Notes on COVID-19 Lockdowns: Emissions, Air Chemistry, and Policy Implications

Overview

• The COVID-19 lockdowns began in China with a strict lockdown in late January, keeping most people at home for about $3 ext{ weeks}$. Direct emissions of air pollution fell sharply at a rate and scale never observed before.
• Researchers used this period as an unintended large-scale experiment: rooftop spectrometers and sensor networks were deployed to capture how skies changed as human activity, especially car traffic, dropped rapidly.
• Key researchers emphasized the scientific value of the moment: "The pandemic is a tragedy, of course—but the science we can do is important" (Joost de Gouw).
• The immediate effect of reduced traffic was lower emissions of $CO<em>2$ (a greenhouse gas) and $NO</em>2$ (which participates in atmospheric chemistry that forms ozone and fine particulates, harming health).
• The situation yielded striking real-world observations such as clearer skies in multiple cities (New Delhi, Los Angeles, etc.) during early lockdowns.
• The overall message: lockdowns show that rapid changes in emissions are observable, measurable, and relevant for policy evaluation, even though the global atmosphere integrates all sources over long timescales.

Regional patterns and timing during lockdowns

China (early lockdown period)

• The average exposure and NO2 signals over eastern China dropped sharply during late January–February; NOAA/CAMS data indicated NO2 fell by about $65\%$ in that period compared with the same period in 2019.
• The reduction was largely due to a sharp decrease in traffic and other combustion sources.
• A notable finding was a nonlinear response in air chemistry: when NOx emissions fall, ozone can actually rise in some areas due to complex photochemical interactions.
• Urban NOx chemistry also affected PM formation, with nuanced impacts on particulate matter such as $PM_{2.5}$ (see PM section).
• There were reports of unusually intense ozone spikes in some urban areas (e.g., northern China, Los Angeles) during strict lockdown periods, highlighting the nonlinear NOx–ozone relationship.
• Wuhan saw particularly large NO2 reductions, with some studies indicating drops as high as $93\%$ in NO2, but PM2.5 responses were more muted and variable.

New Zealand

• Lockdown constraints led to a fall in traffic volume of about $80\%$ at the height of the lockdown.
• Fossil fuel CO2 emissions were inferred to have fallen at roughly the same rate, about $80\%$, according to work by Jocelyn Turnbull (GNS Science).
• Roadside CO2 sensor data in Wellington are still being analyzed to confirm these trends.

San Francisco Bay Area (BEACON network)

• In the SF Bay Area, early weeks of strict orders showed traffic down by about $45\%$.
• BEACON data indicated total CO2 emissions fell by about $25\%$ during that first six-week period relative to the preceding six weeks; results were under journal review at the time.
• The BEACON network consists of rooftop CO2 sensors, with one highlighted as an example in reporting: "One node in a network of rooftop CO2 sensors".

Global spread and qualitative observations

• Similar declines in NO2 and CO2 were observed worldwide during lockdowns (Delhi, Los Angeles, northern Italy, etc.), though the magnitude varied by region and local sources.
• The lockdown period offered a relatively clean window to observe how emission reductions translate into atmospheric signals, contrasted with the longer-term persistence of some pollutants.

Emissions, chemistry, and nonlinearity in the atmosphere

Primary and secondary pollutants

• Primary pollutants (e.g., $NO_2$, sulfur compounds) dropped significantly when traffic and industrial activity declined.
• However, secondary pollutants (e.g., $O<em>3$, $PM</em>{2.5}$) showed nonlinear responses to emission changes, complicating interpretation.
• The relationship among NO2, ozone, and PM2.5 is nonlinear and region-dependent, driven by photochemistry and radical chemistry.

