Performance evaluation of spectral indices to estimate equivalent water thickness

Abstract

Water is important for the growth of plants. In agriculture, Equivalent Water Thickness (EWT) is commonly used to express vegetation moisture status. Due to its simplicity, spectral indices are widely used for EWT prediction in the operational context. Currently, many indices have been proposed for estimating EWT. However, the performances of spectral indices may vary under different conditions. In this paper, we mainly investigated the performances of five spectral indices for determination of EWT with different variations and at different water levels. According to the results, all five spectral indices show good estimation for EWT with a large variation; with the decrease of sample variations, the performances of all spectral indices decline; and the strengths of spectral indices for EWT estimation vary at different water levels. Generally, the performances of spectral indices show a declining tendency as EWT increases. Of five indices, MSI produces the best estimation for low-level and medium-level EWT and WI performs best for estimating high-level EWT.