Title: Time series analysis of Earth's Northern Hemisphere temperature reconstructions using non-linear tools and statistical tests

Authors: George E. Halkos; Kyriaki D. Tsilika; Aggelos C. Iliopoulos

Addresses: Laboratory of Operations Research, Department of Economics, University of Thessaly, Korai 43, 38333 Volos, Greece ' Laboratory of Operations Research, Department of Economics, University of Thessaly, Korai 43, 38333 Volos, Greece ' Laboratory of Operations Research, Department of Economics, University of Thessaly, Korai 43, 38333 Volos, Greece

Abstract: This paper presents results concerning the non-linear analysis of the Earth's Northern Hemisphere mean annual land temperature reconstructions for the time period of 713-2004. The non-linear analysis consists of the application of several filtering methods, the estimation of geometrical and dynamical characteristics in the reconstructed phase space, techniques of discrimination between non-linear low-dimensional and linear high-dimensional (stochastic) dynamics and tests for serial dependence and non-linear structure. The results of this study indicate the efficiency of the filtering methods in revealing non-linear structures and correlations in short noisy time series, such as Earth's Northern Hemisphere temperature reconstructions. No low-dimensional chaotic behaviour evidence was found from the analysis.

Keywords: non-linear time series analysis; temperature reconstructions; filtering methods; discriminating statistics; environmental economics computation.

DOI: 10.1504/IJISD.2018.089267

International Journal of Innovation and Sustainable Development, 2018 Vol.12 No.1/2, pp.220 - 245

Received: 09 Sep 2016
Accepted: 12 Nov 2016
Published online: 11 Jan 2018 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article