Monitoring screen of the computer piloting the Phenoarch platform. © INRA, SLAGMULDER Christian

Modelling and agrosystems

By Pascale Mollier - Catherine Foucaud-Scheunemann, translated by Daniel McKinnon
Updated on 09/11/2013
Published on 05/29/2013

To make agriculture more sustainable, new crop systems must be developed that meet the widest-possible array of economic, environmental, and social objectives. Modelling has become an indispensable tool for recognising, understanding, developing, and sharing these new production strategies.

Developing new crop systems is a major focus area for agricultural research at present. Doing so is very challenging because it involves fundamental changes to complex systems that have many interdependent parameters.

Innovating in complex systems

As environmental objectives become graver and more ambitious, the need for fundamental change in our agricultural models is clear. France, for example, set a goal in 2007 to reduce phytosanitary product use by 50% and to do so, if possible, by 2018 as a part of its Ecophyto 2018 plan. To meet an objective such as this, radical innovation is needed because it calls into question the rationale of existing systems. This was demonstrated in the Ecophyto R&D study coordinated by INRA (2007–2009). Indeed, a sharp reduction in pesticide use has impacts across the entire production chain with regard to the choice of varieties, the question of nitrogen fertilisation, and the organisation of crop rotations, which includes diversifying crops and requires developing new distribution channels to make use of such crops.

Furthermore, agronomy criteria are interdependent and are linked to economic, social, and environmental criteria, such as the protection of natural resources or of biodiversity. Moreover, a crop system is no longer considered solely at farm level, but is seen across a number of farms, at the landscape and industry levels as well.

Modelling offers a way to innovate within this complexity.

The era of systems modelling

In agriculture as in biology, an increasingly large number of variables are being taken into consideration as our understanding of the subject deepens and as the need to synthesise this diversity builds. Now, not only is there a need to describe plants and their biophysical environments, but also to describe less quantifiable or predictable variables as well, such as pollen dispersal, trends in pest populations, and even farmer behaviour. Modelling is able to give structure to this host of information, both to understand the soil–plant–climate system and to develop and evaluate new crop systems that would be impossible using experimentation alone. Modelling does not replace experimentation however. Experimentation is still needed to evaluate the strength and the appropriateness of the models.

Modelling is omnipresent and necessary

To give an idea of the importance of agricultural modelling at INRA, a recent study revealed that one of INRA’s 13 research divisions, Environment and Agronomy, had more than 40 models at work. The division’s models can describe, evaluate, and manage the many variables of a crop system, such as water, nitrogen, organic matter, pests, and phytosanitary products. Among these models, approximately 20 of them drive tools that are widespread and widely used in the agricultural world.

This document offers a non-exhaustive review of a number of models developed with differing aims: knowledge models (STICS, Nitroscape), exploration models (DEXiPM), learning models (Rami Fourrager), and companion models (ComMod).

Modelling in major research programmes on innovative agricultural systems

Modelling plays a key role in major research programmes currently being established to develop new crop systems. For certain programmes modelling is necessary even before work can begin.

Accordingly, the first step of the European Union’s FACCE-JPI* initiative is to compare models in order to evaluate the risks posed to European agriculture by climate change, then to apply the models to the regional level, and lastly to create integrated models.

*FACCE-JPI (Joint Programming Initiative for Agriculture, Climate Change, and Food Security) was launched in 2010 and brings together 21 European countries in an effort to align European food security and sustainable agriculture interests in the light of climate change.

At the international level, the AGMIP  (Agricultural Model Intercomparison and Improvement Project) programme compares no less than 27 wheat models, 26 maize models, and 13 rice models to estimate global production variability based on yearly climate data.