Thesis: Carbon-14 (14C) is, with tritium (3H), one of the radionuclides which are released in greatest amounts into rivers by nuclear power plants in routine operation. In France, these two radionuclides contribute to a major fraction of the estimated annual dose, absorbed by localpopulations (around 1 µSv/year), by ingestion of contaminated fish and water. Current transfer models used for 14C in aquatic ecosystems are unable to explain the observed variability in 14C concentration in sampled fish because these models are very simplistic and do not take account of temporal changes in 14C releases and of various key metabolic processes (assimilation, respiration, incorporation, etc.). In this context, the present PhD project aims to produce physiologically and ecologically robust predictions, using the conceptual approach known as Dynamic Energetic Budget (DEB). This individual-based model describes how an organism acquires and uses the energy from food for its metabolic functions (growth, reproduction…), depending on food availability and temperature. To address this goal, a DEB model for the common carp, Cyprinus carpio, was reparameterized to improve growth and reproduction predictions. This model was applied to the case of 14C transfer, through experimental data acquisition. The aim of these experiments was to measure the kinetics of ingestion, defecation, assimilation, and incorporation of 14C incommon carp exposed to 14C labeled metabolites, via direct and trophic routes. Once the Using the model applied to the case of 14C, scenarios of contamination in 14C under representative conditions of a French river (La Vienne) are simulated in the common carp.