中文摘要：

目前，越来越多的遥感数据被用来监测大面积植物物候的动态变化。利用长时间序列的SPOT／NDVI旬合成数据，通过doublelogistic模型获取了1999～2008年长白山地区植被的3个关键物候参数：生长季始期、生长季末期和生长季长度的多年平均值，并绘制了它们的变化趋势空间格局图。结果表明，林地的生长季开始日期为第100-120天，草地和耕地相对较晚，分别为第130～140天和第140～150天；林地和草地生长季的结束日期为第275-285天，耕地的相对较早，为第265～275天；林地、草地和耕地的生长季长度范围分别为160-180d、140-160d和110--130d。植被物候期的变化趋势表现为一定的空间差异性，生长季长度延长区域主要分布在长白山地区的中东部，平均每年延长约0．7d；缩短的区域在西北地区，平均每年缩短1．1d。最后通过部分物候观测数据及前人在相同研究区的结果验证了利用doublelogistic模型提取预测长白山植被物候期的可行性。

英文摘要：

A wealth of remotely sensed image time series is now available to monitor vegetation dynamics over large＇areas. In this paper, the spatial pattern of three key vegetation phenological mettles, the start of growing season（SOS）, the end of growing season（EOS）, and the length of growing season（LOS） were inferred from the time series SPOT/NDVI data based on double logistic model in Changbai mountains, and the trends of them are assessed. The results show that the SOS of forests focus on dekad 10-12,that of grasslands and croplands start relatively late, they are in decad 13-14 and decad 14-15, respectively. The EOS of forests and grasslands range from 275^th to 285^th day, that of croplands are early relatively, which are between 265^th and 275^th day. Therefore, the LOS of forests, grasslands and croplands are 160-180 days, 140-160 days and 110-130 days, respectively. The trends ofphenophases have certain spatial discrepancy, which express that the LOS prolong in middle-eastern areas and shorten in northwestern areas. Then the derived phenologieal metdes were validated by a few field observed data and some previous research achievements in the same area.