Simulation and Mapping of Leaf Area Index (LAI) using Hyperspectral, Multiangular PROBA/CHRIS Data

Monitoring and mapping of vegetation at global and regional scales is essential since they serve as critical inputs to the general circulation models that predict future trends in global climate. Accurate and quantitative vegetation information extraction from remotely-sensed data relies on the satellite and airborne sensors and the efficient processing applied to such data. To characterise the vegetation function and anisotropy, one of the requirements is the availability of Multiple View Angle (MVA) information. A unique, current satellite configuration that offers both Hyperspectral (HS) and MVA data is from PRoject for On-Board Autonomy (PROBA)/Compact High Resolution Imaging Spectrometer (CHRIS) sensor. The main aim of this study is to evaluate the potential of HS and MVA remote sensing from CHRIS to characterise vegetation through the retrieval of bio-physical and biochemical parameters. Key parameters such as LAI, Leaf Chlorophyll Content (LCC), Leaf Angle Distribution (LAD) and Canopy Chlorophyll Content (CCC) that define the structural and biochemical composition of plant canopies were estimated. Spectrodirectional reflectances from CHRIS sites, Chilbolton, UK are obtained after the pre-processing of CHRIS using a range of techniques from noise removal, cloud and cloudshadow masking, atmospheric correction and geometric registration. A physically based Canopy Reflectance (CR) model, SAIL along with a Leaf Reflectance (LR) model, PROSPECT is coupled to simulate the spectrodirectional reflectances of a spring barley canopy in Chilbolton. Sensitivity analysis of two parameters at leaf level: Leaf Protein and Leaf Cellulose+Lignin content to the simulated reflectances are carried out. The modelled reflectances, thus obtained, are stored in a LookUp Table (LUT) and are inverted against the CHRIS reflectances. The potential contribution of HS and MVA in comparison to multispectral (with less than 10 spectral channels) and single view angle (nadir-or near-nadir view) configurations are evaluated in terms of the retrieval of the biophysical and biochemical parameters. Despite problems in the radiometric calibration of CHRIS data, results of this research shows that plausible vegetation parameters can be derived that agree well with the field measurements of the parameters. The capabilities of the CHRIS sensor to provide quantitative estimates of vegetation are established. This study demonstrates the capability of HS and MVA remote sensing for the effective characterisation of vegetation parameters.