Abstract: With the combination of 1982〖KG-*2〗-〖KG-*7〗2003 GIMMS-NDVI datase ts and GIS technology, and applying correlation analysis based on each pixel, th e responses of the vegetation NDVI in Northeast China to climate change were ana lyzed. In 1982〖KG-*2〗-〖KG-*7〗2003, the mean annual air temperature in study area had an increasing trend, while the annual precipitation was in adverse. The pixels whose NDVI had significant positive correlations with mean annual temper ature occupied 1284% of the total, mainly distributed in south Songnen Plain, central Sanjiang Plain, and west Liaohe Plain. Their vegetation types were cropl and, broadleaf forest, and steppe, and the NDVI had less negative correlation wi th mean annual air temperature. The pixels whose NDVI significantly positively c orrelated with annual precipitation occupied 455%, and their main vegetation t ypes were steppe and cropland. The pixels whose NDVI significantly negatively co rrelated with annual precipitation occupied 752%, and the main vegetation type s were coniferous forest and broadleaf forest. The vegetations significantly pos itively and negatively correlated with the air temperature in growth season occu pied 396% and 435% of the all, while those having significant positive and n egative correlations with the precipitation in growth season occupied 881% and 854%, respectively. There were 5821% pixels whose NDVI significantly correl ated with the air temperature in spring, mainly distributed in central Great Xin gan Mountains, Small Xingan Mountains, Changbai Mountains, and Wanda-Zhangguang cai Mountains. Their main vegetation types were broadleaf forest, cropland, coni ferous forest and meadow, and the NDVI had less negative correlation with the ai r temperature in spring. The pixels whose NDVI significantly positively and nega tively correlated with the precipitation in spring was 481% and 167%, while those whose NDVI having significant positive correlation with the air temperatur e in summer and negative correlation with the precipitation in summer were 716 % and 629%, respectively. The air temperature and precipitation in autumn had smaller effects on the vegetation NDVI. There were 605% of pixels whose NDVI s ignificantly positively correlated with the air temperature in autumn, and few h ad significant negative correlation with it. The proportion of the pixels whose NDVI significantly negatively correlated with the p recipitation in autumn was 543%, and hardly any pixels whose NDVI had signific ant positive correlation with autumn precipitation.

Key words: Haloxylon ammodendron community, Habitat type, Species structure, Species diversity, Biomass, Distribution pattern