Abstract:

Plant architecture and topology are important determinants of crop performance and agro-ecological adaptation, and should thus be taken into account in crop modelling approaches. This has not been done in most crop models designed to answer agronomic questions.
GREENLAB-Maize model combines the dynamic simulation of the complete plant architecture with simple algorithms of biomass formation, for which the growing organs are competing. Plant is considered at organ level (roots, leaves, internodes, cobs, etc.) and thus as a set of sinks competing for assimilates. Morphogenetic processes are governed by generic organ expansion laws associated with organ-specific parameters supposed to be independent from environmental conditions. Environment will then define the carbon supply available to the plant at any given time step. An advantage of the model is that it can be used to get the values of parameters for a given species (target file) to retrieve morphogenetic and organogenetic history of the plant.
Based on the above reasoning, multi-fitting technique was introduced in GREENLAB-Maize model to compute the values of endogenous parameters, which can trace back the dynamical process between source and sink as plant growth. The aim of this study was to make a first evaluation of the ability of GREENLAB-Maize to get the values of endogenous parameters taking both field heterogeneity and inter-annual trials according to the variability in environmental condition. This study focused on the case of one maize genotype cropped in three years in non-limiting conditions but with natural seasonal variation.
Four field experiments were conducted at the China Agricultural University (latitude N，longitude E). ND108 cultivar seeds were sown in rows in north-south direction with a row spacing 0.6 m and plant spacing 0.6 m within the row (27 778 plants·ha^-1). Water and nutrients were supplied to maintain non-limiting conditions. Meteorological data were acquired from a field station located on the site. Tillers were pruned systematically when they appeared to maintain ‘mono-culm’ architecture. The average number of leaves in a plant of this genotype at maturity was 21. Four plants were taken to measure the fresh weights and dimension of individual organs (i.e. internodes, leaf sheaths, leaf blades, cobs and tassels). Leaf blade area was characterized using a LI-COR Model 3100 Area Meter (Lincoln, NB, USA).
Statistical analysis of the results showed that (1) multi-fitting clearly improves the stability of parameters more among experiments and growths than single fitting; (2) the biomass of the plant and its parts were significantly different between years and seasons (except leaf sheath biomass), but not among plants sampled at the same time; (3) the differences of endogenous parameters were little between different years, growth stages and individual maize plant.
The parameters optimized with multi-fitting of year 2000 experiment were then considered as the reference set of parameters and used to simulate plant growth for other experiments. There was a good agreement between simulated and measured data for organ biomass and geometry. Then three-dimensional visualization of maize plant among different growth stages were brought about based on these processes to study the growth of individual maize plant.

Key words: Maize, Model, Architecture, Biomass partitioning, Visualization, Virtual plants