Warm Pool SST Forecast Skill in S2S Models: Mean State Drift versus Anomaly Patterns

Tropical sea surface temperature (SST) forecast skill is assessed using the subseasonal-to-seasonal (S2S) forecast model archive. Two types of skill are examined. The first is mean SST drift with lead time, which is associated with patterns of mean tropical column water vapor (CWV) drift. These SST-linked CWV drift patterns have implications for Madden-Julian oscillation (MJO) forecast skill. Ocean dynamics play a dominant role in Warm Pool SST drift in some models, suggesting a degree of oceanic influence on MJO forecast skill.

The second type of skill is SST anomaly skill, where the anomalies are defined as departures from a given model’s lead-dependent climatology. We quantify how the amplitude and pattern of forecast SST anomalies affects rainfall-surface flux feedbacks as a function of lead time over the Warm Pool. We then examine SST anomaly forecast skill associated with two types of climatologically important ocean weather events: eastward extension of the Warm Pool eastern edge, and ocean heat waves associated with coral bleaching events.

We close with a discussion of how such studies can be improved with newly available ocean output variables in the S2S database.