Complexity-based approach for El Niño magnitude forecasting before the spring predictability barrier

Abstract

The El Niño Southern Oscillation (ENSO) is one of the most prominent interannual climate phenomena. An early and reliable ENSO forecasting remains a crucial goal, due to its serious implications for economy, society, and ecosystem. Despite the development of various dynamical and statistical prediction models in the recent decades, the "spring predictability barrier" (SPB) remains a great challenge for long (over 6-month) lead-time forecasting. To overcome this barrier, here we develop an analysis tool, the System Sample Entropy (SysSampEn), to measure the complexity (disorder) of the system composed of temperature anomaly time series in the Niño 3.4 region. When applying this tool to several near surface air-temperature and sea surface temperature datasets, we find that in all datasets a strong positive correlation exists between the magnitude of El Niño and the previous calendar year's SysSampEn (complexity). We show that this correlation allows to forecast the magnitude of an El Niño with a prediction horizon of 1 year and high accuracy (i.e., Root Mean Square Error = 0.23 • C for the average of the individual datasets forecasts). For the on-going 2018 El Niño event, our method forecasts a weak El Niño with a magnitude of 1.11 ± 0.23 • C. Our framework presented here not only facilitates a long-term forecasting of the El Niño magnitude but can potentially also be used as a measure for the complexity of other natural or engineering complex systems. arXiv:1909.03789v1 [physics.ao-ph] 9 Sep 2019 2/12

Statistical physics approaches to the complex Earth system

“…There exist also many other types of entropy, such as Gibbs, Residual, Approximate, Sinai-Kolmogorov, Sample, Multiscale. Entropy has been proven useful in many real-world systems, including analysis of DNA sequences [233] , cosmology and astrophysics [234] , [235] , [236] , economics [237] , [238] , and climate systems [239] , [240] , [241] . Each definition of entropy could give better results for some systems but fails for others.…”

“…proposed the so called and applied it to study the climate system. Based on the , they could measure the complexity (disorder) of a system composed of temperature anomaly time series and forecast the magnitude of an El Niñoevent with a prediction horizon of year and high accuracy [241] .…”

Classifying El Niño-Southern Oscillation Combining Network Science and Machine Learning

“…In the case of the El Niño phenomenon, some works focused on understanding the effects of the ENSO phases around the world [13], [14], [24] and possible connections with other climate or different phenomena [1]- [3], [12], [24], [25]. In terms of ML applications, several works seek to predict or forecast long-term ENSO episodes [15]- [19], [26], [27]. They select a group of spatiotemporal datasets and information to feed the forecasting model and perform some regression analyses.…”

“…We used the global daily near-surface (1000 hPa) Air Temperature data (SAT), from the National Oceanic and Atmospheric Administration (NOAA) and the Reanalysis I project of the National Center for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) [47]. SAT has been used by several authors to study and analyze the ENSO phenomenon [26], [33], [35]. We opt for using this dataset instead of the SST due to the inclusion of the land areas, which is extra information that we consider important for the analyses.…”

“…Initially, we divide the spatiotemporal SATa dataset into overlapping windows of t days for all the spatial points. Given that the complete climatic seasonal cycle of the planet is one year (four seasons) and the interannual nature of the ENSO, we set t = 365 days [26]. The sliding window is moved every month, in which the reference is the last month in the interval, e.g., ( The temporal CN (G) is the ensemble of CNs from all the months in the dataset.…”