Using the CESM1-CAM5 Large Ensemble to study the predictability of midlatitude atmospheric dynamics

Abstract:

Assessing the internal variability of climate models is an important step for the detection of forced trends, in any metric of climate change. In a pioneering study, using a large ensemble of future simulations with CCSM3 (the Community Climate System Model, version 3), Deser et al (2012) found that surface air temperature was easily predictable, whereas sea level pressure was considerably more difficult to predict (i.e. a very large number of ensemble members was required). We reassess this conclusions using the new, recently completed, Large Ensemble of coupled model simulations using CESM1-CAM5 (the Community Earth System Model, with the Community Atmosphere Model, version 5). In contrast to the earlier work, we here demonstrate that forced response of more widely used metrics of the atmospheric circulation, the latitude and strength of the midlatitude jets, emerges from internal variability with as few as ten ensemble members, over the course of the 21st Century.