Discriminating trees species from the relationship between spectral reflectance and chlorophyll contents of mangrove forest in Malaysia

Abstract

Abstract In situ spectral signatures of 19 mangrove species were measured to investigate whether mangrove species could be discriminated through spectral reflectance data. The study was conducted at Matang Mangrove Forest Reserve and the spectral signatures were recorded by using a handheld spectroradiometer. The reflectance data was analysed using one-way ANOVA to identify bands that exhibit significant difference (at 99.99% level) across the mangrove species. Potential important wave bands that can be used to discriminate mangrove species were identified by using linear discriminant analysis (LDA) (at 99.95% level). The study successfully discriminated 7 wave bands within visible region (400–700 nm), 9 wave bands within NIR region (701–1000 nm), 16 wave bands within SWIR-1 region (1001–1830 nm), and 19 wave bands within SWIR-2 region (1831–2500 nm). Previous studies indicated that the leaf spectral reflectance for mangrove species was reported to provide poor reflectance at visible region (400–700 nm) due to high chlorophyll concentration. Leaf surface reflectance appeared to be the most important factor in this variation. By conducting the laboratory measurement of leaf chlorophyll contents at three different absorbance viz. 1) A662, 2) A663, and 3) A645, the relationship with spectral reflectance of individual mangrove species was identified. Overall, spectral reflectance measurement pairing with leaf chlorophyll measurement provides a sound basis for classifying mangrove trees species (R2 > 80%). The study has shown that there are possibilities for discriminating mangrove trees species from chlorophyll content-to spectra linkages.