Abstract:Precipitation change could exert considerable impacts on the soil and water conservation and ecological construction in the Loess Plateau region. In this study a dataset of daily precipitation series (1961-2010) from 214 surface meteorological stations was used to investigate the spatiotemporal change trends in annual precipitation (PTOT), annual erosive rainfall (R12mm), annual flood season rainfall (RJJAS) and annual torrential rainfall (R50mm) over the Loess Plateau region (LPR) during the period 1961-2010. The examination was made at three hierarchical levels: the entire LPR, the Loess Plateau (LP) and the five divisions of integrated management of the LPR, and the Mann-Kendall and Mann-Whitney methods were used to test the monotonic and step trends, respectively. Results showed that (1) a similar spatial pattern of distribution of change trends at all the stations was found for PTOT, R12mm and RJJAS over the LPR, with a decreasing magnitude of negative change from the southeast to the northwest, besides a slightly positive change in the far northwest and far west areas. However, no clear spatial pattern was observed for R50mm over the region. The spatial patterns for PTOT, R12mm and RJJAS were more profound on the LP than on the LPR. (2) On the LPR, approximately 83% of stations showed a negative change in PTOT, and 69% of stations showed a negative change in R12mm and RJJAS, while 20% of stations showed a significant decrease in PTOT, and 10% of stations showed a significant decrease in R12mm and RJJAS. However, approximately 68% of stations showed a zero slope in R50mm. The proportion of stations showing a negative change or a significant negative change on the LP was higher than on the LPR, about 92% of stations showed a negative change in PTOT, and 80% of stations showed a negative change in R12mm and RJJAS, while 24% of stations showed a significant decrease in PTOT, and 12% of stations showed a significant decrease in R12mm and RJJAS. Approximately 62% of stations showed a zero slope in R50mm on the LP. (3) As a whole, a significant negative trend (P<0.05) in PTOT and a nearly significant negative trend (P<0.10) in R12mm were found on the LPR, with a linear trend of -9.9mm/10a and -5.9mm/10a, respectively; while the decrease was not significant for RJJAS and R50mm over the LPR. A significant negative trend for PTOT and R12mm, and a nearly significant negative trend for RJJAS were found on the LP, with a linear trend of -13.4mm/10a, -8.1mm/10a and -7.6mm/ 10a, respectively; while the decrease was not significant for R50mm over the LP. (4) The first and the second divisions of the LPR showed a significant negative trend in PTOT, and they showed a nearly significant negative trend and a significant negative trend for R12mm, respectively, while the rest three divisions showed a non-significant positive trend for PTOT and R12mm. A significant negative trend in RJJAS was only found in the second division among the five divisions. The first, second and fourth divisions showed a non-significant decrease in R50mm, while the third and fifth ones showed a zero slope in R50mm. (5) One abrupt change was detected in PTOT for the LPR, LP, and the first and second divisions, but not for the rest three divisions, with a significant decrease occurred since 1986 for the LPR, LP, and the second division, and since 1991 for the first division. Overall, the PTOT, R12mm and RJJAS decreased substantially but the R50mm did not change significantly over the LP, especially over the key area of soil and water conservation of the LPR (i.e. the second division) over the past five decades. These results indicated that if the trends are hold in future, although the soil erosion resulted from R12mm may decrease as a whole, the shortage of water resource will become severer, and the situation of the extremely intense soil and water losses caused by R50mm will not have considerable change.