We investigated the use of output from Bayesian stable isotope mixing models asconstraints for a linear inverse food web model of a temperate intertidal seagrass systemin the Marennes-Oléron Bay, France. Linear inverse modeling (LIM) is a technique thatestimates a complete network of flows in an under-determined system using acombination of site-specific data and relevant literature data. This estimation of completeflow networks of food webs in marine ecosystems is becoming more recognized for itsutility in understanding ecosystem functioning. However, diets and consumption rates oforganisms are often difficult or impossible to accurately and reliably measure in the field,resulting in a large amount of uncertainty in the magnitude of consumption flows andresource partitioning in ecosystems. In order to address this issue, this study utilizedstable isotope data to help aid in estimating these unknown flows. δ13C and δ15N isotopedata of consumers and producers in the Marennes-Oléron seagrass system was used inBayesian mixing models. The output of these mixing models was then translated asinequality constraints (minimum and maximum of relative diet contributions) into aninverse analysis model of the seagrass ecosystem. We hypothesized that incorporatingthe diet information gained from the stable isotope mixing models would result in a moreconstrained food web model. In order to test this, two inverse food web models werebuilt to track the flow of carbon through the seagrass food web on an annual basis, withunits of mg C m-2 d-1. The first model (Traditional LIM) included all available data, withthe exception of the diet constraints formed from the stable isotope mixing models. Thesecond model (Isotope LIM) was identical to the Traditional LIM, but included the SIARdiet constraints. Both models were identical in structure, and intended to model the sameMarennes-Oléron intertidal seagrass bed. Each model consisted of 27 compartments (24living, 3 detrital) and 175 flows. Comparisons between the outputs of the models showedthe addition of the SIAR-derived isotopic diet constraints further constrained the solutionrange of all food web flows on average by 26%. Flows that were directly affected by anisotopic diet constraint were 45% further constrained on average. These resultsconfirmed our hypothesis that incorporation of the isotope information would result in amore constrained food web model, and demonstrated the benefit of utilizing multi-tracerstable isotope information in ecosystem models.