Resource information
AIM: The geographical expansion of white‐winged doves (Zenaida asiatica) in North America has attracted the attention of biologists and sportsmen because of their recreational and aesthetic value; however, data on factors driving the spatial spread of this species are lacking. We examined spatial and temporal patterns of range expansion for white‐winged doves along the northern edge of their geographical range from 1979 to 2007 and used a dynamic occupancy model to estimate when and where doves would be found along an expansion gradient. LOCATION: Southern half of the USA. METHODS: We modelled spatial variation in the range expansion of white‐winged doves from 1979 to 2007 using data from the North American Breeding Bird Surveys (BBS) and distance from initial population centre, area of urban land cover, and ecoregion. We used a robust design occupancy analysis to estimate species expansion rate based on a model set grounded in broad‐scale conditions likely to drive dove distribution. We evaluated our best fitting model (using Akaike's information criterion adjusted for small sample size) by comparing estimates to actual observations in 1993 and 2007 using area under the curve (AUC) metrics from receiver‐operating characteristic plots. RESULTS: Our spatial distribution model indicated that range expansion of white‐winged doves was primarily influenced by distance from the core population centre, with additive effects of ecoregion and land cover. In 1979, white‐winged doves were present in 12% of survey locations, whereas presence increased to 41% by 2007. Across all ecoregions, urban land cover was positively related to probability of occupancy. We evaluated our model for 1993 and 2007 by comparing derived occupancy estimates to sightings of white‐winged doves within our sample hexagons using BBS survey data. Based on high AUC statistics (>� 0.85), we concluded that our model was useful for accurately predicting range expansion. MAIN CONCLUSIONS: Estimating occurrence and range expansion of white‐winged doves and other expanding/invasive avian species at large spatial scales can be effectively conducted using BBS data. Using these nationwide, long‐term survey data to relate expansion to a suite of conditions likely to drive population dynamics is an effective approach to developing predictive models of range expansion.
