GLM-Nowcast: uma Aplicação Web para Previsão Probabilística de Risco de Relâmpagos em Tempo Real
Resumo
Lightning is one of the most severe natural hazards in Brazil, the country with the highest absolute lightning incidence in the world. This paper presents GLM-Nowcast, a web application that delivers probabilistic, short-term (nowcasting) predictions of lightning risk with lead times of up to 30 minutes, using exclusively data from the Geostationary Lightning Mapper sensor and a spatiotemporal 3D U-Net deep learning model. The tool is implemented as a containerized microservices system (Next.js front-end, Node.js back-end, and a Python inference worker), currently deployed on commodity home-lab hardware and exposed to the public Internet through a Cloudflare Tunnel. We describe the tool’s objectives, architecture, functionalities, and deployment trade-offs, report its computational performance, and discuss its limitations and potential for academic, social, and operational use. The application is publicly available at https://glm.wagnerloch.com.br/.
Referências
John L. Cintineo, Michael J. Pavolonis, and Justin M. Sieglaff. 2022. ProbSevere LightningCast: A Deep-Learning Model for Satellite-Based Lightning Nowcasting. Weather and Forecasting 37, 7 (2022), 1239–1257.
Jean Pierre Dias da Silva, Fernanda Sumika H. Souza, Carlos Henrique Gomes Ferreira, and Vander L. S. Freitas. 2025. DengueMG: Um Dashboard Interativo para o Monitoramento de Dengue em Minas Gerais. In Anais Estendidos do XXXI Simpósio Brasileiro de Sistemas Multimídia e Web (WebMedia). SBC, 151–154.
Nicola Dragoni, Saverio Giallorenzo, Alberto Lluch Lafuente, Manuel Mazzara, Fabrizio Montesi, Ruslan Mustafin, and Larisa Safina. 2017. Microservices: Yesterday, Today, and Tomorrow. In Present and Ulterior Software Engineering. Springer, 195–216.
ELAT/INPE. 2023. Monitoramento de Relâmpagos no Brasil. [link]. Acesso em: fev. 2026.
Steven J. Goodman, Richard J. Blakeslee, William J. Koshak, Douglas Mach, Jeff Bailey, Dennis Buechler, Lawrence Carey, Chris Schultz, Monte Bateman, Eugene McCaul, and Geoffrey Stano. 2013. The GOES-R Geostationary Lightning Mapper (GLM). Atmospheric Research 125-126 (2013), 34–49.
InMeteo. 2026. Ventusky: Weather Forecast Maps. [link]. Acesso em: fev. 2026.
Moser Zeferino Vicente José, Pedro de Almeida Mutter, and Julio Cezar Estrella. 2025. Connected Health in Action: An IoT-Driven Web Dashboard for Real-Time HRV and Stress Monitoring in a Private Cloud. In Proceedings of the Brazilian Symposium on Multimedia and the Web (WebMedia). SBC, 149–157. [link]. org/10.5753/webmedia.2025.16161
LightningMaps. 2026. LightningMaps.org - Real-time Lightning Map. [link]. Acesso em: fev. 2026.
Elyas Mansouri, Amirhossein Mostajabi, Weibin Tong, Marcos Rubinstein, and Farhad Rachidi. 2023. Lightning Nowcasting Using Solely Lightning Data. Atmosphere 14, 12 (2023), 1713.
Dirk Merkel. 2014. Docker: Lightweight Linux Containers for Consistent Development and Deployment. Linux Journal 2014, 239 (2014).
Sannyer Cardoso Carvalho Nery, Fani Tamires de Souza Batista, and Sergio Augusto Coelho Bezerra. 2025. ColabSomGeo: Ferramenta de Monitoramento de Poluição Sonora Subsidiada por Modelo de Aprendizagem. In Anais Estendidos do XXXI Simpósio Brasileiro de Sistemas Multimídia eWeb (WebMedia). SBC, 99–102.
Claus Pahl and Brian Lee. 2015. Containers and Clusters for Edge Cloud Architectures: A Technology Review. In 2015 3rd International Conference on Future Internet of Things and Cloud. 379–386.
Olaf Ronneberger, Philipp Fischer, and Thomas Brox. 2015. U-Net: Convolutional Networks for Biomedical Image Segmentation. In International Conference on Medical Image Computing and Computer-Assisted Intervention. 234–241.
Windy.com. 2026. Windy as forecasted. [link]. Acesso em: fev. 2026.
