Pacific Decadal Variability and Its Hydroclimate Teleconnections in CMIP6 Models

Abstract Natural decadal climate variability in the Pacific, such as the Pacific decadal oscillation (PDO) or the interdecadal Pacific oscillation (IPO), plays a powerful role in evolving global hydroclimate on decadal time scales. Recent generations of general circulation models (GCMs) have been found to simulate the spatial pattern of the PDO well but struggle to capture temporal variability on decadal time scales. To use GCMs to project future climate, we must understand the degree to which climate models can successfully reproduce historical PDO and IPO spatial patterns, temporal behavior, and influence on hydroclimate. We calculate PDO and IPO spatial patterns and time series using 16 models within the CMIP6 archive, all with large ( n ≥ 10) ensembles, and compare them to observations in an integrated assessment of models’ ability to represent Pacific decadal variability spatiotemporally. All models underestimate decadal variability in the PDO and IPO and have a westward bias in their PDO and IPO North Pacific SST anomalies. We also evaluate hydroclimate teleconnections of the PDO and IPO in models using PDO- and IPO-associated precipitation, circulation, low-cloud, and vapor pressure deficit anomalies. We show that models’ underpowered decadal variability in the Pacific is consistent with their inability to reproduce large-amplitude decadal swings in precipitation in southwestern North America and that models are virtually unable to produce a 30-yr precipitation trend in the southwest of the magnitude observed from 1982 to 2011. We emphasize the importance of model fidelity in simulating Pacific decadal variability for accurate representation of decadal-scale hydroclimate change in Pacific-teleconnected land regions.