← Back to Explore

ILLINOIS-CHICAGO Predicting Large Wildfires Using Machine Learning Approach towards Environmental Justice via Remote Sensing

JSHS · 2023

Overview

Wildfires pose severe health and ecological consequences. In 2021 alone, 58,968 wildfires burned 7.1 million acres across the United States. Large wildfires (> 300 acres) in the United States, account for more than 95% of the burned area in a given year. Predicting large wildfires is imperative, however, current wildfire predictive models are localized and computationally expensive. My research aims to accurately predict large wildfire occurrences across the United States based on easily available environmental data and using a scalable model. The USDA data for 1326 wildfire occurrences over 20 years, representing 35 million acres burned, and NASA MODIS remote sensing data consisting of 925 million satellite observations were used. First, six key environmental variables were identified and annual averages over three years leading up to each wildfire occurrence were computed. Next, the resulting dataset of 18 environmental variables was tested on six different machine learning classification models (Logistic Regression, Decision Tree, Random Forest, XGBoost, KNN, and SVM) to determine their accuracy in predicting large wildfires. Finally, model validation tests and permutation feature importance analysis to identify important variables was performed. The XGBoost Classification model performed the best in predicting large wildfires, with an accuracy of 87.81%. Furthermore, towards Environmental Justice (Justice40 Initiative), an analysis was performed to identify disadvantaged communities that are also vulnerable to large wildfire occurrences. My model can be used by wildfire safety organizations to predict large wildfire occurrences with high accuracy and employ protective safeguards to prioritize resource allocation for socioeconomically disadvantaged communities.

Competition history

  • JSHS 2023 Category not listed

Resources

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Source: Junior Science and Humanities Symposium

Save projects to your library

Sign in with Google to keep track of projects you find interesting, organized into folders. An account also raises your daily allowance for “Has this been done?”, and lets you create a key for the MCP server with a much higher limit than anonymous use. Browsing stays public.

Continue with Google