A0174
Title: Forecasting ENSO: A frequency band factor model approach
Authors: Alessandro Giovannelli - University of L'Aquila (Italy) [presenting]
Tommaso Proietti - University of Roma Tor Vergata (Italy)
Abstract: The El Nino-Southern Oscillation (ENSO) is a major driver of interannual climate variability, with significant impacts on global climate patterns. The aim is to introduce a novel forecasting methodology based on dynamic factor models, where the unobserved factors are estimated separately for two distinct frequency bands: One capturing ENSO-related medium-to-long-term fluctuations (17 years), and another isolating short-term variability. The rationale for this frequency decomposition is that separating signals by frequency enhances medium-term predictive information, while still retaining potentially useful information from short-term fluctuations. Once the factors are estimated, independently for each frequency band, they are incorporated into a sparse regression forecasting model, which dynamically selects the most predictive components. An extensive real-time forecasting application using over 2000 climate variables illustrates the potential benefits of this frequency-specific approach for ENSO prediction.
