Dorsiventral and angular resistant spectral indices for estimating leaf biochemical traits: scaling from leaf to plant canopy levels

Leaf biochemical trait estimation is often limited by the dorsiventral asymmetry in leaf surface and internal structures, which causes marked differences in reflection between the adaxial and abaxial sides. In high spatial resolution remote sensing, this challenge is further complicated by leaf orientations, varying solar geometry, and sensor fields of view, which introduce multiangular reflection information. Therefore, mitigating the asymmetry effects of multiangular reflection in dorsiventral leaves is critical for accurate estimation of biochemical traits at both leaf and canopy levels. Here, we analyzed the multiangular reflection characteristics between two sides of dorsiventral leaves and proposed dorsiventral and angular resistant spectral indices (DARSIs) to estimate leaf Chl content (LCC) and equivalent water thickness based on a theoretical wavelength selection method. We found substantial variations in the accuracy of existing indices across viewing angles and leaf sides, while DARSIs achieved high accuracy by mitigating the asymmetry effect of multiangular reflection in dorsiventral leaves. When upscaling to the plant canopy level using close-range hyperspectral images, DARSIs mapped the spatial distribution of LCC more accurately. Our findings indicate that DARSIs are practical tools for estimating leaf biochemical traits and can provide valuable insights for monitoring plant physiological status.