Northern bobwhite occupancy patterns on multiple spatial scales across Arkansas

Northern bobwhite Colinus virginianus populations have been rapidly declining in the eastern, central, and southern United States for decades. Land use change and an incompatibility between northern bobwhite resource needs and human land use practices have driven declines. Here, we applied occupancy analyses on two spatial scales (state level and ecoregion level) to more than 5,000 northern bobwhite surveys conducted over 6 y across the entire state of Arkansas to explore patterns in occupancy and land use variables, and to identify priority areas for management and conservation. At the state level, northern bobwhite occupied 29% of sites and northern bobwhite were most likely to occur in areas with a high percentage of early successional habitat (grassland, pasture, and shrubland). The statewide model predicted that northern bobwhite were likely to occur (≥ 75% predicted occupancy) in < 20% of the state. Arkansas is comprised of five distinct ecoregions, and analyses at the ecoregion spatial scale showed that habitat associations of northern bobwhite could vary between ecoregions. For example, early successional habitat best predicted northern bobwhite occupancy in both the Arkansas River Valley and Ozark Mountains ecoregions, and other habitat associations such as the proportion of herbaceous habitat and hay-pasture habitat, respectively, further refined predictions. Contrastingly, richness of land cover classes alone best predicted northern bobwhite occupancy in the Ouachita Mountains ecoregion. Ecoregion-level models were thus more discerning than the state-level model and should be more helpful to managers in identifying priority conservation areas. However, in two of five ecoregions, surveys too rarely encountered northern bobwhite to accurately predict their occurrence. We found that likely occupied northern bobwhite habitat lay primarily on private properties (95%), but that numerous public entities own and manage land identified as suitable or likely occupied. We conclude that management of northern bobwhite in Arkansas could benefit from cooperation among state, federal, and military partners, as well as surrounding private landowners and that ecoregion-specific models may be more useful in identifying priority areas for management. Our approach incorporates multiple landscape scales when using remote sensing technology in conjunction with monitoring data and could have important application for the management of northern bobwhite and other grassland bird species.