ENSO Dynamics, Trends, and Prediction Using Machine Learning

Abstract

The El Niño - Southern Oscillation (ENSO) has global effects on the hydrological cycle, agriculture, ecosystems, health, and society. We present a novel non-homogeneous Hidden Markov model (NHMM) for studying the underlying dynamics of sea surface temperature anomalies (SSTA) over the region, 150E -80W, 15N-15S from Jan-1856 to Dec-2019, using the monthly SSTA data from the Kaplan Extended SST v2 product. This non-parametric Machine Learning scheme dynamically simulates and predicts the spatiotemporal evolution of ENSO patterns, including their asymmetry, long-term trends, persistence, and seasonal evolution. The model identifies five hidden states whose spatial SSTA patterns are similar to the so-called “ENSO Flavors” in the literature. From the fitted NHMM, the model shows that there are systematic trends in the frequency and persistence of the regimes over the last 160 years that may be related to changes in the mean state of basin temperature and/or global warming. We evaluated the ability of NHMM to make out of sample probabilistic predictions of the spatial structure of temperature anomalies for the period 1995-2016 using a training period from Jan-1856 to Dec-1994. The results show that NHMMs can simulate the behavior of the Niño 3.4 and Niño 1.2 regions quite well. The NHMM results over this period are comparable or superior to the commonly available ENSO prediction models, with the additional advantage of directly providing insights as to the space patterns, seasonal and longer-term trends of the SSTA in the equatorial Pacific region.

Hidden Tropical Pacific Sea Surface Temperature States Reveal Global Predictability for Monthly Precipitation for Sub‐Season to Annual Scales

“…Through the analysis using the NHMM (see Holsclaw et al., 2017; Rojo‐Hernandez et al., 2020, and Supporting Information for details of the construction of NHMM and parameter estimation), five hidden states (Figure S1 in Supporting Information ) are identified. Hidden states one–five are similar to the classical La Nina pattern, mild La Nina pattern, neutral pattern, Modoki ENSO pattern and classical El Niño pattern, respectively.…”

A deep learning model for forecasting global monthly mean sea surface temperature anomalies

“…The CPC consolidated predictions are 3-month averages for up to a 9-month lead. Note that the consolidated prediction data are for 1-9 months leads and are only available after 2015. Rojo Hernández et al (2020 provided a comprehensive analysis of the forecasting skills of various models archived on the CPC website.…”

“…Similarly, an artificial neural network model has been trained to forecast the SST variations at the peak season of the IOD events (Ratnam et al, 2020). Rojo Hernández et al (2020) used a nonhomogeneous hidden Markov model to achieve superior prediction skills of Nino3.4 SST variability compared to dynamic forecasting models at up to a 9-month lead time. Given the phase-locking characteristics of the climate modes, these models aim to make single-season predictions, and for a single climate index.…”

Climate Change Trends in Colombia: A Case Study in Facatativá, Cundinamarca

“…Even though a trend of increasing precipitation might imply a greater availability of water, this has not occurred over the years. Rather, the phenomena that modify average conditions have meant that the effects of the phenomena related to the El Niño Southern Oscillation [3,24] have resulted in intense and more frequent emergencies related to floods, forest fires, and landslides, which have affected this municipality and Colombia [25]; therefore, a disproportionate distribution of precipitation facilitates the development of extreme water deficits in some seasons and excess water in others.…”