Spectrodirectional Minnaert-k retrieval using CHRIS-PROBA data

Abstract

We report on a detailed analysis of hyperspectral and multidirectional remote sensing data acquired using the Compact High Resolution Imaging Spectrometer (CHRIS) mounted onboard the Project for On-Board Autonomy (PROBA) spacecraft. This instrument is capable of sampling reflected radiation over the visible and near-infrared (NIR) region of the solar spectrum at a spatial resolution (approx. 17 m) intermediary between sensors traditionally used in land applications (such as Landsat and Satellite Pour l'Observation de la Terre (SPOT), 30 m–50 m) and the latest instruments delivering a nominal resolution of 1 m or less. The spectral anisotropic signature of an Alpine coniferous forest during winter in relation to canopy cover was investigated using the Minnaert-k parameter obtained by inverting the Rahman–Pinty–Verstraete (RPV) model against CHRIS data. Although earlier studies have demonstrated that Minnaert-k can be used to characterize surface heterogeneity at subpixel scale, its spectral dependency has not yet been fully assessed in an imaging spectrometry context. Minnaert-k parameter retrievals across CHRIS bands revealed that a switch from bell-shaped to bowl-shaped anisotropic reflectance patterns occurs when comparing visible to NIR responses. Specifically, the degree of canopy cover and background brightness determine where in the spectral domain this switch in reflectance anisotropy occurs. For a bright snow cover background Minnaert-k values correlated best with canopy cover at the end of the red edge (e.g., around 735 nm). In this spectral region, pixels with medium canopy cover (40%–70%) typically produced bell-shaped anisotropy patterns, while pixels with sparse (<30%) or dense (>70%) canopy covers typically produced bowl-shaped reflectance anisotropy patterns.