Long memory effect of past climate change in Vostok ice core records

Time series analysis of Vostok ice core data has been done for understanding of palaeoclimate change from a stochastic perspective. The Vostok ice core is one of the proxy data for palaeoclimate in which local temperature and precipitation rate, moisture source conditions, wind strength and aerosol fluxes of marine, volcanic, terrestrial, cosmogenic and anthropogenic origin are indirectly stored. Palaeoclimate data has a periodic feature and a stochastic feature. For the proxy data, spectrum analysis and detrended fluctuation analysis (DFA) were conducted to characterize periodicity and scaling property (long memory effect) in the climate change. The result of spectrum analysis indicates there exist periodicities corresponding to the Milankovitch cycle in past climate change occurred. DFA clarified time variability of scaling exponents (Hurst exponent) is associated with abrupt warming in past climate.