For most of Ghana's history, knowing how polluted the air was in any given community came down to one question: Is there a monitor nearby? For the vast majority of the country, the answer has been no.

That is about to change.

Ghana's Environmental Protection Authority has formalised a partnership with Columbia University's Lamont-Doherty Earth Observatory to map PM2.5 levels, the tiny particles most dangerous to human health, across the entire country, going back as far as 2005. The announcement, made on March 29, 2026, fills a gap that has quietly enabled a crisis to deepen unseen.

Air pollution kills approximately 32,000 Ghanaians every year. It costs the country an estimated 2.5 billion US dollars annually, about 4.5 per cent of GDP, according to World Bank figures. Speaking on behalf of EPA Chief Executive Professor Nana Ama Browne Klutse at a recent air quality workshop in Accra, Head of Air Quality Selina Amoah was direct.

"Results from the Environmental Protection Authority's air quality monitoring activities reveal particulate matter concentrations that exceeds the national air quality standards in most of the monitoring locations."

Yet for decades, Ghana has lacked the data infrastructure to fully see the scale of what is happening, let alone prove whether any policy is actually working.

The tool at the centre of this partnership is GRASP, the Gridded Africa Surface Pollution dataset, built by Dr. Daniel Westervelt and his team at Columbia University. It combines NASA satellite data passing over Ghana at least twice a day with ground-level sensor readings gathered over the past decade, using machine learning to produce a unified picture of air quality across every part of the country, including areas that have never had a single monitor.

Westervelt, who travelled to Ghana earlier in 2026 to train the EPA on how to use the data, explains why building it was not straightforward.

"Going back 20-plus years, there's not a lot of air quality data at the surface in that time period in Ghana. There was some work by EPA and others over the years, but it's been a little bit sparse. So we had to find another way."

The result covers the whole country from 2005 to 2024, will be updated regularly, and is available free at the GRASP Website with an interactive map, downloadable data, and tools for generating charts by city or region.

Desmond Appiah, Country Lead for the Clean Air Fund in Ghana, says data infrastructure like this has been missing for too long, and the consequences are real.

"The level of air quality data collection in Africa is very low. Because it's low, we are not really putting the information in the hands of the policymakers in the state that the policymakers can understand."

He adds that the problem does not stop at Ghana's borders.

"Air quality is a transboundary issue. Pollution from one part of the sub-region travels to other parts. It is in the interest of us as Ghanaians that our neighbours are also on the same trajectory to improve air quality."

One of the first tests of GRASP's value was an analysis of Ghana's emissions levy, introduced in February 2024, which charged vehicle owners an additional fee with the hoped-for co-benefit of reducing air pollution. The levy was scrapped by the next government before completing a full year. The question the Columbia team asked was simple: while it lasted, did it work?

The answer was unambiguous.

"We did not see any PM2.5 change between the post-levy period and the pre-levy period. Basically, statistically insignificant changes, a fraction of a percent. So we can say pretty definitively that there was not a substantial reduction due to this emissions policy," Westervelt says.

He is not dismissive of emissions pricing as a concept, New York City's congestion charge has worked, but argues the building blocks were not in place in Ghana.

"You have to first build up the infrastructure for public transit. You have to address poverty, incomes. There's a lot of building blocks that sort of get you to where these type of emission levy policies can be popular. I just don't think that they're currently in that position in Ghana."

His immediate recommendations are more practical: restrict imports of old, high-polluting vehicles, expand cleaner public transport, and subsidise cleaner cooking fuels. Vehicles and household cooking, he says, are the two dominant sources of PM2.5 in Ghana.

"The traditional way of cooking with wood or charcoal is very polluting, to put it frankly. Governments can look towards subsidising alternative methods that would allow folks to accomplish what they need to do in their cooking, but by emitting a lot less pollution."

The GRASP data paints a picture of an air quality situation that is neither improving nor collapsing, but sitting at a level that should not be considered acceptable. Annual average PM2.5 in Accra has hovered between 25 and 26.5 micrograms per cubic metre since 2021. The WHO annual guideline is 5.

"It strikes me as very high. If you look at major cities in Europe or the US, you would see concentrations of maybe less than half of what you're seeing in Accra. It is a cause for concern," Westervelt says.

Ghana has made real progress, first in West Africa to phase out leaded gasoline, first to move to low-sulphur diesel, and now operating 15 regulatory monitors and 28 local sensors across five cities. New air quality management regulations, LI 2507, were published in 2025. But the EPA itself acknowledges that virtually the entire population is still exposed to air above WHO guidelines.

The GRASP partnership gives Ghana something it has never had: a continuous, independent, freely accessible record of how its air has changed over two decades, and a tool to hold future policies accountable.

"I hope that we can start to turn the corner a little bit," Westervelt says, "and everyone can breathe cleaner air in the near future."

The data now exists to know whether that is happening. What comes next is a matter of political will.

This story was a collaboration with New Narratives. Funding was provided by the Clean Air Fund, which had no say in the story’s content.