In this study, we used the movement behavior of female sage-grouse obtained from GPS collar data to identify initiation of incubation and subsequent failure or hatching of the nest. Unlike nest monitoring efforts based on conventional telemetry, the approach we used allowed nests to be monitored (1) remotely without observer influence on incubation and (2) on a daily cycle, so the exact date of nest hatch or failure was known. Based on model weights (w
), there was little model uncertainty (Burnham and Anderson 2002) as to the selection of the best model among all candidate models. Within this landscape, nest-site placement by female sage-grouse was influenced by landscape variables at multiple spatial scales (Dzialak et al. 2011a); however, DSR of nests was most influenced by nest site-specific variables (area ≤ 30 × 30 m), similar to another study by Manzer and Hannon (2005). This finding is in contrast to other studies which found that landscape-level variables were most influential on the success of nests by ground-nesting birds (Stephens et al. 2005; Moynahan et al. 2007). Examining the variables that were included in the final model revealed potential mechanisms (i.e., visual and olfactory) that predators used to locate nests when considering that nest depredation and direct predation of the incubating female were the most common sources of nest failure. Last, the modeling approach used offers a simplified and unified framework for modeling nest- and time-specific covariates, fixed and random effects, complex hierarchical data structures, and multiple relationships (e.g., linear and quadratic) of the independent variables, and to account for the correlation of multiple measurements on the same bird and nest (Appendix 1).
Female movement and activity, collected using GPS collars, allowed researchers to find all nests beginning on day 1 of incubation, a phenomenon that rarely occurs in field studies (Shaffer 2004). This approach offered several advantages. First, we reduced any confounding effects of nest age because all nests were found and observed starting on day 1 of incubation (see Dinsmore et al. 2002 for a discussion on nest age as a confounding effect). Typically, apparent estimates of nest survival are biased (Moynahan et al. 2007), but under the conditions of equal detection probability between active and inactive nests (those that have already failed), apparent nest survival is relatively unbiased (Shaffer 2004), as we saw from our estimates. Therefore, we reduced the bias of estimates of nest survival because we found all nests (once incubation was initiated) before they had a chance to fail. Second, we modeled true DSR (interval = 1 day). Because we modeled true DSR, time-specific covariates, such as weather, were estimated with high precision (Shaffer 2004). Third, observer disturbance was minimized, thereby reducing this potentially confounding factor as a source of nest failure. Fourth, most previous studies used very-high-frequency transmitters to locate birds on nests with variable search schedules, thereby finding nests after the first day of incubation and thus biasing estimates of survival high because nests failing early were not detected. Crawford et al. (2004) reported an average nest survival (defined as the probability of hatching ≥ 1 egg) rate of 47.4% (n = 14 studies). Potentially then, the aforementioned average nest survival estimate could be biased high. Thus, our estimate of nest survival (25%) may be more accurate, albeit lower, because nests were detected on day 1 of incubation. Last, even when nests are rechecked periodically, the GLMM approach we present can still account for variable time intervals by using methods (i.e., the link function contains an exponent (1/t, where t = length of observation interval) in the numerator and denominator) similar to the logistic-exposure model (see Equation 2 in Shaffer 2004).
Other researchers hypothesized that DSR of nests would be lower in the early stages of incubation because vulnerable nests would be depredated earlier (Klett and Johnson 1982; Coates and Delehanty 2010); thus, we incorporated a time-dependent covariate (e.g., nest age) into models. However, we observed the opposite trend; nests had a higher probability of daily survival during early stages of incubation compared with later stages of incubation. This finding supports the idea that predators develop search images whereby predators may learn to cue in on female behavior during the course of incubation. Female attendance (Cao et al. 2009) or activity (Burhans et al. 2002) at the nest might draw the visual attention of predators to the site of the nest. More specifically, nests failing later in incubation may simply be related to the risk associated with exposure (Grant et al. 2005) where eggs exposed to the risk for longer periods will have more time to be detected and depredated. It is also plausible that the relationship between DSR of nests and nest age could be a function of predators cuing in on the nest due to olfaction from the female (Storaas 1988) because more odor will be emitted and bound to nest substrates (Conover 2007) the longer a female remains in one area (i.e., nest site).
In this study, there were several landscape features that influenced DSR of nests, particularly nest site-specific variables (≤ 30 m). Most of these features interacted with predator behavior to reduce or facilitate depredation of the nest. These features are important to consider because nest failure in most avian species, particularly ground-nesting birds, is due primarily to predation (Gregg et al. 1994; Conway and Martin 2000; Chalfoun et al. 2002; Holloran et al. 2005; Stephens et al. 2005; Moynahan et al. 2007), as it was in this study. The amount (i.e., percentage) of shrub cover around the nest site was important for reducing depredation. Sagebrush (Artemisia spp.) was the primary brush species comprising shrub canopy cover in our study, and it is well known that the amount or height of sagebrush around nest sites of sage-grouse is important for survival (Connelly et al. 1991; Schroeder et al. 1999; Coates and Delehanty 2010). DSR of nests increased linearly in relation to canopy cover of shrubs at the nest site. The positive relationship observed in this study offers support that shrubs can provide physical impediments to the nest. In addition, shrubs can serve as screening cover to reduce visual detection by terrestrial predators and canopy cover to reduce depredation by aerial predators. In general, greater concealment at the nest site leads to a lower probability of nest discovery by vision-based predators (Lima 2009). Thus, shrubs function primarily as a visual impediment to most predators of nests, albeit shrubs also can function to create turbulence (measure of variance of wind speed and direction) and reduce the odor plume of the bird nesting within (Conover 2007).
