Shaun Lovejoy, Department of Physics, McGill University
Résumé/abstract:
Physics is an interlocking hierarchy of theories and models. Fundamental theories such as quantum or statistical mechanics are often too difficult to apply directly; so that conventional weather and climate modeling use the higher level laws of thermodynamics and fluid mechanics. However in the atmosphere, nonlinear terms are typically about a trillion times larger than linear ones; we anticipate the emergence of still higher level turbulence laws. The classical turbulence laws were restricted to homogeneous and isotropic systems; to apply them to the atmosphere they must be generalized to account for strong anisotropy (especially stratification) and variability (intermittency). Over the last 30 years, using scaling symmetry principles and multifractal cascades, this has been done. While hitherto they were believed applicable only up to ≈ 100 m, (generalized) turbulence laws now cover spatial scales up planetary in extent and in time well beyond weather scales to include the climate.
In the time domain the emergent laws for fluctuations ΔT (for example in temperature T) have means <ΔT > ≈ ΔtH i.e. they are scaling (power laws). In the weather regime we generally find exponents H>0 so that fluctuations increase with scale: the temperature and other atmospheric variables seem to “wander” like a drunkard’s walk. This continues up to ≈ 10 days which is the lifetime of planetary scale structures (this scale is directly set by the turbulent energy flux due to solar forcing). At larger Δt, the spatial degrees of freedom are “quenched”, there is a “dimensional transition” to a new “macroweather” regime in which on the contrary, H<0, so that successive fluctuations tend cancel each other out, diminishing with scale. Finally, at scales >≈30 years, new low frequency climate processes begin to dominate, leading to H>0: the signal again “wanders” in an “unstable” manner.
Scientific definitions of the climate are close to the dictum: “The climate is what you expect, the weather is what you get”. Indeed - since H<0 - averaging weather over periods increasing to ≈ 30 yrs yields apparently converging values. However this “expected” behavior is macroweather, not the climate. On the contrary, the climate is the new lower frequency regime at scales > 30 yrs and it has statistical properties very similar to the weather. At these scales, “macroweather is what you expect, the climate is what you get”.
We discuss the implications of the new emergent laws for weather, macroweather and climate including: new (stochastic) forecasting techniques, distinguishing anthropogenic and natural variability and the evaluation of Global Climate Models.
Cette conférence s'adresse à tous, y compris les professeurs, les chercheurs et les étudiants des trois cycles.
Le café est servi à partir de 11h20.
Cette conférence est présentée par le Département de physique de l'Université de Montréal.
