Extremes in East African hydroclimate and links to Indo-Pacific variability on interannual to decadal timescales

East African hydroclimate exhibits considerable variability across a range of timescales, with implications for its population that depends on the region’s two rainy seasons. Recent work demonstrated that current state-of-the-art climate models consistently underestimate the long rains in boreal spring over the Horn of Africa while overestimating the short rains in autumn. This inability to represent the seasonal cycle makes it problematic for climate models to project changes in East African precipitation. Here we consider whether this bias also has implications for understanding interannual and decadal variability in the East African long and short rains. Using a consistent framework with an unforced multi-century global coupled climate model simulation, the role of Indo-Pacific variability for East African rainfall is compared across timescales and related to observations. The dominant driver of East African rainfall anomalies critically depends on the timescale under consideration: Interannual variations in East African hydroclimate coincide with significant sea surface temperature (SST) anomalies across the Indo-Pacific, including those associated with the El Niño-Southern Oscillation (ENSO) in the eastern Pacific, and are linked to changes in the Walker circulation, regional winds and vertical velocities over East Africa. Prolonged drought/pluvial periods in contrast exhibit anomalous SST predominantly in the Indian Ocean and Indo-Pacific warm pool (IPWP) region, while eastern Pacific anomalies are insignificant. We assessed dominant frequencies in Indo-Pacific SST and found the eastern equatorial Pacific dominated by higher-frequency variability in the ENSO band, while the tropical Indian Ocean and IPWP exhibit lower-frequency variability beyond 10 years. This is consistent with the different contribution to regional precipitation anomalies for the eastern Pacific versus Indian Ocean and IPWP on interannual and decadal timescales, respectively. In the model, the dominant low-frequency signal seen in the observations in the Indo-Pacific is not well-represented as it instead exhibits overly strong variability on subdecadal timescales. The overly strong ENSO-teleconnection likely contributes to the overestimated role of the short rains in the seasonal cycle in the model compared to observations.