高光谱与拟合多光谱植被指数反演武夷山亚高山草甸LAI的对比研究  

A comparison of the LAI retrieval for subalpine meadows in Wuyishan by hyperspectral and simulated multispectral vegetation index

作　　者：安德帅 徐丹丹 刘月[3] 朱建琴 AN Deshuai;XU Dandan;LIU Yue;ZHU Jianqin(College of Biology and Environment,Nanjing Forestry University,Nanjing 210037,China;Co-Innovation Center for Sustainable Forestry in Southern China,Nanjing Forestry University,Nanjing 210037,China;College of Forestry,Nanjing Forestry University,Nanjing 210037,China;Research and Monitoring Center,Wuyishan National Park,Wuyishan 354300,China)

机构地区：[1]南京林业大学生物与环境学院,南京210037 [2]南京林业大学南方现代林业协同创新中心,南京210037 [3]南京林业大学林学院,南京210037 [4]武夷山国家公园科研监测中心,武夷山市354300

出　　处：《生态科学》2022年第5期187-196,共10页Ecological Science

基　　金：国家自然科学基金(41901361);福建省林业厅资助项目(闽林科便函[(2018)26号]);江苏省“六大人才高峰”创新人才团队项目(TD-XYDXX-006)。

摘　　要：植被叶面积指数(Leaf Area Index,LAI)是重要的生态学参数,被广泛用于指示植被密度、生物量、碳、氮物质循环以及气候变化对生态系统的影响,也作为生态过程模型的重要输入参数。地面实测高光谱遥感数据能以更高的空间分辨率及更高的光谱分辨率监测植物的光谱特征,为精准反演LAI提供了基础。本项研究以武夷山国家公园黄岗山顶的亚高山草甸为研究对象,通过建立多种高光谱植被指数和拟合多光谱植被指数反演叶面积指数的统计模型,并比较高光谱与多光谱对叶面积指数反演的效果,阐明用于反演高覆盖率亚高山草甸的最适高光谱和拟合多光谱植被指数。结果表明:高光谱新植被指数(NVI)对于反演LAI有最好的效果,R^(2)=0.85,P<0.01;依据高光谱NVI拟合而成的多光谱NVI反演结果次之,R^(2)=0.82,P<0.01。几种常用比值植被指数NDVI、MSR、RVI和GNDVI在高光谱和拟合多光谱反演结果中相差不大,表现较好,R^(2)都在0.65以上。通过对比高光谱和拟合Sentinel-2A和Landsat-8两种多光谱卫星波段的反演结果发现,光谱响应函数中具有更窄波段范围的近红外、红、绿、蓝波段构成的植被指数可以得到更好的反演结果,而固定波段的高光谱植被指数未必在每种植被指数中都具有最好的反演效果。同时,发现当某种植被指数反演LAI的线性回归方程的斜率越大,说明这种植被指数越有可能随LAI的增大而出现饱和现象,相反的,斜率越小则说明该种植被指数没有出现饱和现象。此外,在研究区内使用高光谱和拟合多光谱波段植被指数法反演LAI,NDVI都获得了较好的效果,存在很好的线性关系,之前的很多研究和判断都认为NDVI不适用于反演高覆盖植被的LAI,这个发现是具有意义的,表明高覆盖植被的叶面积指数在一定范围内是能够被NDVI(应用最广泛的植被指数)较好的反演,进一步扩展了NDVI反演LAI的�The Leaf Area Index(LAI)is an important ecological parameter and is widely used as an indicator for vegetation density,biomass,carbon and nitrogen cycle,and response of ecosystems to climate change.It is also a key biophysical parameter for ecological process models.Field measured hyper-spectra has the potential to monitor plant spectral features with higher spatial resolution,which supports accurate inversion of LAI.Therefore,we establish a variety of hyperspectral vegetation index and simulated multispectral vegetation index inversion statistical models in subalpine meadows in Wuyishan National Park,and then compare the hyperspectral and multispectral leaf area index.Our results clarify the optimal hyperspectral and simulated multispectral vegetation index for LAI estimation in subalpine meadows with high vegetation coverage.The results show that Hyperspectral New Vegetation Index(NVI)has the best effect on LAI inversion,R^(2)=0.85,P<0.01;the results of simulated multispectral NVI inversion based on hyperspectral NVI fitting are the second,R^(2)=0.82,P<0.01.Several commonly used ratio vegetation indices,NDVI,MSR,RVI and GNDVI,have little difference in the results of hyperspectral and simulated multispectral inversion,and perform well,with R^(2)all above 0.65.By comparing the inversion results of hyperspectral and fitting Sentinel-2A and Landsat-8 two multispectral satellite bands,it is found that the vegetation index composed of the near-infrared,red,green,and blue bands with a narrower band in the spectral response function have better inversion results,and the fixed-band hyperspectral vegetation index may not have the best inversion effect in every vegetation index.At the same time,it was found that when the slope of the linear regression equation for the inversion of LAI by a vegetation index is larger,it indicates that the vegetation index is more likely to saturate with the increase of LAI.On the contrary,the smaller the slope,the more the vegetation index is not saturation occurs.In the literature,most st

关 键 词：高光谱 模拟多光谱 遥感反演 叶面积指数(LAI) 归一化植被指数(NDVI) 武夷山国家公园 

分 类 号：S157.2[农业科学—土壤学]