Dmitry Sonechkin

Abstract

TO the study of climatic changes in the Europe-North Atlantic region some methods classical to the meteorological community, such as an empirical orthogonal function (EOF) decomposition and analogues searching were combined with methods used by the contemporary theory of dynamical systems. In this paper we try to resolve a question about the nature of changes in the climatic system, starting from an investigation as to whether features of self-similarity typical for strange attractors (SA) remain invariant or break. The essence of our approach is a reconstruction of a coarse-grained dimension of a strange attractor on the basis of time series of meteorological data. The best results were obtained by joint application of the EOF and Takens methods with a local approximation of a strange attractor trajectory by the fixed mass (analogue) method. The presented method has value not only as a research tool and has since October 1944 been implemented operationally at IMWM (Poland) to forecast a half-year air temperature anomaly within the mid-latitude belt.

Keywords: climate modelling, climate change, chaotic strange attractor, empirical long-term forecasting

Bogumił Jakubiak, Interdisciplinary Centre for Mathematical and Computational Modelling, Warsaw University Pawińskiego 5a, 02-106 Warsaw, Poland
Dmitry Sonechkin, Hydrometeorological Research Centre of Russia Bolshoy Predtechensky Lane 9/13, 123242 Moscow, Russia
Nadjezda Ivachtchenko, Hydrometeorological Research Centre of Russia Bolshoy Predtechensky Lane 9/13, 123242 Moscow, Russia