FurkanBaris/antalya-manavgat-wildfire-analysis-2021

Spatial Analysis of 2021 Manavgat Wildfire Hotspots

This repository contains a comprehensive spatial analysis of the 2021 Manavgat (Turkey) forest fire hotspots. The project focuses on evaluating the performance of the FN-DBSCAN (Fuzzy Neighborhood DBSCAN) algorithm and its computational variants in identifying fire clusters from high-resolution satellite data.

Project Overview

In late July 2021, Turkey experienced one of its most devastating wildfire seasons. This project utilizes NASA VIIRS (Visible Infrared Imaging Radiometer Suite) satellite data to analyze the thermal anomalies (hotspots) during the Manavgat fire.

The core of this project is the implementation and comparison of four density-based clustering approaches:

1. 1.FN-DBSCAN: Standard Fuzzy Neighborhood density-based clustering.
2. 2.AFN-DBSCAN (Approximate): Optimized for speed using approximate neighborhood calculations.
3. 3.SFN-DBSCAN (Sampled): Designed for large-scale datasets through strategic sampling.
4. 4.Landmark FN-DBSCAN: Uses representative landmark points to maintain accuracy while reducing complexity.

Dataset Features

The cleaned dataset (cleaned_turkey_wildfires_2021.csv) includes the following key features:

• —latitude & longitude: Geospatial coordinates of the hotspots.
• —brightness: Brightness temperature (Kelvin).
• —bright_t31: Channel 31 brightness temperature.
• —frp: Fire Radiative Power (MW), indicating the fire's intensity.

Visualizing Clusters

The analysis generates comparative visualizations to show how different algorithms perceive the fire's spread.

[image]

Key Findings

The implementation highlights the trade-off between computational efficiency and clustering sensitivity:

• —FN-DBSCAN provides the most granular clusters but requires higher processing time.
• —Landmark-based and Approximate versions significantly reduce execution time (up to 70-80%) while maintaining the general structure of the fire clusters.

Installation & Usage

To replicate the analysis, ensure you have the required Python libraries installed:

pip install numpy pandas matplotlib scikit-learn
python main.py