Hot days in the Congo Basin nearly doubled, paper weather logs show

Hot days in the Congo Basin became nearly twice as common between the 1960s and the 1980s, according to a peer-reviewed study built on handwritten weather sheets pulled from archives in the Democratic Republic of the Congo and Belgium.

The paper, titled "From paper to proof: revealing Congo Basin warming through rescued climate archives," was published in Earth System Dynamics on October 12, 2026 ^[3]. Its number is the headline, but the achievement underneath is the data itself.

Central Africa is one of the few regions where the Intergovernmental Panel on Climate Change (IPCC), the UN-backed body that assesses climate science, had not assessed changes in hot extremes or extreme rainfall since the 1950s ^[3]. The study says the main reason is sparse observations ^[3].

Thousands of sheets, photographed one by one

The records sat in two places: the archives of INERA, the agricultural research institute at Yangambi in the DRC, and the Belgian State Archives ^[2][3]. During two field campaigns in 2023, a team working with INERA staff and researchers from Ghent University photographed 9,885 sheets ^[2][3].

The sheets came from 37 weather stations across the DRC. Back at the desk, the team used a software tool called MeteoSaver to turn the photographs into digital time series, then ran quality checks before analyzing anything ^[2][3]. The result is more than one million temperature and precipitation observations, now published as an open-access dataset ^[2][3].

This is measurement, not modelling. Climate models fill gaps with physics and assumptions. These are thermometer readings taken at specific places on specific days, which is why they are so valuable in a region that has had so few of them in the scientific record.

What the sheets show

Only 21 of the 37 stations had enough data for the temperature analysis. Across them, the median warming trend was 0.22°C per decade for daily maximum temperatures, 0.10°C for minimums and 0.15°C for averages ^[3]. Over 1961 to 1990, that adds up to roughly 0.7°C, 0.3°C and 0.5°C respectively ^[3].

Daytime highs climbed more than nights warmed. That fits what the authors found about extremes: with each successive decade, hot extremes grew more frequent and cold extremes less so ^[3]. The latest decade in their comparison had nearly twice as many hot days per year as the earliest ^[3]. The university release describes that comparison as the 1980s against the 1960s ^[2].

Heavy rainfall also appears to have risen. Of the three stations with enough precipitation data, two showed more heavy-rain events and the third showed no substantial change ^[2][3]. Three stations are too few to say anything about rainfall across a country the size of the DRC, and the authors say so ^[2][3].

Saved from the fire

The archive nearly did not survive to be read. Lead author Derrick Muheki said part of the INERA collection had been damaged by fire just months before his visit ^[1][2].

"We could have lost decades of irreplaceable data," Muheki said ^[1].

It is an unglamorous kind of rescue. Nobody pulls anyone from a burning building.

A researcher photographs a page of columns of numbers, written by someone decades ago who presumably believed the readings mattered, and a trend appears that the world had no other way to see. A record is only a memory that somebody bothered to write down, and paper memory is fragile.

What the study does not say

The study documents observed warming and more frequent hot extremes. It does not itself attribute the regional changes to human influence ^[3]. The published accounts of the work leave several things open: how the paper defines a hot day, who kept each set of records, why they remained on paper, how much of the INERA archive burned, and what the heat has meant for health and daily life in the communities of the basin.

The nearly doubled count is a comparison between the earliest and most recent decades in the study, and the data end in 1990. The paper is a picture of the late twentieth century, not of the heat of recent years.

Why it matters beyond the archive

The researchers say the dataset can help establish climate context for the region and be used to evaluate and improve climate models ^[2][3]. Models are tested by asking whether they can reproduce what actually happened. For the Congo Basin, there was little actual to test them against.

The authors also call for more data-rescue work ^[2][3]. The logic is simple. If 37 stations' worth of sheets can fill a blank on the map, then other boxes of paper in other archives may do the same, provided somebody gets to them before damp, insects or fire do.

The next test is whether anyone does. The dataset is open to all, and the authors' call for more rescue campaigns is the part of the study that remains to be answered ^[2][3].