Imputação Multivariada de Séries Temporais Meteorológicas do INMET com Reanálise ERA5-Land e Self-Attention
Resumo
Valores ausentes em séries temporais meteorológicas comprometem análises climáticas e modelos preditivos. Este trabalho investiga um pipeline de imputação multivariada para dados observacionais do INMET, utilizando o ERA5-Land como fonte de atributos auxiliares. O SAITS foi adotado como modelo principal e comparado com Median, LOCF, MissForest, XGBoost, Transformer e ImputeFormer. Foram criados cenários MCAR de 10% a 80% de ausência artificial, com avaliação por MAE, RMSE e MASE. Os resultados mostram que o SAITS apresenta melhor desempenho em cenários de baixa ausência, enquanto o XGBoost obtém os menores erros globais de 40% a 80% de mascaramento, indicando maior robustez sob degradação severa dos dados.Referências
Alabadla, M., Sidi, F., Ishak, I., Ibrahim, H., Affendey, L. S., Che Ani, Z., Jabar, M. A., Bukar, U. A., Devaraj, N. K., Muda, A. S., Tharek, A., Omar, N., and Jaya, M. I. M. (2022). Systematic Review of Using Machine Learning in Imputing Missing Values. IEEE Access, 10:44483–44502.
Alejo-Sanchez, L. E., Márquez-Grajales, A., Salas-Martínez, F., Franco-Arcega, A., López-Morales, V., Acevedo-Sandoval, O. A., González-Ramírez, C. A., and Villegas-Vega, R. (2025). Missing data imputation of climate time series: A review. MethodsX, 15:103455. 1Código-fonte disponível em: [link]
Araújo, C. S. P. D., Silva, I. A. C. E., Ippolito, M., and Almeida, C. D. G. C. D. (2022). Evaluation of air temperature estimated by ERA5-Land reanalysis using surface data in Pernambuco, Brazil. Environmental Monitoring and Assessment, 194(5):381.
Chen, T. and Guestrin, C. (2016). XGBoost: A Scalable Tree Boosting System. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD ’16, pages 785–794, New York, NY, USA. Association for Computing Machinery.
Du, W., Côté, D., and Liu, Y. (2023). SAITS: Self-attention-based imputation for time series. Expert Systems with Applications, 219:119619.
Engels, J. (2003). Imputation of missing longitudinal data: a comparison of methods. Journal of Clinical Epidemiology, 56(10):968–976.
Instituto Nacional de Meteorologia (2024). Sobre o instituto nacional de meteorologia. Acesso em: 28 nov. 2025.
Lachin, J. M. (2016). Fallacies of last observation carried forward analyses. Clinical Trials, 13(2):161–168.
Lien, P. L., Do, T. T., and Nguyen, T. (2023). Data Imputation for Multivariate Time-series Data. In 2023 15th International Conference on Knowledge and Systems Engineering (KSE), pages 1–6. ISSN: 2694-4804.
Muñoz-Sabater, J., Dutra, E., Agustí-Panareda, A., Albergel, C., Arduini, G., Balsamo, G., Boussetta, S., Choulga, M., Harrigan, S., Hersbach, H., Martens, B., Miralles, D. G., Piles, M., Rodríguez-Fernández, N. J., Zsoter, E., Buontempo, C., and Thépaut, J.-N. (2021). ERA5-Land: a state-of-the-art global reanalysis dataset for land applications. Earth System Science Data, 13(9):4349–4383. Publisher: Copernicus GmbH.
Nie, T., Qin, G., Ma, W., Mei, Y., and Sun, J. (2024). ImputeFormer: Low Rankness-Induced Transformers for Generalizable Spatiotemporal Imputation. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD ’24, pages 2260–2271, New York, NY, USA. Association for Computing Machinery.
Stekhoven, D. J. and Bühlmann, P. (2012). MissForest—non-parametric missing value imputation for mixed-type data. Bioinformatics, 28(1):112–118.
Troyanskaya, O., Cantor, M., Sherlock, G., Brown, P., Hastie, T., Tibshirani, R., Botstein, D., and Altman, R. B. (2001). Missing value estimation methods for DNA microarrays. Bioinformatics, 17(6):520–525.
Varotsos, K. V., Katavoutas, G., and Giannakopoulos, C. (2023). On the Use of Reanalysis Data to Reconstruct Missing Observed Daily Temperatures in Europe over a Lengthy Period of Time. Sustainability, 15(9):7081.