Nonlinear NOx–ozone chemistry

• A key nonlinearity: NOx (NO + $NO_2$) can both promote and suppress ozone formation depending on the regime.
• Abundant NOx can scavenge hydroxyl (OH) radicals, reducing the formation of $O_3$ from volatile organic compounds (VOCs). Conversely, at certain NOx/VOC balances, NOx can facilitate ozone production.
• When NOx is reduced in urban areas, ozone can rebound because titration of $O<em>3$ by NO decreases, allowing $O</em>3$ to accumulate under sunlight-driven chemistry.
• This nonlinear behavior helps explain observed ozone spikes during some strict lockdown periods in cities like northern China and Los Angeles.

NOx, NO, ozone, and PM2.5 interactions

• NO2 can react with other gases to form particulates and influence secondary aerosol formation; reductions in NO2 do not automatically translate to proportional reductions in $PM_{2.5}$.
• In Wuhan, NO2 dropped dramatically (up to $93\%$), but PM2.5 didn't fall in a linear or uniform way; hotspots and regional transport continued to influence particulate matter levels.
• The nonlinear chemistry means that reducing urban NO2 alone is not a guaranteed path to clean air; broader regional and sectoral emissions reductions are often required.

PM2.5 observations and regional sources

• PM2.5 declined only moderately despite large NOx and NO reductions, with some spikes linked to biomass burning during wheat harvest in southeast Delhi.
• There were notable observations about the composition of PM2.5 changing little in some cities (e.g., Delhi), suggesting that a large fraction of PM2.5 precursors originate from sources outside the city (regional and agricultural contributions).
• IIT Kanpur researchers and collaborators monitored PM2.5 particles of size $2.5\,\mu\text{m}$ and smaller ($PM_{2.5}$) and highlighted that reductions in primary pollutants did not automatically translate to equivalent changes in PM2.5 composition.

Implications for air-quality goals

• A major takeaway for regulators is that simply reducing urban traffic emissions is not sufficient; policies must address broader regional emissions and non-traffic sources.
• For areas like Delhi, rural heating and cooking systems, as well as agricultural emissions, are significant contributors to $PM_{2.5}$ and require attention for meaningful air-quality improvements.

Data sources, methods, and implications for policy

• Data collection methods included rooftop spectrometers, sensor networks (e.g., BEACON) for CO2, and satellite-based monitoring (e.g., CAMS—Copernicus Atmosphere Monitoring Service) for NO2 and other pollutants.
• The combination of real-world data and models strengthened the case that emission reductions can be monitored and quantified, which is important for evaluating regulatory effectiveness.
• The NO2 reductions observed during lockdowns provided a practical demonstration of how transportation and combustion sources drive urban air quality and how policy interventions might achieve similar results.

Key researchers and insights captured in the transcript

• Joost de Gouw (Chemist, Cooperative Institute for Research in Environmental Sciences): emphasizes the scientific value of the observed emission changes during the pandemic.
• Nga Lee Ng (Georgia Institute of Technology): describes the lockdown as a "weird opportunity" to run an experiment of interest that had long been dreamed of.
• Vincent-Henri Peuch (Copernicus Atmosphere Monitoring Service, CAMS): notes that NO2 is predominantly from traffic (about $45\%$ to $50\%$ of NO2 emissions) and documents observed NO2 declines.
• Yuan Wang (NASA JPL): highlights nonlinear chemistry, including ozone behavior under reduced NOx conditions, and notes significant NO2 reductions (e.g., in Wuhan) with complex PM responses.
• Rima Habre (USC): emphasizes that the relationship between emissions reductions and pollutant levels is not linear; ozone can rise in some contexts when NOx declines.
• Sachchida Nand Tripathi (IIT Kanpur) and collaborators: monitor PM2.5 and observe that changes in PM2.5 composition were modest during lockdowns, suggesting regional sources beyond urban centers.
• Joshua Apte (UC Berkeley): stresses regulatory implications, arguing for broader emission reductions beyond traffic in cities like Delhi and for addressing regional and agricultural sources.
• Cohen (researcher referenced for policy impact): suggests the lockdown data could be persuasive to policymakers, by providing a real-world test of how emissions reductions translate to air-quality improvements.

