Multiple-constraint inversion of SCOPE. Evaluating the potential of GPP and SIF for the retrieval of plant functional traits

The most recent efforts to provide remote sensing (RS) estimates of plant function rely on the combination of Radiative Transfer Models (RTM) and Soil-Vegetation-Atmosphere Transfer (SVAT) models, such as the SoilCanopy Observation Photosynthesis and Energy fluxes (SCOPE) model. In this work we used ground spectroradiometric and chamber-based CO2 flux measurements in a nutrient manipulated Mediterranean grassland in order to: 1) develop a multiple-constraint inversion approach of SCOPE able to retrieve vegetation biochemical, structural as well as key functional traits, such as chlorophyll concentration (Cab), leaf area index (LAI), maximum carboxylation rate (Vcmax) and the Ball-Berry sensitivity parameter (m); and 2) compare the potential of the of gross primary production (GPP) and sun-induced fluorescence (SIF), together with up-welling Thermal Infrared (TIR) radiance and optical reflectance factors (RF), to estimate such parameters. The performance of the proposed inversion method as well as of the different sets of constraints was assessed with contemporary measurements of water and heat fluxes and leaf nitrogen content, using pattern-oriented model evaluation. The multiple-constraint inversion approach proposed together with the combination of optical RF and diel GPP and TIR data provided reliable estimates of parameters, and improved predicted water and heat fluxes. The addition of SIF to this scheme slightly improved the estimation of m. Parameter estimates were coherent with the variability imposed by the fertilization and the seasonality of the grassland. Results revealed that fertilization had an impact on Vcmax, while no significant differences were found for m. The combination of RF, SIF and diel TIR data weakly constrained functional traits. Approaches not including GPP failed to estimate LAI; however GPP overestimated Cab in the dry period. These problems might be related to the presence of high fractions of senescent leaves in the grassland. The proposed inversion approach together with pattern-oriented model evaluation open new perspectives for the retrieval of plant functional traits relevant for land surface models, and can be utilized at various research sites where hyperspectral remote sensing imagery and eddy covariance flux measurements are simultaneously taken
JPL, MM, and MR acknowledge the EnMAP project MoReDEHESHyReS “Modelling Responses of Dehesas with Hyperspectral Remote Sensing” (Contract No. 50EE1621, German Aerospace Center (DLR) and the German Federal Ministry of Economic Affairs and Energy). DM, MM and MR received funding from the European Union’s Horizon 2020 research and innovation programme via the TRUSTEE project under the Marie Skłodowska-Curie grant agreement No. 721995. Authors acknowledge the Alexander von Humboldt Foundation for supporting this research with the Max-Planck Prize to Markus Reichstein; SynerTGE “Landsat-8 + Sentinel-2: exploring sensor synergies for monitoring and modelling key vegetation biophysical variables in tree-grass ecosystems” (CGL2015-69095-R, Spanish Ministry of Science, Innovation and Universities); and FLUχPEC “Monitoring changes in water and carbon fluxes from remote and proximal sensing in Mediterranean ‘dehesa’ ecosystem” (CGL2012- 34383, Spanish Ministry of Economy and Competitiveness). Authors are very thankful to Dr. Karl Segl and Prof. Dr. Luis Guanter for their support with the EnMAP end-to-end simulator; as well as the MPI-BGC Freiland Group and especially Martin Hertel as well as Ramón LópezJiménez (CEAM) for technical assistance. We are grateful to all the colleagues from MPI-BGC, University of Extremadura, University of Milano-Bicocca, SpecLab-CSIC, INIA and CEAM which have collaborated in any of the field and laboratory works. We acknowledge the Majadas de Tiétar city council for its support.