Refine
Year of publication
Document Type
- Article (16)
- Preprint (3)
- Monograph/Edited Volume (1)
- Doctoral Thesis (1)
Institute
Climate variability is triggered by several solar and orbital cycles as well as by the intern ocean dynamics. Consequently, paleoclimate proxy records are expected to vary on very different time scales ranging from subdecadal to millennial duration. We demonstrate, that Foster's (Foster, 1996) wavelet analysis technique is an appropriate tool for investigating temporarily changing spectral properties of records characterized by awkward sampling quality, which is a typical feature of climate proxy records. By applying it to the Holocene part of different glaciochemical records of Greenland ice cores we proof evidence for a significant contribution of the 1.47 kiloyears cycle over alomst the entire Holocene
The dynamics of noisy bistable systems is analyzed by means of Lyapunov exponents and measures of complexity. We consider both the classical Kramers problem with additive white noise and the case when the barrier fluctuates due to additional external colored noise. In case of additive noise we calculate the Lyapunov exponents and all measures of complexity analytically as functions of the noise intensity resp. the mean escape time. For the problem of fluctuating barrier the usual description of the dynamics with the mean escape time is not sufficient. The application of the concept of measures of complexity allows to describe the structures of motion in more detail. Most complexity measures sign the value of correlation time at which the phenomenon of resonant activation occurs with an extremum.
We have used techniques of nonlinear dynamics to compare a special model for the reversals of the Earth's magnetic field with the observational data. Although this model is rather simple, there is no essential difference to the data by means of well-known characteristics, such as correlation function and probability distribution. Applying methods of symbolic dynamics we have found that the considered model is not able to describe the dynamical properties of the observed process. These significant differences are expressed by algorithmic complexity and Renyi information.