September Research Roundup: Select Papers

Every month, researchers at Lamont-Doherty Earth Observatory and other parts of the Columbia Climate School publish work that deepens our understanding of the planet.

This month, we wrote about the impact of livestock on Madagascar’s native animals, robot sailboats in the Southern Ocean, New York City’s $300 million methane problem and how extreme heat is spreading beyond summer.

Here are additional studies by our researchers that are worth noting.

When U.S. Embassies Stopped Reporting Air Quality, Pollution Rose

Dense smog blankets Metro Manila skyline. Credit: MDV Edwards

For more than a decade, U.S. embassies and consulates provided independent, publicly available air-quality readings in cities around the world. When that program was suspended in 2025, pollution increased in those cities, according to new Climate School research.

Using satellite observations across 46 cities, researchers found that the loss of embassy reporting was followed by about a 33% increase in aerosol pollution. The increases were largest and most persistent in places without another publicly available national air-quality monitoring system.

To estimate the impact of losing embassy monitoring on fine-particle pollution, the researchers compared the satellite data with embassy readings collected before the shutdown. They found an increase of about 1.9 micrograms per cubic meter of PM2.5, the tiny airborne particles that can penetrate deep into the lungs. The authors also tested alternative explanations and found little evidence that pollution was already rising in those cities before the reporting stopped.

“Public access to trusted environmental data appears to be part of what creates accountability for pollution,” says lead author Zoey Zhou, an assistant professor at the Columbia Climate School. “Data is evidence, and we’re seeing a trend not just internationally but in the U.S. of ending programs that gather climate and environmental data.”

Why it matters: The findings suggest that environmental transparency can function as a form of environmental governance. When credible, real-time pollution information disappears, some of the environmental gains associated with public disclosure may weaken.

Other Columbia-affiliated authors: Fanyu Wang and Douglas Almond, Columbia University.

Hidden Urban Deprivation Extends Far Beyond Megacities

Hundreds of millions of people live in densely built, poorly connected neighborhoods that can signal serious living-condition deficits. Mapping these patterns across 5,132 cities in Africa, Asia, and Latin America and the Caribbean, a new study estimates that about 395 million people live in such neighborhoods. Africa had the highest relative prevalence of people living in deprivation, at 43.6% of urban residents. About 136 million people, or more than one-third of the total, live in small and medium-sized cities (less than 1 million residents) that are often overlooked in global urban data and policy discussions.

Kisumu, Kenya, one of thousands of cities included in the analysis. Credit: Victor Ochieng

The study builds on the City Segments Layer developed by Dana Thomson of the Climate School’s Center for Integrated Earth System Information (CIESIN) and colleagues. The layer divides cities into standardized, neighborhood-sized units using open data on roads, waterways, buildings and population.

Along with her colleagues, Thomson, a co-author on the new study, trained a machine-learning model on locally mapped neighborhoods in eight large cities, including Nairobi, Mumbai and Mexico City, then applied it to cities across the three regions.

“Small and medium-sized cities are absorbing a growing share of urban population growth, especially in the Global South, but often have the fewest resources to manage it,” says Thomson. “Without comparable neighborhood-scale data, many vulnerable communities remain invisible when planning and investment decisions are made. Making these neighborhoods visible is a first step toward ensuring that basic services and resources reach the people and places that need them most.”

Why it matters: Smaller cities often have fewer resources for planning, even as many face rapid growth. The new mapping approach could help identify places that need closer assessment and support efforts toward adequate housing and basic services.

A Forgotten 1941 Heat Wave Offers a New Comparison for 2021

The 1941 heat wave spread from western to eastern North America over several weeks, bringing unusually high temperatures, dry conditions and low soil moisture across much of the continent.

A newly published study reconstructs a major North American heat wave from 1941 that was comparable in magnitude and location to a record-shattering 2021 event, but whose impacts had never been studied in detail. Although the 1941 event was about 1.7°C cooler, it occurred in a much cooler climate and was still extraordinarily rare for its time. The comparison offers a look at how two similarly extreme events, separated by 80 years, affected people and communities very differently.

Sam Anderson, a visiting postdoctoral research scientist at Lamont, and Shawn Chartrand of Simon Fraser University reviewed 357 issues of 25 newspapers and combined those accounts with climate, streamflow, crop, mortality and other data. They found widespread impacts in 1941, including deaths, water shortages, crop damage, wildfires and flooding. The study identified 344 deaths reported as directly or indirectly related to the heat, with drowning a much more prominent cause than in 2021, when most heat-attributed deaths in British Columbia occurred indoors.

“The impacts of any heatwave are shaped not only by its meteorology, but by the social, cultural and historical context in which it occurs,” says Anderson. “Our comparison shows that the impacts of future extremes will depend in large part on the world we choose to create.”

Why it matters: Comparing two extreme heat waves separated by 80 years shows how changes in infrastructure, technology, behavior and policy can shape who is affected and how.