冠层水平互花米草叶片光合色素含量的高光谱遥感估算模型

以闽江河口鳝鱼滩湿地互花米草(Spartina alterniflora)的实测冠层高光谱反射率和叶片光合色素含量(LPPC)为数据源,在分析LPPC与原始光谱反射率、一阶导数光谱反射率、22种已报道光谱指数和14种新构建的植被指数相关性的基础上,利用直线回归、指数回归、对数回归以及乘幂回归方法,系统地比较了36种植被指数在估算互花米草LPPC中的表现。研究表明:(1)一阶导数光谱反射率组合的植被指数用于估算互花米草的LPPC优于原始光谱反射率;(2)红边区域一阶导数光谱是估测互花米草LPPC的最佳波段;(3)对于单一色素含量的估算,叶绿素a(Chla)的最佳估算指数为FDNDVI[723,703];叶绿素b(Chlb)的最佳估算指数为FDRVI[723,525];类胡萝卜素(Cars)的最佳估算指数为FDNDVI[723,703];(4)对于使用统一参量同时估算Chla、Chlb、Cars,由FDRVI[723,703]建立的对数估算模型效果最佳。研究成果可为湿地植物生化参量反演提供参考,也可为闽江河口湿地入侵种互花米草的动态监测和生态评估管理提供有力的科学依据。

Hyperspectral remote sensing estimation models for foliar photosynthetic pigment contents at canopy level in an invasive species, Spartina alterniflora

Foliar photosynthetic pigments are the most important biochemical parameters relative to the physiological function of wetland plants. Quantitative estimation of photosynthetic pigments can provide important information about the dynamics of the vegetation productivity, vegetation stress, or nutrient cycles within wetland ecosystems. However, the estimation of foliar photosynthetic pigments is complicated because canopy reflectance in the visible and near infrared wavelengths is affected by confounding effects that come not only from foliar photosynthetic pigment contents variation but also from the changing environmental conditions of wetland ecosystem. Our objective was to address the question by establishing hyperspectral remote sensing estimation models for foliar photosynthetic pigments at canopy level in an invasive species, Spartina alterniflora. In this study, the hyperspectral reflectance of canopy leaf and leaf photosynthetic pigment contents (LPPC) from S. alterniflora in Min River Shanyu beach were recorded as data source. The correlation between LPPC and raw spectral reflectance, the first derivative reflectance, 22 reported vegetable indices and 14 new formed indices were determined. Based on the results of correlation analysis, a total of 36 indices were tested by linear regression, exponential regression, logarithm regression and the power of regression to explore their potentials in LPPC estimation in S. alterniflora. The results showed that: (1) This study selected 5 wavebands in the region of 400-900 nm, which appeared to be the optimal wavebands for the S. alterniflora foliar photosynthetic pigments estimation. Of the selected wavebands, the most frequently occurring wavebands were 723 nm, 703 nm, 525 nm, 752 nm, 668 nm. (2) Vegetation index portfolio by the first derivative reflectance was evidently better than raw reflectance for estimating LPCC in S. alterniflora. Combining the optimal wavebands, the results indicated that the first derivative of reflectance in the red edge region (680-760 nm) was the optimal band for estimating LPCC. (3) For a single pigment content, the best indexes for estimating chlorophyll a (Chla), chlorophyll b (Chlb) carotenoids (Cars) were FDNDVI[723,703], FDRVI[723,525], and FDNDVI[723,703], respectively. The three new formed indices were proved to have better linearity with corresponding photosynthetic pigment. (4) Using the same index to estimate different pigments, the best model was the logarithmic model using FDRVI[723,703], with high predicted correlation coefficients R² of 0.6997, 0.7187, and 0.7132, respectively. This study would not only provide a good reference for hyperspectral remote sensing retrieval of biochemical variables in wetland vegetation, but also provide a strong scientific basis for the dynamic monitoring of S. alterniflora and management of ecological assessment in Min River estuary.