Wildfire Monitor: Real-Time Prediction and Interpretable Alerts from Fused Satellite and Meteorological Data

Luís Fabrício de Freitas Souza; Guilherme F. B. Severiano; José Jerovane da C. Nascimento; Pedro Hugo Ursulino Fernandes; Ícaro de Sousa Rodrigues; Jesus Ossian Cunha Silva; Francisco Italo G. da Silva; Osvaldo Soares Landim Junior

doi:10.5753/sbsi.2026.248736

Luís Fabrício de Freitas Souza UFCA / LISIA
Guilherme F. B. Severiano IFCE / LISIA
José Jerovane da C. Nascimento UFC / LISIA
Pedro Hugo Ursulino Fernandes UFCA / Instituto Atlântico / LISIA
Ícaro de Sousa Rodrigues UFC / LISIA
Jesus Ossian Cunha Silva UFC
Francisco Italo G. da Silva UFC / Instituto Atlântico / LISIA
Osvaldo Soares Landim Junior UFCA / IFCE / LISIA

DOI: https://doi.org/10.5753/sbsi.2026.248736

Resumo

Research Context: Wildfires cause significant environmental, economic, and social damage, requiring intelligent systems capable of supporting prevention and mitigation actions. Scientific and/or Practical Problem: The increasing frequency of wildfires in Brazil and worldwide demands innovative technological approaches for real-time monitoring and accurate prediction of fire outbreaks using reliable and official data sources. Proposed Solution and/or Analysis: This paper presents the Cariri Wildfire Monitor, a system for mapping and monitoring wildfires through a web portal that integrates Artificial Intelligence (AI) and official datasets. The proposal includes a technological solution for continuous monitoring and a predictive module for fire hotspots. Related IS Theory: The system design is grounded in Information Systems (IS) theories related to data-driven decision support, predictive analytics, and socio-environmental information integration for sustainable management. Research Method: The system performs data fusion from INPE and INMET sources, applies regression models to predict wildfire hotspots, and integrates a decision layer with Large Language Models (LLMs) to produce auditable alerts and qualitative insights. Summary of Results: Among the results, the ExtraTrees regressor achieved R²=0,9843. In the stratified audit of the LLM module with live threshold 0,758 (proxy 0,500; N_pop=1000; M_audit used = 12), we obtained Accuracy (w) 0,7973±0,1165, and p95 latencies between 4,75s and 5,95s. Contributions and Impact to IS area: The system demonstrates the potential of predictive computational models and LLMs in environmental monitoring, contributing to the development of intelligent, auditable, and sustainable Information Systems solutions for wildfire management.

Referências

da Costa Nascimento, J. J., Marques, A. G., do Nascimento Souza, L., de Mattos Dourado, C. M. J., da Silva Barros, A. C., de Albuquerque, V. H. C., de Freitas Sousa, L. F., et al. (2025). A novel generative model for brain tumor detection using magnetic resonance imaging. Computerized Medical Imaging and Graphics, 121:102498.

Da Silva, P. M., Lima, M. N., Soares, W. L., Silva, I. R., Fagundes, R. A. d. A., and De Souza, F. F. (2019). Ensemble regression models applied to dropout in higher education. In 2019 8th Brazilian conference on intelligent systems (BRACIS), pages 120–125. IEEE.

de Andrades, R. S., Grellert, M., and Fonseca, M. B. (2019). Hyperparameter tuning and its effects on cardiac arrhythmia prediction. In 2019 8th Brazilian conference on intelligent systems (BRACIS), pages 562–567. IEEE.

Mantovani, R. G., Horváth, T., Cerri, R., Vanschoren, J., and De Carvalho, A. C. (2016). Hyper-parameter tuning of a decision tree induction algorithm. In 2016 5th Brazilian Conference on Intelligent Systems (BRACIS), pages 37–42. IEEE.

MapBiomas (2025). Área queimada no brasil cresce 79% em 2024 e supera os 30 milhões de hectares. Acesso em: 28 abr. 2025.

