Articles
STRATEGY FOR MODELLING PLANT GROWTH
For this purpose algorithms are to be developed and tied together in a formal program.
Growth is expressed in different states of the plant (state variables) and the growth rate (rate variables) is determined by the intensity of different growth factors (driving variables).
In Hannover the goal of modelling was to elaborate predictive tools in order to provide – at the end – a production manager with a basis for decision-making applicable to his specific conditions.
Therefore growth models are submodels of a more comprehensive bio-economic-model and their design is highly influenced by this fact.
In fig. 1 the structure of the bio-economic-model for production planning and its control is shown.
There are at least two ways a growth model might be used:
- To calculate the whole duration of growth or the yield and its quality for crops and crop rotations under different temperature regimes in greenhouses.
These models are called planning models, they are static ones. - To calculate the crop growth in order to maintain a production target, that means set point control.
These models are called control models and they are dynamic ones.
Starting point of modelling was to collect data from greenhouse experiments with several vegetable crops grown under different temperature regimes.
Planting or sowing dates were varied to obtain the effects of radiation.
All other growing conditions were kept as normal as possible.
The time course of growth was determined by successive harvestings (about 10 times per crop, fig.2). Fresh weight and dry weight as well as leaf area were measured.
If possible, nondestructive measurements were preferred (e.g. tuber diameter of kohlrabi). From these data the growth period from planting or sowing until a defined weight or tuber diameter was attained could be calculated by interpolation.
No specific effect of plant age or changing weather conditions or processes of plant growth like photosynthesis or respiration were taken into account.
Therefore modelling was started from the highest abstraction level neglecting all details of growth processes (black box).
There are several reasons that forced the Hannover group