Improving Chlorophyll Fluorescence Retrieval Using Reflectance Reconstruction Based on Principal Components Analysis

Abstract

Methods based on Fraunhofer line discrimination (FLD) for solar-induced chlorophyll fluorescence retrieval require making assumptions or estimations about true reflectance, which is a large source of error, particularly at the O <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> -B band. This letter presents an alternative solution, which is named the pFLD method, based on principal components analysis to reconstruct the reflectance spectra. The principal components were generated with simulated reflectance spectra covering the most common conditions of vegetation. The pFLD method has been tested with both simulated and field experiment data sets. Compared with the widely used three-band FLD and improved FLD methods, pFLD showed better performance at the O <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> -B band, particularly when the spectral resolution (SR) or the signal-to-noise ratio was relatively low. The pFLD method can be also successfully applied to field measurements with credible accuracy even at low SR.