Exploring the feasibility of early stress detection with sun-induced chlorophyll fluorescence from tower to satellite

Terrestrial ecosystems are undergoing increasingly frequent periods of stress as the climate is changing. Monitoring ecosystem health efficiently requires global spatial coverage and high temporal resolution, which are assets of satellite-based remote sensing. However, conventional optical remote sensing (RS) approaches offer limited potential for the early detection of ecosystem stress, as changes in ecosystem structure and function often need to be substantial in order to be detectable when using reflectance in the visible and near-infrared range of the energy spectrum. Satellite-based RS of sun-induced chlorophyll fluorescence (SIF) offers much greater promise to that end, but there is still no dedicated SIF mission in-orbit, leaving coarser instruments designed for atmospheric measurements, such as Sentinel-5P TROPOMI, as the only option. The use of SIF from such instruments is challenged by the confounding effects of canopy structure and biochemistry. Furthermore, to correctly diagnose whether plants are under stress, SIF needs to be quantified jointly with the energy that is dissipated as heat. This can be potentially done through monitoring changes in reflectance around the green peak, exploited by the so-called photochemical reflectance index (PRI). Another option may lie in constraining the system with information on land surface temperature (LST), but such measurements should be ideally made at the same time as the instantaneous SIF retrievals. Another challenge is that, combining these measurements requires the proper confrontation of very different spatial footprints over potentially heterogeneous landscapes.

This present work reflects some of the results emerging from the AustroSIF project. The overarching goal of AustroSIF is to make present and future satellite-based SIF measurements a sensitive and reliable means for the early detection of ecosystem stress by combining remotely sensed SIF and PRI. In this project, we have gathered time series of ground-based active and passive chlorophyll fluorescence and hyperspectral reflectance from 7 eddy-covariance flux tower sites. At the same time, we collected respective time series of Sentinel-5P TROPOMI SIF data and MODIS MAIAC PRI data, which we complemented with sub-daily LST measurements from MSG SEVIRI. We also collocated datacubes of Sentinel-2 data to quantify the spatio-temporal heterogeneity within the large TROPOMI and MSG footprints. Finally, we also derived a series of senSCOPE simulations enabling us to place all variables in a synthetic environment to test the strength of the SIF-PRI relationship under stress conditions. By combining together the ground measurements, the satellite measurements, and the simulations, we are able to provide a first glimpse of how well the SIF-PRI relationship can be applied in practice over a variety of  ecosystems.