Section: Partnerships and Cooperations

National Initiatives

SACADEAU-APPEAU : Decision-aid to improve streamwater quality

Participants : Marie-Odile Cordier, Véronique Masson, Christine Largouët.

The Sacadeau project (Système d'Acquisition de Connaissances pour l'Aide à la Décision pour la qualité de l'EAU - Knowledge Acquisition System for Decision-Aid to Improve Streamwater Quality) was funded by Inra (French institute for agronomy research) from October 2002 to October 2005. The main partners were from Inra (Sas from Rennes and Bia from Toulouse) and from Irisa . We have continued to develop the Sacadeau model with our partners until now and a PdD thesis has been funded by Inra (ASC).

We were then involved in a project, named Appeau and funded by anr/add , which started in February 2007 and ended in December 2010. The Appeau project aimed at studying which politics, for which agronomic systems, are best adapted to improve water management. It included our previous partners as well as new ones, mainly from Inra . A paper has been written in 2011 in cooperation with all the partners and submitted to Environmental Modelling and Software, currently in revision.

Our work aims at building a decision-aid tool to help specialists in charge of the catchment management to preserve the streamwater quality. The Sacadeau simulation model couples two qualitative models, a transfer model describing the pesticide transfer through the catchment and a management model describing the farmer decisions. The simulation results are analyzed, thanks to classification and symbolic learning techniques, in order to discover rules explaining the pesticide concentration in the stream by the climate, the farmer strategy, the catchment topology, etc., and, finally, in order to build recommendation actions for a given situation. In the Appeau context, the idea is to study how this kind of model can be used to simulate scenarios in a more generic way and to compare, and possibly unify, our work with what is done by our partners from SAS concerning nitrate transfer. (http://wwwagir.toulouse.inra.fr/agir )

ACASSYA : Supporting the agro ecological evolution of breeding systems in coastal watersheds

Participants : Marie-Odile Cordier, Véronique Masson, René Quiniou, Christine Largouët.

The Acassya project (ACcompagner l'évolution Agro-écologique deS SYstèmes d'élevage dans les bassins versants côtiers) is funded by ANR/ADD and started at the beginning of 2009. The main partners are our colleagues from Inra (Sas from Rennes. One of the objectives is to develop modeling tools supporting the management of ecosystems, and more precisely the agro ecological evolution of breeding systems in coastal watersheds. In this context, the challenge is to transform existing simulation tools (as Sacadeau or TNT2 into decision-aid tools, able to answer queries or scenarios about the future evolution of ecosystems. (http://www.rennes.inra.fr/umrsas/programmes/acassya_accompagner_l_evolution_agro_ecologique_des_systemes_d_elevage )

PayOTe-II : characterizing agricultural landscapes via data mining

Participants : Thomas Guyet, Christine Largouët, René Quiniou.

The PayOTe-II project (Paysage Ou Territoire) is funded by AIP INRA/INRIA and started at the end of 2010. The project associates INRIA Teams (Orpailleur and Dream) with INRA Team (UBIA, MIAJ and SAD-Paysage).

One of the objectives of the PayOTe project is to provide tools to generate “realistic” agricultural landscapes. This kind of generator is expected by expert to study the impact of the landscape on agro-ecological systems. The main approach of this project is to use data mining to automatically construct a neutral model of a landscape. Then, the model of a landscape may be used to generate new landscapes with same spatial properties.

In this context, the challenge is to develop spatio-temporal data mining algorithms to analyse the spatial organization of agricultural landscapes.

PAYTAL : Mining spatial correlations between urban sprawl and landscape

Participant : Thomas Guyet.

The PayTal project (Paysage et Etalement Urbain - Landscape and urban sprawl) is funded for 3 years by the french Ministry of Ecology and Sustainable Developement. This project started in september 2011. It involves our colleagues from INRA/SAS and Agrocampus-Ouest.

This project proposes a multidisciplinary approach, firstly, to describe the fine forms of urban sprawl and the dynamics of the landscape and, secondly, to study the links between urban sprawl and landscape evolution. The first aim of the project is to develop a methodology to acquire a spatial description of both the landscape and the urban extent. This spatial information will be acquired by our colleagues from remote sensing images and available official documents (local development plans, landscape repository, etc.) related to several “conurbations” in western France (Rennes, Angers, Lorient, Brest).

Our work aims at using and proposing spatio-temporal data mining tools to extract landscape patterns, i.e. a set of landscape elements linked through spatial relationships. Using symbolic learning techniques, we expect to extract landscape spatial patterns that may explain the urban sprawl (evolution barrier or facilitator).