Our physical-based ENSO prediction model (EPM) employs a multivariate linear regression framework utilizing four key predictors:
Our EPM allows for the evaluation of individual factor contributions to ENSO events by analyzing different combinations of predictors. The model has been trained on data from 1982 to the most recent year of observations. For more information, please see our publications below.
Niño3 (orange line) and Niño4 (yellow line) indices using the same approach previously. Note that the North and South hemisphere of EPISLP are used separately.
SST patterns reconstructed using a multivariate linear regression model based on ERSST and the Niño3 and Niño4 indices from different months. Future SST patterns are derived from our predicted Niño3 and Niño4 indices using this regression model.
Niño3.4 (black line) and Niño1+2 (gray line) indices derived from OISST (climo.: 1991-2020), 3-month running mean
EPISST, EPIOA and EPISLP. EPISLP includes the North Hemisphere (NH, solid line) and South Hemisphere (SH, dashed line) partitions.
EPIWWV and EPIOA.
All variables are averaged between 5°S and 5°N.
All variables are averaged between 5°S and 5°N.
Correlation coefficients between 3-month running mean observed and predicted Niño3.4 index.
Chen, H. C., Tseng, Y. H., Huang, J. H., & Juang, P. H. (2025). Understanding the driving mechanisms behind triple-dip La Niñas: insights from the prediction perspective. npj Climate and Atmospheric Science, 8(1), 143. [doi link]
Tseng, Y. H., Huang, J. H., & Chen, H. C. (2022). Improving the predictability of two types of ENSO by the characteristics of extratropical precursors. Geophysical Research Letters, 49(3), e2021GL097190. [doi link]
Chen, H. C., Tseng, Y. H., Hu, Z. Z., & Ding, R. (2020). Enhancing the ENSO predictability beyond the spring barrier. Scientific Reports, 10(1), 984. [doi link]
Tseng, Y. H., Hu, Z. Z., Ding, R., & Chen, H. C. (2017). An ENSO prediction approach based on ocean conditions and ocean–atmosphere coupling. Climate Dynamics, 48, 2025-2044. [doi link]