Delineating mangrove forest zone using spectral reflectance

The zonation of mangrove species has to be clearly identified and demarcated using geospatial data. This will help to update the zonation patterns of mangroves species associated with the current anthropogenic threat to mangrove. By using geospatial techniques and remotely sensed imagery data, the distribution of mangrove tree species associating with anthropogenic matrix can clearly be identified and mapped. The objectives of this study were to: (1) examine the variation of electromagnetic spectral reflectance on trees species and colonizing mangrove forest, and (2) demarcate the zonation of tree species in mangrove forests associated with anthropogenic activities. To identify individual mangrove species, in-situ measurement was conducted using handheld optical sensors of spectroradiometer to examine the most effective wave bands and spectral regions for discriminating mangrove tree species. To determine the significant wave bands, one-way ANOVA was applied. Later, linear discriminant analysis (LDA) was used to discriminate mangrove species. Also, laboratory tests for chlorophyll were conducted to determine the total chlorophyll contents using the same leave samples. The relationship between chlorophyll content and spectral reflectance of individual mangrove species was later identified. In order to determine the anthropogenic effect across the entire range of study area, four temporal satellite imageries Landsat 7 and Landsat 8 were analysed and compared with Boolean algebra. Mangrove loss because of anthropogenic activities was observed across the study area. Later, electromagnetic wave bands derived from the in-situ measurements were used as spectral libraries to classify individual mangrove species. Species identification with spectral library derived from in-situ measurements using SID algorithm and derived from Landsat 8 using SAM algorithm was done. Variation of species distribution associated with anthropogenic matrix was also examined. The significant species having a relationship with distance to anthropogenic activities were tested with one-way ANOVA. The study successfully discriminated 7 wave bands within visible region (400-700nm), 9 wave bands within NIR region (701-1000nm), 16 wave bands within SWIR-1 region (1001-1830nm), and 19 wave bands within SWIR-2 region (1831-2500nm). The study indicated that the leaf spectral reflectance for mangrove species provided poor reflectance at visible region due to high chlorophyll concentration. By conducting the laboratory measurement of leaf chlorophyll contents at three different observances, viz. 1) A662, 2) A663, and 3) A645, the relationship with the spectral reflectance of individual mangrove species was identified. Overall, the spectral reflectance measurement pairing with leaf chlorophyll measurement provides a sound basis for classifying mangrove tree species (R2>80%). Mangrove loss resulting from anthropogenic activities was observed across the study area. The primary driver of anthropogenic mangrove loss was found to be the conversion of mangrove to aquaculture, however logging activity showed continuous decrease of land use and land changes at 53%. Classification accuracy was observed at 84.95% and 85.21% respectively for the in-situ measurements, and Landsat 8 spectral library. The mangrove species distribution was found to be correlated with anthropogenic activities, which were randomly distributed without specific zones (SID classification- Moran index: 0.019; z-score: 0.361; p-value: 0.718; SAM classification- Moran index: 0.010014; z-score: 0.731010; p-value: 0.464773). Based on the findings, this study has shown the possibilities of discriminating mangrove trees species through chlorophyll content-to spectra linkages. The use of SID and SAM may provide the most promising classification algorithm for improving mangrove species mapping. In addition, the characteristics of mangrove zonation is better to understand the mangrove species appearance and conservation. Therefore, mangrove zonation study will remain as an important challenge for ecologists in the future.

Text FPAS 2020 18 IR.pdf Download (2MB)

Actions (login required)

View Item