【摘要】：Due to the topography and local nonuniform distribution of heating, extratropical cyclones in the lower troposphere usually have complex shapes and structures, and there remain some uncertainties in identifying them. Using a modified cyclone area automatic objective recognition algorithm for cyclones, we investigated the patterns of spring cyclone activities affecting Changjiang River-Huaihe River valleys(CHV) of China during the previous 37 years. The results indicated that the algorithm performs well in reproducing the cyclogenesis, movement, and cyclolysis of cyclones in CHV. Three new perspectives were found.(1) Most influential cyclones have relatively short travel distances and lifetimes, and they are typically excluded when conducting synoptic-scale cyclone tracking.(2) The cyclogenesis location of influential cyclones in spring displays multi-source characteristics. In particular, the influential cyclones originated in northern China account for 43% with more marked mobility compared to the locally generated cyclones,although most of their centers do not enter CHV.(3) Multi-center cyclones appear mainly in Da Xingan Mountains which is on the east side of the Mongolian Plateau and over the East China Sea. These cyclones are relatively large in horizontal scale and have relatively long lifetimes.