Help wanted: inconsistent SDM methods seeking new management

Link To Share This Poster: https://cdmcd.co/zY94a6
Live Discussion Link: https://cdmcd.co/4dqGXA

Hannah R. Bevan and David G. Jenkins, Department of Biology, University of Central Florida, Orlando, FL

Background/Question/Methods
Species distribution models (SDMs) are widely used to help inform management of both native and non-native species. Existing SDM frameworks describe the standards of a well-performing, reproducible distribution model (e.g., Araujo et al. 2019; Feng et al. 2019; Zurell et al. 2020), but consistent methodological details to achieve these standards are lacking. Here, we reviewed 150, randomly selected, empirical SDMs for the years 2017-2019 (50/year) to better understand the recent state of the science as a basis for recommendations of SDM methods. We collected data on 32 categories related to SDM source information, response variables, predictor variables, model building, and model evaluation. Categories reflect existing SDM framework standards.
Results/Conclusions
SDMs are biased toward terrestrial, native species but represent all seven continents with a strong conservation focus. Overall, SDMs provided wide-ranging results, especially for model building: 22 different algorithms used in single and multi-algorithmic groupings, many without methods explanation or justification. Other inconsistencies included the number of: presence points (9 - 1 million); spatial autocorrelation mitigations (< 50%); predictors (1 - 44); model resolutions (<1km2 - 50km2); and performance metrics (15), again will little explanation or justification. Model quality and reproducibility of SDMs in general cannot follow existing SDM frameworks with so much procedural variation. SDM methods must become more transparent and consistent for reliable conservation use, such as in habitat evaluations, restorations, and prioritizations, species translocations, global change predictions, and biological invasion mitigation. We provide detailed recommendations toward that goal.