Precipitation and winter temperature predict long‐term range‐scale abundance changes in Western North American birds

摘要: Predicting biodiversity responses to climate change remains a difficult challenge, especially in climatically complex regions where precipitation is limiting factor. Though statistical climatic envelope models are frequently used project future scenarios for species distributions under change, these rarely tested using empirical data. We long-term data on bird and abundance covering five states the western US Canadian province of British Columbia test capacity predict temporal changes populations over 32-year period. Using boosted regression trees, we built presence-absence that related presence 132 spatial variation conditions. Presence/ absence 1970–1974 forecast majority later time period, 1998–2002 (mean AUC = 0.79 � 0.01). Hindcast performed equivalently 0.82 Correlations between observed predicted abundances were also statistically significant most (forecast mean Spearman 0 s q 0.34 0.02, hindcast 0.39 0.02). The stringent geographic patterns through time. Observed significantly positively correlated with those 59% 0.28 across all species). Three variables (for wettest month, breeding season, driest month) minimum temperature coldest month important predictors this region, hence Our results suggest describing associations can some species, winter appear have already driven shifts North