3D modelling is set to play an ever greater role in research. It’s an important tool in the quest to design better LED modules, better greenhouses and better cultivation systems. With the calculations produced in this way, researchers can pre-test the designs and then try out the best options in greenhouse trials.

“Just as an engineer designs a new industrial product and tests it using CAD/CAM, we too can model the ideal plant and ‘grow’ it in a simulated greenhouse under various conditions,” says Pieter de Visser of Wageningen University & Research in the Netherlands. The simulated greenhouse can then be optimised, bringing the ideal crop ever closer.

New research questions

De Visser and his colleague Gert-Jan Swinkels receive requests from both colleagues in the Kas als Energiebron research project and suppliers of greenhouse roof materials to calculate which aspects of a greenhouse design can be improved in order to achieve a more productive crop. Issues they look at include how diffuse the light should be to achieve the best result in both summer and winter. De Visser can’t comment on the outcomes for businesses since this type of research is strictly secret.

3D modelling has been rapidly gaining ground at Wageningen over the past two or three years. The models have steadily improved over time, which is just as well: new greenhouse and agricultural technologies raise so many questions that it would take vast numbers of practical trials to answer them, at a time when budgets are being slashed. Understanding of plant processes has also improved greatly. This benefits the model calculations on the one hand, but it also raises a whole set of new research questions.

Temperature distribution in the crop

In recent years, for example, it has become increasingly clear that the shape of the plant is a very important factor, particularly in the early stages of the crop. So what is the ideal shape? Does a cucumber crop with smaller leaves perform better in winter? Are leaves in a horizontal position better? What root system is best? What aspects of plant shape can you steer using light colours? And is there any point in steering in this way if the LAI (leaf area index, or the total leaf area per ground surface area) is already at the optimum level?

With so many questions, the 3D calculations are an effective tool for separating the meaningful from the less relevant ones and sorting them by their potential outcomes.

“Utilisation of light is still very much at the top of the list,” de Visser says. “We are doing a lot of modelling in that area, both with natural and artificial light. But we have gradually started to focus more on temperature distribution in the crop. After all, changing the light often changes the way the temperature is distributed. And if you do a lot of tinkering with a greenhouse, you get other temperature gradients and the crop really does start growing differently. You invariably get places where the crop grows more slowly or more rapidly. We can identify those places.”

Another area being given increasing attention is the shape of the root system, both in open-field systems and in greenhouse crops grown in the ground.

Lambertian distribution

The researcher is involved in a lot of light-related research at WUR’s Bleiswijk site, but commercial companies also call on his services. One example is Philips, who have developed a new type of higher-performance LED based on his 3D calculations. The greenhouse trial with the new LEDs in Bleiswijk was screened off on all sides to prevent people from looking inside.

But de Visser is happy to share some general principles for improving LED lighting: “With LED interlighting, it’s easy to work out the optimum height of the modules. We now know that the light incidence on the upper and lower surfaces of the leaf impacts differently on photosynthesis and growth. That is something we never used to take into account. Another important factor is the LEDs’ emission patterns. With a Lambertian distribution (evenly decreasing light output to the side), the modules are positioned quite low down. With an optimised emission pattern, they are hung half a metre higher and the plant makes better use of the light. There is also less light loss.”

Better models lead to new insights

Light loss is an important criterion, particularly in the Winterlight greenhouse. In winter it may make sense to grow a more open crop. According to de Visser’s calculations, the aisles should then be as narrow as possible. It also makes sense to space the plants slightly further apart in the row so that they are evenly distributed over the surface.

“Another important issue is light penetration,” he says. “When plenty of light reaches the lower leaves, they photosynthesise better. But the more light there is at the bottom of the crop, the more falls on the aisles between the rows and is therefore wasted. So increasing light penetration is not always better.”

The improved model calculations are leading to new insights. Three years ago, the researcher was able to demonstrate that interlighting was more effective than top lighting because the light from top lighting reflects off the crop and therefore can’t be used for assimilation. But the picture has since become even more nuanced. “The model shows that lighting from above delivers slightly more photosynthesis than interlighting, provided you can limit the light loss,” de Visser says. That loss can consist of reflection but also of light that falls unused on the ground.

More photosynthesis

A particular criterion to consider is the colour of the light, such as the ratio between red and far-red light. In a trial with far-red light in tomatoes, the model calculations supported the conclusion that the increase in production under far-red light was to a very small extent due to slightly higher photosynthesis, to a slightly greater extent due to the changed plant shape, and largely due to dry matter being distributed differently, probably as a result of hormonal changes. Hormones can’t yet be modelled, by the way.

In addition, the trials in which red light was alternated with pure green or blue light for a few hours were also analysed with model calculations. “The colour changes the shape of the plant,” de Visser says. “This has a dramatic effect on light interception to begin with, but there is almost no difference at all above a LAI of 3. What’s more, green is absorbed less than blue but green delivers more photosynthesis, so there is virtually no difference between green and blue at the crop level. Red scores much better, both in terms of absorption and photosynthesis, so red light is the best choice for assimilation lighting.”

Ideal plant

The model constantly needs to be updated with new knowledge, so measurements have to be taken on plants on an ongoing basis. Not enough is known about green light as yet to enable everything to be predicted on the basis of a model, for example. De Visser again: “We calculate the results before the greenhouse trials, and this enables our colleagues to structure the trials better. An important question is what plant shape is best for intercepting light from LEDs. We are currently studying this in a project in collaboration with Bayer CropScience. You need to know what the ideal plant is and whether a new technique could help achieve it. So the model informs the trials and the trials inform the model.”


New greenhouse and agricultural technologies are producing so many options that it is impractical to investigate them all in greenhouse trials. Calculations with 3D computer models act as a filter so that only the most promising options are studied in the greenhouse setting. The main focal points are the shape of the plant and its roots, utilisation of light and temperature distribution in the greenhouse and the crop. One example being studied is the optimum position and radiation of LED modules.

Text: Tijs Kierkels.
Images: Wilma Slegers and WUR.