Plant production (biomass) modelling may be complex, especially if structural aspects are also considered.


Plant production modelling simplified

Sound simplifications, demonstrated by agronomic experiments, are of major interest for defining efficient models.

At crop level, taking into account the architecture of the plant is not always necessary.
However, knowing the distribution of the various organs at respective occurrence times and physiological ages is a key point.
In other words, geometrical aspects may be ignored, but not the quantitative composition of the structure.
We postulate the existence of a common biomass pool.

Similarly, we can consider only net photosynthesis, i.e. the proportion of sugars involved as construction materials in the dry matter of the organs.

Plant modelling principles

In general,dry matter production is divided into compartments (leaves, stems, roots, fruits) whose proportions are sinks for storage of the biomass produced.
In fact, some organs produce biomass, such as leabes and seeds, are sources, while others consume, and are sinks, such as internodes and fruits.

In models predicting crops, equation (1) defines the basic framework.
Indeed, it accumulates dry matter in the compartment of interest, day after day, processing the energy received by the canopy, and sharing the dry matter among sinks.

Functional Structural Plant models (FSPM)

Functional structural plant models are designed to model and simulate both plant structure establishment and biomass production.
They are therefore centered on individual plants.

Development of the architecture is controlled by a grammar or automaton that pilot organogenesis vs. time (ideally vs thermal time).
Branches consist of phytomers, the elementary axis entity, consisting of an internode, one or more axillary buds and its leaves.
Axillary buds may give rise to branches or fruit.

Most FSPM approaches use the (dynamically computed) plant structure as a biomass transportation pathway.
Production is thus not acomputed at organ compartment level, but at each individual organ level.
Depending on its nature, each source organ may diffuse matter within the structure pathway, and/or, conversely, may consume biomass.
In such approaches, local pools of biomass are locally stored and managed.

FSPM implementation is based on a mesh that allows a geometric representation of the plant with polygons (these leaves will intercept light), or other shapes.
An underlying graph structure reflects the plant architecture in order to compute the propagation of sugar and water in the structure. Source organs send the biomass they produce into the network and the sink organs intercept it according to their sink strength.

the FSPM approach mimics the morphogenesis process in its entirety.
Such systems provide a better understanding of the plant, but cannot address simulations on the scale of agricultural production.
Indeed, simulation time and memory costs usually become substantial, which, added to a large number of parameters to be identified, may usually not be lead to convenient model calibration under true field conditions.