Quantitative takeaways (summarized)

• Lockdown duration cited: $ext{about } 3 ext{ weeks}$ in China at the start; other regions experienced multiple weeks of restrictions.
• Traffic reductions observed: China (not quantified globally here), New Zealand: $80\%$ drop in traffic; San Francisco Bay Area: $45\%$ decline in traffic over the first 6 weeks; CO2 emissions in SF Bay Area fell by about $25\%$ during that period.
• NO2 reductions: eastern China dropped by approximately $65\%$ (compared to 2019) in the early lockdown period.
• Urban NO2 reductions: in Wuhan, NO2 declined by as much as $93\%$.
• Ozone observations: documented ozone spikes during some lockdown periods due to nonlinear NOx–ozone chemistry; not a universal trend.
• PM2.5 observations: primary pollutants (NOx, sulfates) dropped substantially, but $PM_{2.5}$ decreased only modestly; some regional spikes occurred due to biomass burning during agricultural periods (e.g., wheat harvest in SE Delhi).
• Regional vs urban sources: many regions saw NO2 and CO2 declines tied to traffic reductions, but PM2.5 and ozone formed in broader atmospheric chemistry were influenced by multiple sources beyond urban traffic, including rural heating, cooking, and agricultural emissions.

Connections to broader principles and real-world relevance

• The lockdowns provided a natural experiment illustrating how quickly atmospheric composition responds to emission reductions and how nonlinear chemical processes can modify expected outcomes.
• The findings support the importance of long-range transport and regional coordination when designing air-quality policies, not just city-level controls.
• The results underscore the value of robust monitoring networks (rooftop CO2 sensors, spectrometers, and satellite data) for evaluating emissions regulations and for informing policy decisions.
• Ethical and practical implications: while the pandemic is a tragedy, the data generated offer actionable insights for improving public health through emissions reductions and cleaner urban environments.

Takeaways for exam preparation

• Understand the distinction between primary emissions reductions (e.g., $NO<em>2$, $CO</em>2$ from traffic) and secondary pollutant responses (e.g., $O<em>3$, $PM</em>{2.5}$) and why they may diverge due to nonlinear atmospheric chemistry.
• Be able to explain why NOx reductions can lead to ozone increases in some contexts and ozone reductions in others, depending on the NOx-VOC regime.
• Recognize that PM2.5 responds to multiple sources (urban traffic, regional transport, biomass burning, agricultural emissions) and that urban reductions may only partially reduce $PM_{2.5}$ if regional sources remain high.
• Acknowledge the value of real-world data (rooftop sensors, BEACON, CAMS) in validating models and informing policy.
• Remember key numerical benchmarks from the transcript: $3\text{ weeks}$ (early lockdown), $80\%$ traffic/CO2 reduction in NZ, $45\%$ traffic reduction in SF Bay Area, $25\%$ CO2 reduction in SF Bay Area over 6 weeks, NO2 reductions of $65\%$ in eastern China and up to $93\%$ in Wuhan.

References to figures and terms

• CO2: $CO_2$, greenhouse gas.
• NO2: $NO_2$, key pollutant linked to ozone and fine particulates.
• Ozone: $O_3$, secondary pollutant formed via photochemical reactions.
• PM2.5: $PM_{2.5}$, fine particulate matter, with size 2.5 micrometers and smaller.
• Regional networks: BEACON (Berkeley Environmental Air-quality and CO2 Network), CAMS (Copernicus Atmosphere Monitoring Service).
• Key institutions: JPL, USC, IIT Kanpur, UC Berkeley, Georgia Tech, CIERES, GNS Science.

Summary statement

• The lockdown period demonstrated that rapid, substantial reductions in traffic and fossil-fuel combustion lead to observable changes in atmospheric composition, especially for $NO<em>2$ and $CO</em>2$, but the overall air-quality response is governed by nonlinear chemistry and regional emissions. This has direct implications for designing effective, comprehensive air-quality policies that address both urban and regional sources, including rural heating, cooking, and agricultural emissions.