The effect of using different vegetation indices for mangrove leaf area index modelling

Abstract

Spectral vegetation index (SVI) is a mathematical combination of image spectral bands, especially ranges from visible to near infrared portion of the light spectrum. The purpose of SVI is to emphasize the vegetation content information from an image and doesn’t directly correlate with any physical or bio-physical characteristics of vegetation. One of the important biophysical parameters of vegetation that can be derived from SVI is Leaf Area Index (LAI). LAI can be defined as one half the total green leaf area per unit horizontal ground surface area and considered as an indicator to determine the level of mangrove health. Various SVIs have been developed and different SVI affects the accuracy of the LAI model. This study aimed to (1) compare and contrast the performance of several SVIs applied to WorldView-2 (WV-2) image to estimate the LAI, and (2) find the most accurate model for estimating LAI of mangrove forest in Perancak Estuary, Bali. The SVI used are Simple Ratio (SR), Normalized Differenced Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Green Atmospherically Resistant Vegetation Index (GARI), dan Wide Dynamic Range Vegetation Index (WDRVI). The LAI models developed were based on the semi-empirical relationships between SVIs and field LAI measured from hemispherical photograph. The corresponding values of both parameters were correlated to find the regression function for the modelling. The results show that the best accuracy was obtained from NDVI which has an R2 value of 0.83 and an estimation accuracy of 89.10 %.