COS 66-1
Transferring species distribution models between native and non-native ranges: Implications for the importance of biotic interactions

Wednesday, August 13, 2014: 8:00 AM
Carmel AB, Hyatt Regency Hotel
Jennifer E. Weaver, Environmental Science, Policy and Management, UC Berkeley, Berkeley, CA
George K. Roderick, Environmental Science, Policy and Management, UC Berkeley, Berkeley, CA
Background/Question/Methods

The relationship between a species’ distribution, climate and land use may differ between its native and non-native ranges. The extent to which it differs can give us some insights into the importance of variables that are not included in the model such as biotic interactions. The objective of this study is to test the transferability of species distribution models between the native and non-native range of avian species. Our study examines twenty species of birds that have their native ranges in Europe and their introduced ranges in North America. Species distribution data was obtained from the EBCC Atlas of European breeding birds, the United States Breeding Bird Survey route data and GBIF. The relationships between species’ distributions and landscape variables were determined using an ensemble of correlative models, including GLMs, GAMs and ANNs. Habitat suitability maps were created for both the validation and reciprocal ranges and transferability success was assessed using AUC and maximum Kappa values. 

Results/Conclusions

Preliminary results indicate that of the twenty species, some of the species’ models exhibited transferability, while others exhibited asymmetrical transferability, i.e. models built in one range were not able to accurately predict species distributions in the reciprocal range, or vice versa. While causation cannot be inferred from these results, asymmetrical transferability may be a function of time since introduction, or for those species that have had the opportunity to disperse to all viable habitats, asymmetrical transferability may indicate the effect of differing biotic relationships in the species’ native and non-native ranges. These results further add to the discussion of the importance of biotic interactions at the broader scale, and whether biotic interactions should be considered for inclusion in species distribution models to attain accurate habitat suitability maps and future predictions.