancies were substantial in the temporal distribution of emission drops. For total NOx in April 2020 (peak restrictions), differences between near-real-time inventories reached up to a factor of 1.5.
- Specific Sector Findings:
- Public Power: Near-real-time estimates generally showed lower SO2 reductions than official ones, often due to pre-existing decarbonization trends.
- Residential Combustion: Changes were largely driven by meteorology, not COVID-19.
- Manufacturing Industry: Differences were notable, with some near-real-time estimates showing much larger NMVOC reductions than official data.
- Road Transport: guevaraetal and liuetal showed better agreement with official estimates than doumbiaetal and forsteretal.
- Aviation: Showed the largest drops across all estimates, with relatively consistent results.
- NMVOC from Pandemic Products: Heterogeneous reporting in official inventories makes it difficult to capture the impact of increased use of hand sanitizers, for example.
Limitations and Future Work
- No Absolute Ground Truth: The study acknowledges that official emission inventories themselves have uncertainties and are subject to revisions. Therefore, direct cross-verification of near-real-time data against official data doesn't establish absolute accuracy but rather provides a measure of consistency and highlights discrepancies.
- Data Availability and Consistency: Obtaining accurate and consistent local information for near-real-time emission proxies remains a challenge, especially at a global level. While European data (like guevaraetal) showed better alignment, global datasets exhibited larger discrepancies.
- Limited Scope of Proxies: While mobility and energy data are emerging, near-real-time monitoring for sectors like solvent use, residential combustion, and agricultural activities remains scarce.
- Future Work: The authors suggest that combining new data sources (mobility, satellite data) with traditional statistics is crucial. They also emphasize the need for improved emission science that integrates near-real-time data into modeling and forecasting applications, moving beyond traditional inventory-based approaches. The increasing frequency of disruptive events (like the war in Ukraine) further underscores the importance of this shift.
Summary updated