基于多时相CBERS/CCD数据的塔里木河下游植被变化

Vegetation change based on temporal trajectory analysis with multi-temporal CBERS/CCD data in the lower reaches of Tarim River

自2000年实施生态输水以后,塔里木河下游断流区沿岸自然植被得到了不同程度的恢复。根据逐年中等分辨率遥感数据,利用时间轨迹分析方法,对植被变化的过程和趋势进行分析。选择2000-2005年各年的最佳CBERS/CCD图像,根据土壤调整植被指数值,将各年土地覆被分为植被和非植被;基于分类结果逐像元构建土地覆被变化轨迹;并按照过程和趋势特征将变化轨迹划分为恒定非植被、恒定植被、转变为植被、转变为非植被和非稳定变化5种变化类型。通过分析植被面积变化与生态输水之间的关系、以及不同趋势类型的数量、空间特征,并结合输水资料和地面调查结果,了解各种变化趋势的变化过程、植被结构和变化原因。结果表明:植被面积随生态输水的进行呈逐年增加趋势;输水过程中,7.94%的区域呈现转变为植被的趋势,主要是草本植物恢复和灌木丛的复苏所致;同时仅有2.66%的像元转变为非植被;由于生态脆弱性较高,尚有10.43%的像元表现为非稳定变化。变化轨迹所揭示的过程特征显示,由非植被向植被转变的变化主要出现在2003年最大规模输水以后;在河水漫溢区尤为明显。实践还表明,干旱区植被变化波动频繁,时间轨迹分析方法有利于揭示植被的变化过程和趋势;而CBERS/CCD数据的免费使用,为利用中等空间分辨率遥感数据进行时间轨迹分析节约了成本。

After ecological water transfusion, restoration has taken place for vegetation to some extent along the dried-up lower reaches of the Tarim River. This study aimed to analyze the processes and trends of the vegetation change during the period of intermittent water transfusion, based on temporal trajectory analysis using multi-temporal middle spatial resolution remote sensing data. According to some conditions, we chose optimal CBERTS/CCD images of every year; the land-cover in every year was classified into vegetation and non-vegetation by the SAVI threshold respectively; then change trajectory of every pixel was constructed based on the classification results. According to the change trends, all 64 trajectories were characterized into five trend groups： stable non-vegetation, stable vegetation, change into vegetation, change into non-vegetation and unstable change classes. This paper analyzed the relationship between vegetation area change and water transfusion, the quantitative and spatial feature of different change trends and the change processes discovered by the change trajectory. Integrating ground survey data, the paper inferred the vegetation composition during the study and the reason of change. The results indicated that vegetation area increased gradually after ecological water transfusion ; and 7.94% of the study area showed the trend of changing into vegetation resulted from the re-growth of grass and recovering of shrubs, 2.66% changed into non-vegetation; at the same time, 10.43% was unstable change because of the frangibility of vegetation. The analysis of change processes indicated that the change into vegetation mainly took place after 2003, in that year the volume of water transfusion was the biggest one. The result showed that a temporal trajectory analysis is effective to monitor the processes and trends of vegetation change in arid inland where the vegetation changed frequently.