OBSERVATIONS-BASED COMPUTATIONAL ANALYTICS ON LOCAL CLIMATE DYNAMICS: CHANGE-POINTS
The paper concerns the problem on the computational decision making on evolution of local climate dynamics taking into account inevitable nonlinear nature of such systems and deficiency of reliable data on climate dynamics. With this purpose we consider annual temperature variation in the context of the bifurcation phenomena under the hysteresis regulation with double synchronization. The corresponding conceptual model (HDS-model) provides the homogeniety of the analysed states as well as the ability to derive some constituent of regional (external) impacts on a local climate system. In accordance with the peculiarities of HDS-model dynamics, we formalize three mechanisms of local temperature changes and introduce the corresponding indicators of change-points. It allows to substitute, at least partly, expert analytics concerning identification of qualitative changes in local climate dynamics. Examples of applications of the indicators are presented on the basis of processing the time series of temperature observations on daily mean surface air temperature made over last century. We believe that the results could be applied in order to increase the confidence of estimations about local climate changes.