3DREAMS@SG

Elevated aerosol layers (haze, transported smoke, dust) sit aloft for hours before they mix down to street level. By the time a ground monitor reads the spike, the event is already happening. The signals that see it coming live in Doppler-LiDAR backscatter and wind profiles, but those feeds are noisy, ragged, and arrive on different clocks across sites, and clouds routinely masquerade as aerosol.

The brief: make that early signal trustworthy and actionable. Fuse the sources onto one timeline, score the likelihood of an episode with a physical model an operator can interrogate, forecast when it reaches the surface, and push that judgement to the people who act on it, without drowning them in false alarms.

One pipeline, five stages, sensors to surface. Raw range gates are deduplicated and state-tracked in SQLite, wind is recovered by a least-squares retrieval, and every feed is resampled onto a single ten-minute clock before any detector reasons over it.

Detection is a calibrated, glass-box score, not a black box. A two-hour rolling window is rated 0–100% from a six-component physical model, each component weighted and individually inspectable, so an operator can see exactly why a number moved.

A score nobody sees is worthless. The system is proactive: when the lead detector crosses threshold, conversational agents reach out over Microsoft Teams and Telegram with the full signal set attached, and the same agents answer questions on demand.