Estimating avian wingspan from wing length: an order-level regression approach
Fox, M. R.; Brunton, E. A.; Shephard, J. M.
Show abstract
Wingspan is a key parameter in avian collision risk models but is rarely measured systematically, meaning for many species, collision risk remains largely unquantified due to a lack of morphological and flight data required to parameterise models. Folded wing length scales predictably with wingspan and is available for all extant bird species, yet no widely available method has been published for predicting wingspan from wing length; here we compiled a global wingspan dataset for 1,442 species across 25 orders and fitted order-level linear regression models predicting wingspan from wing length. Order-level models performed well, with 85% achieving R2 [≥] 0.95 (global model R2 0.96); cross-validation identified that model predictions generally fell within the intraspecific variation in wingspan, resulting in wingspan estimates for 9,194 species. Our models and accompanying wingspan data provide a transparent, reproducible, and validated method for generating species-level wingspan estimates from widely available wing length data, with direct applications to collision risk modelling, flight performance research, and macroecological analyses of avian morphology.
Matching journals
The top 14 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
- Natal philopatry, dispersal and age of first breeding in relation to size and sex of Arctic Terns Sterna paradisaea 92%
- Habitat partitioning and spatial segregation at multiple scales promotes year-round coexistence in a guild of forest songbirds 92%
- Re-assessing the phylogenetic status and evolutionary relationship of Forest Owlet using genomic data 91%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.