摘要
草地冠层高度是指示植被生长和草地覆盖利用变化的重要物候动态及地表空间异质性指示因子。研究以内蒙古自治区锡林郭勒盟天然草地为例,使用MODIS遥感光学多角度冠层观测数据,参考RTLSR辐射传输模型完成了草地植被群落BRDF二向反射分布特征的重构,实现了冠层热点和暗点反射率的计算,构建了植被冠层归一化热点和暗点植被指数(NDHD)。研究基于GRNN神经网络模型实现了以NDHD、NDVI及EVI多种植被指数为驱动参数的草地冠层高度参数动态反演模型,结果表明:(1)基于RTLSR模型的f_(iso)、f_(vol)和f_(geo)散射核系数,可分别完成草地植被冠层热点和暗点的反射率计算,并可将体散射、几何光学散射特性用于定量描述植被结构特征的光谱响应变化。(2)使用冠层BRDF数据构建的红光、近红外NDHD植被指数可以较好地指示草地植被时序生长所表现的结构变化差异及空间覆盖分布异质性特征。(3)研究区草地冠层高度反演结果空间分布呈现西部低、中部过渡、东北部高的格局,时间序列结果表现为随草地植被物候过程而动态变化。(4)经与NDVI、EVI光谱植被指数反演对比可获知,NDHD时间和空间反演结果的R2分别为0.58和0.89,较NDVI指数(R2=0.44)和EVI指数(R2=0.36)明显改进,且红光和近红外NDHD植被指数对冠层高度反演的贡献比例数值为3.32%和3.57%,显示出较好的植被结构特征指示优势和数值反演驱动潜力。综上,基于RTLSR模型及NDHD植被指数构建的草地冠层GRNN高度反演模型在时间序列和空间分布过程均具有较理想的数值反演精度,并可为实际应用提供理论方法参考。
Grassland canopy height is a key indicator of phenological dynamic and surface spatial heterogeneity,reflecting vegetation growth and grassland cover changes.The study focused on the natural grasslands of Xilin Gol League,Inner Mongolia Autonomous Region,China.Moderate resolution imaging spectroradiometer(MODIS)optical multi-angle canopy observation data were used to reconstruct the plant structural characteristics of the grassland population′s bidirectional reflectance distribution function(BRDF)based on the RTLSR kernel-driven model.Canopy hotspot and darkspot reflectance were calculated using BRDF data,and the normalized difference hotspot and darkspot(NDHD)vegetation index was derived for the study region.Further,a dynamic inversion model of grassland canopy height(GCH)was developed using the NDHD,NDVI and EVI vegetation indices and a general regression neural network(GRNN).The GCH inversion results showed as follows:(1)Based on the kernel-driven model coefficients(f_(iso),f_(vol)and f_(geo)),the reflectance of grassland vegetation canopy at hotspot and darkspot was calculated.The volume and geometric optical scattering characteristics of the vegetation population effectively described the spectral dynamic response associated with vegetation growth.(2)The red and near-infrared NDHD vegetation index,derived from canopy BRDF data,effectively captured structural changes and spatial heterogeneity in grassland growth dynamics.(3)The retrieved GCH spatial distribution in the study area exhibited a pattern of low-to-moderate values in the west and higher values in the northeast.Temporal variations in GCH aligned with the phenological progression of grassland growth.(4)The R2 values of the NDHD-based temporal and spatial model were 0.58 and 0.89,respectively,significantly higher than those obtained from traditional vegetation indices(0.44 and 0.36).The contribution ratio of the red and nearinfrared NDHD vegetation indices were 3.32%and 3.57%,demonstrating their improved spatial heterogeneity representation and retrieval potential.In summary,the grassland canopy height inversion model,developed using the RTLSR model and NDHD vegetation index with GRNN,achieved high numerical inversion accuracy in both temporal and spatial dimensions.This approach provides a valuable theoretical and methodological reference for practical applications in grassland monitoring and management.
作者
兰春阳
郭利彪
谭维贤
黄平平
李苏和
马铭泽
LAN Chunyang;GUO Libiao;TAN Weixian;HUANG Pingping;LI Suhe;MA Mingze(School of Information Engineering,Inner Mongolia University of Technology,Hohhot 010080,China;Key Laboratory of Radar Technology and Application of Inner Mongolia Autonomous Region,Hohhot 010080,China)
出处
《中国草地学报》
北大核心
2025年第3期31-44,共14页
Chinese Journal of Grassland
基金
内蒙古自然科学基金面上项目(2021LHMS04002)
内蒙古教育厅基本科研业务费项目(JY20220198,JY20220072)
国家自然科学基金联合基金项目(U22A2010)
内蒙古工业大学科研项目(BS2021078)。
关键词
草地植被
多角度遥感
二向反射分布函数
归一化热点和暗点植被指数
冠层高度模型
Grassland vegetation
Multi-angle remote sensing
Bidirectional reflection distribution function
Normalized difference hotspot and darkspot vegetation index
Canopy height model
