The Signal — Issue #006: Testing Perenco's Congo gas flare claim

A company facing pollution findings will usually give you two statements, and only one of them is worth your time. Perenco has just given Human Rights Watch both: a flat denial that its operations cause pollution or acute health effects, which you cannot test, and a claim that 220 flare points are permanently out, which you can. This week you learn to test a completion claim against free thermal satellite detections and to file the gap when the two do not line up.

An audit commissioned in December 2024 has produced nothing the public can read

Human Rights Watch published its findings on Perenco's operations in Muanda on 27 July 2026. Agathe Bounfour, its senior fossil fuel researcher, and Lewis Mudge, its Central Africa director, are the named voices on the release. A team interviewed 45 residents, oil workers, health professionals, officials and experts in Muanda and Kinshasa in January 2026, and analysed satellite imagery alongside geolocated photographs and video. Perenco is Congo's sole oil operator, working onshore and offshore across a coastline of 37 kilometres in Kongo Central province.

The open record carried this before the fieldwork did. A 2013 Congolese Senate investigation found elevated sulphur dioxide and nitrogen dioxide near flaring sites, hydrocarbon contamination in groundwater and surface water, and lead and mercury in water samples above World Health Organization drinking water guidelines. A 2025 study by researchers at universities in Congo and Cameroon found high levels of total petroleum hydrocarbons and volatile organic compounds in soil around extraction sites, and named the villages: Kinkazi, Kitombe, Tshiende. The two documents sit very differently in the open. Human Rights Watch had to obtain the Senate report. The 2025 study is publicly indexed and retrievable, and on its own it names settlements against contaminants. Its authors went further and recommended soil remediation alongside a long-term monitoring system for soil and groundwater quality.

So at what point did the open record contain enough to flag this. The answer is the 2025 soil study, indexed and retrievable through AGRIS. Add the November 2025 announcement that the audit contract had been extended by a year without a stated reason or a release date, and the file was openable eight months before the report landed. The signal was a research index and a ministerial announcement, not a satellite and not a leak.

The wider condition here is not secrecy. Count what is stacked up: a monitoring system recommended by researchers in 2025 and not built, an audit commissioned in December 2024 and never published, and air quality sensors installed by the company whose readings the government's own control office says it cannot access. Nobody is hiding the existence of the audit. The withheld thing is the finding, and that changes what you go looking for. You stop hunting for a leak and start building the timeline of everything that was promised and never published.