Temporal-spatial And Probability Of Distribution Of Precipitation In China

Based on the observed daily precipitation records from540stations and the girded data from a high-resolution regional climate model (CCLM) in1961-2010, by applying the probability method and extreme value theory, analyzed the temporal-spatial distribution and possible trends of annual precipitation and daily precipitation; Hypothesis testing was used to choose the optimal distribution function, and applied the optimal distribution function to estimated the daily maximum precipitation under the different return period. The main conclusions as follows:(1) The annual precipitation did not change significantly, but the inter-annual changed obviously; and significant change had appeared in the regional scale, there were85stations’ annual precipitation appeared significantly increased trend, mainly distributed in Liaohe river, Haihe river, Yangze river basin and Yellow river shelter-forest region; and79stations decreased obviously, distributed in continental river basin in Northwest and Southwest river basin. The abruption was happened in1969, regional abruption was happened in Northwest, North China and Southwest, some region in Northeast and Southeast. The periodic change of annual was very obviously, the most significant and stable cycle was6-years.(2) The distributive patterns of the kurtosis and skewness calculated from the simulated and observed series are consistent with each other, the maximum value was appeared in the most region of Northwest and the Northwest of Inner Mongolia, next was mid-eastern of Inner Mongolia, North China and some region of Northeast, the minimum value was in South China, especially Southeast and Southwest China. It was projected that in some parts of Jianghuai region, mid-eastern of Northeast China and Inner Mongolia, the kurtosis and skewness will rise significantly, and precipitation extremes will also increase simultaneously in2011-2050, and the kurtosis and skewness will decrease significantly in mid-west of West China, the precipitation extremes will also decrease simultaneously. In the past50years, the trend of no rain and strom rain days was increased, but their distribution was opposite the light rain, moderate and heavy rain days was decreased, the days in South China was more than North China, it showed that the precipitation was tended to extreme in temporal-spatial in the past50years.(3) In the43distribution functions, there are26distribution functions pass the hypothesis testing, and the Wakeby distribution function is the best, it could fit the AM series of precipitation from different region of China, this illustrates that the probability distribution of extreme precipitation follows the Wakeby distribution; using the Wakeby distribution, Generalized extreme value distribution and Person-Ⅲ distribution estimate the daily maximum precipitation under the different return period, the result show that the Wakeby distribution is better than Generalized extreme value distribution and Person-Ⅲ distribution on the estimate effect and accuracy.