Moving from pattern towards process for improved forecasts of species range dynamics
This exchange project (2019-2020) is supported by the DAAD (Grant to Damaris Zurell) and the University of Melbourne (Grant to Gurutzeta Guillera-Arroita)
Summary
Forecasts of where species are likely to occur in the future are important for informed species management. Predictions help in conserving populations that may otherwise be lost, or detecting invasive or harmful species that might otherwise be overlooked. However, the correlative species distribution models (SDMs) most commonly used to generate these forecasts have well-known limitations that may make them unreliable. Using both real data (Swiss breeding bird survey records for >170 species) and simulated datasets, we will test whether alternative models that explicitly describe dynamics and incorporate key ecological processes can improve forecasts of species ranges. We will first focus on occupancy dynamic models (ODMs), which show much promise but have not been tested against other approaches. We will then extend our analyses to consider more complex models that require additional ecological knowledge (e.g. information about dispersal, biotic interactions and/or demography). Outputs from correlative SDMs are currently used to inform a range of decisions, including whether to list a species as globally threatened. We will assess outputs from alternative models, asking whether any are better able to correctly guide these decisions. We will provide practical guidelines for methods that prove suitable, making them accessible for typical users.
Collaborators
- Gurutzeta Guillera-Arroita, Jane Elith, Natalie Briscoe, Univ. Melbourne, Australia