Although nest depredation often is a function of visual detection of the nest or the incubating female by predators, we also indirectly considered the importance of olfaction by terrestrial mammals while searching for nests (Conover 2007; Conover et al. 2010; Dritz 2010). Similar to Moynahan et al. (2007), DSR of nests decreased the day after a rainfall event. Higher failure rates the day following precipitation events may have been caused by a combination of factors. First, female activity may have increased the day following precipitation as the result of reduced activity (i.e., greater attendance at the nest site) on the day of precipitation (Moynahan et al. 2007). Increased activity of the female on the day following precipitation could have drawn attention to the female or the nest by predators. Other possible explanations include (1) increased predator activity the day after precipitation events (Moynahan et al. 2007) due to reduced activity on the days of precipitation, and (2) high moisture contents in the air may have facilitated olfaction of predators (Gutzwiller 1990) that led them to the nest. For instance, a wet bird releases more scent than a dry bird because water molecules displace scent molecules on the skin and feathers and allow them to evaporate and enter the air column (Conover 2007). Therefore, it appears reasonable that moisture within the air can heighten the olfactory senses of predators because of increased scent released by the bird.
A second weather variable (maximum wind speed), when added to the landscape model, improved model fit, providing further evidence of an olfactory cue used by predators to locate the source (i.e., the nest) of the scent. Wind speed influences olfaction of predators by carrying the scent over longer distances, allowing predators to track to the source of the scent (i.e., nest site). It has been hypothesized that high wind speeds will dilute odor plumes to undetectable levels or create odor plumes that are more difficult for a predator to follow (Conover 2007). However, in windy conditions, predators may require a consistent wind direction to navigate to the source of the odor (i.e., the nesting bird). It is interesting to note that the two weather variables most responsible for olfaction were included into the final models whereas temperature and humidity were not; a similar finding was observed by Dritz (2010).
There is a growing body of literature that points towards energy development as a factor in reduced demographic performance in certain species (e.g., Sawyer et al. 2009; Harju et al. 2010; Gilbert and Chalfoun 2011; Dzialak et al. 2011b) through means such as increased risk, landscape fragmentation, and altered predator communities and animal behavior. Typically, human-altered landscapes have a greater abundance of predators (Kurki et al. 1997, Kurki et al. 1998; Manzer and Hannon 2005), which is facilitated by infrastructure associated with wells that provide artificial perch sites for avian predators or ambush cover and den sites for terrestrial mammals (i.e., predator subsidization; Manzer and Hannon 2005; Coates and Delehanty 2010). Although predators will exploit human-altered landscapes, it may take several years for the full effects of disturbance to cascade across the landscape and influence predator occurrence. Therefore, it will be important to incorporate lag effects into model building to assess demographic responses related to disturbance that occurred in previous years. Harju et al. (2010) and Walker et al. (2007) report effects of previous development on population and breeding performance of sage-grouse. Therefore, we can infer that disturbance alters landscapes and subsequently influences predator search behavior or efficiency (Stephens et al. 2005).
In a recent study, the risk of losing a nest before hatch was most influenced by distance to mesic areas and wells; distance to roads did not structure nest survival (Dzialak et al. 2011a). We observed that DSR of nests was marginally greater for birds that nested closer to roads. However, the difference in distance to the nearest road between successful (2,568 m) and unsuccessful (2,693 m) birds was 125 m, which could be considered biologically insignificant. Given that birds nested away from roads in general (mean ≥ 2,568 m), the minimal difference in distance to the nearest road between the two groups (125 m) and the finding that roads did not structure the risk of nest failure (Dzialak et al. 2011a) provides further support of the importance of other factors in structuring DSR of nests in this ground-nesting bird of conservation concern.
Many species of grouse select for habitats with mesic areas nearby for various life-history phases (Walker et al. 2007). We found that nest survival was lowest when nests were within 50 m from mesic habitat. This relationship may be caused by the use of mesic habitat by predators (Stephens et al. 2005; Walker et al. 2007), many of which function as travel corridors, edge habitat, or ambush sites for terrestrial predators (Winter et al. 2000). While mesic areas associated with anthropogenic features of the landscape reduced nest survival (Dzialak et al. 2011a), DSR of nests was reduced when in proximity to any mesic area, likely due to the functional habitat value of the area for a wide number of avian and terrestrial predators. In other studies, taller vegetation associated with mesic areas increased the number of perch sites for avian predators, thus facilitating predation of nests by avian predators (Manzer and Hannon 2005). Although any type of mesic area reduced DSR of nests, careful attention should be paid to water discharge associated with anthropogenic activities (e.g., agriculture irrigation, oil and gas extraction, and ranching), which would artificially create and inflate the number of mesic areas (example of predator subsidization).