Quantcast

Air Pollution Dropped During Pandemic Lockdowns

Washington, D.C., New York and Boston Showed Biggest Air Quality Improvements


Sally Strong
May 30, 2022

Washington, D.C., New York and Boston Showed Biggest Air Quality Improvements

As vehicle traffic lightened and industry slowed during the COVID-19 stay-at-home period in 2020, a University of Houston study by the air quality forecasting group led by Yunsoo Choi, associate professor in the Department of Earth and Atmospheric Sciences, estimates levels of potentially-dangerous air pollutants simultaneously decreased in major cities across the country.

Photo of clear skyAir became cleaner during 2020 pandemic stay-at-home orders in 10 of the 11 U.S. cities studied by University of Houston air quality researchers. The team looked specifically at quantities of the pollutant PM2.5, which presents potential health risks that include heart disease and respiratory problems. photo credit: Getty

All but one of the 11 U.S. cities examined experienced reduced levels of the pollutant PM2.5 — tiny particles or droplets in the air that are 2.5 microns or less in diameter. The negative health impacts of increased exposure to the pollutant include cardiovascular diseases, respiratory-related illnesses and similar conditions.

“We hope our study’s findings will provide insight useful to medical researchers and perhaps increase awareness of the need to explore cleaner energy alternatives,” Choi said.

Photo of researchers Yunsoo Choi and Masoud GhahremanlooFrom left: Yunsoo Choi, University of Houston associate professor of Earth and Atmospheric Sciences, leads the air quality research team that documented a potentially dangerous air pollutant decreased during 2020’s COVID-19 stay-at-home period. Masoud Ghaharemanloo, doctoral student and member of the research team, is the lead author of an article that details the recent findings.

Using deep learning approaches, the researchers estimated and then compared PM2.5 levels from March through May 2020 — months when U.S. stay-at-home orders were tightest — to the same period in 2019. They also turned to Google’s COVID-19 Community Mobility reports, which temporarily reported anonymous data about traffic and destinations.

The biggest air quality improvement was in Washington D.C., which experienced a 21% decrease in pollution levels, followed by New York and Boston. The findings are published in the journal Atmospheric Environment.

Change in PM2.5 levels by city between March–May of 2020 and 2019: 

  • Washington DC, -21.1%
  • New York, -20.7%
  • Boston, -18.5%
  • Detroit, -13.53%
  • Chicago, -8.05%
  • Seattle -7.73%
  • Dallas, -6.71%
  • Philadelphia, -4.82%
  • Houston, -3.63%
  • Los Angeles, -3.29%
  • Phoenix +5.5%
Houston’s 3.6% decrease was among the mildest changes, a result the team attributed to the region’s many oil refineries and coal-fired power plant. “We need fuel, so the stay-at-home strategies did not impact oil refineries and power plants. Those facilities continued to operate and therefore continued to emit pollutants,” explained lead study author Masoud Ghahremanloo, a doctoral student at the University of Houston College of Natural Sciences and Mathematics.

In understanding Phoenix’s increase in PM2.5 levels — a 5% jump through the study period — the team noted that its residents had been more resistant to stay-at-home orders than most Americans, but they also found nature played a role, too.

“PM2.5 has many different ingredients — black carbon, organic carbon, nitrate, sulfate, dust, sea salt and so on. The natural ingredients — dust, sea salt and others — are not caused by industry or human mobility,” Ghahremanloo said.

Co-authors of the study from the UH Department of Earth and Atmospheric Sciences include doctoral students Yannic Lops, Jia Jung and Seyedali Mousavinezhad; and Davyda Hammond, project manager with Oak Ridge Associated Universities in Tennessee.

Publication: MasoudGhahremanloo, et al., A comprehensive study of the COVID-19 impact on PM2.5 levels over the contiguous United States: A deep learning approach, Science Direct (2022). DOI: 10.1016/j.atmosenv.2022.118944.

Original Story Source: University of Houston


RECOMMENDED