Analysis of sun-induced chlorophyll fluorescence and biophysical variable patterns in a mixed forest

This work aims to analyse red and far red sun-induced chlorophyll fluorescence (F) and biophysical variable patterns in a plain mixed forest (Forêt de Hardt, France) using ground measurements and airborne data. The airborne data were acquired using the novel airborne imaging spectrometer HyPlant developed by the Forschungszentrum Jülich in collaboration with SPECIM Spectral Imaging Ltd (Finland), covering the visible, near infrared and shortwave infrared spectral regions. The HyPlant sensor is the only currently available airborne imaging spectrometer with a subnanometer spectral resolution in the region from 670 to 780 nm, offering the opportunity to quantify F at the two fluorescence emission peaks (red and far-red). The hyperspectral images were used to obtain three different products: (i) a forest species composition map based on a classification approach (ii) red and far red F maps obtained applying the singular vector decomposition (SVD) method and (iii) maps of leaf chlorophyll content (Chl) and Leaf Area Index (LAI) produced by means of radiative transfer models. Ground data were collected in a field campaign that was conducted concurrently with the airborne data acquisition with the aim to evaluate the accuracy of all the products. The ground measurements were acquired in 22 elementary sampling units of 20 x 20 m spread over an extensive area and consisted on species composition and crown condition assessment, leaf sampling for pigment extraction, digital hemispherical photography acquisition for LAI estimation, and top of canopy F measurements for each of the main forest species (i.e., hornbeam (Carpinus betulus L.), oak (Quercus robur L.), linden (Tilia L.) and pine (Pinus L.)) with high-resolution spectroradiometers manually operated from a mobile hydraulic platform. The spatial variability of the biophysical variables (Chl and LAI) and F retrieved from the airborne data were compared and analysed in relation with species-specific patterns and forest structure. Moreover, the radiative transfer model for vegetation Soil-Canopy Observation of Photochemistry and Energy fluxes (SCOPE) was used to simulate the “potential” F emission expected over the forest setting the biophysical and environmental input variables to their real values and keeping the physiological parameters fixed. The simulated potential F was then compared to the one retrieved from field and airborne measurements to identify deviations between the expected and the observed values. The spatial distribution of these anomalies was analysed in relation to species composition and canopy density and greenness. In this contribution, we show for the first time F maps in both the red and far-red region over a mixed forest. The F maps show reliable ranges of variation when compared to ground canopy F measurements performed concurrently with the overflight. The analysis of the F maps shows that the F intensity depends on the variation of forest species composition, leaf Chl and vegetation density. However, the relationship between F and the biophysical variables (Chl and LAI) is not unique, suggesting that F can provide additional information, particularly in dense and green canopies. Thus, this study highlights the potential of retrieving F maps in combination with reflectance-derived biophysical variables to improve our understanding of F dynamics in complex ecosystems. This is the concept behind the European Space Agency’s (ESA’s) FLuorescence EXplorer (FLEX) mission currently under evaluation