Linking Individual Behavior to Population Dynamics: Insights for Effective Raven Management

Abstract

The expansion of human infrastructure has altered ecosystems worldwide. Anthropogenic changes to the landscape reshape wildlife populations by influencing survival, reproduction, dispersal, connectivity, and predator-prey interactions, causing some wildlife populations to decline while enabling others to thrive and expand their range. Predatory species that benefit from anthropogenic features, known as synanthropic species, can further impact declining prey populations. The common raven (Corvus corax) is an example of such a predator. From 1966¬¬¬-2018, raven populations increased by 1.5% annually across North America. As generalist predators, ravens thrive in human-altered landscapes. As a result, anthropogenic subsidies can lead to high raven densities, ultimately having negative impacts on prey, including species of concern. Predation of these already declining populations requires effective management of raven populations. In this dissertation, I applied novel approaches to understand raven demographics and the mechanisms driving changes in raven abundance and space use.In Chapter 1, I employed a reaction-diffusion model to predict raven density across the western United States and quantify the effects of anthropogenic development on raven dispersal. My results indicated that raven abundance doubled in fewer than 2 decades (2001-2019) with hotspots of raven density in southern California, central Nevada, and northern Utah. I also identified low elevations as a driver of longer raven residence times and find moderate evidence that high road density also contributed to residence times. As ravens expand eastward, areas with these attributes may be particularly susceptible to high raven densities. By identifying areas prone to raven settlement and high densities, raven density projections provide valuable guidance for ongoing monitoring of habitat conditions, as well as raven and prey populations, which are both crucial for effective management. In addition, my results indicated that raven management in these areas might require actions beyond raven removal, as other ravens may continue to move in and settle. In Chapter 2, I developed a mechanistic movement model to provide estimates of age-specific breeding propensity and apply it to a raven population in northern Nevada. Movements of breeding ravens were generally less diffuse and demonstrated a stronger mean-reverting tendency than non-breeding ravens, indicating that breeding ravens concentrated movement in a smaller area close to their respective nests. My results also indicated that adult ravens (≥5 years old) in the region had a breeding propensity of 0.71 (95% CRI = 0.35-0.95) and subadult (3-4 years old) breeding propensity had a breeding propensity of 0.38 (95% CRI = 0.14¬-0.67). This suggests that 29% of adults and 62% of subadults in the central Nevada population were not breeding, despite being sexually mature. Sexually mature, non-breeding individuals can buffer breeder mortality by quickly filling vacant territories and stabilizing reproductive density or acting as a source of future breeders for neighboring populations. In areas with a high proportion of mature, non-breeding ravens, management of the breeding population may not be sufficient to control raven densities. In Chapter 3, I compared breeding and non-breeding raven space use using 3 metrics: step length, home range size, and core area size. I also compare habitat selection within home ranges between individuals of different reproductive classes. Breeding ravens had shorter step lengths than non-breeding ravens (μ = -793.41 m/hr, 95% CRI = -879.49--703.24 m/hr). Although I did not find a difference in home range size between reproductive classes, my results indicated that breeding ravens had a smaller mean core area size, (μ = -268.73 km2, 95% CRI = -355.30--189.96 km2). Ravens in both breeding classes selected high normalized difference vegetation index (NDVI) and low annual grass and shrub cover. However, non-breeding ravens concentrated activity near forest edges, natural water sources, and anthropogenic features, whereas breeding ravens focused activity close to their nests. These results suggest that raven management could be most effective in areas with high NDVI and annual grass and shrub cover, especially in anthropogenically modified landscapes and near forest edges. In addition, preventing the establishment of raven nests near prey populations could benefit species of concern. The research contained in this dissertation advances scientific understanding of the raven behavior and demographics driving abundance patterns. At a landscape level, low elevations and high road densities result in ravens spending more time and possibly establishing territories in these areas. Sexually mature, non-breeding ravens not only contribute to higher raven densities in these areas, but can also act as a buffer for breeding raven mortality. Sexually mature, non-breeding ravens can also maintain high raven densities in areas with abundant resources that lack nesting sites. Furthermore, behavioral differences between non-breeding ravens and breeding ravens contribute to variation in predation intensity, with breeding ravens concentrating foraging around their nest and non-breeding ravens gathering at forest edges, natural water sources, and anthropogenic features throughout their home range. Finally, the methods I develop and employ in this dissertation not only inform raven management but can also contribute to the management and conservation of species in other systems that have similar predator-prey dynamics.