Model-based analytics oriented on the intermittency observed in local temperature dynamics

摘要

In practice of the time series analysis there is a specific problem on how to estimate tendencies of nonlinear dynamics evolution when several periodical processes alternate (so-called intermittency phenomena). Such alternations relate to bifurcations, but particular regularities and mechanisms of intermittency phenomena remain insufficiently understandable due to the height sensitivity of these phenomena to running disturbances and initial conditions. Model-based simulations provide the basic abilities to systematic analytics of this subject. In particular, it occurs in order to analyze intermittency observed in local temperature dynamics. In this case, dynamics description is based conceptually on the model of the hysteresis regulator with double synchronization (so-called HDS-model). With this purpose we introduce a new description of the intermittency based on the notion of an uncertainty zone about a bifurcation point known in the experimental bifurcation analysis in order to analyze deeply a stability margin. Then, a general viewpoint on intermittency can be summarized, in accordance to which it becomes possible to analyze in more detail a certain realization of intermittency and compare different realizations. In particular, we discuss differences between the observed intermittency and the simulated intermittency, and also first consider fast transients under which the observed specific effects in local climate dynamics could be naturally formed. The model-based simulations are verified by the results of processing the real data of the meteorological observations of land surface air temperature. We believe that the presented results could be interesting for scientific researches on tendencies of present-day evolution of local temperature dynamic, including latent effects of transients which occur against the background of intermittency. Correspondingly, it could give novel arguments in the field of practical meteorology and climatology.