According to this model, internal variability explains the bulk of the relative cooling of the equatorial Pacific since 1980 and the drying of southwestern North America.
Attributing recent climate changes to human emissions or natural fluctuations remains sensitive, as both signals overlap in observations. A team including L. Ruby Leung and Shang-Ping Xie has trained a convolutive memory network (ConvLSTM) on large sets of simulations from several climate models, and reports that it surpasses existing statistical and learning methods to isolate forced response.
Applied to the observations, the network attributed a large part of several multi-decade trends to internal variability: it would have dominated the La Niña type warming of the equatorial Pacific and the decline in rains in southwestern North America since 1980, amplified the reversal of temperature asymmetry between the two Atlantic hemispheres, and modulated arctic amplification around an almost constant forced factor of three.
We also note differences with climate models, which simulate a stronger warming of the northern hemisphere and a less pronounced decrease in precipitation on continents.
Read the full article on the original source
Go to the source →