HR: 14:30h
AN: NG43A-03 INVITED     [Abstracts]
TI: A Formula for Estimating the Mean Effects of State Dependent Noise
AU: * Sardeshmukh, P D
EM: Prashant.D.Sardeshmukh@noaa.gov
AF: NOAA-CIRES Climate Diagnostics Center, R/CDC, 325 Broadway, Boulder, CO 80305
AU: Penland, C
EM: Cecile.Penland@noaa.gov
AF: NOAA-CIRES Climate Diagnostics Center, R/CDC, 325 Broadway, Boulder, CO 80305
AU: Newman, M
EM: Matt.Newman@noaa.gov
AF: NOAA-CIRES Climate Diagnostics Center, R/CDC, 325 Broadway, Boulder, CO 80305
AB: It is well known that perturbing a dynamical system with stochastic noise can not only alter its variability but also its time mean, especially if the noise is multiplicative, i.e. if its amplitude depends upon the system state. Analytical expressions for this mean effect can be written down for linear systems perturbed by multiplicative white noise. For the practically more relevant case of linear systems perturbed by multiplicative red noise (i.e. noise with a finite correlation time scale), the problem becomes much more difficult. This is because the moment equations in this case are not closed: equations for the lower-order moments involve higher-order moments. Here we introduce a closure approximation that allows the mean effects of the noise in such systems to be estimated with high accuracy. The approximation is accurate for red noises with a wide range of correlation scales, and approaches the correct limits for both very small and very large correlation scales. We illustrate an application to stochastically perturbed planetary-scale Rossby wave dispersion over the globe, and discuss the implications for weather prediction and climate modeling.
DE: 1620 Climate dynamics (3309)
DE: 3309 Climatology (1620)
DE: 3319 General circulation
DE: 3379 Turbulence
SC: Nonlinear Geophysics [NG]
MN: 2005 Joint Assembly