Based on a monthly dataset of precipitation time series (1961-2010) from 12 meteorological stations across the Three-River Headwater Region (THRHR) of Qinghai Province, China, the spatio-temporal variation and abrupt change analysis of precipitation were examined by using moving average, linear regression, spline interpolation, the Mann-Kendall test and so on. Major conclusions were as follows. (1) The long-term annual and seasonal precipitation in the study area indicated an increasing trend with some oscillations during 1961-2010; however, the summer precipitation in the Lantsang (Lancang) River Headwater Region (LARHR), and the autumn precipitation in the Yangtze River Headwater Region (YERHR) of the THRHR decreased in the same period. (2) The amount of annual precipitation in the THRHR and its three sub-headwater regions was greater in the 1980s and 2000s. The springs were fairly wet after the 1970s, while the summers were relatively wet in the 1960s, 1980s and 2000s. In addition, the amount of precipitation in the autumn was greater in the 1970s and 1980s, but it was relatively less for the winter precipitation, except in the 1990s. (3) The normal values of spring, summer, winter and annual precipitation in the THRHR and its three sub-headwater regions all increased, but the normal value of summer precipitation in the LARHR had a negative trend and the normal value of winter precipitation declined in general. (4) The spring and winter precipitation increased in most of the THRHR. The summer, autumn and annual precipitation increased mainly in the marginal area of the west and north and decreased in the regions of Yushu, Zaduo, Jiuzhi and Banma. (5) The spring and winter precipitation in the THRHR and its three sub-headwater regions showed an abrupt change, except for the spring precipitation in the YARHR. The abrupt changes of spring precipitation were mainly in the late 1980s and early 1990s, while the abrupt changes of winter precipitation were primary in the mid-to late 1970s. This research would be helpful for further understanding the trends and periodicity of precipitation and for watershed-based water resource management in the THRHR.
