For principal meteorologist Grieffy John Stegling, the storerooms at Botswana’s national weather service headquarters in Gaborone hold a rare treasure: floor-to-ceiling shelves containing boxes of old notebooks with carefully recorded weather observations going back more than a century.

Such records offer clues not only to the country’s past, but also to the future of its climate. Like most African countries, Botswana is ill served by global climate models because predictions are based on patchy records of key variables such as temperature, humidity and atmospheric pressure ¹ .

“Historical climate data over Africa are very valuable for understanding climate variability and trends,” says Chris Taylor, a meteorologist at the Centre for Ecology and Hydrology in Wallingford, UK, who studies African climate trends.

In 2017, Taylor and his team found that climate change will increase extreme rainfall in the Sahel, a semiarid region south of the Sahara desert ² . A crucial part of their study involved cobbling together historical records — some of them “locked away in cupboards” — from different national weather services, Taylor says. “Having a historical baseline is a prerequisite for understanding how intense rainfall is changing,” he says.

Since 2015, the World Meteorological Organization in Geneva, Switzerland, and Germany’s weather service, Wetterdienst, have provided training and equipment to help Botswana digitize and share its historical climate data. But because there are no dedicated staff members, progress has been slow. Of 2 million records, only 100,000 have been processed. “If we had more manpower, it would go much faster,” Stegling says.

Cash for access

Whereas Botswana is making some progress, in other meteorological offices across Africa, millions of records are mouldering in cardboard boxes or languishing on obsolete technology. Digitization efforts have been held up because of concerns that giving researchers free access to the data will prevent such offices from making money by selling the information to individuals and companies.

The South African Weather Service (SAWS) has turned down offers by the International Data Rescue (I-DARE) project to help to digitize historical climate data because the agency wants to be able to sell its data. “If unrestricted access to the National Climatological Databank, of which SAWS is the custodian, is allowed, SAWS might not be able to deliver on its commercial mandate,” a spokesperson told Nature .

Similar concerns are holding up the digitization of 2 million surface observations — including temperature, rainfall and humidity — from 48 African countries. Those data are stored at the African Centre of Meteorological Applications for Development (ACMAD) in Niamey, Niger.

“The private sector is progressively being involved in climate services delivery,” says ACMAD director-general Andre Kamga Foamouhoue, and this sometimes creates conflicts of interest with government agencies looking to commercialize data. This is one reason why some African meteorological offices don’t allow files to be shared with projects that rely on citizen science volunteers to digitize the data.

Many of the data-rescue requests come from initiatives led by individuals or institutions from Europe or the United States, says Jane Olwoch, executive director of the Southern African Science Service Centre for Climate and Land Management, a regional climate research centre located in Windhoek, Namibia. And that can be a problem because institutions in African countries aren’t sure how they will benefit if the data expertise comes from outside the continent.

“The trust has been broken when international researchers come here and take the data and just go back,” says Olwoch. She hopes that data-rescue efforts fronted by her own organization, in Angola and Botswana, will be viewed with less suspicion because the organization is backed by four southern African governments and has local headquarters and staff, even though much of the funding comes from the government of Germany.

Recovering old records

Not all of Africa’s climate records are in Africa, however. Many of the oldest ones were collected by professional and amateur meteorologists who came to Africa from Europe during colonial times. Stefan Grab, a geographer at the University of the Witwatersrand in Johannesburg, South Africa, says that these records can, paradoxically, be easier to access than local ones.

South Africa has the Southern Hemisphere’s longest uninterrupted weather observations, recorded at the astronomical observatory in Cape Town. It was thought that these data stretched back to 1841, but Grab, who leads South Africa’s data-rescue efforts, knew that astronomers had been in the Cape since the 1830s. So he contacted staff at the Royal Greenwich Observatory in London, who directed him to the archives at the University of Cambridge, UK. “Lo and behold, they found the earliest records, which go back to 1834,” he says.

ACMAD’s Kamga Foamouhoue says that weather agencies must be persuaded that there are benefits to mining historical data and then sharing these with other scientists — and that the biggest benefit is that it will lead to more accurate climate predictions.

“Anything that’s really old, like from the nineteenth century, is extremely valuable,” Grab emphasizes. “It’s worth far more than gold and diamonds.”