Abstract
Greater sage-grouse (Centrocercus urophasianus, hereafter, sage-grouse) in the Great Basin have experienced loss of habitat due to expansion of western juniper (Juniperus occidentalis; hereafter, juniper) woodlands into sagebrush steppe. Juniper expansion can alter the sagebrush understory by reducing cover and species richness of herbaceous plants and shrubs, which may influence the availability of resources required by sage-grouse. On average, sage-grouse avoid juniper, especially when cover is > 10%, and avoidance of juniper can increase survival rates. However, there is significant variation in habitat selection among sage-grouse individuals when juniper cover is < 10%, and some individuals demonstrate preference for these areas. This pattern is possibly related to condition of the understory; cover of sagebrush shrubs and herbaceous plants may not yet be affected in areas where juniper cover is < 10%. Thus, individuals could select areas with non-zero levels of juniper cover despite potential for higher risk of mortality in those areas because resources required for survival and reproduction are still available.
In this thesis, I sought to evaluate if reproductive status influences habitat selection among female sage-grouse under different reproductive status and if physiological condition among hens is influenced by juniper cover. Female sage-grouse under different reproductive status can vary in habitat selection, however, comparisons of selection among hens in landscapes undergoing juniper expansion have not been evaluated. In addition, effects that juniper may have on hen physiological condition have not been explored. I conducted my study in Owyhee County, Idaho 2017–18 where juniper expansion is considered one of the primary threats to local sage-grouse populations.
In chapter 2, I investigated if reproductive status among hens with and without broods (hereafter, brooding and non-brooding hens, respectively) influences habitat selection at multiple spatial scales. Habitat selection patterns may be a function of reproductive status because specific conditions that support individuals with young may not yield the same benefits for individuals without young. I employed a use and available design and collected data on habitat through field-based surveys and using remotely-sensed layers in a Geographic Information System (GIS). I used resource selection functions to evaluate habitat selection for brooding and non-brooding hens during the brood-rearing period (30 April¬–26 July) and made comparisons between reproductive groups. I conducted field-based habitat surveys at 181 use and available locations from 10 (2017) and 18 (2018) hens. I collected geospatial data at 2,226 use and available locations for 11 (2017) and 21 (2018) hens. At my smallest spatial extent, brooding hens were more likely than non-brooding hens to select habitats with more cover (e.g., taller perennial grass and non-sagebrush shrubs). At greater spatial extents, both reproductive groups generally avoided cover class II (> 10–20% juniper cover) and III (> 20% juniper cover) but selected for cover class I (> 0–10% juniper cover), woody wetlands, and herbaceous wetlands with high perimeter to area ratios. Brooding hens may select for taller vegetation because these areas provide more concealment cover for chicks, thereby providing more protection from predators. In contrast, non-brooding hens may use grouping behavior as an anti-predator strategy and may not have to rely on areas with taller vegetation for protection. Hens avoided cover class II and III because resources that support demographic processes are less available in these areas. Both reproductive groups selected cover class I, possibly because food resources and concealment cover are not yet reduced to levels that result in habitat unsuitable for sage-grouse. Furthermore, brooding and non-brooding hens selected for wetland habitats because these areas may provide high amounts of food sources (i.e., forbs and insects) than the surrounding uplands.
In chapter 3, I investigated relationships between concentrations of stress hormones among hens and ecological factors. Along with possibly reducing the availability of food and concealment cover, juniper trees may create suitable habitat for avian predators, potentially increasing the risk of predation for sage-grouse. In several avian species, habitat characteristics can influence concentrations of stress hormones, and elevated levels of stress hormones can have negative influences on factors related to survival and reproductive success (e.g., suppress immune function, probability of nest and brood abandonment, and slower growth rates in offspring). Hormone concentrations in sage-grouse may be positively associated with juniper cover through decreased resource availability or increased pressure from predators. I collected fecal samples at nighttime roost locations of radio-collared hens during the lekking (4 March–8 May) and brood-rearing period (24 May–26 July) to estimate corticosterone concentrations (i.e., stress hormones; hereafter, FCORTm). I evaluated relationships between vegetation cover (hereafter, ecological variables) and FCORTm in hens. I used remotely-sensed layers to estimate ecological variables within multiple spatial extents centered at breeding grounds (i.e., leks) and within separate, minimum convex polygons (MCP) that surrounded use locations of each hen. I used values from ecological variables estimated within leks and MCPs to evaluate relationships with FCORTm during the lekking and brood-rearing period, respectively. Prior to evaluating relationships with ecological variables, I accounted for factors previously shown to influence FCORTm in other vertebrate species, such as age, temperature, and sample mass. I collected 37 fecal samples from 34 hens during the lekking period (4 March–8 May) and 36 fecal samples from 22 hens during the brood-rearing period (24 May–26 July). During the lekking period, FCORTm had a negative relationship with dry mass of the fecal sample and there was no relationship with ecological variables. During the brood-rearing period, FCORTm had a positive relationship with total area of MCP but a negative relationship with the number of days of reproductive activity, maximum daily temperature (°F), and proportion of cover class I (> 0–10% juniper cover) within MCP. I may not have observed relationships between ecological variables and FCORTm during the lekking period because hens arrive on breeding grounds at different times and could vary temporally and spatially in their use of habitat surrounding each lek. During the brood-rearing period, FCORTm may decrease with greater proportions of cover class I because of density dependent factors and high productivity of shrubs and herbaceous plants in areas with young stands of juniper. Because interpretation of relationships between stress and ecological factors can be influenced by sampling and extraction procedures, my results lay the groundwork for additional studies that employ the same laboratory methods to evaluate FCORTm in sage-grouse.
Although hens preferred cover class I, previous research has demonstrated lower survival among sage-grouse that occupy areas with low levels of juniper cover, and removal of cover class I would likely benefit sage-grouse. My results do suggest lower stress levels among hens that use habitats with cover class I, but this benefit likely does not outweigh the cost to survival. Given the avoidance of cover class II and III, I also suggest targeted removal of juniper around wetlands dominated by woody vegetation, patchy, herbaceous wetlands with high edge ratios, and mesic habitats with taller non-sagebrush shrubs may be the most beneficial because these habitats were preferred by hens. Wetlands and mesic habitats with tall shrubs likely benefit sage-grouse, perhaps by positively influencing survival of chicks and adults. However, additional monitoring is needed to assess benefits and costs to demographic processes among sage-grouse that select woody wetlands and tall shrubs.