s:2540:"TI Building the foundations of a Coffea arabica FSPM AU Dauzat J AU Griffon S AU Roupsard, O. AU Rodrigues G C AU Vaast, P. AB Coffee is the most valuable product worldwide after oil. A quite large literature is dedicated to the agronomy and physiology of coffee tree (DaMatta et al. 2007) but there have been few attempts to model plant growth and production on functional bases. The most integrative process-based model was recently proposed by Rodriguez et al. (2011). This model combines phenology, shoot and root dynamics, light interception, photosynthesis and carbon allocation. However it works with cohorts of branches and leaves and doesn't explicitly consider the plant geometry and topology. Our purpose is thus to lay the foundations of coffee FSPM using information obtained in different experiments. FSPMs mostly lie on four cornerstones: plant topology, plant geometry, carbon acquisition and carbon allocation. Putting together these different components leads to complex models with numerous parameters and, as far as many parameters need to be calibrated, there is then no insurance that proper parameter values can be obtained through automatic optimization methods. A step by step procedure is thus necessary for calibrating and validating the components of the integrated model. Plant topology is the first step to address when aiming to develop a functional growth model with an explicit 3D structure. Basically the plant structure is the result of bud activity (growth, dormancy, death and ramification processes) over time. Unfortunately the functional bases of organogenesis are complex and poorly understood. A way to overcome this problem is to force the plant structure against field observations. This procedure is not straightforward since the plant structure cannot be reasonably measured all along its development and in coffee trees and there is no growth marker that can be used for inferring the age of the elements of a plant structure (Taugourdeau et al. 2012). Dating the elements of a coffee tree therefore needs a precise knowledge of the growth rates of plant sub-structures. Fortunately the growth rates of the main stem and branches are closely related in coffee trees (deReffye, 1990), making possible to retrieve back in time the plant structure in former stages. This procedure is illustrated here for arabica coffee plants at six growth stages from DAP 156 to 797. The experience acquired on modeling the carbon acquisition and the 3D plant structure is also briefly presented. ";