Changes of Precipitation Intensity Spectra in Different Regions of Mainland China During 1961-2006

The spectral characteristics of precipitation intensity during warm and cold years are compared in six regions of China based on precipitation data at 404 meteorological stations during 1961-2006.In all of the studied regions except North China,with the increasing temperature,a decreasing trend is observed in light precipitation and the number of light precipitation days,while an increasing trend appears in heavy precipitation and the heavy precipitation days.Although changes in precipitation days in North China are similar to the changes in the other five regions,heavy precipitation decreases with the increasing temperature in this region.These results indicate that in most parts of China,the amount of precipitation and number of precipitation days have shifted towards heavy precipitation under the background of a warming climate;however,the responses of precipitation distributions to global warming differ from place to place.The number of light precipitation days decreases in the warm and humid regions of China(Jianghuai region,South China,and Southwest China),while the increasing amplitude of heavy precipitation and the number of heavy precipitation days are greater in the warm and humid regions of China than that in the northern regions(North China,Northwest China,and Northeast China).In addition,changes are much more obvious in winter than in summer,indicating that the changes in the precipitation frequency are more affected by the increasing temperature during winter than summer.The shape and scale parameters of the Γ distribution of daily precipitation at most stations of China have increased under the background of global warming.The scale parameter changes are smaller than the shape parameter changes in all regions except Northwest China.This suggests that daily precipitation shifts toward heavy precipitation in China under the warming climate.The number of extreme precipitation events increases slightly,indicating that changes in the Γ distribution fitting parameters reflect changes in the regional pr