Wavelet Analysis on Temperature Variability of Kanto Region, Japan

Abstract

Climate oscillations, such as the El Niño–Southern Oscillation and Indian Ocean Dipole, are known to influence surface temperature and precipitation variability. This thesis investigates intraseasonal (2–12 months) and interannual (13–120 months or 1–10 years) variability in monthly mean temperature data from seven urban and five rural lowland weather stations sited in Kanto region, Japan, over the period January 1973–August 2015. Indices of seven climate oscillations (Madden–Julian Oscillation, El Niño–Southern Oscillation, Indian Ocean Dipole, Quasi-biennial Oscillation, Arctic Oscillation, Pacific North American Oscillation, West Pacific Oscillation) known to influence Japan’s climate are analysed in relation to temperature variability. The wavelet analysis is used to identify oscillations in temperature data at various periods and years, as well as to establish the relationship between temperature variability and climate oscillations. The results show that temperature variability is quasi-periodic and is more pronounced at the intraseasonal timescale. Urbanisation is not found to influence intraseasonal and interannual temperature variability. Rather, temperature variability of stations sited next to Pacific Ocean are remarkably different from those sited inland or next to Tokyo Bay. Temperature variability is found to be most strongly influenced by West Pacific Oscillation, followed by Pacific North American Oscillation. This suggests that the key to reducing the uncertainty in temperature forecasts for the Kanto region may lie on the forecasted teleconnection indices of these two oscillations.