Multidimensional analysis of dynamics of annual warming-cooling cycles on the basis of index model of temperature observations

Trend analysis remains the most useful to show the dominant tendencies in climate dynamics, but the possibilities of such analysis are restricted. In particular, both information losses and distortions appear inevitably while estimating the dynamics of the land surface air temperature. The peculiarity of the viewpoint discussed in the paper is connected with the consideration of the profile of the annual warming-cooling cycles by the adaptation of one of the approach to multidimensional data analysis. That approach (SUC-logic) has already shown good results to analyse the instrumental noisy time series with a variable profile. In according to SUC-logic we transform each instrumental time series into a consecution of homogeneous fragments, where each fragment is determined by the ensemble of traditional, rare considered and unconsidered characteristics in absolute and relative marks. The index model is proposed to describe the transformation. So, the correlated multidimensional data analysis over centenary time scales is computer realized. The examples of local climate systems located within four quasi-homogeneous climatic regions are considered. Since only the instrumental observations are analyzed, then the results are verified concerning the real events. We consider our research, first of all, to find a way how to reveal the indicators of abnormalities in observed temperature dynamics. At the same time, the proposed multidimensional analysis gives the possibility to reveal and study novel correlations in climate dynamics. For example, local and regional temperature oscillations (LTO- and RTO-effects correspondingly) are illustrated to be discussed.