RECORD: an open platform to build, evaluate and simulate integrated models of farming and agroecosystems

Résumé En

The rapid change in the agricultural industry requires the development of new methods of production to guarantee sustainable agriculture. In silico approaches offer the possibility to identify more quickly new systems to tackle current social, political and environmental concerns. Numerous agro-ecosystem functioning models already exist. Nowadays the main issue is more to couple and use them at different spatial and temporal scales rather than developing new ones. INRA set up an integrated modelling framework to gather, link and provide models and companion tools to answer new society questions regarding agriculture. Functional specifications were drawn during a two-day meeting in January 2007. Requirements are: to develop new models as modular components, to re-use and combine them to represent cropping and farming systems at different time and spatial integration steps; to allow the modelling and simulation of farmer's decision making process; to link with statistical packages to perform model calibration or in silico experiment analyses; with economics and optimization software and risk analysis; with databases and GIS; to include random weather generators to work on climate change and uncertainty. After different tests, the RECORD platform was built under VLE, a generic modeling and simulation environment based on the discrete event formalism DEVS It allows the design of atomic and coupled models. It proposes standard mathematical formalisms (e.g. difference equations, differential equations, state charts) and a decision formalism to model agroecosystems. It makes possible the use of different time steps and spatial scales within a model. A graphical user-interface was designed to simplify coding tasks and a dynamic link to the R software has also been realized in order to enable statistical work. A variety of research projects already use RECORD. Examples are given showing the ability to recode simple models, encapsulate more complex ones, link with GIS and databases, and use the R statistical package to run models and analyse simulation outputs.