针对安徽省天长和宿松气象站建站至2010年逐年平均风速资料，分别利用直接检验方法和3种间接检验方法：标准正态均一检验（the standard normalhomogeneity test，SNHT）、惩罚最大t检验（the penalized maximal t test，PMTT）和惩罚最大F检验（the penalized maximal Ftest，PMFT），对这两个气象站逐年平均风速序列进行了均一性检验，并对不同检验方法的效果进行了对比分析。研究结果表明：1）站址迁移、观测场环境变化、仪器变更以及测风手段变化等均能对年平均风速序列的均一性产生影响，其中测风手段变化的影响最为明显。2）由于以气象台站观测记录为依据，直接检验得到的结果最为真实可信；此外SNHT、PMTT和PMFF方法对年平均风速序列的间断点均有一定的检验能力，但遗漏间断点的情况也比较明显。因而在年平均风速序列的均一性检验中，若气象台站观测记录相对较齐全，首先考虑采用直接检验，然后再考虑补充采用其他间接检验手段。%By using the direct homogeneity test （based on the observation data and the t test） and 3 in- direct homogeneity methods （including SNHT, PMTT and PMFT ）, time series of annual mean wind speed at 2 meteorological stations,Tianchang and Susong,from 1957 to 2010,have been tested respec-tively. The results show that station relocation, observational environment change, instruments updating and measurements variation can all influence the homogeneity of the mean wind speed, and the impact of measurements updating is the most prominent factor. Because based on observation data from meteor-ological stations, the results of direct test is more reliable. However, SNHT, PMTT and PMFT are effec-tive in identifying the change-points, but the missing ratio is also of a high degree. Thus, to test mean wind speed homogeneity, if the observation data from meteorological stations are correct and intact, the direct test should be used firstly and then other methods can be used additionally.
展开▼
