Satellite night-time heat detections over the flare field at roughly 57.95N, 107.08E show one persistent hot cluster. Between 28 August and 26 September 2026 it was lit on 18 of the 30 nights, with a peak fire radiative power of 36.73 MW and 157 hot-night detections in the area overall. Independently, the cluster sits 0.54 km from a site listed in the Earth Observation Group's VIIRS Nightfire annual flare summary for 2024, where that site was credited with 0.157 billion cubic metres of flared gas.
What this is not: thermal sensors cannot separate a gas flare from an industrial furnace, refinery or kiln, and heat is not volume — the sensor can saturate, and the flared-volume figure comes only from an earlier annual summary, which proves the site is a known flare but not what burned this month. Unlit venting and cold methane leaks would be invisible here. No operator or owner is identified; this is a place, not a company.
What would change our mind: a cloud-free month with the cluster dark, a ground or registry record showing a different heat source, or methane observations pointing to venting rather than burning.
Written by AI, then checked by a separate reviewer before publication. Every claim points at the evidence below.