Life-history phenology strongly influences population vulnerability to toxicants: a case study with the mudsnail Potamopyrgus antipodarum.

[Departement_IRSTEA]Eaux [TR1_IRSTEA]BELCA

International audience

English. One of the main objectives of ecological risk assessment is to evaluate the effects of toxicants on ecologically relevant biological systems such as populations or communities. However, the effects of toxicants are commonly measured on selected subindividual or individual endpoints due to their specificity against chemical stressors. Introducing these effects into population models is a promising way to predict impacts on populations. The models currently employed are very simplistic, and their environmental relevance needs to be improved to establish the ecological relevance of hazard assessment. The present study with the gastropod Potamopyrgus antipodarum combines a field experimental approach with a modeling framework. It clarifies the role played by seasonal variability of life-history traits in the population's vulnerability to the alteration of individual performance, potentially due to toxic stress. The present study comprised 3 steps: 1) characterization of the seasonal variability in life-history traits of a local population over 1 yr by using in situ experiments on caged snails, coupled with a demographic follow-up; 2) development of a periodic matrix population model that visualizes the monthly variability of population dynamics; and 3) simulation of the demographic consequences of an alteration in life-history traits (i.e., fertility, juvenile, and adult survival). The results revealed that demographic impacts strongly depend on the season when alterations of individual performance occur. Model analysis showed that this seasonal variability in population vulnerability is strongly related to the phenology of the population. The authors emphasize that improving the realism of population models is a major objective for ecological risk assessment, and that taking into account species phenology in modeling approaches should be a priority.