IoT System Based on LoRa WAN for Early Wildfire Detection and Prediction

Abstract

Wildfires pose increasing threats to forest ecosystems, biodiversity, and local communities, particularly in the context of climate change. This paper presents an IoT-based environmental monitoring system for early wildfire detection and prediction, leveraging LoRa WAN technology. The proposed architecture consists of sensor nodes measuring temperature, humidity, luminosity, and GPS coordinates, which transmit data via LoRa gateways connected to a Raspberry Pi. Data streams are processed through a Node-RED workflow, stored in an Influx DB time-series database, and visualized in real-time using Grafana. Historical environmental data obtained from the High Commission for Water and Forests, combined with sensor measurements, were used to train a machine learning model for wildfire prediction. The system achieved over 95% transmission reliability and demonstrated good predictive capacity for early fire risk assessment. This solution combines long-range communication, real-time data supervision, and predictive analytics to support decision-making for the sustainable protection of Moroccan forest ecosystems. Future work will focus on integrating cybersecurity-oriented AI models to enhance system robustness against potential attacks.