Climate Models Underestimate Diversity of Synoptic Conditions Associated With Extreme Precipitation Over California
Climate models are shown to underestimate the variability of precipitation distribution among days of extreme precipitation over California. This is related to underestimates of the variability of the jet stream latitude and of associated low-pressure systems.
This study highlights that on average GCMs fail to represent the synoptic variability over a region with highly variable seasonal precipitation such as California. These comparisons to the reanalyses are highly relevant to regional projections, in particular for dynamical and statistical downscaling efforts.
Days of extreme precipitation over California are evaluated in Coupled Model Intercomparison Project Phase 6 (CMIP6) models and the ERA-Interim reanalysis. In the current climate, the model spread in composited precipitation on extreme precipitation days is closely related to the magnitude of column water vapor transported by atmospheric rivers, which almost all models overestimate. Moreover, most models underestimate the variance in the latitude of maxima of numerous variables among days of extreme California precipitation. That is, in the general circulation models there is a lack of diversity in the latitude of the disturbances bringing winter precipitation to California. In the future climate, most models project a decrease in the frequency of southward-displaced disturbances among California extreme precipitation days. Hence, the greatest increases in extreme precipitation are over northern California. However, the historical underestimate of the latitudinal variance of disturbances calls into question the reliability of these projections. This bias should be especially considered for dynamical downscaling efforts over the region.