Dynamic species distribution models reveal spatiotemporal habitat shifts in native range-expanding versus non-native invasive birds in an urban area

摘要

Urbanisation as a major driver of changes leads to the extinction of some species while others increase in abundance, especially non-native species. Spatiotemporal distribution patterns of these successful species are likely to be shaped by their response and tolerance to urban features. This study assesses the anthropo-ecological requirements of two co-occurring bird species, the native range-shifting jackdaw Corvus monedula and the non-native invasive ring-necked parakeet Psittacula krameri. We built yearly models over an eight-year period using an ensemble modelling approach assessing response differences through time and between species. Predictors describing human-made structures, socio-ecological proxies and resources availability were extracted from temporally coincident databases. Dispersal and habitat constraints were implemented in final models to provide more realistic forecasts of species future distributions. Ensemble models evaluated with a random partition of the training dataset showed a higher accuracy than those evaluated with an independent dataset from another time period. Our results highlight temporal variations in the relative importance of predictors for both studied species. Single-season occurrence data may thus be insufficient to characterize species ecological requirements. The ring-necked parakeet and the jackdaw showed different responses to urban features. Jackdaws preferred the more urbanized part of the city while the distribution of parakeets was strongly positively associated with the density of exotic ornamental trees. We concluded that ring-necked parakeet range expansion is likely to be driven by its effective ability to exploit urban resources which native species do not or under exploit, suggesting an open window of foraging opportunities. However, the jackdaw may be misled by a high cavity availability and a large amount of low-quality anthropogenic food in the urban core. We suggest that dynamic SDMs are a critical tool not only to forecast the future expansion of invasive species but also for a better understanding of processes driving urban biodiversity persistence.