NDVI and fluorescence indicators of seasonal and structural changes in a tropical forest succession

Abstract

In this research, the Sun-induced Chlorophyll Fluorescence (SIF) signal was derived from field-based spectrometers and Hyperion satellite imagery of Hong Kong’s Country Parks by applying the Fraunhofer Line Difference (FLD) method. Due to high levels of atmospheric aerosols and water vapour in the study area, a modified FLD method (FLD-M) was tested for its ability to deliver improved SIF estimates by cancelling out the effects of atmospheric scattering. The SIF results were compared with the Normalised Difference Vegetation Index (NDVI), and analysed according to seasonality, five successional age groups of forest and monoculture exotic plantations. Results indicate that SIF methods are more sensitive to seasonal phenology and diurnal fluctuations than the NDVI, indicating its greater sensitivity to photosynthetic activity. The SIF responds earlier and stronger, to senescence in winter and green-up in summer. Field spectrometer data showed NDVI to be unresponsive to time of day, whereas fluorescence responds to changes in sunlight intensity from 10:15 a.m. to 02:00 p.m. Both seasonal and diurnal results indicate that SIF is better than NDVI in representing the subtle changes in vegetation conditions. Although both NDVI and SIF distinguish between the four woody structural stages of vegetation, only SIF can distinguish between forest and exotic plantations, the difference being greater in the dry season. The study provides an improved operational remote sensing methodology for investigating the health status of tropical secondary forest, where atmospheric turbidity is high.