PIOSL: An improved version of PROSPECT model for simulating leaf optical properties by considering leaf internal structural variations

Abstract

Leaf radiative transfer simulation is an important tool for studying the quantitative relationships between plant biochemical parameters and leaf spectra, at the leaf scale. This type of simulation can be used to investigate leaf structures and their biochemical parameters with regard to reflectance and transmittance, and can produce mathematical representations of the radiative transfers of leaves. Most current radiative transfer models represented by PROSPECT are based on the assumption that all leaf interiors are turbid and homogeneous, ignoring the impacts of leaf layer structures on their spectra. A leaf is composed of two heterogeneous layers, we assume that the optical properties of each layer are different. A PIOSL (PROSPECT considering the internal optical structure of leaves) model was proposed, with the assumption that a leaf is composed of two stacked plates with different biochemical parameters and structures. The PIOSL model was used to simulate the directional-hemispherical reflectance of plant leaves for four publicly available datasets. The structural parameters and biochemical contents of the two layers used in the simulation were determined using the bald eagle search optimization algorithm. The results showed that the structural parameters of the first layer were generally higher than those of the second layer in leaves of different plant species; chlorophyll was mainly distributed in the upper layer, while water and dry matter were concentrated in the lower layer—although some plant species showed different trends. PIOSL outperformed PROSPECT-4 to simulate leaf optical properties of various plant species.We therefore concluded that the PIOSL model better explains the responses of leaf biochemical parameters to leaf spectra, and provides a new strategy for modeling the leaf radiative transfer process.