Abstract: Considered the level of underground water in Qingyang stressed well as a comprehensive observed quantity, authors give the brief introduction of principle, method and processing steps of principal components analysis (PCA) method. By using the underground water level and precipitation observation of stressed well in Qingyang, Jinjiang City, Quanzhou, from March to June 2007, the water level has been analyzed by PCA using the intraday rainfall and the accumulated precipitation separately, and comparison of these two situations indicates that the water level can not be effected by the intraday rainfall and the result reflects the real water level of stressed well; on the other hand, the water level has been effected by accumulated rainfall from March to June 2007 and the result reflect the accumulated effects of rainfall. Finally, authors give the curves of first and second synthetic observation values.

Key words: Water level, Precipitation, Variance-covariance matrix, Eigenvalue and eigenvector, Principal components analysis