ESTIMATING THE 3-DIMENSIONAL GEOMETRY OF A MAIZE CROP AS AN INPUT OF RADIATION MODELS - COMPARISON BETWEEN 3-DIMENSIONAL DIGITIZING AND PLANT PROFILES

The radiation regime of the plant canopies is largely influenced by the geometrical structure, i.e. the angular and spatial distributions of the leaf area. In this paper, two methods for investigating the structure of a maize row crop are compared. The first is a new method which consists of a three-dimensional (3D) digitizing of the foliage using a sonic 3D digitizer. The second one is the plant profile method in which a picture of the plant is taken and then digitized in two dimensions. The latter assumes that the leaves of a plant are in a single vertical plane called the 'azimuthal plane'. The 3D digitizer has the ability to locate a given point with an accuracy of within +/- 1 cm: the representation and the repeatability of a leaf description are satisfactory and not largely influenced by the digitizing density. The azimuthal plane assumption of the plant profile method was then tested using the results of 3D digitizing as the reference. Regarding the individual plant description, the leaf azimuth function is bimodal as expected for the maize phyllotaxy. The theoretical azimuthal plane was computed by a principal component analysis. The plant leaf area projection onto this plane represents about 96% of the 3D spatial distribution. Moreover, this plane is close to the azimuthal plane estimated by eye. That is why the spatial distributions of the leaf area in the azimuthal plane are quite similar for the two methods. However, the leaf area with an azimuth within +/- 15-degrees around the plant azimuth is less than 40% of the total area and the foliage spreads out in a slice 40 cm thick, around the azimuthal plane. Regarding the canopy structure description, the two methods gave similar results for both the leaf inclination and the vertical leaf area density function. In contrast, the distributions in the horizontal plane were different: the plant profile method overestimates the row effect, increasing the foliage clumping around the stem because of the plant azimuth plane assumption. Such discrepancies have repercussions on the modelled radiative balance of a row crop, so the plant profile method should not be used in estimating the spatial distribution of the leaf area in the horizontal plane.