Evaluation of Temporal Aggregation in Sea Surface Temperature Forecasting of the Atlantic Ocean
Abstract
Extreme environmental events such as droughts affect millions of people all around the world. Its prediction enables the mitigation of eventual damages caused by its occurrence. An important variable for identifying occurrences of droughts is the Sea Surface Temperature (SST). In the Tropical Atlantic Ocean, SST data are collected and provided by the PIRATA Project, which is an observation network composed of sensor buoys arranged in this region. Sensors of this type, such as Internet of Things (IoT) sensors, commonly fails, leading to data losses that influence the quality of datasets collected for adjusting prediction models. In this paper, we explore the influence of temporal aggregation in predicting step-ahead SST considering different prediction horizons and different sizes for training datasets. Our results point out scenarios for training datasets and prediction horizons indicating whether or not temporal aggregated SST time series may be beneficial for prediction.
References
Cho, J., Yeh, P. J. F., Lee, Y.-W., Kim, H., Oki, T., Kanae, S., Kim, W., e Otsuki, K. (2010). A study on the relationship between Atlantic sea surface temperature and Amazonian greenness. Ecological Informatics, 5(5):367–378.
Durand, B., Servain, J., Laurent, H., e Machado, L. A. T. (2005). Tropical Atlantic Moisture Flux, Convection over Northeastern Brazil, and Pertinence of the PIRATA Network*. Journal of Climate, 18(12):2093–2101.
EM-DAT (2016). The international disasters database.
GOOS-Brasil (2015). PIRATA Dataset. Technical report, [link].
Gujarati, D. (2002). Basic Econometrics. McGraw-Hill/Irwin, Boston; Montreal, 4 edition edition.
Hertig, E. e Jacobeit, J. (2010). Predictability of Mediterranean climate variables from oceanic variability. Part II: Statistical models for monthly precipitation and temperature in the Mediterranean area. Climate Dynamics, 36(5-6):825–843.
Perera, C., Liu, C. H., Jayawardena, S., e Chen, M. (2014). A survey on internet of things from industrial market perspective. IEEE Access, 2:1660–1679.
Salles, R., Assis, L., Guedes, G., Bezerra, E., Porto, F., e Ogasawara, E. (2017). A framework for benchmarking machine learning methods using linear models for univariate time series prediction. In The 2017 International Joint Conference on Neural Networks (IJCNN), pages 1–8.
Salles, R., Bezerra, E., Soares, J., e Ogasawara, E. (2015). Evaluating Linear Models as a Baseline for Time Series Imputation. In XXX Simpósio Brasileiro de Banco de Dados, Petrópolis, RJ.
Salles, R., Mattos, P., Iorgulescu, A.-M. D., Bezerra, E., Lima, L., e Ogasawara, E. (2016). Evaluating temporal aggregation for predicting the sea surface temperature of the Atlantic Ocean. Ecological Informatics, 36:94–105.
Salles, R. P. e Ogasawara, E. (2015). TSPred: Functions for Baseline-Based Time Series Prediction. Technical report, [link].
Shumway, R. H. e Stoffer, D. S. (2010). Time Series Analysis and Its Applications: With R Examples. Springer, New York, 3rd ed. 2011 edition edition.
Teixeira, P. M., Dubois-Iorgulescu, A.-M., Salles, R., Bezerra, E., Ogasawara, E., e Lima, L. S. d. (2015). Uso de Redes Neurais para Previsão da Temperatura da Superfície do Mar do Oceano Atlântico Tropical. In IX Brazilian e-Science Workshop, Recife, PE.
Tiao, G. C. (1972). Asymptotic behaviour of temporal aggregates of time series. Biometrika, 59(3):525–531.
Wei, W. W. S. (2005). Time Series Analysis : Univariate and Multivariate Methods. Pearson, Boston, 2 edition edition.
