Abstract

The reintroduction of locally extinct species is one method of conservation action, but successful reintroductions remain low because conservation managers do not adequately assess for suitability of habitat. In particular, much of the traditional site selection approaches fail to address the relationship between environmental variables, the species' ecological niche, and climate changes. These limitations necessitate a more accurate, predictive assessment of reintroduction suitability, in order to both ensure the recovery of the species population, and the liability of the reintroduced area over the long-term. This paper demonstrates one computationally-appropriate method of Ecological Niche Modelling (ENM) for species reintroduction planning by employing high-resolution bioclimatic, topographic, and land-use datasets. We merge historical species occurrence records from existing archives, with present-day environmental data, and employed machine learning data-mining approaches (MaxEnt, Random Forest) to model potential habitat within the study area. The modelling pipeline incorporates climate change data (RCP 4.5 and RCP 8.5 scenarios) to predict possible habitat changes in the future. Model outputs were compared to baseline suitability assessments constructed through traditional expert-based mapping, and comparisons evaluated predictive ability, spatial overlap, and sensitivity to environmental factors in each assessment, and therefore, we can interrogate the value of data-based modelling vs traditional relational or qualitative approaches. The ENM approach produced priority reintroduction landscapes with >85% predicted accuracy, and identified habitat patches previously unknown that showed resilience to climatic shifting. These results build on evidence-based planning for reintroduction efforts, in order to enhance conservation managers' allocation of limited resources and ensure population density longevity post-reintroduction. Ultimately, the results support the notion that natural science–based engineering practices can transform species recovery strategies under dynamic environmental conditions.