Remote Sensing of Monitoring Forest Fire Dynamics in Near/Ultra Real Time
Date
2024-12-04
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Forest fires cause extensive environmental damage, significant carbon emissions, and economic losses, posing major challenges for disaster management and emergency response. Accurate and timely delineation of fire perimeters is essential for minimizing these impacts. The overall aim of this study was to develop an early forest fire monitoring model that enables automatic clustering of active fire points and precise delineation of fire perimeters using satellite-based active fire data from VIIRS and MODIS datasets. Various algorithms—buffer, concave, convex, and combination methods—are assessed for their efficiency in near real-time (NRT), real-time (RT), and ultra-real-time (URT) fire perimeter delineation. Additionally, this study also introduced the automated Timely Active Fire Progression (TAFP) model for clustering and calculating active fire perimeter consistently. The results indicate that increasing the concave α (alpha) values (e.g., 0.1 to 0.5) enhances the matching percentage with ground fire areas but also leads to higher commission errors, signifying a risk of overestimation. Combination methods achieved the highest matching percentages but were also associated with increased commission errors. However, the TAFP model demonstrated an 85.13% matching rate for fire perimeters across various size classes, with a 95.95% clustering accuracy for fires larger than 100 hectares. This research was the first to scientifically evaluate multiple algorithms, both individually and synergistically, using NRT/RT/URT active fire data. The insights gained underscore the trade-offs between enhancing perimeter accuracy and avoiding overestimation, providing a critical foundation for optimizing sensor data alignment techniques. By improving the timeliness and accuracy of forest fire monitoring, the proposed models can significantly bolster operational responses by fire management agencies, aiding in resource allocation, evacuation planning, and mitigation strategies. This study lays the groundwork for future advancements in automated forest fire perimeter assessment, essential for effective disaster management and rapid response efforts.
Description
Keywords
forest fire, remote sensing, active Fire, fire monitoring, wildfire, VIIRS, MODIS, satellite, gis, model
Citation
Bhuian, M. (2024). Remote sensing of monitoring forest fire dynamics in near/ultra real time (Master's thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca.