Night-time satellite passes over the flare field at roughly 19.45N, 92.09W picked up 7 clusters of heat that reached at least 5 MW and reappeared on 5 or more of the last 30 nights (28 August to 26 September 2026). The most persistent was lit on 17 nights, with a peak fire radiative power of 28.89 MW, and 252 hot-night detections were logged in total. All 7 clusters sit on sites listed in an independent 2024 annual flare summary from the Earth Observation Group, the largest of them 0.14 km away and credited with 0.217 billion cubic metres for that year.
What this is not: heat alone cannot separate a gas flare from a refinery furnace, a kiln or another industrial source — only the repetition and the known-flare match point to flaring. The brightness figure is not a volume; the sensor saturates on the biggest flares. Unlit venting and cold methane leaks would not show up at all. The finding names a place, not an operator.
What would change our mind: a month of heavy cloud can hide nights and mimic a shutdown, so a gap is not proof of stoppage. Daytime imagery, methane sounding, or an updated flare inventory could recast these clusters.
Written by AI, then checked by a separate reviewer before publication. Every claim points at the evidence below.