Relating South Pacific Atmospheric Internal Variability to ENSO Predictability and Forecast Skill in Operational Subseasonal-to-Seasonal Models

While substantial advances in forecasting the El Niño-Southern Oscillation (ENSO) phenomenon have occurred in recent decades, skillful long-lead (i.e., 6-9 month) predictions of the strength and flavor (i.e., eastern vs. central Pacific) of ENSO events remains an open challenge. Past studies have considered several extratropical and tropical precursors in the atmosphere and ocean to improve the skill of these forecasts, but operational subseasonal-to-seasonal (S2S) models struggle with these specifics of ENSO. Emerging studies suggest that South Pacific atmospheric internal variability during the late austral fall and winter may hold clues to the evolution and subsequent strength of an ensuing ENSO event, but this has yet to be explicitly tested.

This presentation illustrates that austral winter (June-August; JJA) South Pacific atmospheric internal variability associated with the South Pacific Oscillation (SPO), a meridional dipole of opposing sea level pressure anomalies in the subtropical and extratropical Southeast Pacific Ocean, contributes significantly to the structure and strength of an ENSO event peaking 4-6 months later (i.e., the following November – January; NDJ(+1)). We employ both reanalysis and hindcasts from several operational S2S models (and their ensemble members) initialized in late austral fall to test this hypothesis. Our findings indicate that while tropical Pacific sea surface temperature (SST) anomalies have a remote and concurrent influence on the extratropical South Pacific atmosphere, variability in the SPO and specifically the strength of the South Pacific subtropical high (i.e., the northern node of the SPO) is highly intrinsic. This intrinsic component of the SPO (denoted as SPO’) generates stochastic zonal wind stresses in the equatorial Pacific that influence ENSO development during JJA by either promoting or inhibiting Kelvin wave propagation and consequently the activation of the Bjerknes feedback. Sensitivity analyses with the multiple S2S models and their ensemble members illustrate that the SPO’ strongly regulates anomalies in subsurface temperature fields in the eastern and central equatorial Pacific through suppressed upwelling and possibly zonal advection feedbacks. As such, we illustrate that the SPO’ is the source of much of the ensemble-spread in NDJ(+1) forecasts of tropical Pacific SST anomalies in both magnitude and where the peak anomaly occurs. As such, we argue that better simulations and monitoring of the SPO in the models could be a significant source to benchmarking operational model performance for ENSO forecast skill. Specific cases and an application to the 2018-19 season are also discussed.