HR: 10:25h
AN: U32B-01 INVITED     [Abstracts]
TI: Extreme events: Some theoretical and practical considerations
AU: * Ghil, M
EM: ghil@atmos.ucla.edu
AF: Department Terre-Atmosphere-Ocean, Ecole Normale Supérieure, Paris Cedex 05, 75231, France
AU: * Ghil, M
EM: ghil@atmos.ucla.edu
AF: Department of Atmospheric and Oceanic Sciences, University of California, Los Angeles, CA 90095-1567, United States
AU: Zaliapin, I
EM: zal@unr.edu
AF: Department of Mathematics and Statistics, University of Nevada, Reno, NV 89577, United States
AB: Predictive understanding of how extreme events arise is an important goal of the natural sciences that is tightly connected with adaptation and mitigation of natural hazards. This talk reports first on selected results from a highly interdisciplinary, multi-national, European-Union sponsored project on "Extreme Events: Causes and Consequences (E2-C2)." These results will include some of the more practical aspects of extreme events and natural hazards. Second, it reports on preliminary results about the key mechanisms that generate extreme events in nonlinear, periodically driven systems with a delayed feedback. Toy models of this type were shown to capture major features of the El Nino-Southern Oscillation (ENSO) phenomenon [Jin et al., Science, 1994; Tziperman et al., Science, 1994]; they may be relevant to other natural systems in which internal instabilities interact with external forcing and give rise to extreme events. We focus here on a one-dimensional differential- delay model with a single delay and periodic forcing. This simple model is characterized by unstable solutions in a broad range of parameters, and we describe the mechanisms of the observed instabilities, connect our findings to climate phenomena, and formulate questions for further theoretical and numerical analysis. We report on two classes of instability: one prevails at a small amplitude of the external forcing and is caused by frequency locking, as in the Devil's staircase scenario connected with ENSO already. The other one, more relevant for the applications of interest here, is seen at stronger external forcing and leads to abrupt overall changes in the solutions' behavior. These changes include the near-periodic appearance of extreme events (as in the observed warmings and coolings of the Tropical Pacific), as well as sudden changes of the mean (as in global climate shifts associated with rapid transitions between glaciated and ice-free Quaternary climates). The nature of these two types of instability will be analyzed and implications for the predictability of extreme events will be discussed.
DE: 3215 Instability analysis
DE: 3339 Ocean/atmosphere interactions (0312, 4504)
DE: 4425 Critical phenomena
DE: 4445 Nonlinear differential equations
DE: 4522 ENSO (4922)
SC: Union [U]
MN: 2007 Joint Assembly