Abstract:  Based on the daily air temperature data in Shangqiu weather station from 1955 to 2011, the annual, seasonal and inter-annual variations of air temperature and their trends, the abrupt changes were analyzed using methods of a Morlet wavelet analysis, a Mann-Kendall analysis, a regression analysis and a variation coefficient. The results indicate that the annual mean air temperature is in an increasing trend by a regression analysis and a 5-year slipping mean analysis, and its rate is 0.17 ℃ per decade. The trends of air temperature in four seasons differ significantly. Air temperature in summer is in a decreasing trend with a rate of -0.08 ℃ per decade, while that in the other three seasons is in an increasing trend, especially in winter with a rate of 0.35 ℃ per decade. An abrupt change of mean annual air temperature appeared in 1992 according to the Mann-Kendall analysis, while that for winter is in 1986. There is not the abrupt change for summer air temperature. According to the Morlet analysis, there is a 17-year cycle for the mean annual air temperature from 1950s to 1980s and then it disappears afterwards, while there is a 32-year cycle from 1980s to 1990s. There are a 5-year cycle and a 2-year cycle from 1955 to 2011 in the study area. It suggests that low temperature could appear in the future according to the cycle signals. The inter-annual variation coefficient of the mean air temperature is 0.04 and it is stable, and its variation rate is weak. The analysis of temperature variation characteristic and trend prediction has an important guiding significance for agricultural planting activities of Shangqiu.

Key words:  Morlet wavelet analysis, Mann-Kendal analysis method, Time series of temperatures, Shangqiu