ABSTRACT Aim Protecting phylogenetic diversity (PD) has been advocated as a basis for spatial conservation planning. However, phylogenetic branch lengths can vary substantially depending on methodological choices made during the inference of phylogeny and divergence times, and little is known about the sensitivity of spatial patterns of PD to these choices. We aimed to test the variability of spatial PD patterns, including the size and locations of ‘hotspots’ of evolutionary history, under alternative phylogenetic methods. We then aimed to test the downstream effects for conservation priority areas selected using Systematic Conservation Planning methods. Location Australia. Methods Using phylogenomic and fossil data for the plant clade Grevilleoideae (663 species), we inferred divergence times under 26 combinations of methods relating to topology and divergence time estimation, choice of fossil calibration, calibration constraints, and evolutionary model parameterization. We compared spatial patterns of PD and conservation priority areas selected using a reserve selection algorithm across the resulting 26 phylogeny treatments. Results Continent‐wide, there was substantial incongruence in spatial PD patterns among treatments, with large areas identified as PD hotspots or priority areas under fewer than 26 treatments. Areas of incongruence in priority areas amounted to approximately 30% of the land budget allocated to expansion of the protected area network in our simulations (~8% of Australia's land area). Spatial patterns were especially sensitive to the choice of Bayesian (MCMCtree) vs. fast rate‐smoothing (RelTime) methods of divergence dating. Main Conclusions Spatial patterns of PD and choice of priority areas for conservation of PD can vary substantially under alternative, routine, methodological choices in phylogenetics. This introduces uncertainty into conservation planning that aims to maximize protection of evolutionary history. Uncritical use of phylogenetic information in conservation may lead to wasted resources or poor conservation outcomes.
Kearns et al. (Sun,) studied this question.