Abstract
Woody plant expansion is an ongoing threat in many grassland and shrubland habitats. Over time woody plant expansion leads to changes in habitat structure and complexity which can alter interactions among species, including predation risk. One prominent example of woody plant expansion in North America is expansion of conifer trees into sagebrush steppe habitats. Over time, increasing conifer cover can reduce sagebrush and herbaceous cover and leads to broad-scale changes in composition and structure across the landscape. These landscape-level changes can lead to an ecosystem shift, which has implications for many shrubland-habitat species and may result in reduced rates of survival or reproduction. Numerous conifer removal projects are currently underway to improve habitat for declining shrubland-habitat species, most notably the greater sage-grouse (Centrocercus urophasianus; hereafter sage-grouse). In areas where conifer trees are present, sage-grouse have higher annual mortality rates and conifer cover as little as 4% is associated with lek abandonment. Removal of conifer cover leads to increased adult and nest survival for sage-grouse, suggesting the presence of conifer trees contributes to higher rates of nest failure. Sage-grouse, like many ground-nesting birds, experience relatively high rates of nest failure primarily due to predation. Common ravens (Corvus corax) are often identified as one of the primary predators of sage-grouse nests and are expected to benefit from the presence of trees in previously treeless landscapes. However, mammalian nest predators, which can account for up to half of all nest depredations in some geographic areas of sage-grouse range, may also respond to conifer dynamics in shrublands. We sought to inform how conifer expansion and removal affect predator-prey dynamics using sage-grouse as a model species. We conducted our study in southwestern Idaho where a large-scale western juniper (Juniperus occidentalis; hereafter juniper) removal project began in 2019. We first examined how landscape features, including measures of juniper expansion and removal, influence habitat use by mammalian mesopredators. Second, we sought to identify what landscape-level and fine-scale features influence rates of depredation for sage-grouse nests. To estimate rates of habitat use for mesopredators, we deployed trail cameras (n = 80 cameras in 2019 and 2020, and n = 76 cameras in 2021) from April–July in 2019–2021. We stratified the landscape by juniper cover categories and randomly placed cameras within each stratum. We estimated probability of habitat use for five mesopredator species: coyote (Canis latrans), red fox (Vulpes vulpes), American badger (Taxidea taxus), striped skunk (Mephitis mephitis), and bobcat (Lynx rufus) using mark-recapture occupancy models. Additionally, we estimated the relative abundance of avian predators through surveys within the vicinity of active sage-grouse nests. We used single-season models to examine effects of juniper cover on habitat use of mesopredators. We created each capture history using unique camera-year combinations and “stacked” the yearly detection histories for all cameras. Thus, the total sample size for analysis of habitat use for each species of mesopredator was 236 camera-year combinations. We detected strong associations between juniper cover and rates of habitat use for four species of mesopredator. As juniper cover increased, rates of habitat use for coyotes and red foxes decreased, whereas rates of habitat use for striped skunks and bobcats increased. Rates of habitat use for American badgers were unaffected by amount of juniper cover. We saw no direct effect of juniper removal on rates of habitat use for any mesopredator species, however, changes in habitat use may take years to manifest following large-scale changes on the landscape. Additionally, increased use of conifer removal areas by prey species, as has been observed in other conifer removal studies, could precipitate increased use of the same areas by mesopredators . Mixed community-level responses may result in unpredictable or no net decrease of exposure to predators for species targeted for management by conifer removal. However, prey may still benefit from tree removal via indirect effects mediated by habitat changes (e.g., increased shrub or herbaceous cover). To estimate the effects of landscape features, including habitat use by mesopredators, on rates of depredation for sage-grouse nests, we used logistic-exposure models with an implemented Schaffer link. Proportion of class 2 juniper cover and clustering of juniper trees affected predation rates, as did various features of shrub and herbaceous vegetation at fine spatial scales. We found no direct support for an effect of a single species of mesopredator on nest predation. However, as observed in our first chapter, mesopredator species at our study area respond to conifer expansion in sagebrush systems, suggesting that predator-prey dynamics are affected by woody plant expansion, and that compensatory predation is a possibility with tree removal. Our results further highlight the importance of removing class 2 juniper cover and clustered stands of trees for sage-grouse, especially near intact nesting habitat. Identifying thresholds of remaining woody plant cover that allow shrubland-habitat specialists to persist and experience improved demographic rates will be the most efficient way to restore populations. Managers should focus conifer removal efforts where there is a robust herbaceous understory and should monitor populations of interest following removal efforts to ensure that improved demographic rates are realized.