Using modeling to improve models

We evaluated the error of estimate brought while: (1) undertaking some classical hypotheses in modeling plant growth in a dynamic way, (2) using imprecise measurements of physiological characteristics of the plant or in the environmental data, and (3) not taking into account plant acclimation ability. We synthesised classical models based on a carbon mass-balance approach and run for populations of Ranunculus peltatus, a spreading macrophyte in rivers of Northeastern France. Simulations were performed for five contrasting combinations of environmental parameters. Among the 54 models tested, we demonstrate highly variable results in terms of maximum biomass reached and plant temporal biomass production. The way plant architecture was approximated contributed significantly to biomass results. From the simulations, we selected a family of models using simple decision rules. Using these models, we underlined that: (1) light availability and to a lesser extent, temperature and current velocity were key main environmental factors for plant growth and should be measured with accuracy; (2) the optimum temperature for photosynthesis and the maximum activities for all physiological processes were the most sensitive constants entering the model; (3) taking into account plant plasticity (i.e. their capabilities to modify their physiological and morphological characteristics to adapt to lack of resources or seasonal environmental changes) greatly modifies biomass production, especially when adapting to nutrient stress or to seasonal temperature variation. All these results may contribute significantly to the improvement of existing dynamic models and especially of the validation process.