Networks of Interactions in Intercropping

Intercropping (IC) consists in growing various crops on one single plot at the same time.

Despite the wide range of advantages, it is associated to, its adoption in Belgium remains low. In fact, many interactions can take place between two associated crops, which lead either to facilitation or competition. After several decades of research, the exact relationships leading to facilitation or competition and the conditions in which they occur are still unclear. Consequently, it is hard to predict the outcome of a given IC design in a given environment and to advise farmers.

In order to (i) better understand the mechanisms at work in intercropping and to (ii) identify the gaps in current knowledge, we did an extensive literature review and compiled the information in form of networks of interaction.

The first step was to create the rough network. To do so, we read 142 papers from major scientific journals operating in various fields, from plant physiology to soil biology and ecology. From these papers, we identify the interactions that can take place between two associated crops. In order to be as exhaustive as possible, we have taken into account both proven and assumed interactions. We then classified the gathered information into nodes, – which represent elements, e.g. “Leaf of plant A”, or processes, e.g. “Photosynthesis” -, and links, – which represent relationships between nodes, e.g. “Increases” -, and we created a global network. In order to give some consistency to this network derived from sound scientific literature, we added some links, based on implicit knowledge (e.g. Light -> Photosynthesis) and common sense (e.g. Crops 1 -> Height of crop 1).

The second step, still in progress, is the simplification of the network. As such, the global network is not usable. We are therefore currently trying to simplify it in three ways:

• By splitting the global network in smaller networks, each focusing on a particular subject (e.g. “Pests and diseases in intercropping”);
• By identifying the ten most used words in the global network and create an individual network for each word, in order to identify the relationships of main importance;
• By creating an ontology that should allow us to clarify the nature of the nodes and of the links between nodes and to assign to each node a certain level of accuracy, thus simplifying the global network by using only part of the accuracy levels.

The objective of the simplification is to get network(s) that could be used, among others, to:

• Identify the gaps in current knowledge about intercropping;
• Advise researchers and research funds on what to study in intercropping;
• Serve as a base of discussion with farmers.

The final version of the network(s) will be made available online via a RShiny application.