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., and Polosukhin, I. (2017). Attention Is All You Need.
Zainuddin, A., Hairuddin, M. A., Yassin, A. I. M., Latiff, Z. I. A., and Azhar, A. (2022). Time Series Data and Recent Imputation Techniques for Missing Data: A Review. In 2022 International Conference on Green Energy, Computing and Sustainable Technology (GECOST), pages 346–350.
Alejo-Sanchez, L. E., Márquez-Grajales, A., Salas-Martínez, F., Franco-Arcega, A., López-Morales, V., Acevedo-Sandoval, O. A., González-Ramírez, C. A., and Villegas-Vega, R. (2025). Missing data imputation of climate time series: A review. MethodsX, 15:103455. 1Código-fonte disponível em: [link]
Araújo, C. S. P. D., Silva, I. A. C. E., Ippolito, M., and Almeida, C. D. G. C. D. (2022). Evaluation of air temperature estimated by ERA5-Land reanalysis using surface data in Pernambuco, Brazil. Environmental Monitoring and Assessment, 194(5):381.
Chen, T. and Guestrin, C. (2016). XGBoost: A Scalable Tree Boosting System. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD ’16, pages 785–794, New York, NY, USA. Association for Computing Machinery.
Du, W., Côté, D., and Liu, Y. (2023). SAITS: Self-attention-based imputation for time series. Expert Systems with Applications, 219:119619.
Engels, J. (2003). Imputation of missing longitudinal data: a comparison of methods. Journal of Clinical Epidemiology, 56(10):968–976.
Instituto Nacional de Meteorologia (2024). Sobre o instituto nacional de meteorologia. Acesso em: 28 nov. 2025.
Lachin, J. M. (2016). Fallacies of last observation carried forward analyses. Clinical Trials, 13(2):161–168.
Lien, P. L., Do, T. T., and Nguyen, T. (2023). Data Imputation for Multivariate Time-series Data. In 2023 15th International Conference on Knowledge and Systems Engineering (KSE), pages 1–6. ISSN: 2694-4804.
Muñoz-Sabater, J., Dutra, E., Agustí-Panareda, A., Albergel, C., Arduini, G., Balsamo, G., Boussetta, S., Choulga, M., Harrigan, S., Hersbach, H., Martens, B., Miralles, D. G., Piles, M., Rodríguez-Fernández, N. J., Zsoter, E., Buontempo, C., and Thépaut, J.-N. (2021). ERA5-Land: a state-of-the-art global reanalysis dataset for land applications. Earth System Science Data, 13(9):4349–4383. Publisher: Copernicus GmbH.
Nie, T., Qin, G., Ma, W., Mei, Y., and Sun, J. (2024). ImputeFormer: Low Rankness-Induced Transformers for Generalizable Spatiotemporal Imputation. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD ’24, pages 2260–2271, New York, NY, USA. Association for Computing Machinery.
Stekhoven, D. J. and Bühlmann, P. (2012). MissForest—non-parametric missing value imputation for mixed-type data. Bioinformatics, 28(1):112–118.
Troyanskaya, O., Cantor, M., Sherlock, G., Brown, P., Hastie, T., Tibshirani, R., Botstein, D., and Altman, R. B. (2001). Missing value estimation methods for DNA microarrays. Bioinformatics, 17(6):520–525.
Varotsos, K. V., Katavoutas, G., and Giannakopoulos, C. (2023). On the Use of Reanalysis Data to Reconstruct Missing Observed Daily Temperatures in Europe over a Lengthy Period of Time. Sustainability, 15(9):7081.
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., and Polosukhin, I. (2017). Attention Is All You Need.
Zainuddin, A., Hairuddin, M. A., Yassin, A. I. M., Latiff, Z. I. A., and Azhar, A. (2022). Time Series Data and Recent Imputation Techniques for Missing Data: A Review. In 2022 International Conference on Green Energy, Computing and Sustainable Technology (GECOST), pages 346–350.
Publicado
19/07/2026
Como Citar
DIAS, Davi Aquila de Souza; DANTAS, Maria José Pereira.
Imputação Multivariada de Séries Temporais Meteorológicas do INMET com Reanálise ERA5-Land e Self-Attention. In: SEMINÁRIO INTEGRADO DE SOFTWARE E HARDWARE (SEMISH), 53. , 2026, Gramado/RS.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2026
.
p. 956-961.
ISSN 2595-6205.
DOI: https://doi.org/10.5753/semish.2026.23855.
