A multi-angular invariant spectral index for the estimation of leaf water content across a wide range of plant species in different growth stages

Abstract

Plant leaf water content plays a key role in several biogeochemical processes, such as photosynthesis, evapotranspiration, and net primary production. Yet, the accurate estimation of leaf water content using multi-angular reflectance measurements across different plant species is still challenging. This study aims to propose a generic spectral index for accurately estimating equivalent water thickness (EWT) when multi-angular spectral reflection is considered. The index was selected to have the format of a difference ratio using three reflectance factors. The reflectance factor at 410 nm was used to reduce the specular reflection from the leaf surface in the 400–2500 nm range, and the ratio of the wavelengths in the 1268–1285 nm range (at non-water absorption wavelengths) to the wavelengths in the 1339–1346 nm range (at water absorption wavelengths) strengthened the relationship with EWT for all of the sampled plant species. The modified difference ratio (MDR) index, (R1271-R410)/(R1342-R410), was linearly proportional to EWT, with R2 > 0.90, when the leaves reflection data were collected from various viewing angles. However, the relationships between some existing indices (simple difference, simple ratio, normalized difference, double difference index and difference ratio indices) and EWT at the leaf level were weak and unstable for all of the 18 plant species (including 14 broadleaf, 3 shrub, and 1 liana species) at different angles under laboratory and field conditions. Moreover, validation results from six independent datasets (n = 1800) and one modeled dataset (n = 2375) further confirmed that the algorithm derived from the proposed index (based on multi-angular reflectance factors of leaves) was not only effective for EWT estimation across a diverse set of plant species with widely variable leaf structure and water content, but also insensitive to different measurement conditions (leaf clip, integrating spheres or multi-angle measurements). The algorithm developed from this new index is generic, does not require reparameterization for each species, and can be accurately used for nondestructive EWT estimations using a simple handheld laboratory or field instrument, and thus, is convenient for agricultural and ecological studies.