On the morning of September 8, 2025, Austin Public Health's emergency preparedness coordinator, Tanisha Reyes, received a Trace AQ automated alert: a high-confidence PM2.5 spike was forecast to reach the Austin metro area within 72 hours, driven by wildfire smoke from the Llano County fires burning 80 miles to the northwest.
What happened in the following 72 hours became a model for proactive air quality emergency management — and a demonstration of what's possible when forecasting data is actually integrated into public health operations rather than being a passive information source.
The Setup: Air Quality Without Advance Warning
Before partnering with Trace AQ in early 2025, Austin Public Health relied on the same tools available to any resident: the AirNow website, EPA monitoring station data, and occasionally, voluntary air quality reports from the Texas Commission on Environmental Quality. These tools are valuable for understanding current conditions but offer limited predictive capability.
"We were essentially reactive," Reyes explained in an interview. "A sensor would cross the 'unhealthy' threshold and we'd scramble to issue an advisory. By the time residents saw the advisory, the exposure had already started. For elderly residents, for people with asthma or COPD, that lag time wasn't just inconvenient — it was a clinical problem."
Austin Public Health began piloting Trace AQ's API-based forecast service in March 2025, integrating the forecast data into their existing emergency notification infrastructure. When a forecast exceeds a configurable PM2.5 threshold for any zip code in their coverage area, the system automatically drafts an alert and queues it for human review before distribution.
September 2025: The Event
The Llano County fires started on September 6, 2025, driven by record-low relative humidity and sustained southwesterly winds. By September 7, fire behavior analysts at the Texas A&M Forest Service were projecting continued spread. Trace AQ's forecast model, ingesting satellite-derived fire radiative power and updated atmospheric dynamics from the GFS model, generated a 72-hour PM2.5 forecast showing values reaching 85–110 µg/m³ in central Austin — firmly in the "Unhealthy for Sensitive Groups" to "Unhealthy" AQI categories.
Tanisha Reyes received the automated alert on the morning of September 8, when Austin skies were still blue and the fires were invisible to the naked eye. She escalated to the Austin Public Health director, who authorized an immediate response.
Within six hours, Austin Public Health had issued a formal air quality advisory through city emergency notification channels. They coordinated with Austin ISD to cancel outdoor physical education and recess activities at all 130 schools district-wide for the following two days. They opened three cooling and clean-air centers at community facilities with upgraded HEPA filtration. They pushed targeted text messages to registered users in the Austin Public Health alert system, emphasizing the 72-hour window and recommending that high-risk individuals stock up on medications and prepare to limit outdoor exposure.
What Happened When Smoke Arrived
On the afternoon of September 10, exactly as forecast, smoke began moving into the Austin metro area. AQI readings at the EPA monitoring stations climbed steadily through the afternoon and evening, peaking at 147 (Unhealthy) by midnight. The Trace AQ forecast had predicted a peak of 138–152, placing the actual observed value well within the predicted range.
Austin Public Health's pre-positioned response meant that clean air centers were already staffed and operating when the smoke arrived, rather than being set up reactively. School administrators didn't have to make last-minute decisions about outdoor activities. High-risk residents had an extra 60+ hours to prepare compared to a scenario where they would have received the advisory only when conditions deteriorated.
Emergency department respiratory admission data from the event showed a 17% lower-than-expected surge compared to a comparable smoke event in September 2023, when no advance warning infrastructure was in place. Austin Public Health attributes much of this difference to the effectiveness of the preparatory advisories in prompting high-risk individuals to take protective measures.
Lessons and Next Steps
The September 2025 event validated Austin Public Health's investment in forecast-driven air quality management and produced several learnings that have since been incorporated into their protocols. Most importantly, they found that the 72-hour window is genuinely actionable — it is long enough to mobilize institutional responses but short enough that the forecast remains accurate and the public takes it seriously.
Austin Public Health is now expanding their Trace AQ integration to include wildfire smoke impact forecasts disaggregated at the zip-code level, allowing more targeted advisories for the neighborhoods most likely to experience the worst air quality during any given event.
They're also working with Austin ISD and Austin Energy to use Trace AQ data to automate certain low-stakes decisions — like routing electric school buses to use routes that minimize total occupant PM2.5 exposure during smoky periods — without requiring human review for every decision point.
This is what an air quality-aware city looks like in practice: not just monitoring, but acting. Trace AQ is committed to providing the forecasting foundation that makes this kind of proactive response possible.
