News from ISGlobal
Although a number of operational climate models are capable of predicting El Niño events, they cannot perform long-term forecasts more than half a year in advance. Now, a team from the Barcelona Institute for Global Health (ISGlobal) has developed a new statistical climate model able for the first time to predict El Niño episodes up to two-and-a-half years in advance.
The El Niño Southern Oscilation (ENSO) is a climate phenomenon that represents a variation of atmospheric and oceanic features over the equatorial Pacific. It occurs every 2-7 years, but has an irregular periodicity.
The study, published in The Journal of Climate, uses a series of predictor variables including sea temperatures at different depths, as well as winds in the tropical Pacific, in a flexible statistical dynamic components model to make retrospective predictions of ENSO events between 1970 and 2016 . The model is capable of predicting all the major El Niño episodes that occurred within that period, including the extreme event of 2015-2016, up to two-and-a-half years in advance.
The computational tool developed in this study is an improved version of a statistical dynamic components model already proposed two years ago by the same ISGlobal researchers. Desislava Petrova, first author of the two studies, says that this is an important advance in the area of climate sciences and ENSO research.
Desislava Petrova, Joan Ballester, Siem Jan Koopman, Xavier Rodó. Multi-year statistical prediction of ENSO enhanced by the Tropical Pacific Observing System. The Journal of Climate, October 2019. https://doi.org/10.1175/JCLI-D-18-0877.1