Moderate-resolution imaging spectroradiometer-based vegetation indices and their fidelity in the tropics

Abstract

Moderate-resolution imaging spectroradiometer (MODIS)-derived vegetation indices (VIs)-the Normalized Difference Vegetation Index (NDVI), the Enhanced Vegetation Index (EVI), and the Land Surface Water Index (LSWI)-are evaluated in terms of their sensitivity to the seasonal greenness pattern and moisture regime of the tropical forests of Hawaii. The annual mean NDVI and EVI signals most rapidly saturated as total annual rainfall increased to a mesic condition, but LSWI was responsive to a much wetter environment. Ecoregional analyses of biweekly VI time series revealed that all three VIs followed the typical pattern of summer-low and winter-high rainfall for dry forests and shrublands. However, NDVI and EVI did not show any significant seasonality of wet forests while LSWI represented a summer-high and winter-low greenness pattern. The three VIs did not respond to the Leaf Area Index (LAI) very well as LAI reached 4, but they sensitively responded to the fraction of photosynthetically active radiation (fPAR) and leaf moisture content. Especially, LSWI responded most sensitively to fPAR and leaf water content in the wet environment, where fPAR and leaf water content were >0.6 and >40%. Greenness seasonality was more strongly represented by LSWI than by NDVI and EVI for all ecoregions considered in the study. In short, it is believed that LSWI is more appropriate for a canopy phenology study than the other two VIs in wet forests of the tropical environment.