新疆阜康绿洲生态系统生物量遥感估算分析

Estimation of Ecosystem Biomass in Fukang Oasis of Xinjiang Using Remote Sensing

利用阜康绿洲地区2003年8月野外实测53个样方的生物量干重数据和同期的陆地卫星MODIS的1通道,2通道250m遥感图像数据,分析了植被指数NDVI与绿洲植被生物量的相关关系,进而建立该遥感植被指数与绿洲生物量的一元线性和非线性回归模型,并对不同回归模型进行分析比较.研究表明:植被指数NDVI与绿洲生态系统植被生物量之间存在较好的相关性;所建遥感植被指数与绿洲植被生物量的回归模型中,三次方程得到的非线性回归模型(y=-5 593.3NDVI3+7 509.7NDVI2-1 268.9NDVI+191),是最适合用于监测阜康绿洲生态系统植被生物量的监测;利用该模型进行反演同年4月～9月研究区的生物量,并分析得出阜康绿洲生态系统内不同植被生物量在同一时期的空间分布特征;阜康绿洲生态系统内不同植被生物量在不同时期生物量的独特时空分布特征.

The paper analyzes regression between ecosystem biomass and the vegetation index （NDVI） and establishes linear and non-linear regression models between them in Semi-Arid and Arid Land of Fukang Oasis of Xinjiang using MODIS Thematic Mapper data in its 1 - 2 band with 250 meter resolution and data from the field survey of 53 sampling sites in August 2003, which were interpreted from GPS, and therefore explores the relationship between ecosystem biomass and NDVI in Semi-Arid and Arid Land. The major conclusions obtained from this study are： 1 ） fitting accuracies from the non-linear regression models are much higher than those from linear regression models, and the result could represent ecosystem biomass measured in Fukang Oasis; 2） comparison of the relations between the vegetation index and the field measured datum through non-linear regression analysis indicates that the cubic equation is the best one for the study area; 3） the non-linear regression model： Y = -5 593.3NDVI^3 ＋ 7 509.7NDVI^2 - 1 268.9 NDVI ＋ 191 is the best model which can be used to monitor vegetation biomass of the Oasis ecosystem in the region; 4） biomass from April to September is estimated using the model, and spatial and temporal distribution of biomass of different vegetations at both the same time period and different time period is identified.