Classification of time series of temperature variations from climatically homogeneous regions based on long-term persistence

Abstract

We used detrended methods for scaling analysis (DFA2 and DMA) and wavelet transform spectral analysis (WTS) to characterize long-term persistence (LTP) properties of temperature anomalies time series from observational stations from climatically and topologically homogeneous HadCRUT4 grid cells. We were interested to investigate the validity of the expectation that their LTP scaling properties remain preserved in the process of HadCRUT4 spatially interpolated and bias adjusted averaging, which was indeed the case in our selection. We additionally utilized the Hurst space analysis (HSA), a methodological solution that we recently developed, to classify climatic records from our dataset according to their LTP character and similarity of their WTS cyclical presentation. We were able to use HSA to observe four distinct patterns of climate dynamic behaviour according to the 'preferred' characteristic that those do not 'belong to the ocean'. In this way, our results suggested that there probably exists a necessity to examine cycles in climate records as important elements of natural variability. To illustrate how the procedure developed in this article can be extended, we used HSA developed by the use of a dataset from climatically and topologically homogeneous regions to classified time series of teleconnection indices that may have influence on their regional climate. HSA can be further systematically utilized in this way, to link LTP properties of temperature anomalies with their possible spatially remote sources.