• Reduce text

    Reduce text
  • Restore text size

    Restore text size
  • Increase the text

    Increase the text
  • Print

Vue d'ensemble de parcelles de différentes variétés de céréales : catalogue variétal des blés tendres d'hiver et des orges. La Minière (Yvelines).. © INRA, FOUCHARD Marc

Innovative Agricultural Systems: System Experiments

Birth of an idea

Jean-Marc Meynard, director of Inra’s Science for Action and Sustainable Development department from 2003 to 2012, was an early proponent of system experiments. Here, he revisits the development of this new way to think about experimentation.

System experiments are being scaled up to the level of an entire farm, or even of a landscape.. © inra, Christian Slagmulder
System experiments are being scaled up to the level of an entire farm, or even of a landscape. © inra, Christian Slagmulder

What makes rule-based system experiments unique is that they test a set of techniques that form a cohesive crop and/or livestock system rather than testing techniques in isolation. This way, the system experiment simulates the way such techniques would be put into practice on a farm.


System experiments versus factorial experiments

Agricultural experiments, whether on research stations or on farms, have certainly always been a source of essential data for Inra. Until recently, however, it is largely factorial experiments that have been performed – seeking to learn more about a poorly understood process, to model the effects of a practice, or to compare the effects of different environments on varieties or breeds (genotype–environment interaction). This approach was, and remains, the basis for an increasingly refined understanding of factors affecting production and the impact thereof on the environment.

Jean-Marc Meynard. © © DR
Jean-Marc Meynard © © DR

Idea developed in the 1980s

The development of rule-based system experiments in the mid-1980s can be seen as a response to the limits of the then widely prevalent idea of reductionism. Agronomists and animal science technicians were beginning to feel that not everything could be understood, or predicted, by dissecting reality into smaller and smaller pieces. The attention paid to the practices and decision-making strategies of crop and livestock farmers, to feedback loops, and to new notions of a system being more than the sum of its parts, led to a theorisation of crop systems in the 1970s and of livestock systems in the early 1980s. Quickly thereafter the first innovative system experiments were carried out on wheat and beet crop management techniques at Grignon in the Île-de-France region and on sheep per hectare ratios at Redon in Auvergne. This new way to conduct experiments was initially given short shrift. Results from the first system experiments were published only in second tier journals or as book chapters.


Consolidation in the 1990s

Over time, scientists began to make use of system experiments, as highlighted by the 2006 Inra mission report Conception de Systèmes Agricoles Innovants (Designing Innovative Agricultural Systems). In the 1990s, system experiments were given a rule-based framework. Thus, system experiments came to evaluate the suitability of the rules to meet a stated objective.


Prototype approach versus step-by-step approach

In the 2000s, there were discussions around the rules of multiyear experiments and whether they should remain unchanged from year to year to ensure the experiment’s uniformity, or whether ineffectual rules should be changed for the betterment of the system. The former creates a predefined prototype system, whereas the latter leads to a step-by-step system.

Today, system experiments are being put in place on a growing number of research stations and on wide range of issues. They put agroecology ideas and knowledge to practice. System experiments are also being scaled up to the level of an entire farm, or even of a landscape.

Scientific contact(s):

Associated Division(s):
Science for Action and Development