H11P-04
The role of internal variability in prolonging the California drought

Monday, 14 December 2015: 08:45
2022-2024 (Moscone West)
Nikolaus H Buenning and Lowell D Stott, University of Southern California, Los Angeles, CA, United States
Abstract:
The current drought in California has been one of the driest on record. Using atmospheric general circulation models (AGCMs), recent studies have demonstrated that the low precipitation anomalies observed during the first three winters of the current drought are mostly attributable to changes in sea surface temperature (SST) and sea ice forcing. Here we show through AGCM simulations that the fourth and latest winter of the current drought is not attributable to SST and sea ice forcing, but instead a consequence of higher internal variability. Using the Global Spectral Model (GSM) we demonstrate how the surface forcing reproduces dry conditions over California for the first three winters of the current drought, similar to what other models produced. However, when forced with the SST and sea ice conditions for the winter of 2014-2015, GSM robustly simulates high precipitation conditions over California. This significantly differs with observed precipitation anomalies, which suggests a model deficiency or large influence of internal variability within the climate system during the winter of 2014-2015. Ensemble simulations with 234 realizations reveal that the surface forcing created a broader range of precipitation possibilities over California. Thus, the surface forcing caused a greater degree of internal variations, which was driven by a reduced latitudinal temperature gradient and amplified planetary waves over the Pacific. Similar amplified waves are also seen in 21st century climate projections of upper-level geopotential heights, suggesting that 21st century precipitation over California will become more variable and increasingly difficult to predict on seasonal timescales. When an El Nino pattern is applied to the surface forcing the precipitation further increases and the variance amongst model realizations is reduced, which indicates a strong likelihood of an anomalously wet 2015-2016 winter season.