• Reduce text

    Reduce text
  • Restore text size

    Restore text size
  • Increase the text

    Increase the text
  • Print

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

Modelling and agrosystems

The challenge of spatialisation: GENESYS and MAPOD

GENESYS and MAPOD study the effects of crop systems on gene flows at landscape level.

By Pascale Mollier, translated by Daniel McKinnon
Updated on 09/10/2013
Published on 05/30/2013

Grains of maize pollen.. © INRA, FORMISANO Sophie
Grains of maize pollen. © INRA, FORMISANO Sophie

GENESYS and MAPOD are gene flow models. Since their creation in 1999 and in 2001 respectively their performance has constantly improved. They are, at present, the only models in Europe studying the effects of crop systems on gene flows at landscape level.

Managing coexistence among field varieties

GENESYS (rapeseed, beet) and MAPOD (maize) simulate flow rates between plots for seeds or for pollen based on various crop parameters, such as plot size, borders, crop rotations, flowering dates, and crop management (soil tilling, seeding date, variety selection, etc.). The models allow crops to be organised in time and in space to avoid contamination among varieties.

Each species has unique issues:

The major problem for rapeseed is managing volunteer plants in fields using agricultural techniques. The aim is for coexistence over time and across landscapes of both industrial-use varieties and food varieties (lower in erucic acid or, more recently, lower in linolenic acid, which is responsible for nuisance odours when used for frying).

In the case of beets, the issue is controlling, through the use of agricultural techniques, the spread of weed beet, a pest plant related to the sugar beet.

The MAPOD model for maize was developed to manage the coexistence of GMO and non-GMO varieties, but it can also be used when growing different varieties of one plant, as in the case of waxy maize for example.

Particle dispersion: an issue for both models

Gene transfer mechanisms also vary among the plant species. For rapeseed and beet, gene flows happen both in space, with seeds and pollen, and in time, with volunteer plants from previous crops. For corn, gene flows are largely limited to the transfer of pollen in space.

Whether gene flows happen through seeds or through pollen, the modelling principals remain the same. What makes these models unique is the fact that they simulate flows at agricultural landscape level. They take account of a block of plots and of diverse crop practices applied thereto, rather than a field-by-field approach seen in many other models. The models are therefore able to make a number of technical recommendations, on isolation distances, buffer zones, and staggered flowering times for example, that are adapted toparticular crop conditions, such as crop density at landscape level, and plot size. These models have been used in a number of studies to assist in public decision-making processes with regard to the coexistence of GMO and non-GMO crops. They have also been used in research projects, such as the EU SIGMEA project, studying product segregation from field to silo.

 Creating an effective algorithm for spatial dispersal

GENESYS and MAPOD were implemented using computer coding. This allows for simulations in diverse conditions at level of a block of agricultural plots. A new algorithm was created to deal with spatial dispersion. The major change involves calculating flows between pairs of polygon shapes rather than between source and target points. This is more representative of actual shapes and of distances between transmitting and receiving areas, such as plots or subplots. This method is unique and draws on computational geometry and integration techniques for polygons (for cubature). It has been made available through a free software programme known as CaliFloPP. CaliFloPP is expected to be integrated into the RECORD modelling platform developed in Toulouse, France (see Section 8 of this report).

The GENESYS model now uses CaliFloPPfor spatial dispersion modelling. CaliFloPP is an adaptable tool that can be used not only for pollen and seed flows but also in plant epidemiology to model spore pathogens (1).

(1) Papaïx, J., David, O., Lannou, C. and Monod, H., 2013. "Dynamics of Adaptation in Spatially Heterogeneous Metapopulations”,PLoS ONE, Public Library of Science, 8, e54697.

Improving GENESYS

The GENESYS model now uses CaliFloPP to model spatial dispersal. According to Nathalie Colbach, “GENESYS can correctly forecast volunteer plant density, but it underestimates gene dispersal beyond 50 metres because some factors cannot yet be modelled, such a gene dispersal by bees. We must add a 30% correction factor.” The team continues to make improvements to the model. It can now be used to forecast linolenic acid levels, and the ability to forecast imidazoline resistance was recently added as well.

Improving MAPOD

Applied mathematics come into play when dealing with a number of other gene dispersal issues. For example, what is the effect to pollen flows of landscape heterogeneity such as hedges and roads? Work of this type must be carried out in close collaboration among agronomists, physicists, and statisticians, as was the case for the ANR GCOM2AP project led by Frédérique Angevin from 2007 to 2011.
As a part of the European Union’s PRICE project, other modelling work, using Baysian methods, has been carried out to develop a decision-making tool allowing the potential for corn varieties to coexist to be quickly tested before sowing.

Scientific contact(s):

Associated Division(s):
Applied Mathematics and Informatics, Environment and Agronomy, Plant Biology and Breeding
Associated Centre(s):
Jouy-en-Josas, Bourgogne-Franche-Comté, Versailles-Grignon


Bouvier A. et al. 2009. Computation of integrated flow of particles between polygons. Environmental Modelling & Software, 24:843-849

Colbach N. 2009. How to model and simulate the effects of cropping systems on population dynamics and gene flow at the landscape level. Example of oilseed rape volunteers and their role for co-existence of GM and non-GM crops. Environmental Sciences & Pollution Research 16, 348–360

Angevin, F. et al. 2008. Modelling impacts of cropping systems and climate on maize cross-pollination in agricultural landscapes: The MAPOD model. European Journal of Agronomy, 28, 471-484. DOI: 10.1016/j.eja.2007.11.010

Managing weeds on a large scale

In Nathalie Colbach’s team, researchers are working on another model, named FLORSYS, that is able to simulate the weed flora dynamics of different cultivation methods. The core element of the FLORSYS model is not seed dispersal, although this is taken into account, but rather soil seed stocks. FLORSYS models weed life cycles and calculates weed density and biomass. “We are able to simulate the crop yield losses weeds will cause, but also to calculate the positive, biodiversity aspects of weeds as food sources for bees, with their flowers, and for birds, with their seeds on the ground in the winter” says Colbach. “Our work at the moment is at the level of a block of plots, although we are not yet able to predict dispersal caused by agricultural equipment.”