Introduction

This project investigates the correlation between climate change and wildfires, with a specific focus on California. Utilizing machine learning techniques, we aim to predict wildfire occurrences and intensity based on weather data, helping mitigate risks associated with climate change and wildfires.

Key Findings

• Southern California (e.g., Los Angeles and San Diego) has experienced a significant increase in maximum temperatures over the last 20 years, while Northern California (e.g., San Francisco) has remained relatively stable.
• Key factors influencing wildfires include average firepower (FRP), surface temperature, and geographic location (latitude and longitude).
• Regression models like Random Forest achieved an R² value of 0.76 for predicting wildfire counts, indicating strong predictive power.
• Logistic regression models binned fire intensity into low, medium, and high categories, achieving moderate accuracy with an R² value of 0.5.

Future Predictions

Using our predictive models, we forecast strong wildfire counts for upcoming years:

• 2023: 1,817 fires
• 2024: 2,003 fires
• 2025: 2,350 fires

Project Report

Explore our detailed analysis and methodology in the embedded report: