Quantifying Chaos in the Atmosphere: the use of ensembles of General Circulation Model Integrations

The atmosphere is known to be forced by a variety of energy sources, including radiation and heat fluxes emanating from the boundary layer associated with sea surface temperature anomalies. The atmosphere is also subject to internal variability which is essentially unforced and is thought to be a basic characteristics of fluids. Whereas much work has been done in quantifying the links between external forcing of the atmosphere and its long term response as well as the influence of boundary layer forcing in determining organised large scale modes of planetary scale circulation, less is known about the importance of internal variability or chaos in determining the evolution of weather and climate. In part this derives from the inability of scientists to run experiments on the weather. General Circulation Models (GCMs) now provide for this possibility. Multiple evolutions of the climate system may be computed in GCM simulations. Where these simulations are similar except for the conditions by which the model is initialised, the degree of departure in the evolution of climate from one model run to the next corresponds precisely to the degree of internal variability or chaos present in the model atmosphere. A methodology for quantifying this chaotic forcing is presented and is applied to century long intergrations of the U K Meteorological Office model HADAM2A.