PubAg

Main content area

Effect of dust deposition on spectrum-based estimation of leaf water content in urban plant

Author:
Lin, Wenpeng, Li, Yuan, Du, Shiqiang, Zheng, Yuanfan, Gao, Jun, Sun, Tengteng
Source:
Ecological indicators 2019 v.104 pp. 41-47
ISSN:
1470-160X
Subject:
dust, environmental indicators, fuels, hyperspectral imagery, leaves, pollution, trees, water content, China
Abstract:
Hyperspectral technology has equipped large-scale efforts to estimate leaf water content. In an urban environmental context, these efforts may be hampered by dust deposition on leaves. However, the influence of dust deposition on the accuracy of estimating water content is still to be elucidated. This study aimed to comprehensively analyze the influence of dust deposition on the spectrum-based estimation of water content in Shanghai, China. We measured the amount of dust deposition, water content, and the spectrum of 707 sample leaves. We then derived the spectrum-based indexes of water content and assessed the influence of dust deposition on the performance of such indexes in estimating leaf water content. The results revealed that dust-polluted leaves retained the correlations between the observed water content indexes, which includes Equivalent Water Thickness (EWT) and Fuel Water Content (FWC), with spectrum-based indexes. However, these correlation coefficients dramatically varied with the changes in dust deposition amount. EWT had weaker while FWC got stronger correlation coefficients with spectral indexes for dusty leaves compared with clean cases. The spectral indexes were correlated with EWT at a consistently stronger extent than with FWC, suggesting an advantage of using EWT over FWC to represent the leaf water content. The correlations between EWT and spectral indexes were stronger with leaf samples from the same tree species compared with the case of using leaf samples from mixed species. These findings can strengthen the understanding of the influence of dust pollution on leaf water content estimation using spectral indexes.
Agid:
6394785