中文摘要：

根据野外观测,运用统计建模的方法,研究了不同郁闭度条件下青海云杉林冠截留与降水的关系,确定了最佳降水截留模型。结果表明：在郁闭度为0.4的情况下,模拟模型为S曲线函数,相关系数（R2）可达0.698,在郁闭度为0.6和0.8的情况下,模拟模型为幂函数,相关系数（R2）分别为0.832和0.886,降水量较小时,模拟的精度较高,雨量较大时模型的误差很大。林冠对降水的再分配是干旱区水资源评估的一个重要方面,准确描述降水截留过程是干旱区生态水文过程研究的重要内容,因此考虑林冠特征和截留机制是未来降水截留研究的重点,以期精确定量评价该地区青海云杉林的生态水文效应。

英文摘要：

The rainfall interception loss from forests is an important component in the water balance of a catchment.The quantitative effects of forests on water resources are largely dependent on interception losses,since tree canopies typically intercept the majority of rainfall,and control its subsequent evaporation and drainage.Interception rates thus have clear local scale effects on runoff volumes and runoff times.In areas with scarce water resources,a good understanding of these effects is clearly essential for effective water resources management and landuse planning.The study area（the eastern Qilianshan mountains） is located in the arid region of northwestern China.To predict the interception loss from these woodland types,simulation models will be effective tool.Our objective is to develop relationship between interception loss and precipitation in this study.Three kinds of canopy coverage such as 0.4,0.6 and 0.8 were selected.Based on the measurement,the relationships between rainfall interception and precipitation were built using statistics analysis in the eastern Qilianshan Mountains.The optimal models of three kinds of canopy coverage were selected by comparing many kinds of statistic models.Some conclusion could be drawn from the research： S-curve model was selected for 0.4 canopy coverage with correlation coefficient（R2） of 0.698.Power model were the best models for predicting interception amount under the 0.6 and 0.8 canopy coverage.There were higher correlation coefficients as 0.832 and 0.886 respectively.When canopy coverage is small,interception amount can easily reach canopy storage capacity with small precipitation.Then interception amount is stable with the increase of precipitation.Therefore,S-curve model is good fit model for low canopy coverage.However,interception amount keeps rapid increase with the increase of precipitation under high canopy coverage.So power model is optimal one to estimate the intercept.By comparing observed value and predicted value,we found that precipitation is s