Incorporating Methodologies from Astronomy to Wildfire Detection in An Incident Management Platform

Abstract:

Wildfires are the third-most expensive type of disaster, with an annualized economic burden estimated between $71.1 billion and $347.8 billion in the US (Thomas et al. 2017). Improving early warning systems via richer and more reliable wildfire characterization can prevent incident escalation, thereby reducing economic impact and saving lives.

The Visible Infrared Imaging Radiometer Suite (VIIRS) and MODerate resolution Imaging Spectroradiometer (MODIS) satellite-based instruments provide spectroscopic observations of the Earth’s surface in the infrared and visible spectra. These instruments are used for hotspot detection by NASA Fire Information for Resource Management System (FIRMS). However, the FIRMS algorithm produces numerous false positives linked to human industrial and recreational activity. Further, FIRMS does not currently allow for hot spot tracking or characterization. Large events are represented by hundreds of discrete hot spots, rather than grouped in a single object. This incomplete and inaccurate situational assessment impedes emergency managers from responding in a timely and efficient manner to wildfires.

We propose source detection algorithms which group heat observations into a single object whose evolution can be tracked across orbits using these near real-time analytics and modeling techniques. We leverage aperture photometry techniques to construct the spectrum, progression, and extent of wildfires. These observations allow the accurate classification of events based on their severity, trajectory, and combustion profiles, dramatically reducing false positives.

This reliable detection method could improve early warning systems and enhance decision support to communities. Furthermore, this would enable the development of a wildfire characterization library to improve scientific understanding of wildfires across the disaster cycle. DTI, in conjunction with academic, industry, and government partners, is developing a platform comprised of multi-hazard decision-support tools for emergency managers and community stakeholders. These solutions, informed by Sendai Framework guiding principles, allow practitioners to mitigate, prepare, and respond to incidents by improving situational awareness through context-specific visualization.