Dealing with non-equilibrium bias and survey effort in presence-only invasive Species Distribution Models (iSDM); predicting the range of muntjac deer in Britain and Ireland

Invasive species managers utilise records to inform management. These data can also be used in Species Distribution Models (SDM) predict future spread or potential invasion of new areas. However, issues with non-equilibrium (also called disequilibrium) cause difficulties modelling invasive that have not fully colonised their distribution and, addition, sampling bias result from a lack information on survey effort, particular issue for presence only techniques. Geographical confounds are unavoidable when building iSDMs but there methods allow prediction optimised. We maximum entropy (Maxent) model suitable habitat Reeve's muntjac deer (Muntiacus reevesi) throughout Great Britain and Ireland comparing several aimed address Modelling (iSDM) including spatial filtering, weighted background points targeted built at varying extents. Model evaluation metrics suggested the model, which explicitly failed account full extent using random points, predicted species' current range best. This highlighted negative environmental relationships likely represent uncolonised areas rather than selection thus, low suitability was misleading. Of models dealt conceptually best, by restricting training core range, utilised accounting effort (cells other recorded as present yet no muntjac) best metric, yielded relatively poor predictive performance. implied limited selectivity avoidance within which, spatially extrapolated, virtually all regions may vulnerable invasion.