Group seminar on 26. October, 14:15 CET
Correlated vorticity extremes as fractional Poisson processes
Dr. Richard Blender
Observations show that the return times of relative vorticity extremes in the exit regions of the storm tracks are not exponentially distributed. We consider these extremes as fractional Poisson processes (FPPs) with a return time distribution given by the Mittag–Leffler function and approximate this function by a Weibull distribution. The Weibull shape parameter yields the dispersion parameter which agrees with results obtained independently for mid-latitude cyclones. The memory in the vorticity time series (determined by detrended fluctuation analysis, DFA), provides an independent estimate of the shape parameter. The analysis yields a concise view of the deviation from Poisson statistics, correlation and non-exponential return time distributions, in terms of a single parameter.