High-spectral resolution data for determining leaf water content in Eucalyptus species: leaf level experiments

Abstract

Laboratory measurements of the spectral reflectance of six Eucalyptus species were carried out over the 400–2500 nm range using a spectroradiometer. The relationship between leaf spectral reflectance and water content was investigated to determine the best bands for predicting water content of leaves. Water availability is a critical factor in plant survival and development and water stress is one of the most common limitations of primary productivity. Estimation of water content of forests is an important component of fire modelling as it determines the likelihood and intensities of fire. For raw reflectance spectra, strong correlations were observed at 1175 nm and 1650 nm. For the first derivative of reflectance spectra, high correlations were obtained at 940 nm, 1000 nm and 2100 nm. It is suggested that the first derivative spectra is more reliable for the remote estimation of leaf water content as it is highly correlated with leaf water content and insensitive to differences in leaf structure. Statistically significant correlations were obtained for all species in the 2050–2090 nm region of the first derivative spectra, suggesting that this region could be used to develop a general index for remote estimation of water content in a mixed forest environment.

Keywords:

• Hyperspectral
• Eucalyptus
• Leaf water content
• Leaf reflectance

Acknowledgements

We wish to thank Dr Jon Huntington (CSIRO, North Sydney) for the use of IRIS spectroradiometer and Xspectra software for the analysis of hyperspectral reflectance data. Thanks are also due to Professor Andrew Skidmore (ITC, The Netherlands) for reading the manuscript and providing helpful comments.