A Multi-Modal Wildfire Prediction and Personalized Early-Warning System

Overview

Wildfires are increasingly impacting the environment and human health. Among the top 20 California wildfires, those in 2020-2021 burned more acres than the last century combined. Lack of an adequate early warning system impacts the health and safety of vulnerable populations disproportionately and widens the inequality gap. In this project, a multi-modal wildfire prediction and early warning system has been developed based on a novel spatio-temporal machine learning architecture. A comprehensive wildfire database with over 37 million data points was created, including the historical wildfires, environmental and meteorological sensor data from the Environmental Protection Agency and PurpleAir, and geological data. The data was augmented into 2.53 km x 2.53 km square grids to overcome the sensor network coverage limitations. Leading and trailing indicators for the wildfires are proposed, classified, and tested. The leading indicators are correlated to the risks of wildfire conception. The trailing indicators are correlated to the byproducts of the wildfires. Additionally, geological data was incorporated to provide additional information for better assessment on wildfire risks and propagation. Next, a novel U-Convolutional Long Short-Term Memory (ULSTM) neural network was developed to extract key spatial and temporal features of the dataset, specifically to address the spatial nature of the location of the wildfire and time-progression temporal nature of the wildfire evolution. Through over 50 iterations of improvements and optimization, the final ULSTM network architecture achieved >97% accuracy for predicting large wildfires, as compared to ∼76% using traditional Convolutional Neural Network (CNN) techniques. The final model was applied to conduct a retrospective study for the 2018-2022 wildfire seasons, and successfully predicted 85.7% of wildfires >300K acres in size. This technique would be used to enable fire departments to anticipate and prevent wildfires before they strike and provide early warnings for at-risk individuals for better preparation, thereby saving lives, protecting the environment, and avoiding economic damages.

A Multi-Modal Wildfire Prediction and Personalized Early-Warning System

Overview

Source coverage

Awards (1)

Competition history

Resources

Related projects

BlazeAlert: A Novel Spaciotemporal Deep Convolutional Neural Network Model for Real Time Wildfire Risk Assessment

Risk-Aware Early Wildfire Detection from Satellite Imagery Guided by Fire Weather Indices

Stop the Flame in Its Infancy! Multivariable Early-Warning System for Low-Cost Prevention of Wildfire Proliferation

A Novel AI Pipeline for Predicting Next Day Wildfire Spread

It's Flaming Out: Using Artificial Intelligence To Emulate Critical Aspects of Wildfire Growth

SparkNet: Transparent Prediction of Wildfire Potential in Southern California

An Accessible, Low Cost Tool for Citizen Scientists: Using Remote Sensing Techniques to Predict Fire Damage Propensity

A Data Analysis Approach to Wildfire Prediction

Browse more like this