Aim

We compare the performance of habitat suitability models using climate data only or climate data together with water chemistry, land cover and predation pressure data to model the distribution of European grayling (Thymallus thymallus). From these models, we (1) investigate the relationship between habitat suitability and genetic diversity; (2) project the distribution of grayling under future climate change and (3) model the effects of habitat mitigation on future distributions.  

Location

United Kingdom

Methods

Maxent species distribution modelling was implemented using a Simple model (only climate parameters) or a Full model (climate, water chemistry, land-use and predation pressure parameters).  Areas of high and low habitat suitability were designated. Associations between habitat suitability and genetic diversity for both neutral and adaptive markers were examined. Distribution under minimal and maximal future climate change scenarios was modelled for 2050, incorporating projections of future flow scenarios obtained from the Centre for Ecology and Hydrology. To examine potential mitigation effects within habitats, models were run with manipulation of orthophosphate, nitrite and copper concentrations.

Results

We mapped suitable habitat for grayling in the present and the future. The full model achieved substantially higher discriminative power than the Simple model. For low suitability habitat, higher levels of inbreeding were observed for adaptive, but not neutral loci.  Future projections predict a significant contraction of highly suitable areas. Under habitat mitigation, modelling suggests that recovery of suitable habitat of up to 10% is possible.

Main conclusions

Extending the climate-only model improves estimates of habitat suitability. Significantly higher inbreeding coefficients were found at immune genes, but not neutral markers in low suitability habitat indicating a possible impact of environmental stress on evolutionary potential. The potential for habitat mitigation to alleviate distributional changes under future climate change is demonstrated and specific recommendations are made for habitat recovery on a regional basis.

Species occurence records were accessed from GBIF in 2014 considering all records since 1990, which had environmental data for the rivers in the study area. Records were spatially rarefied at a scale of 30 km² across the latitudinal range and at a scale of 1 km² for the final subset study area. Furhter, a R-script is provided used for the fine tuning of modelling parameters.