When and where to move: Dynamic occupancy models explain the range dynamics of a food nomadic bird under climate and land cover change.

Globally, long-term research is critical to monitor the responses of tropical species to climate and land cover change at the range scale. Citizen science surveys can reveal the long-term persistence of poorly known nomadic tropical birds occupying fragmented forest patches. We applied dynamic occupancy models to 13 years (2002-2014) of citizen science-driven presence/absence data on Cape parrot (Poicephalus robustus), a food nomadic bird endemic to South Africa. We modeled its underlying range dynamics as a function of resource distribution, and change in climate and land cover through the estimation of colonization and extinction patterns. The range occupancy of Cape parrot changed little over time (ψ = 0.75-0.83) because extinction was balanced by recolonization. Yet, there was considerable regional variability in occupancy and detection probability increased over the years. Colonizations increased with warmer temperature and area of orchards, thus explaining their range shifts southeastwards in recent years. Although colonizations were higher in the presence of nests and yellowwood trees (Afrocarpus and Podocarpus spp.), the extinctions in small forest patches (≤227 ha) and during low precipitation (≤41 mm) are attributed to resource constraints and unsuitable climatic conditions. Loss of indigenous forest cover and artificial lake/water bodies increased extinction probabilities of Cape parrot. The land use matrix (fruit farms, gardens, and cultivations) surrounding forest patches provides alternative food sources, thereby facilitating spatiotemporal colonization and extinction in the human-modified matrix. Our models show that Cape parrots are vulnerable to extreme climatic conditions such as drought which is predicted to increase under climate change. Therefore, management of optimum sized high-quality forest patches is essential for long-term survival of Cape parrot populations. Our novel application of dynamic occupancy models to long-term citizen science monitoring data unfolds the complex relationships between the environmental dynamics and range fluctuations of this food nomadic species.