Predictability and Prediction of Southern California Rains during Strong El Niño Events: A Focus on the Failed 2016 Winter Rains

The failed Southern California (SCAL) winter rains during the 2015/16 strong El Nino came as a surprise and a disappointment. Similarities were drawn to very wet winters during several historical strong El Nino events, leading to heightened expectations that SCAL's multiyear drought would abate in 2016. Ensembles of atmospheric model simulations and coupled model seasonal forecasts are diagnosed to determine both the potential predictability and actual prediction skill of the failed rains, with a focus on understanding the striking contrast of SCAL precipitation between the 2016 and 1998 strong El Nino events. The ensemble mean of simulations indicates that the December-February 2016 winter dryness was not a response to global boundary forcings, which instead generated a wet SCAL signal. Nor was the extreme magnitude of observed 1998 wetness entirely reconcilable with a boundary-forced signal, indicating it was not a particularly precise analog for 2016. Furthermore, model simulations indicate the SCAL 2016 wet signal was 20%-50% less intense than its simulated 1998 counterpart. Such a weaker signal was captured in November 2015 initialized seasonal forecasts, indicating dynamical model skill in predicting a less prolific 2016 rainy season and a capability to forewarn that 2016 would not likely experience the flooding rains of 1998. Analysis of ensemble spread indicates that 2016 dryness was an extreme climate event having less than 5% likelihood in the presence of 2016 global forcings, even though its probability of occurrence was 3-4 times greater in 2016 compared to 1998. Therefore, the failed seasonal rains themselves are argued to be primarily a symptom of subseasonal variability unrelated to boundary forcings whose predictability remains to be explored.