Texas · sulfur dioxide · 2021 to 2025
The air keeps its own records.
Facilities must report their accidental releases. Public monitors record the air hour by hour. Nobody checks one against the other.
We did, in both directions, and the two records rarely agree.
Monitor ledger · Unexplained
pollution episodes recorded by the monitors have no matching report from any facility in the same county.
Facility ledger · Unseen
releases the facilities reported themselves left no clear trace at the nearest monitor.
The case that started it
8,830 pounds. A city told to stay indoors. The monitor read 1.5.
On 27 June 2023 the Galveston Bay Refinery reported releasing 8,830 lb of sulfur dioxide, and Texas City issued its first shelter-in-place in over a decade. The county’s only sulfur dioxide monitor, 1.4 km away, peaked at 1.5 ppb. Two hundred and twenty-eight other hours that year read higher. If the paperwork and the air can disagree that completely about an event everyone noticed, we wondered what else they disagree about.
Texas City
GALVESTON County
Open the record →
Port Arthur and Beaumont
JEFFERSON · ORANGE County
Open the record →
Big Spring
HOWARD County
Open the record →
How it works
Three public records, put on one clock.
Every verdict is grounded in physical evidence and explicit rules, so every match and mismatch can be audited.
What the air recorded
Hourly sulfur dioxide from EPA monitors, with the wind measured at the monitor itself wherever it exists.
What was reported
Every emission event facilities filed with the state, parsed down to the pound and the cause they gave for it.
Which way the wind blew
A facility counts as upwind only when the wind arrived from its direction, allowing for the width of the plant and the uncertainty of the reading.
Where the two disagree
Rules decide each verdict, so every one of them can be explained to a resident and argued with by a regulator.
What this does not claim
No matching report does not mean an illegal release. Releases below the reportable quantity, permitted emissions in poor dispersion, mobile sources and sources outside the searched counties all look the same here. Read the methods and limits →