Recipient Organization
HELIOS POMPANO INC
2226 N CYPRESS BEND DR
POMPANO BEACH,FL 33069
Performing Department
(N/A)
Non Technical Summary
In the US, lightning is the leading cause of wildfires in terms of area burned, responsible for 5.5 million acres lost in 2020. Over 70% of the area burned during wildfires in the Western US was due to lightning ignitions, and lightning is the number one cause of wildfires in Florida. On average, lightning-ignited wildfires become nine times larger than human-initiated fires because they often go unnoticed during ignition and early growth, and existing technology available to fire management agencies can neither quickly nor accurately identify them. Thousands of lightning strikes can occur each day within a wildfire management area, but only less than 10% of these are capable of initiating a fire, known as High-Risk-LightningTM (HRLTM). It is difficult to identify which strikes will result in a fire, and impractical to send out personnel to search each strike site. Fire Neural NetworkTM (FNNTM) technology can identify these strikes within 40 seconds, but not all of these strikes will result in a wildfire ignition because this is also dependent on other factors, such as ground and weather conditions. Therefore, it is essential to quickly verify these strikes to allow first responders to assess and act while the risk spread is still small. There is an unmet need for rapid wildfire ignition verification and real-time tools to mitigate risks to vulnerable communities.FNNTM will develop a novel method to protect vulnerable rural communities from lightning-initiated wildfires by creating a functional operating model to; (i) pinpoint and monitor a wildfire by verifying HRLTM ignition using an Unmanned Aerial Vehicle (UAV) based system called FirebirdTM, reducing wildfire localization time from around 24 hours to 40 minutes, and (ii) integrate real-time wildfire data with local Community Wildfire Protection Plans (CWPPs) within a Geographic Information System (GIS) dashboard for use by emergency services, planning authorities, community leaders and local governments. It will expand on proprietary detector technology developed by FNNTM, and will use an underlying novel and open-source geocode. The research is closely aligned with the John D. Dingell, Jr. Conservation, Management and Recreation Act of 2019 and the federal Healthy Forests Restoration Act of 2003. In this Phase I project, eight vulnerable Florida counties and their associated CWPPs will act as a trial for the technological development, to include aspects such as communities at risk and associated fire risk ratings; critical facilities locations and their vulnerability; and wildland-urban interfaces including wildland fuel types.The overall goal of the project is to produce a CWPP-led GIS dashboard that will identify, within 40 minutes and within 3 meters, a lightning-initiated wildfire, which can then be monitored and acted on by emergency responders. The GIS dashboard will also identify vulnerable communities and infrastructure, helping emergency responders make informed decisions on how best to control the wildfire. The use of this system across the US will have significant positive impacts on human health, homes and businesses, small communities, farms and farmland, livestock, wildlife and habitats, forests, and important infrastructure such as power lines and roadways.
Animal Health Component
35%
Research Effort Categories
Basic
15%
Applied
35%
Developmental
50%
Goals / Objectives
The major goal of this project is to produce a Community Wildfire Protection Plan (CWPP) led Geographic Information System (GIS) dashboard to quickly identify, locate, and monitor lightning-initiated wildfires. The objectives are:Assess hardware components and the software interface between them. To include lightning detectors, Unmanned Aerial Vehicles (UAVs), visual and infrared cameras, GNSS positioning systems, cellular connectivity modules, and VHF radio transmitters.Explore UAV search algorithms. Find the best available combination of response speed and accuracy to identify and locate wildfires based on the FNNTM HRLTM coordinates, using visual and infrared images taken from the UAV.Determine and itemize the most relevant CWPP-based information necessary for identifying specific risks and vulnerabilities for each of the eight Florida rural communities. This information will be incorporated into GIS layers.Develop a GIS dashboard with a grid-based geocode to include critical CWPP information on vulnerable communities and infrastructure, incoming live data from wildfire locations, and predictions of wildfire growth based on the data collected by the UAV.Assess the use and direct impact of the developed technology on vulnerable rural communities, protecting the environment, climate change mitigation, and the socioeconomic development of rural areas. This will be done through a participatory process with firefighters, community leaders, and local government representatives.
Project Methods
Efforts: The work will be carried out near the existing eight FNNTM detectors in central Florida which will serve as the field test sites. Thereby, the Florida Forest Service (FFS) will be involved, from the start, in this experiential learning process of the project. Various hardware components will be specified, calibrated, and checked for operational compatibility. These include FNN™ lightning detectors, UAVs, visual and infrared cameras, GNSS positioning systems, cellular connectivity modules, and VHF radio transmitters. The data from the FNN™ detectors will be stored in a cloud server and an API connection will be established with the UAV control module. Similarly, the UAV control module will be connected with the camera systems and GNSS positioning system to ensure that the footage is georeferenced with 3m or better location accuracy. Finally, data connectivity will be established and tested using both cellular and VHF technology to ensure a continuous connection with the UAV in flight.FNNTM will assess several UAV flight path planning search patterns, to optimize response speed and accuracy for precise location of possible lightning-ignitions triggered by the FNN™ instruments at the FFS locations. For this part of the project, FNN™ will collaborate with the University of Florida Unmanned Aircraft Systems Research Program (UFUASRP). FNN™ will hold seminars about the project there to enrich the lab participants' minds and engage with potential new researchers in the field. A thermal camera will allow the drone to identify the fire from the background forest with high contrast and location precision on the order of a few meters. The time needed to identify the exact location of the ignition source will be a function of the altitude above ground of the drone, and the flight pattern employed. In addition, FNN™ will also perform positional accuracy assessment of UAS-derived ignition location estimates by comparing them to high-accuracy field observations. These field observations will be made using relative-solution Global Navigation Satellite System (GNSS) observations such as those from Real-time Kinematic (RTK) positioning techniques.FNN™ is an official partner of Esri and the ArcGIS platform will be utilized to build the GIS dashboard that integrates community vulnerability data, FNN™ HRL™ data, and fire perimeter data from the UAV. FNN™ will showcase this technology developed in the project at various GIS conferences to spread the knowledge on this important topic. Raster data regarding the vegetation type, wildland-urban interface, and fire spread projections with our partner Technosylva will be integrated alongside vector data showing the locations of vulnerable community members and sensitive resources. The GIS dashboard will be based on a novel geocode system which divides the entire world into a fixed 3x3m grid each with a unique identifier of three words. This builds on Task 2 to use GNSS positioning to correlate the position that the UAV has imaged with the visual and infrared cameras to a three-word address, and then to apply the fire search pattern to label the square as either 'on fire' or not. Once the 'on fire' squares have been identified, the fire perimeter is sent through an API to FNN™'s partner Technosylva which performs 8-hour fire spread modeling in seconds using weather and vegetation data.Evaluation: During this project, the focus will be on North Central Florida. Representatives from these communities will be actively involved throughout the project to assess and reflect upon the development and use of the technology. There will be interviews and feedback sessions conducted with several FFS members along with local rural community leaders to determine how the developed system helps build rural wildfire resilience and how the technology could be further improved. The number of lightning-initiated wildfires that the FFS members responded to based on the new information available to them will be quantified. The rural community leaders will assess how their wildfire resilience levels have changed due to their involvement in the project.