A dynamic model of maize 3D architecture: Application to the parameterisation of the clumpiness of the canopy

Abstract

A dynamic 3D maize canopy architecture model is proposed for radiative transfer computation required for canopy functioning or remote sensing applications. It is based on a previous static model describing the 3D architecture of fully developed plants observed at the male anthesis stage. Laws of development and growth in dimension of the stem and the leaves are established based on experimental observations, in order to infer plant architecture at any stage from that of fully developed plants. The leaf curvature and shape are assumed to be the same over the whole leaf duration, with the exception when leaves are still within the top leafy cone at younger stages. The time is described by the number of visible leaves, which can easily be deduced from the cumulated growth degree days. The model requires only four input variables: the sowing pattern (row distance, plant density), the final number of leaves produced, the maximum height at anthesis, and the cumulated leaf area for the fully developed plants. It was validated on independent data sets and provides globally good performances. The model is later used to parameterise the canopy gap fraction which is one of the main variables governing radiative transfer processes. The gap fraction P(o)(θ) for the observation direction is classically described by an exponential function of the leaf area index, L: P0(θ) = e-λ0G)/cos(θ)·L where λ0 is the clumping parameter describing the non-random leaf arrangement and G is the projection function that depends on the leaf inclination distribution function. The gap fraction model was adjusted over a time series of maize canopies simulated using our 3D dynamic canopy architecture model. We showed that maize canopies have a marked clumped character, with an average clumping parameter of λ0 ≃ 0.8. However, results suggest that the clumping parameter depends on the developmental stage of the canopy, and, to a lesser degree, on the observation direction θ.