Investigating decadal variations of the seasonal predictability limit of sea surface temperature in the tropical Pacific

Abstract El Niño and the Southern Oscillation (ENSO) have a worldwide impact on seasonal to yearly climate. However, there are decadal variations in prediction skill of ENSO dynamical statistical models; particular, has declined since 2000. The shortcomings models mean that it is very important study predictability its variation using observational/reanalysis data. Here we quantitatively estimate limit (PL) from 1900 2015 Nonlinear local Lyapunov exponent (NLLE) theory with an dataset explore variations. PL sea surface temperature (SST) high central/eastern tropical Pacific low western Pacific, reaching 12–15 7–8 months, respectively. varies timescale, interdecadal standard deviation up 2 months central similar spatial structure PL. Taking SST 3.4 region as representative clearly higher values 1900s, mid-1930s, mid-1960s, mid-1990s, lower 1920s, mid-1940s, mid-2010s. Meanwhile, 6 region—whose average value 7 months—is good agreement most regions 1910s, 1940s, 1980s 1930s, 1950s, mid-1990s. In framework NLLE theory, determined by error growth rate (representing dissipation predictable signal) saturation relative signal intensity). We reveal mainly rate. variability affected more equatorial whereas cannot be ignored some regions. As source dynamics, relationship between warm water volume critical role through error. This strong reduces initial period increases saturated error, contributing