基于Landsat 8 OLI遥感影像的天山北坡草地地上生物量估算  

Estimation of Grassland Aboveground Biomass Using Landsat 8 OLI Satellite Image in the Northern Hillside of Tianshan Mountain

作　　者：张雅[1,2,3] 尹小君 王伟强[1,2,3] 汪传建 鲁为华[4] 孙世泽 高军[1,2,3] 

机构地区：[1]石河子大学信息科学与技术学院,新疆石河子832000 [2]兵团空间信息工程技术研究中心,新疆石河子832000 [3]兵团空间信息工程实验室,新疆石河子832000 [4]石河子大学动物科技学院,新疆石河子832000

出　　处：《遥感技术与应用》2017年第6期1012-1021,共10页Remote Sensing Technology and Application

基　　金：国家自然科学基金项目"基于北斗终端时空轨迹和遥感的天然草地利用评估方法研究"(41461088);国家自然科学基金项目"基于绵羊放牧轨迹的绢蒿荒漠草地植物种子消化道传播研究"(31560659);石河子大学自然科学基金项目"加工番茄主要病害高光谱遥感生化反演研究"(RCZX201226);石河子大学自然科学基金项目"时空融合的荒漠土壤盐渍化动态监测"(2013ZRKXYQ18)资助

摘　　要：利用Landsat 8OLI遥感数据获得NDVI、RVI、DVI、EVI、GNDVI和SAVI等6种常用植被指数,同时结合研究区草地地面实测数据,再根据坡向将研究区划分为阴坡和阳坡两类,利用统计分析方法分别建立紫泥泉牧场阴坡和阳坡的草原生物量遥感估算模型,并进行生物量空间反演和验证。相关分析结果表明:所选植被指数与牧场生物量显著相关,依据坡向分类后数据与未分类数据相关性存在较大差异,其中NDVI相关性最高,EVI相关性最低;紫泥泉草场生物量最优反演模型为基于SAVI的二次多项式模型,精度达80%。利用该模型反演得到2015年紫泥泉牧场草原平均鲜草产草量为113g/m2,折合干草产草量41.85g/m2。研究表明:坡向是影响生物量分布变化的重要因素;利用遥感数据、地面实测生物量数据并结合研究区阴阳坡地形特征,提出的生物量估算模型精度较高,可为该牧区草原生物量合理估算和草地放牧管理提供科学依据。The Landsat 80LI remote sensing data was used to obtain six kinds of commonly vegetation in- dices including NDVI, RVI, DVI, EVI, GNDVI and SAVI. Meanwhile, combining with the measured data of grassland in the research area, the research area was divided into two kinds of shady and sunny slope ac- cording to the slope. Then the biomass remote sensing estimation models of shady and sunny slope in Ziniquan Ranch were created by Statistical analysis method and biomass space inversion and verification was implemented.The results of correlation analysis showed that the selected vegetation indices were sig- nificantly correlated with pasture biomass and there was a significant difference between the correlation of the classified data and the non classified data by slope,in which NDVI was the highest and EVI was the lowest.The optimal inversion model of Ziniquan Ranch biomass was based on the two order polynomial model of SAVI with the accuracy 80%. By using this model reversion, the grassland average yield of Ziniquan Ranch in 2015 was 113 g/mz ,which equaled to dry grass yield 41.85 g/m2.The research shows that the slope direction is an important factor affecting the distribution of biomass. Using remote sensing data and ground measured biomass data and combining with the characteristics of the topography of shady and sunny slope of the research area,the biomass estimation model has higher accuracy,which could pro- vide scientific basis for the reasonable estimation of grassland biomass and management of grassland graz- ing in the pastoral area.

关 键 词：遥感 LANDSAT 8 草地生物量 坡向 反演模型 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]