Modelling of rare flood meadow species distribution by a combined habitat surface water-groundwater model

Floodplains are highly complex and dynamic systems in terms of their hydrology. Thus, they comprise a wide habitat heterogeneity and therefore harbour highly specialized species. For future projections of habitat and species diversity, process-based models simulating ecohydrological conditions and resulting habitat and species distributions are needed. We present a new modelling framework that includes a physically based, surface water-groundwater model coupled with a habitat model. Using the model framework, we simulate the occurrence of 23 flood meadow plant species in a Rhine River floodplain. To benchmark the data, results are compared with a conventional approach with simple spatial hydrological information. Our results show that models with predictors obtained from the surface water-groundwater model are significantly more accurate for rare and endangered species, as well as for typical flood meadow species. A total of 15 hydrological predictors were defined, which look relevant and promising for a good prediction, but at the same time reflect very different hydrological conditions. The standard deviation of the groundwater level, wet soil conditions, and inundation belong to the most relevant predictors for an accurate prediction. Therefore, we recommend including more specific hydrological information in habitat models of species in complex floodplain ecosystems. Such spatial explicit habitat models can also open up further possibilities, such as the projection of global change impact studies or nature conservation planning.