Density vs. Cover: Which is the better choice as the proxy for plant community species diversity estimated by spectral indexes?

Abstract

Recent studies have shown the crucial role of plant cover in the spectral signals of remote sensing images. However, how plant species cover affects the spectral-species diversity relationship remains largely unknown. To clarify this, we conducted a study in wetlands and obtained high spatial resolution multispectral images collected by unmanned aerial vehicles (UAVs). Quadrat information was obtained through a field survey. The traditional species diversity indexes and Hill numbers of each quadrat were computed by species density and cover. The mean, standard deviation, and coefficient of variation of the normalized difference vegetation index (NDVI) were computed based on the multispectral images. We employed univariate and multivariate regression models to assess the ability of spectral information to predict species diversity and the explanatory degree of variables to the multivariate regression models were decomposed. The results showed that rather than species density, spectral indexes were more strongly related to species diversity indexes computed by species cover, and the robustness and universality of the methods that predict plant species diversity based on spectral indexes have been improved by considering species cover. Meanwhile, in light of the flexibility of Hill numbers that assign different weights to dominant and rare species in the community, we believe that Hill numbers are the better choice in the spectral-species diversity studies, which can suitable to various communities with different density-cover relationships. In addition, our study also provided an insightful perspective for future spectral-plant species diversity research.