https://doi.org/10.3897/oneeco.3.e28143 (13 Sep 2018)
- ContentsContents
- Article InfoArticle Info
- CiteCite
- MetricsMetrics
- CommentComment
- RelatedRelated
- FigsFigs
- TabsTabs
- TaxaTaxa
- DataData
- RefsRefs
- CitedCited
Pages 2 to 48: two maps on each page, showing input species occupancies and predicted probability of occurrence, for 47 species of bumblebees.
Occupancy maps are based on records available from 1991 to 2012 inclusive, identified to the level of species and with a spatial accuracy of at least 10 km.
The predicted probability of occurrence must be interpreted as the potential environmental suitability based on the set of environmental predictors and the input occupancy. Additional records and updated environmental information can lead to different predictions.
Page 49: two maps showing the overall probability of occurrence for those species predicted to be present in each grid cell and predicted species richness derived therefrom.
The presence of a species is derived from the predicted probability of occurrence, after applying a threshold to discriminate presence from absence. No common rules exist to choose the threshold. A combination of statistical and ecological criteria might be used. For this study, we adopted 'Minimum training presence' as a threshold rule, which uses the suitability associated with the least suitable training presence record as the threshold. While this rule might lead to optimistic predictions (also assigning presence to areas at the margin of the species' ecological requirements), it also shows the potential suitability of the environment for the species. This information can be used, for instance, to select target areas for specific pollinator-friendly measures aiming at improving local environmental suitability.
See main text for additional details and for the data sources.