Estimation of leaf chlorophyll content and density in Populus euphratica based on hyperspectral characteristic variables

Abstract

Chlorophylls are the most important pigments for photosynthesis and are related to the phenology and health status of plants. Monitoring the chlorophyll content and density of Populus euphratica leaves via remote sensing technology has great significance for restoring and reconstructing damaged desert ecosystems. In this study, Populus euphratica leaves were collected in the public welfare forest reserve of Awati county, Xinjiang, China. Spectral variables and the chlorophyll content and density of leaves were obtained, simple linear regression and multiple linear regression models were built to predict leaf chlorophyll content and density. The results showed that partial least-squares regression models are the most appropriate for chlorophyll content estimation and its determination coefficient and root mean square error were 0.630 ∼ 0.745 and 0.031 ∼ 0.240, respectively. The standard deviation to root mean square error ratio was greater than 1.8 and the ratio of the standard error of the laboratory to the standard error of prediction was close to 1 for the partial least-squares regression models. These results confirmed that chlorophyll content can be accurately predicted and the proposed partial least-squares regression models were adequate for estimating Populus euphratica chlorophyll content rapidly and nondestructively, representing an alternative to conventional measurement methods. The rapid monitoring of chlorophyll content can reflect the health condition of Populus euphratica in a timely manner and provide technical support for ecological reconstruction in arid areas.