Elucidating physiology of plant mediated exchange processes using airborne hyperspectral reflectance measurements in synopsis with eddy covariance data

International audience; The Carbo Europe Regional Experiment Strategy (CERES), conducted between May and June 2005, aims to determine a spatially resolved regional balance of carbon dioxide fluxes using different methodological approaches bundled within an international framework. In this study we elucidated several meteorological and physiological parameters determining plant mediated exchange processes using airborne hyperspectral reflectance measurements in synopsis with meteorological information like eddy flux covariance data. We used an airborne hyperspectral system to record spatial and temporal transects of vegetated areas surrounding flux tower sites in the Bordeaux / Landes Region (France). Alongside hyperspectral measurements a comprehensive range of meteorological and surface flux parameters were measured at a flux tower site within the same time frame. Unsupervised data analysis using cluster analysis was performed on predefined spectral wavelength windows. Additionally multiblock principal component analysis was used as explanatory data driven tool to reveal underlying data structures and to elucidate potential interdependencies between airborne hyperspectral reflectance data and meteorological ground measurements. In detail temporal, physiological changes in relevant ground information like temperature, humidity, vegetation mediated CO2 flux, latent heat flux or photosynthetic light use efficiency on the one hand and selected regions of the hyperspectral signatures, such as reflectance in the region of 531nm, which determines variations in the photochemical reflectance index (PRI) or in the spectral regions of chlorophyll fluorescence were examined. Unsupervised cluster analysis revealed coupled dependencies between changes in the derivate spectra in the range of 720 to 740 nm with gross photosynthetic uptake rate, global radiation and time of day. Multiblock principal component analysis revealed that first derivative reflectance in wavelengths from 500 to 540 nm and between 680 and 750 nm had higher loading values in respect to observed ground variables global radiation and gross photosynthetic uptake rate respectively. The significance of these findings is discussed.