稀疏的植被覆盖是干旱和半干旱地区最主要的环境特征,因此长期定量的植被分布和变化观测能够分析干旱和半干旱地区的环境变化。在以干旱和半干旱地区为主要的中国西北地区存在着森林减少、土地侵蚀、盐碱化和沙漠扩张等严重的环境问题,生态环境十分脆弱。通过NOAA/AVHRR建立近20年来中国西北地区NDVI变化序列,利用差分法、斜率变化和主成分分析3种方法分析植被变化。3种方法显示出基本一致的结果,即大部分地区植被状况恶化,局部地区有所好转。通过分析植被变化与温度、降水变化的关系,发现NDVI与降水存在明显的正相关关系,而与温度变化的关系并不明显,表明降水是影响西北地区植被变化最主要的自然因素。

Land vegetation plays amajorrole inthe global climate change through carbon cycle,and climate change inturn af\nfects vegetation growth and its photosynthetic activity.In arid and semi-arid area,sparse vegetation cover characterizes environ\nments.Thus quantitativetemporal series analysisofvegetation distribution and itsvariations enablesobservingannualtrends,helps\nto find out the reason for environment variability.There are serious environmental problems,such as deforestation,soil erosion,\nsalinization,and desert encroachment in the northwestern China,its natural conditions is very delicate.In this paper,we built th\ntime series of vegetation change by theNDVI(normalized difference vegetation index)covered northwest regions in recent 2\nyears,and analyzd the time serialNDVIvariability using three methods,which are simple differencing,slope map ofNDVI(an\nbiomass),and principal component analysis(PCA).The change image demonstrated thatthe area of vegetation covers decrease i\nrecent 20 years innorthwesternChina,especially intheLoessPlateauwhere precipitation decrease andtemperature increase.Only\nlittle part,such as theTianshanMoruntain area,vegetation is increased.The character of vegetation change in northwest showe\nthat the ecological environment inthis area is becomingmore fragile,suchas shrinkingof lakes,oasis andwetland,area extending\nof soil erosion,salinization land,and desert because human activities and globalwarming.The correlation analysis betweenNDVI\nwith the temperature and precipitation in northwestern China was carried out.There was significant positive correlation betwee\nNDVIand precipitation,and temperature was not strong correlation relatively withNDVIin the northwest China.The results\nshowed AVHRR is reliable forregional vegetation change research after comparingwiththe change of precipitation and tempoera\nture.