Projecting the dynamics of invading deer with pattern-oriented modelling to support management decision-making

Avoiding the undesirable impacts of invasive species requires robustly evaluating the effects of alternative management scenarios. Such evaluations depend on reliable spatio-temporal projections of changes in the distribution and abundance of the invasive species under different scenarios. However, commonly used modelling approaches are constrained for this purpose because they do not incorporate demographic processes and are seldom validated. We develop a spatially and temporally explicit grid-based population model for invasive fallow deer Dama dama in Tasmania, Australia, validated against observed distribution and abundance ‘targets’ using a pattern-oriented approach. We use this validated model to project the future invasion patterns of deer, including encroachment into areas of high conservation value, under eight harvest and habitat suitability scenarios. Projected population sizes of deer differ greatly depending on harvest and habitat suitability scenarios. Without harvest, the population grows sixfold, from about 80,000 deer in 2019 to over 500,000 by 2100. In contrast, the population stabilises at 32,000–41,000 deer with 25% annual harvest. Abundance in the environmentally sensitive Tasmanian Wilderness World Heritage Area (TWWHA), a region of high conservation value, increases in all scenarios. We identify likely areas in the expanding range of deer where targeted removal should be most effective at minimising deer numbers in the TWWHA. Synthesis and applications: Using a pattern-oriented model validated with spatio-temporal data, we show how the future distribution and abundance of invasive fallow deer in Tasmania can be substantially reduced by targeted increases in harvest and prioritising areas for removal around high conservation value regions. Our approach can be used to project likely effects of management interventions on future distributions and abundances for a range of invasive taxa.