Marques, A. G., de Figueiredo, M. V. C., da Costa Nascimento, J. J., de Souza, C. T., de Mattos Dourado, C. M. J., de Albuquerque, V. H. C., de Freitas Souzal, L. F., et al. (2024). New approach generative ai melanoma data fusion for classification in dermoscopic images with large language model. In 2024 37th SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI), pages 1–6. IEEE.

Melo, M. A. d. (2024). Desenvolvimento de um plugin no qgis para tratamento de dados disponibilizados pelo inmet.

Mohnish, S., Kannan, B. D., Vasuhi, S., et al. (2023). Vision transformer based forest fire detection for smart alert systems. In 2023 2nd International Conference on Applied Artificial Intelligence and Computing (ICAAIC), pages 891–896. IEEE.

Murilo, L., Oliveira, G., and Martins, L. (2024). Evolutionary adjustment of a cellular automata-basedmodel for wildfire spreading. In Anais da XXXIV Brazilian Conference on Intelligent Systems, pages 260–275, Porto Alegre, RS, Brasil. SBC.

Nieuwenhuijsen, M. J. (2024). Climate crisis, cities, and health. The Lancet, 404(10463):1693–1700.

Park, M., Jeon, Y., Bak, J., Park, S., et al. (2022). Forest-fire response system using deep-learning-based approaches with cctv images and weather data. IEEE access, 10:66061–66071.

Peng, W., Wei, Y., Chen, G., Lu, G., Ye, Q., Ding, R., Hu, P., and Cheng, Z. (2023). Analysis of wildfire danger level using logistic regression model in sichuan province, china. Forests, 14(12):2352.

Povo, O. (2023). Incêndios florestais geram nuvem de fumaça em cidades do cariri. Acesso em: 28 abr. 2025.

Severiano, G. F. B., Marques, A. G., Nascimento, J. J. d. C., and Rodrigues, Y. O. A. (2024). Alpr system perspective adjustment: New automatic license plate. Intelligent Systems Design and Applications: Real World Applications, Volume 5, 1050:295.

Silva, A. P., Genaro, A. F., and Branco, R. H. (2024). A contribuição de sistemas de informação para o processo de gestão do portfólio de projetos e programas do inpe. In Workshop de Computação Aplicada em Governo Eletrônico (WCGE), pages 210–221. SBC.

Sousa, M. J., Moutinho, A., and Almeida, M. (2019). Classification of potential fire outbreaks: A fuzzy modeling approach based on thermal images. Expert systems with applications, 129:216–232.

Suklabaidya, S. and Das, I. (2023). Processing iot sensor fire dataset using machine learning techniques. In 2023 International Conference on Intelligent Systems, Advanced Computing and Communication (ISACC), pages 1–7. IEEE.

Umoh, U. A., Eyoh, I. J., Murugesan, V. S., and Nyoho, E. E. (2022). Fuzzy-machine learning models for the prediction of fire outbreaks: A comparative analysis. In Artificial intelligence and machine learning for EDGE computing, pages 207–233. Elsevier.

Viganó, H. H. d. G., Souza, C. C. d., Reis Neto, J. F., Cristaldo, M. F., and Jesus, L. d. (2018). Prediction and modeling of forest fires in the pantanal. Revista Brasileira de Meteorologia, 33(2):306–316.

Weisse, M. and Goldman, E. (2023). Latest analysis of deforestation trends. Accessed on: March 13, 2025.

Xu, R., Lin, H., Lu, K., Cao, L., and Liu, Y. (2021). A forest fire detection system based on ensemble learning. Forests, 12(2):217.

Zhang, J., Cai, S., Jiang, Z., Xiao, J., and Ming, Z. (2024). Firerobbrain: Planning for a firefighting robot using knowledge graph and large language model. In 2024 10th IEEE International Conference on Intelligent Data and Security (IDS), pages 37–41. IEEE.