Decadal-to-multidecadal variability of seasonal land precipitation in northern hemisphere in observation and CMIP6 historical simulations

Abstract

Based on the centennial-scale observations and CMIP6 historical simulations, this paper employs the ensemble empiricalmodedecomposition to extract the decadal-to-multidecadal variability of land precipitation (DMVLP) in the northern hemisphere. The spatial distributions of the dominant mode from the empirical orthogonal function are different in four seasons. Regions with the same sign of precipitation anomalies are likely to be teleconnected through oceanic forcing. The temporal evolutions of the leading modes are similar in winter and spring, with an amplitude increasing after the late 1970s, probably related to the overlap of oceanic multidecadal signals. In winter and spring, the Interdecadal Pacific Oscillation (IPO) and the Atlantic Multidecadal Oscillation (AMO) play a joint role. They were in phase before late 1970s and out of phase after then, weakening/strengthening the impacts of the North Pacific and North Atlantic on the DMVLP before/after late 1970s. In summer and autumn, AMO alone plays a part and the amplitude of time series does not vary as in winter and spring. The ability of the coupled models from CMIP6 historical simulations is also evaluated. The good-models average largely captures the spatial structure in four seasons and the associated oceanic signals. The poor-models average is hardly or weakly correlated with observation.