Graph theory in analysis of paleoclimatic time series

Abstract

At the current work, we present the results of diagnostic of similarities of various paleo reconstructions of temperatures with the help of network approaches. The first step of data processing is transformations of correlational patterns of time series into the geometry of the corresponding network, which is an input for further processing steps by methods of algebraic topology. To detect the possible non-linear connections between climatic series and solar activity, we propose the approach based on network generation with the help of embedding time series into the feature space of the corresponding dimension. In conclusion, we give markovian chains for climatic reconstructions and Wolf numbers.