HyDiaD model: A tool to explore the effect of climate change on diadromous species assemblage in Western Europe

International audience

Английский. Diadromous fishes and lampreys are migratory species that share their life cycle between marine and freshwater environments. For some species, climate change is a new threat that may weaken populations already in an alarming situation. For some other species, climate change may provide an opportunity to settle in new catchments.During the European-funded DiadES project (Interreg Atlantic Area), we developed a hybrid species distribution model, named HyDiaD 1.2 for Hybrid model for Diadromous species Distribution. This approach, halfway between correlative and mechanistic modelling, integrates two components: i) the correlation of species observations with physical, and climactic predictors at the catchment scale, ii) the simulation of biotic processes relevant for the distribution of the species, such as population demography and dispersal dynamics. This single modelling framework was applied for eleven diadromous species present in the European Atlantic coast.The correlative component provides information on shifts in habitat suitability related to climate scenarios. It was calibrated with boosted regression trees applied to the presence / absence data from 290 European catchments for the 1850-1950 period (EuroDiad 4.0 database). The model was used to simulate population trends in a subset of catchment representative of the Atlantic Area for two time horizons (2050 and 2100) according to two climatic scenarios (RCP4.5 and RCP8.5).Population dynamics are based on hockey-stick stock-recruitment relationship with a depensatory effect and the maximum production of recruitment limited by catchment suitability. No direct anthropogenic impacts (habitat degradation, mortality) were included. The dispersal dynamics integrated only accessibility between catchments (kernel function of inter-catchment distance). The calibration of such biotic processes is still challenging, especially for data-poor species. To overcome this difficulty, the parameters set were calibrated based on elicitation of expert knowledge.The current outputs of the single-species models were summarised into maps of: (i) species turnover in the catchments, and (ii) species sensitivity ranking (cumulated changes of abundance in catchments). This result interpretation at the species assemblage level (all species pooled/considered together) allowed the definition of species or catchments to prioritize. The relative reliability of the HyDiaD applications between species was discussed, according to modellers’ confidence in the abiotic and biotic components, and other modelling choices. New developments for HyDiaD were proposed in response.