Forest Fire Prediction Model: a Comprehensive Approach Using Remote Sensing, Gis, and Machine Learning for Effective Fire Risk Management
Year: 2025
Type: article
Abstract: Forests are often referred to as the lungs of the Earth due to their vital role in maintaining ecological balance.However, forest fires pose a significant environmental challenge, particularly in regions with diverse ecosystems and limited forest cover, such as Gujarat in India.These fires threaten biodiversity, local communities, and the economy while contributing to climate change through substantial carbon emissions.Given the critical importance of forests and the severe consequences of forest fires, the development of proactive prediction models is essential to mitigate their impacts effectively.This study focuses on creating a comprehensive forest fire prediction system tailored to India's distinct geographical, climatic, and ecological conditions, with Gujarat serving as a case study.Advanced machine learning algorithms, remote sensing data, and image processing techniques are utilized to analyze key meteorological factors such as temperature, humidity, wind speed, and rainfall, which significantly influence forest fire risks.The research integrates satellite imagery, GIS-based mapping, and historical fire data to build predictive models capable of generating accurate early warnings.Special emphasis is placed on Gujarat-specific factors, including forest types and topography, to identify areas most susceptible to fires.This study specifically targets the Polo Forest, aiming to develop a robust prediction model using satellite imagery, GIS mapping, and machine learning techniques.By examining the region's unique ecological and climatic conditions, the research provides valuable insights into fire behavior, helping forest authorities and policymakers design effective prevention and mitigation strategies.For this study, the Random Forest algorithm was employed alongside vegetation indices Normalized Difference Vegetation Index (NDVI) and the Normalized Burn Ratio (NBR) to predict the severity of forest fires.The results demonstrate approximately 90% accuracy in assessing the impact of fires on different land use classes, underscoring the model's reliability and potential for practical application in forest fire management.
Authors: Zalak Vyas, Vibha Patel
Institution: Gujarat Technological University
Language: English
Topic: Fire effects on ecosystems
Subfield: Global and Planetary Change
Field: Environmental Science
Domain: Physical Sciences
SDGs: Life on land, Climate action
Open Access status: diamond
OpenAlex ID: W7117367136