摘要
基于不同地域月降水量的特点,采用聚类方法选取两个具有代表性地域的气象数据。将交互结构数据转化为嵌套分组结构形式,从而将地域因素和季节因素作为层次效应纳入分层线性模型(HLM)。据此建立相应的三层线性模型(HLM3),按照模型理论方法进行模型计算、优化和分析以解释层次效应和检验残差。结果表明,各层解释变量(气象因素、季节效应、地域效应)能很好地解释月降水量的差异,定量指导不同地域不同季节月降水量的回归。
According to the hierarchical characteristics of monthly rainfall in different regions,we select the meteorological data in two representative regions by clustering methods.After transforming the interactive date into nested data,we can take the geographical factors and seasonal factors into the hierarchical linear model（HLM） as the level effects.Then three level hierarchical linear model（HLM3） is established.Based on the model theory,we perform the corresponding model calculation,optimization and analysis to interpret the level effects and test residual errors.The results show that the difference of monthly rainfall is explained by variables（meteorological factors,seasonal effects and geographic effects） in different levels and it would quantitatively guide the regression of monthly rainfall in different seasons and regions.
出处
《水电能源科学》
北大核心
2012年第10期1-5,共5页
Water Resources and Power
