在定义生长季节长度(GSL)、生长季节开始时间(LSF)以及生长季节结束时间(FFF)三个生长季节指数的基础上,采用模糊聚类、稳健回归及Mann-Kendall 非参数检验等方法分析新疆1961—2010 年生长季节指数时空变化特征,并针对作物生长季节指数变化对区域农业生产的影响进行了深入探讨.结果表明:新疆20 世纪90 年代中期,GSL变化幅度很大,低纬地区生长季节受低温影响较小,而随纬度上升,中高纬地区生长季节受低温影响越大.生长季节长度的变化与秋季低温(生长季节结束时间)变化情况紧密相连,全疆生长季节长度在80%的保证率下,较多年平均生长季节长度缩短5~14 d.生长季节长度的延长,会导致净初级生产力以及年蒸散发总量上升,继而影响水文、生态系统以及生物地球化学过程.

Three growing season indices, the growing season length, the starting/ending time of the growing season were defined based on the daily minimum and the daily average temperature data across Xinjiang, China. Spatiotemporal properties of the growing season indices were analyzed using Fuzzy Clustering Technique, robust regressive method and nonparametric Mann-Kendall trend test method. Besides, the implications of change of growing season indices on agricultural activities were discussed. The results indicate that there was a great change of growing season length in mid-1990s across Xinjiang. The agricultural activities in southern Xinjiang were not heavily influenced by the low temperature. However hazards were examined in the northern parts of Xinjiang. It is also found that the growing season length is in close relations with the occurrence of low temperature at the end of spring (before and after April) and in the early autumn (before and after October). The increase of the growing season length may improve the preliminary production and annual total evaporation, and it may further impact the hydrological, ecological and geochemical processes across Xinjiang. In this case, our study is of practical and theoretical merits in terms of planning and management of agricultural activities over Xinjiang, China.