Retrieving seasonal variation in chlorophyll content of overstory and understory sugar maple leaves from leaf-level hyperspectral data

Abstract

Leaf chlorophyll content is a useful parameter for assessing vegetation physiological status and dominates the spectral signal of leaf and canopy reflectance at visible wavelengths. Using hyperspectral instruments, we quantified leaf chlorophyll content and optical properties for 255 overstory and understory leaf samples through the growing season in a mature sugar maple (Acer saccharum) stand. Strong seasonal and canopy-height-related differences were observed in both leaf chlorophyll content and leaf reflectance and transmittance spectra. Seasonal and canopy-height-related variation in leaf spectra were closely related to leaf chlorophyll content. We estimated leaf chlorophyll content using two approaches, namely empirical spectral indices, and a mathematical inversion of the leaf optical model PROSPECT. Both estimates were highly correlated with the measured leaf chlorophyll content; however, the spectral indices resulted in greater accuracy, with the best-performing index (modified simple ratio) showing an accuracy of R2 = 0.88 and RMSE = 3.94 µg/cm2. A leaf thickness factor was introduced in the PROSPECT model to take into account the effects of changes in leaf structure on light absorption. The model inversion was improved after incorporating leaf thickness factors based on observed seasonal and canopy-height-related variation in leaf thickness. The improved model had the best performance, with an accuracy of R2 = 0.93 and RMSE = 3.09 µg/cm2 in retrieved leaf chlorophyll concentration in comparison with laboratory measurements.