Exploring nonstationarity of ENSO teleconnections using state-of-the-art Linear Inverse Models

The El Niño-Southern Oscillation (ENSO) is the leading source of predictable, internally generated, large-scale climate variability. Centered in the tropical Pacific, ENSO influences global precipitation, atmospheric circulation, and temperature patterns via atmospheric teleconnections. ENSO events can trigger climate extremes, in turn devastating communities and costing billions of dollars. Over recent decades, substantial progress has been made in early and accurate seasonal forecasting of ENSO events. However, relatively less is known about how ENSO teleconnections vary in space and time, so called nonstationarity, which limits our ability to confidently relate these forecasts to expected impacts.

Assessing ENSO teleconnection nonstationarity is challenging because the instrumental record is relatively short. Comprehensive physical climate models help to address these limitations, but intrinsic biases undermine their utility for this purpose. By contrast, statistical climate models are trained on observations and can therefore provide a valuable complementary perspective.

Linear Inverse Models (LIMs) are efficient, linear statistical climate models that are computationally inexpensive and straightforward to modify. Here we develop state-of-the-art LIMs that simulate ENSO asymmetry and diversity (Martinez-Villalobos et al., 2025), the seasonal cycle (Shin et al., 2021), and ENSO’s teleconnected impacts in the extratropics (Ault et al. 2018). Utilizing the LIMs we provide the most confident estimates to-date of intrinsic nonstationarity in ENSO teleconnections. Furthermore, by altering the LIM we assess the role for various processes in driving nonstationarity, with a particular focus on the influence of ENSO asymmetry, the seasonal cycle and phase locking, and inter-basin interactions. Our results have implications for seasonal forecasting, characterizing ENSO impacts in a changing climate, and validating the comprehensive physical climate models that are the basis of future projections.