During 1979-2019, the seasonal mean precipitation averaged over the U.S. mid-Atlantic coastal region show significant increases in summer but no changes in winter, while that over California show significant drying in winter. Clarifying the drivers of these observed changes is important for understanding and interpreting future projections of regional precipitation by climate models. This study examines the relative effects of external forcing and internal variability in the observed multi-decadal changes, as well as their uncertainty. We use historical and RCP8.5/SSP585 simulations from climate models in Coupled Model Intercomparison Project phase 5 (CMIP5), CMIP6, as well as large ensemble simulations from 6 single model, each including more than 20 members.

For California, uncertainty from internal variability accounts for >80% of the total uncertainty in winter precipitation change during 1979-2019. The Interdecadal Pacific Oscillation (IPO) plays a key role in this internal contribution. Linearly removing the IPO’s influence can reduce the uncertainty by up to 26%, which is arguably critical in planning for future water security. Accounting for the positive-to-negative phase transition of the IPO during 1979-2019 by linear regression, the simulated California precipitation trends are comparable to the observed drying trends.

For a sub-region of the U.S. mid-Atlantic, enhanced warming over land relative to ocean due to external forcing induces low pressure anomalies over land that contribute to the wetting summer. While similar wetting also occurs during winter, such effect is offset by the drying associated with the positive phase of the North Atlantic Oscillation (NAO) in recent decades. Internal variability produces large uncertainty in the multi-decadal precipitation changes simulated by climate models through changes in large-scale circulation associated with the NAO in winter and the North Atlantic Subtropical High in summer.