Sensitivity of leaf chlorophyll empirical estimators obtained at Sentinel-2 spectral resolution for different canopy structures

摘要

A comparison of the sensitivity of canopy scale estimators of leaf chlorophyll, obtainable with Sentinel-2 spectral resolution, to soil, canopy and leaf mesophyll factors, was addressed. The analysis of a synthetic dataset, generated simulating the reflectance in the 1-4 LAI range of canopies for the main general classes of leaf inclination (i.e. erectophile, plagiophile, spherical, planophile and extremophile) and for different soil types was used for such a purpose. The synthetic dataset was obtained using the PROSPECT5-4SAIL model in the direct mode with a large variety of soil backgrounds. Additionally an experimental dataset including airborne hyperspectral data gathered during ESA (European Space Agency) campaigns SPARC and AGRISAR, was employed to simulate Sentinel-2 spectral and spatial resolution, to confirm model results. Analysis of the synthetic and experimental datasets indicated that: (i) the CVI (Chlorophyll Vegetation Index), relying only on visible and NIR (Near Infra-Red) bands and obtainable at 10 m spatial resolution, can be used as leaf chlorophyll estimator, at growth stages suitable for nitrogen fertilizer topdressings, for all canopy structures except for erectophile canopies; (ii) better results can be obtained by using different indices for different leaf architectures, with TCI/OSAVI (Triangular Chlorophyll Index/Optimized Soil Adjusted Vegetation Index) performing better for erectophile canopies, whereas MTCI (MERIS Terrestrial Chlorophyll Index) provides better results for planophile canopies, despite the fact that these indices require bands obtainable at 20 m spatial resolution from Sentinel-2 data.